Citation for published version (APA): Bruinenberg, V. M. (2017). Phenylketonuria in mice and men [Groningen]: Rijksuniversiteit Groningen

Transcription

1 University of Groningen Phenylketonuria in mice and men Bruinenberg, Vibeke Marijn IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2017 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Bruinenberg, V. M. (2017). Phenylketonuria in mice and men [Groningen]: Rijksuniversiteit Groningen Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 Phenylketonuria in mice and men Vibeke Marijn Bruinenberg

3 The studies described in this thesis were carried out at the Groningen Institute for Evolutionary Life Sciences (GELIFES), University of Groningen, Groningen and the at the University Medical Center Groningen, Beatrix Children s Hospital, Groningen, The Netherlands. This research has been supported by a grant from Nutricia Research. Layout: Alex Wesselink (Persoonlijk Proefschrift.nl) Cover design: Vibeke Bruinenberg Printed by: Ipskamp Printing ( ISBN:

4 Phenylketonuria in mice and men Proefschrift ter verkrijging van de graad van doctor aan de Rijksuniversiteit Groningen op gezag van de rector magnificus prof. dr. E. Sterken en volgens besluit van het College voor Promoties. De openbare verdediging zal plaatsvinden op vrijdag 22 september 2017 om 14:30 uur door Vibeke Marijn Bruinenberg geboren op 19 juni 1988 te Groningen

5 Promotores Prof. dr. E.A. van der Zee Prof. dr. F.J. van Spronsen Beoordelingscommissie Prof. dr. C.O. Harding Prof. dr. S. Spijker Prof. dr. G.J. van Dijk

6

7

8 TABLE OF CONTENTS Chapter 1 General introduction 9 Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6 Chapter 7 The Behavioral Consequence of Phenylketonuria in Mice Depends on the Genetic Background The behavioral phenotype of female phenylketonuria mice differs partially from male phenylketonuria mice Sleep disturbances in Phenylketonuria: an explorative study in men and mice A novel treatment strategy for phenylketonuria: exploring the possibilities of nutrients to improve brain function A specific nutrient combination attenuates the reduced expression of PSD-95 in the proximal dendrites of hippocampal cell body layers in a mouse model of phenylketonuria Long-term treatment with a specific nutrient combination in phenylketonuria mice improves recognition memory Chapter 8 Chapter 9 Large neutral amino acid supplementation exerts its effect through three synergistic mechanisms: proof of principle in phenylketonuria mice Therapeutic brain modulation with targeted Large Neutral Amino Acid supplements in the Pah-enu2 phenylketonuria mouse model Chapter 10 Summary and general conclusion Appendix I Nederlandse samenvatting 203 Appendix II Dankwoord 213 Appendix III Curriculum Vitae 219 Appendix IV Publications 223

9

10 CHAPTER 1 General introduction

11 Chapter 1 1. PHENYLKETONURIA The heritable metabolic disorder phenylketonuria (PKU) is caused by an inborn error in phenylalanine (Phe) metabolism. This inborn error causes a dysfunction in the hepatic enzyme phenylalanine hydroxylase (PAH) affecting the hydroxylation of Phe to tyrosine. As a consequence, Phe obtained from dietary protein intake is not converted, resulting in a vast rise of Phe in blood and brain. Furthermore, as the primary catabolic pathway of Phe relies on functional PAH, an alternative pathway is initiated which causes an increase in metabolites of Phe (phenylethylamine, phenylpyruvate, phenylacetate, 2-hydroxyphenylacetate, and phenyllactate). The distinct odor of one of these metabolites (phenyllactate) in the urine of patients was key in the first discovery of the disease in The name was later changed to phenylketonuria considering the presence of these phenylketones in urine 2. The first descriptions of a PKU child was by a mother describing her normally developing child whom at one point did not develop anymore (The Child Who Never Grew by Pearl Buck, ). In the end, the buildup of Phe in the blood and brain affected the developing brain causing mental disablement, problems with movement, and seizures THE AFFECTED PHENYLKETONURIC BRAIN. Although the conversion of Phe is disrupted in the liver, the most dramatic consequences of raised Phe concentrations are found in the brain. Within the brain, these high concentrations of Phe can have detrimental effects via direct and indirect mechanisms (Figure 1). The first manner in which Phe can have an impairing effect is already with the entry of Phe in the brain (Figure1 (1)). The LAT1 transporter responsible for this process is the predominant transport system of large neutral amino acids (LNAA s) over the blood-brain barrier 5. As the affinity of Phe is very high to this transporter, high concentrations of Phe can outcompete other LNAA s in transport, causing reduced concentrations of non-phe LNAA s in brain 6,7 (Figure 1 (2)). Among these LNAA s are tyrosine and tryptophan, precursors of the neurotransmitters dopamine and serotonin. Together with the poor intrinsic capacity to produce tyrosine, this can result in reduced concentrations of these precursors in the brain. In turn, this can result in reduced concentrations of the above mentioned neurotransmitters 8 (Figure 1 (4)). Furthermore, high levels of Phe can inhibit the enzymatic activity of tyrosine hydroxylase and tryptophan hydroxylase, important in the conversion of these precursors to their subsequent neurotransmitters 9 (Figure 1 (3)). Not only neurotransmitter metabolism is affected by high Phe. Increased Phe concentrations and its metabolites can increase the generation of reactive species and inhibit antioxidant enzymes 10,11 (Figure 1 (5)(6)). An imbalance between this production and/or removal of reactive species is defined as oxidative stress that is frequently found in PKU models and PKU patients (reviewed by Ribas 12 ). Oxidative stress can damage proteins, lipids, carbohydrates, and DNA. This can result in cell damage or death, f.e. in neurons. Indeed, reduced functioning of neurons or synapses are found in PKU models and 10

12 Chapter 1 post-mortem tissue of PKU patients (Figure 1 (8)). However, not only oxidative stress can have these neurotoxic effects. In 2012, Adler-abramovich and colleagues showed that high concentrations of Phe can result in the assembly of toxic fibrils with an amyloid-like structure 23 (Figure 1 (7)). When added to cell culture, these structures have clear neurotoxic effects. Finally, myelin alternations are consistently reported in PKU (found in in vitro and in vivo models, and in untreated and treated PKU patients (Figure 1 (9)). However, the exact mechanism in which Phe affects myelin is not clear yet. A possible mechanism could be the effect of Phe on cholesterol and protein synthesis, processes key in the production and maintenance of myelin sheaths The increased concentrations of Phe have an extensive effect on the brain, affecting neurtransmitter metabolism, oxidative stress, synaptic functioning, and myelin. These are domains that are highly interrelated and are at the basis of cognitive functioning. 1 Figure 1 The affected phenylketonuric brain. (1) The LAT1 transporter has a high affinity to phenylalanine. The increased Phe concentrations compared to other non-phe large neutral amino acids, such as tryptophan (Tryp) and tyrosine (Tyr) outcompete these non-phe LNAA for entry into the brain. (2) This causes high concentrations of Phe in the brain and low concentration of other non-phe LNAA s, such as Tryp and Tyr. Furthermore, the dysfunction of phenylalanine hydroxylase in the periphery causes poor intrinsic capacity to produce Tyr. (3) High concentrations of Phe inhibit tyrosine hydroxylase and tryptophan hydroxylase important in the conversion of Tyr to 3,4-dihydroxyphenylalanine (DOPA) and Tryp to 5-hydrotryptophan (5-HTP) respectively. (4) Together this causes a decrease in serotonin 11

13 Chapter 1 and dopamine. (5) High Phe concentrations can inhibit anti-oxidant enzymes and (6) increase reactive species. The disrupted balance between production and removal of reactive species is oxidative stress. (7) High concentrations of Phe can assembly in neurotoxic fibrils with an amyloid-like structure. (8) Increased Phe concentrations can affect neuronal functioning trough affecting synaptic morphology, and proteins related to synaptic functioning. (9) In PKU, alternations in myelin/ white matter integrity are consistently found. A possible mechanism could lay in the indirect effect of Phe on cholesterol and protein synthesis, processes important in the production and maintenance of myelin sheaths. 3. THE OUTCOME OF PHENYLKETONURIA PATIENTS. In the 1950 s, it became clear that the treatment of PKU should focus on reducing Phe intake 38. In subsequent years, the first studies showed that low-phe intake could reduce Phe in blood and cerebrospinal fluid and improve behavior in PKU patients 39,40. Currently, the treatment of PKU patients still aims to reduce Phe intake that is now achieved by a protein-restricted diet supplemented with artificial amino acids, vitamins, and minerals. New born screening facilitates the early introduction of this diet which is prescribed as a diet-for-life 41. In early-treated patients, this diet could overcome the severe mental disabilities experienced by untreated PKU patients. However, in recent years, it became clear that the current treatment, the difficulties maintaining diet, and the world-wide misalignment of Phe targets of treatment causes a suboptimal outcome 42,43. Early-treated PKU patients still experience reduced psychosocial outcome, quality of life, cognitive functioning, such as in processing speed, attention and working memory, and more internalizing problems such as depression and anxiety Therefore, the development of new and/or additional treatment strategies in PKU is of great importance. 4. THESIS OUTLINE In this current thesis we investigate two new treatment strategies in PKU, namely a specific nutrient combination and large neutral amino acids supplementation. This research consists of preclinical studies in the PKU mouse model. A major challenge in this preclinical research is the translational value of the model. The translational value implies and demands a link between the model and the modelled condition (the PKU mouse model to PKU patients). Therefore, the first part of this thesis aims for a completer understanding of the PKU mouse model, whereas the second and third part relate to the two new treatment strategies. As such, the outline of this thesis is divided in three parts: I) Characterization and translational value of the PKU mouse model, II) The effect of a specific nutrient combination in PKU and III) Large neutral amino acids supplementation in PKU. 12

14 Chapter 1 Part I: Characterization and translational value of the PKU mouse model In PKU research, several models are used to investigate underlying mechanisms and new treatment strategies. A model often used is the PKU mouse model. This model, first described in 1993 by Shedlovsky and colleagues, was developed via germline mutagenesis with N-ethyl-N-nitroso-ureum (ENU) 46. This technique caused a mutation in the PAH gene resulting in very little enzymatic activity and, consequently, increased concentrations of Phe in blood and brain 46,47. This model, Pah enu2, was developed in the BTBR genetic background. However, outside of PKU research, this genetic background is known for their abnormalities in behavior and brain morphology Therefore, the BTBR Pah enu2 was crossed back on to the more commonly used C57Bl/6 background. At this moment, both backgrounds of the mouse model are used as equals despite clear indications from other research fields that genetic background can influence phenotypical behavior 52. In chapter 2, we directly compare the adult male wild-type (WT) and PKU individuals of both strains (BTBR and C57Bl/6) in four PKU-related domains; activity, motor performance, anxiety-and depressivelike behavior and learning and memory. As genotyping with specific probes and primers confirmed that the PKU mice of both strains have an identical point mutation, this study further investigates if the biochemical profiles of the PKU mice of both genetic backgrounds is similar. The biochemical profiles consist of neurotransmitter concentrations in brain, and amino acid concentration in blood and brain. 1 In chapter 2, we examined male mice, as preclinical research most often focusses on the males. However, it is progressively recognized that males can respond differently than females Therefore, chapter 3 investigates the four PKU-related domains in adult female WT and PKU individuals of both strains. In the previous two chapters we have tested four PKU-related domains described for PKU. A phenotypical consequence of PKU that is not described in literature is sleep-related problems. The lack in sleep research is somewhat surprising to us, as several modulators of sleep and wakefulness are affected in PKU 56,57 (f.e. serotonin, dopamine, norepinephrine and orexin). For this reason, chapter 4 investigates sleep characteristics in PKU patients and sleep/wake patterns in male and female PKU mice of both genetic backgrounds. In PKU patients and first degree relatives, four questionnaires in an electronic survey are used. Together with ten questions to characterize the subjects, the questionnaires are used to examine the possible occurrence of sleep disorders, sleep quality, sleepiness during the day, and chronotype. In PKU and WT mice, rest/wake patterns are monitored via passive infrared recorders. From these patterns, the fragmentation score (the frequency of switching between active and nonactive behavior) and diurnality (e.g. night active animals are active in the dark phase which gives a negative diurnality score) is calculated. 13

15 Chapter 1 Part II: The effect of a specific nutrient combination in PKU Current dietary treatment is often difficult to maintain by PKU patients. This can result in reduced compliance causing Phe concentrations to rise and fluctuate 58. In contrast to the traditional focus of treatment to reduce Phe, counteracting the detrimental effects of Phe on the brain could be of great interest to improve PKU treatment. As previously described, several domains are affected in PKU, namely neurotransmitter metabolism, oxidative stress, white matter integrity, and synaptic functioning. Chapter 5 reviews nutrients that can positively affect theses domains of the brain. In this review, a combination of specific nutrients is postulated as an additional new treatment strategy for PKU. This specific nutrient combination (SNC) consists of uridine monophosphate (UMP), docosahexaenoic acid (DHA), eicosapentaenoic acid (EPA), choline, phospholipids, folic acid, vitamins B12, B6, C, and E, and selenium. The synergistic approach was originally designed to improve the synthesis of phospholipids, a major component of (synaptic) membranes. Together and beyond the original focus, SNC supplementation have shown positive effects on synaptic functioning, e.g. pre- and postsynaptic proteins and neurite outgrowth 61,63, neurotransmitter release and signaling 61,64, and memory performance 65,66. In chapter 6, for the first time, SNC supplementation is implemented in male and female C57Bl/6 WT and PKU mice. In this proof-of-concept study, we examine the effect of SNC supplementation on a post-synaptic marker, postsynaptic density protein 95, in specific subregions of the hippocampus. In chapter 7, we continue this work in the BTBR WT and PKU mice to investigate the behavioral outcome of this treatment. In a long-term intervention study, we examine whether SNC supplementation can improve motor performanc, novel- and spatial object recognition memory in high Phe and low Phe conditions. Part III: Large neutral amino acid supplementation in PKU As previously described, high Phe concentrations in blood can outcompete other non- Phe LNAA s (tyrosine, tryptophan, valine, isoleucine, leucine, methionine, histidine, and threonine) in the transport over the blood-brain barrier. Consequently, Phe concentrations are increased in brain, non-phe- LNAA s concentrations are reduced, and neurotransmitter metabolism impaired. Restoring the balance between Phe and non-phe LNAA s in blood with supplementing additional non-phe LNAA s could counteract these consequences. Indeed, clinical studies support this hypothesis However, the optimal composition and the effect on all three biochemical disturbances are still unknown. Therefore, as a start, the acute diet regime of Pietz and colleagues (1999) 67 in PKU patients was transformed to a continuous treatment of supplementation of equal amounts of all non-phe LNAA s except for threonine. In this study, described in chapter 8, we offer this LNAA s regime for six weeks to male and female C57Bl/6 WT and PKU mice. After six weeks, blood and brain LNAA s and neurotransmitter concentrations are examined to investigate the three biochemical treatment objectives. In chapter 9, the results found in chapter 8 were used to optimize the LNAA regimes. In this study six specific LNAA regimes were offered to C57Bl/6 PKU mice 14

16 Chapter 1 and compared to a normal diet in C57Bl/6 PKU and WT mice of both sexes. Again, blood and brain LNAA s and neurotransmitter concentrations were examined to investigate the three biochemical consequences. To conclude, in Chapter 10, the results of all chapters will be summarized and discussed. Together with additional data, suggestions are made for future research. 1 15

17 Chapter 1 5. REFERENCES 1 Følling. Über Ausscheidung von Phenylbrenztraubensäure in den Harn als Stoffwechselanomalie in Verbindung mit Imbezillität. Hoppe-Seylers Ztschr Physiol Chem 1934; 227: Penrose. Inheritance of phenylpyruvic amentia (phenylketonuria). Lancet 1935; 2: Buck P. The Child Who Never Grew Blau N, van Spronsen FJ, Levy HL. Phenylketonuria. Lancet 2010; 376: Kanai Y, Segawa H, Miyamoto K i, Uchino H, Takeda E, Endou H. Expression cloning and characterization of a transporter for large neutral amino acids activated by the heavy chain of 4F2 antigen (CD98). J Biol Chem 1998; 273: van Spronsen FJ, Hoeksma M, Reijngoud D-J. Brain dysfunction in phenylketonuria: is phenylalanine toxicity the only possible cause? J Inherit Metab Dis 2009; 32: de Groot MJ, Sijens PE, Reijngoud D-J, Paans AM, van Spronsen FJ. Phenylketonuria: brain phenylalanine concentrations relate inversely to cerebral protein synthesis. J Cereb Blood Flow Metab 2015; 35: Hommes FA, Lee JS. The control of 5-hydroxytryptamine and dopamine synthesis in the brain: a theoretical approach. J Inherit Metab Dis 1990; 13: Ogawa S, Ichinose H. Effect of metals and phenylalanine on the activity of human tryptophan hydroxylase-2: comparison with that on tyrosine hydroxylase activity. Neurosci Lett 2006; 401: Mazumder MK, Paul R, Borah A. β-phenethylamine--a phenylalanine derivative in brain--contributes to oxidative stress by inhibiting mitochondrial complexes and DT-diaphorase: an in silico study. CNS Neurosci Ther 2013; 19: Rosa AP, Jacques CED, Moraes TB, Wannmacher CMD, Dutra A de M, Dutra- Filho CS. Phenylpyruvic acid decreases glucose-6-phosphate dehydrogenase activity in rat brain. Cell Mol Neurobiol 2012; 32: Ribas GS, Sitta A, Wajner M, Vargas CR. Oxidative stress in phenylketonuria: what is the evidence? Cell Mol Neurobiol 2011; 31: Christ SE, Price MH, Bodner KE, Saville C, Moffitt AJ, Peck D. Morphometric analysis of gray matter integrity in individuals with early-treated phenylketonuria. Mol Genet Metab 2016; 118: Andolina D, Conversi D, Cabib S, Trabalza A, Ventura R, Puglisi-Allegra S et al. 5-Hydroxytryptophan during critical postnatal period improves cognitive performances and promotes dendritic spine maturation in genetic mouse model of phenylketonuria. Int J Neuropsychopharmacol 2011; 14: Liang L, Gu X, Lu L, Li D, Zhang X. Phenylketonuria-related synaptic changes in a BTBR-Pah(enu2) mouse model. Neuroreport 2011; 22: Cordero ME, Trejo M, Colombo M, Aranda V. Histological maturation of the neocortex in phenylketonuric rats. Early Hum Dev 1983; 8:

18 Chapter 1 17 Hörster F, Schwab MA, Sauer SW, Pietz J, Hoffmann GF, Okun JG et al. Phenylalanine reduces synaptic density in mixed cortical cultures from mice. Pediatr Res 2006; 59: Schlegel G, Scholz R, Ullrich K, Santer R, Rune GM. Phenylketonuria: Direct and indirect effects of phenylalanine. Exp Neurol 2016; 281: Zhang Y, Zhang H, Yuan X, Gu X. Differential effects of phenylalanine on Rac1, Cdc42, and RhoA expression and activity in cultured cortical neurons. Pediatr Res 2007; 62: Imperlini E, Orrù S, Corbo C, Daniele A, Salvatore F. Altered brain protein expression profiles are associated with molecular neurological dysfunction in the PKU mouse model. J Neurochem 2014; 129: Horling K, Schlegel G, Schulz S, Vierk R, Ullrich K, Santer R et al. Hippocampal synaptic connectivity in phenylketonuria. Hum Mol Genet doi: /hmg/ ddu Cabib S, Pascucci T, Ventura R, Romano V, Puglisi-Allegra S. The behavioral profile of severe mental retardation in a genetic mouse model of phenylketonuria. Behav Genet 2003; 33: Adler-Abramovich L, Vaks L, Carny O, Trudler D, Magno A, Caflisch A et al. Phenylalanine assembly into toxic fibrils suggests amyloid etiology in phenylketonuria. Nat Chem Biol 2012; 8: ALVORD EC, STEVENSON LD, VOGEL FS, ENGLE RL. Neuropathological findings in phenyl-pyruvic oligophrenia (phenylketonuria). J Neuropathol Exp Neurol 1950; 9: Huttenlocher PR. The neuropathology of phenylketonuria: human and animal studies. Eur J Pediatr 2000; 159 Suppl: S Anderson PJ, Leuzzi V. White matter pathology in phenylketonuria. Mol Genet Metab 2010; 99 Suppl 1: S Hood A, Antenor-Dorsey JA V, Hershey T, Rutlin J, Shimony JS, McKinstry RC et al. White matter integrity and executive abilities in individuals with phenylketonuria. Mol Genet Metab 2013; 109: Anderson PJ, Wood SJ, Francis DE, Coleman L, Warwick L, Casanelia S et al. Neuropsychological functioning in children with early-treated phenylketonuria: impact of white matter abnormalities. Dev Med Child Neurol 2004; 46: Bick U, Ullrich K, Stöber U, Möller H, Schuierer G, Ludolph AC et al. White matter abnormalities in patients with treated hyperphenylalaninaemia: magnetic resonance relaxometry and proton spectroscopy findings. Eur J Pediatr 1993; 152: Leuzzi V, Tosetti M, Montanaro D, Carducci C, Artiola C, Antonozzi I et al. The pathogenesis of the white matter abnormalities in phenylketonuria. A multimodal 3.0 tesla MRI and magnetic resonance spectroscopy (1H MRS) study. J Inherit Metab Dis 2007; 30: Mastrangelo M, Chiarotti F, Berillo L, Caputi C, Carducci C, Di Biasi C et al. The outcome of white matter abnormalities in early treated phenylketonuric patients: A retrospective longitudinal long-term study. Mol Genet Metab doi: /j. ymgme Dyer CA, Kendler A, Philibotte T, Gardiner P, Cruz J, Levy HL. Evidence for central nervous system glial cell plasticity in phenylketonuria. J Neuropathol Exp Neurol 1996; 55:

19 Chapter 1 33 Shefer S, Tint GS, Jean-Guillaume D, Daikhin E, Kendler A, Nguyen LB et al. Is there a relationship between 3-hydroxy- 3-methylglutaryl coenzyme a reductase activity and forebrain pathology in the PKU mouse? J Neurosci Res 2000; 61: Smith I, Knowles J. Behaviour in early treated phenylketonuria: a systematic review. Eur J Pediatr 2000; 159 Suppl: S Hoeksma M, Reijngoud D-J, Pruim J, de Valk HW, Paans AMJ, van Spronsen FJ. Phenylketonuria: High plasma phenylalanine decreases cerebral protein synthesis. Mol Genet Metab 2009; 96: Hughes J V, Johnson TC. The effects of phenylalanine on amino acid metabolism and protein synthesis in brain cells in vitro. J Neurochem 1976; 26: Binek-Singer P, Johnson TC. The effects of chronic hyperphenylalaninaemia on mouse brain protein synthesis can be prevented by other amino acids. Biochem J 1982; 206: WOOLF LI, VULLIAMY DG. Phenylketonuria with a study of the effect upon it of glutamic acid. Arch Dis Child 1951; 26: BICKEL H, GERRARD J, HICKMANS EM. Influence of phenylalanine intake on phenylketonuria. Lancet (London, England) 1953; 265: BICKEL H, GERRARD J, HICKMANS EM. The influence of phenylalanine intake on the chemistry and behaviour of a phenylketonuric child. Acta Paediatr 1954; 43: van Spronsen FJ, van Wegberg AM, Ahring K, Bélanger-Quintana A, Blau N, Bosch AM et al. Key European guidelines for the diagnosis and management of patients with phenylketonuria. Lancet Diabetes Endocrinol doi: /s (16) Moyle JJ, Fox AM, Arthur M, Bynevelt M, Burnett JR. Meta-analysis of neuropsychological symptoms of adolescents and adults with PKU. Neuropsychol Rev 2007; 17: Enns GM, Koch R, Brumm V, Blakely E, Suter R, Jurecki E. Suboptimal outcomes in patients with PKU treated early with diet alone: revisiting the evidence. Mol Genet Metab; 101: Jahja R, van Spronsen FJ, de Sonneville LMJ, van der Meere JJ, Bosch AM, Hollak CEM et al. Social-cognitive functioning and social skills in patients with early treated phenylketonuria: a PKU-COBESO study. J Inherit Metab Dis 2016; 39: Jahja R, Huijbregts SCJ, de Sonneville LMJ, van der Meere JJ, Bosch AM, Hollak CEM et al. Mental health and social functioning in early treated Phenylketonuria: the PKU- COBESO study. Mol Genet Metab 2013; 110 Suppl: S Shedlovsky A, McDonald JD, Symula D, Dove WF. Mouse models of human phenylketonuria. Genetics 1993; 134: van Vliet D, Bruinenberg VM, Mazzola PN, van Faassen MHJR, de Blaauw P, Kema IP et al. Large Neutral Amino Acid Supplementation Exerts Its Effect through Three Synergistic Mechanisms: Proof of Principle in Phenylketonuria Mice. PLoS One 2015; 10: e

20 Chapter 1 48 Wahlsten D, Metten P, Crabbe JC. Survey of 21 inbred mouse strains in two laboratories reveals that BTBR T/+ tf/tf has severely reduced hippocampal commissure and absent corpus callosum. Brain Res 2003; 971: Ding Z, Georgiev P, Thöny B. Administration-route and genderindependent long-term therapeutic correction of phenylketonuria (PKU) in a mouse model by recombinant adenoassociated virus 8 pseudotyped vectormediated gene transfer. Gene Ther 2006; 13: MacPherson P, McGaffigan R, Wahlsten D, Nguyen P V. Impaired fear memory, altered object memory and modified hippocampal synaptic plasticity in split-brain mice. Brain Res 2008; 1210: Jones-Davis DM, Yang M, Rider E, Osbun NC, da Gente GJ, Li J et al. Quantitative trait loci for interhemispheric commissure development and social behaviors in the BTBR T + tf/j mouse model of autism. PLoS One 2013; 8: e Sittig LJ, Carbonetto P, Engel KA, Krauss KS, Barrios-Camacho CM, Palmer AA. Genetic Background Limits Generalizability of Genotype-Phenotype Relationships. Neuron 2016; 91: Miller LR, Marks C, Becker JB, Hurn PD, Chen W-J, Woodruff T et al. Considering sex as a biological variable in preclinical research. FASEB J 2017; 31: Soldin OP, Mattison DR. Sex differences in pharmacokinetics and pharmacodynamics. Clin Pharmacokinet 2009; 48: Sandberg K, Umans JG, Georgetown Consensus Conference Work Group the GCCW. Recommendations concerning the new U.S. National Institutes of Health initiative to balance the sex of cells and animals in preclinical research. FASEB J 2015; 29: Holst SC, Valomon A, Landolt H-P. Sleep Pharmacogenetics: Personalized Sleep-Wake Therapy. Annu Rev Pharmacol Toxicol 2016; 56: Eban-Rothschild A, Rothschild G, Giardino WJ, Jones JR, de Lecea L. VTA dopaminergic neurons regulate ethologically relevant sleep wake behaviors. Nat Neurosci 2016; 19: MacDonald A, Gokmen-Ozel H, van Rijn M, Burgard P. The reality of dietary compliance in the management of phenylketonuria. J Inherit Metab Dis 2010; 33: Wurtman RJ, Cansev M, Sakamoto T, Ulus IH. Administration of docosahexaenoic acid, uridine and choline increases levels of synaptic membranes and dendritic spines in rodent brain. World Rev Nutr Diet 2009; 99: Cansev M, Wurtman RJ. Chronic administration of docosahexaenoic acid or eicosapentaenoic acid, but not arachidonic acid, alone or in combination with uridine, increases brain phosphatide and synaptic protein levels in gerbils. Neuroscience 2007; 148: Cansev M, van Wijk N, Turkyilmaz M, Orhan F, Sijben JWC, Broersen LM. A specific multi-nutrient enriched diet enhances hippocampal cholinergic transmission in aged rats. Neurobiol Aging 2015; 36: Sakamoto T, Cansev M, Wurtman RJ. Oral supplementation with docosahexaenoic acid and uridine-5 -monophosphate increases dendritic spine density in adult gerbil hippocampus. Brain Res 2007; 1182: Pooler AM, Guez DH, Benedictus R, Wurtman RJ. Uridine enhances neurite outgrowth in nerve growth factor-differentiated PC12 [corrected]. Neuroscience 2005; 134:

21 Chapter 1 64 Savelkoul PJM, Janickova H, Kuipers AAM, Hageman RJJ, Kamphuis PJ, Dolezal V et al. A specific multi-nutrient formulation enhances M1 muscarinic acetylcholine receptor responses in vitro. J Neurochem 2012; 120: Scheltens P, Kamphuis PJGH, Verhey FRJ, Olde Rikkert MGM, Wurtman RJ, Wilkinson D et al. Efficacy of a medical food in mild Alzheimer s disease: A randomized, controlled trial. Alzheimers Dement 2010; 6: 1 10.e1. 66 Scheltens P, Twisk JWR, Blesa R, Scarpini E, von Arnim CAF, Bongers A et al. Efficacy of Souvenaid in mild Alzheimer s disease: results from a randomized, controlled trial. J Alzheimers Dis 2012; 31: Pietz J, Kreis R, Rupp A, Mayatepek E, Rating D, Boesch C et al. Large neutral amino acids block phenylalanine transport into brain tissue in patients with phenylketonuria. J Clin Invest 1999; 103: Schindeler S, Ghosh-Jerath S, Thompson S, Rocca A, Joy P, Kemp A et al. The effects of large neutral amino acid supplements in PKU: an MRS and neuropsychological study. Mol Genet Metab 2007; 91: Koch R, Moseley KD, Yano S, Nelson M, Moats RA. Large neutral amino acid therapy and phenylketonuria: a promising approach to treatment. Mol Genet Metab 2003; 79: Yano S, Moseley K, Azen C. Large neutral amino acid supplementation increases melatonin synthesis in phenylketonuria: a new biomarker. J Pediatr 2013; 162: Kalkanoğlu HS, Ahring KK, Sertkaya D, Møller LB, Romstad A, Mikkelsen I et al. Behavioural effects of phenylalaninefree amino acid tablet supplementation in intellectually disabled adults with untreated phenylketonuria. Acta Paediatr 2005; 94:

22 Chapter

23

24 CHAPTER 2 The Behavioral Consequence of Phenylketonuria in Mice Depends on the Genetic Background Vibeke M. Bruinenberg 1, Els van der Goot 1, Danique van Vliet 2, Martijn J. de Groot 2, Priscila N. Mazzola 1,2, M. Rebecca Heiner-Fokkema 3, Martijn van Faassen 3, Francjan J. van Spronsen 2, Eddy A. van der Zee 1 *. 1 Molecular Neurobiology, Groningen Institute for Evolutionary Life Sciences (GELIFES), University of Groningen, Groningen, the Netherlands, 2 Beatrix Children s Hospital, University Medical Center Groningen, Groningen, the Netherlands, 3 Laboratory Medicine, University of Groningen, University Medical Center, Groningen, the Netherlands Front Behav Neurosci Dec 20;10:233. doi: /fnbeh ecollection 2016.

25 Chapter 2 ABSTRACT To unravel the role of gene mutations in the healthy and the diseased state, countless studies have tried to link genotype with phenotype. However, over the years, it became clear that the strain of mice can influence these results. Nevertheless, identical gene mutations in different strains are often still considered equals. An example of this, is the research done in phenylketonuria (PKU), an inheritable metabolic disorder. In this field, a PKU mouse model (either on a BTBR or C57Bl/6 background) is often used to examine underlying mechanisms of the disease and/or new treatment strategies. Both strains have a point mutation in the gene coding for the enzyme phenylalanine hydroxylase which causes toxic concentrations of the amino acid phenylalanine in blood and brain, as found in PKU patients. Although the mutation is identical and therefore assumed to equally affect physiology and behavior in both strains, no studies directly compared the two genetic backgrounds to test this assumption. Therefore, this study compared the BTBR and C57Bl/6 wild-type and PKU mice on PKUrelevant amino acid- and neurotransmitter levels and at a behavioral level. The behavioral paradigms were selected from previous literature on the PKU mouse model and address four domains, namely 1) activity levels, 2) motor performance, 3) anxiety and/or depressionlike behavior, and 4) learning and memory. The results of this study showed comparable biochemical changes in phenylalanine and neurotransmitter concentrations. In contrast, clear differences in behavioral outcome between the strains in all four above-mentioned domains were found, most notably in the learning and memory domain. The outcome in this domain seem to be primarily due to factors inherent to the genetic background of the mouse and much less by differences in PKU-specific biochemical parameters in blood and brain. The difference in behavioral outcome between PKU of both strains emphasizes that the consequence of the PAH mutation is influenced by other factors than Phe levels alone. Therefore, future research should consider these differences when choosing one of the genetic strains to investigate the pathophysiological mechanism underlying PKU-related behavior, especially when combined with new treatment strategies. 24

26 Chapter 2 1. INTRODUCTION Transgenic and knockout/ knock-in mice are used to investigate the consequence of genetic mutations to understand the human biological system, especially in a diseased condition. It is progressively acknowledged that the strain of these mice highly influences the outcome of the gene mutation 1 4. Nevertheless, identical gene mutations in different strains are often still considered equals in various disciplines. A striking example is the mouse model used in the field of phenylketonuria (PKU, OMIM ). PKU is an inheritable metabolic disorder characterized by high concentrations of the amino acid phenylalanine (Phe) in blood and brain caused by mutations in the gene that encodes for the enzyme Phe hydroxylase (PAH, EC ). This mutation results in a loss of catalytic activity of the enzyme and, as a consequence, the conversion of Phe to tyrosine is disrupted. In untreated patients, these raised concentrations of Phe are associated with symptoms such as a severe intellectual disability, disruptions in motor performance, mood swings, anxiety, depression disorders, and epilepsy 5. The mouse model of PKU mimics the PKU patients through a chemically induced point mutation in the gene encoding for the enzyme PAH. Originally, this point mutation was described for the black and tan, brachyury (BTBR) mouse 6. However, the wild-type (WT) mice of the BTBR strain were found to have difficulties with breeding and displayed abnormalities in brain morphology and behavior, thus limiting their suitability for preclinical research Therefore, The BTBR PKU mouse was crossed back on a C57Bl/6JRj (referred to as B6 hereafter) background. As a result, both strains are currently used in PKU studies, often without justification. Without fully understanding the influence of the genetic background on behavior and physiology, notably behavioral results in these PKU studies can be difficult to interpret. 2 Various studies highlight the phenotypical difference in behavior between BTBR and B6 WT, as the BTBR is often used in autism research. For example, novelty induced activity is greater in BTBR WT than B6 WT but decreased in home-cage conditions 11. Furthermore, motor performance of the BTBR WT is inferior to the B6 WT on the rotarod 12,14. However, mixed results are described for the differences between the strains in anxiety-related behavior and learning and memory. For anxiety-related behavior, no differences 13, a reduction 11, and an increase of anxiety-related behavior of BTBR WT 12 compared to B6 WT are reported. Similar contradicting results are described for learning and memory. Some articles show an intact ability to master a short-term or long-term memory task by both backgrounds 11,13. Others report memory deficits in for instance short-term novel object memory in the B6 WT 13, reversal learning in the BTBR 11, and cued and contextual fear conditioning in BTBR 9,15. The deficits found in the BTBR could be restored with an increase in training 15 and cage enrichment 9. These results clearly indicate differences between the BTBR and B6 WT individuals in domains important in PKU research. It highlights that the fundaments on which the PKU genotype is induced are already different. Therefore, it is important to 25

27 Chapter 2 characterize the different strains biochemical profile along with behavioral outcome in order to highlight similarities and, most of all, differences between the strains, providing better translational insight in the use of the PKU mouse model. For this reason, this study aims to directly compare male BTBR and B6 mice in terms of amino acid-and neurotransmitter levels, and behavioral and cognitive performance under identical laboratory conditions (e.g. food regime, housing conditions, and experimental design). We particularly aimed to answer two specific questions: 1) Do WT and PKU mice of the two strains differ from each other? and 2) Do PKU mice differ from WT mice within a strain? To this aim, the mice were behaviorally tested in four PKU relevant domains previously described in the literature for at least one of the strains, namely 1) activity levels, 2) motor performance, 3) anxiety and/or depression-like behavior, and 4) learning and memory. 2. MATERIAL AND METHODS 2.1 Animals Heterozygous mating pairs of either BTBR or B6 mice were bred to obtain male BTBR WT, BTBR PKU, B6 WT, and B6 PKU mice. Original breeding pairs of B6 were obtained from the lab of Prof. Dr. Thöny, University of Zürich, Switzerland and in our hands crossed back every fifth generation with the C57Bl/6JRj (Janvier). The breeding pairs of the BTBR were kindly provided by Prof. Puglisi-Allegra, Sapienza University of Rome, Italy. They obtained the original BTBR-Pahenu2/J parental pairs from Jackson Laboratories (Bar Harbor, ME, USA) 16 All individuals were weaned on postnatal day 28 and tissue obtained at weaning was used to establish genotype with quantitative PCR (forward primer: 5 CCGTCCTGTTGCTGGCTTAC 3, reverse primer: 3 CAGGTGTGTACATGGGCTTAGATC 5, WT probe: CCGAGTCZZLCALTGCA, PKU probe: CCGAGTCZLLCACTGCA, aimed at exon 7 of the PAH gene (Eurogentec, Fremont, USA). Mice were group housed until the start of the experiment in a cage with a paper role and nesting material. Animals were tested around the age of 4-5 months. All mice were housed individually in cages with a paper role and nesting material seven days before the experiment. Mice were handled by the researcher for two minutes on the three consecutive days before the start of the first behavioral paradigm. The reported behavioral paradigms were obtained from two separate cohorts of 10 males for each group. In the first cohort, in chronological order, we examined the open field (OF), long-term novel object recognition (NOR), long-term spatial object recognition (SOR), and the forced swim test (FST). In the second cohort, in chronological order, we examined home-cage activity, the elevated plus maze (EPM), and the balance beam (BB). An overview of the time line is given in figure 1A. During the experiment, animals had ad libitum access to water and normal chow (RMH-B 2181, ABdiets, Phe: 8.7 g/kg food) and were kept on a 12/12 light/dark cycle. All experimental measurements, except for home-cage activity, were performed between Zeitgeber Time 1 (ZT1) and ZT6. All proceedings were carried out in accordance with the 26

28 Chapter 2 recommendation of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (The ARRIVE Guidelines Checklist) and protocols were approved by the Institutional Animal Care and Use Committee of the University of Groningen (Permit No 6731A and 6731D). 2 Figure 1 Overview of behavioral paradigms. (A) Time line of the study. The first cohort started with an open field (OF) and novel object recognition test (NOR), 5 days later followed by a spatial object recognition test (SOR) and three weeks later a Forced swim test (FST). After the FST, the mice were sacrificed. The second cohort was monitored for five days with passive infrared sensors (PIR), subsequently 24 hours later tested in the elevated plus maze (EPM) and 24 hours later on the balance beam (BB). (B) OF and NOR setup. The habituation phase of the NOR was used as OF. For the analysis 27

29 Chapter 2 with ethovision, the arena was divided in three regions; 1) corner, 2) border, and 3) center. The NOR was started 24 hours after the habituation phase, in which the animals could freely explore two identical objects. Again 24 hours later, one object was replaced for a new object. (C) The SOR was performed in two days. The first day the mice were exposed to four trials of six min with a two min break in between. The first trial was a habituation phase and the second to fourth trial the mice could explore three different objects in a specific configuration. 24 hours later, one object was moved to another position. 2.2 Amino acid and neurotransmitter analyses Two hours after the FST, animals were anesthetized with isoflurane. When the hind paw reflex was no longer present, blood samples were taken via heart puncture and collected in heparin tubes (temporarily stored at 4ºC). Brains were removed from the skull and cerebrum was immediately flash frozen. To obtain blood plasma, heparin tubes with the blood samples were centrifuged at 1,500 rcf for ten minutes and the supernatant was taken. Plasma and brain samples were stored at -80ºC until further processing. Amino acid and neurotransmitter analyses were performed as described in van Vliet et al Open field test Activity and anxiety-like behavior were assessed in a square OF test (50x50x35 cm) with a white Plexiglas floor and gray Plexiglas walls with a checkerboard cue on one of the walls. This same arena was used for the NOR and SOR. In all tests with this arena, dim lighting was used (10 lux in the center of the arena). At the start of the trial, the animal was placed in the middle of the OF and left to explore the arena freely for ten minutes. All trials were video recorded and were analyzed at a later time with Ethovision v.11. In this analysis, the arena was divided into a center zone, four border zones, and four corner zones 18. Activity was quantified by the distance moved and anxiety-like behavior was examined by the preference of the animal to seek out the more sheltered zones. 2.4 Novel object recognition NOR was used to examine the ability of the mice to recall a familiar object after a delay of 24 hours. This widely used learning and memory paradigm and the SOR task (see section 2.5 below) are both based on the innate exploratory behavior of mice towards novel stimuli. If mice successfully recall the previous event, they will increase exploration towards the novel stimuli and in the SOR a displaced object 19,20. In figure 1B is shown that the NOR test consists of three phases; the habituation phase, the familiarization phase, and the test phase 20. The OF test was used as the habituation phase. The familiarization phase was performed twenty-four hours after the habituation phase. Within this phase, the mice were allowed to freely explore two identical objects. Twenty-four Hours later, mice were exposed to one previously encountered object and one novel object in the test phase. The objects were randomized over the trials. All phases were ten minutes long and video recorded. The readout parameter of the NOR test was the time spent exploring the novel object compared to 28

30 Chapter 2 the previously encountered object, expressed as the ratio of novel object exploration time to total exploration time of both objects. The exploratory behavior was scored manually with the program ELINE (developed in house). 2.5 Spatial object recognition In the SOR test, mice were tested for their ability to recognize the displaced object after twenty-four hours. Within this test, the displaced object that caused the new configuration of objects was the novel stimulus. SOR was tested five days after the NOR test. Comparable to the NOR test, the SOR test consisted of a habituation phase, a familiarization phase, and a test phase (figure 1C). In the habituation phase, the mice could explore the arena without objects, similar to the previous OF. In the familiarization phases 2-4, the mice were placed in the same arena with three objects with different shapes and different materials (pottery, RVS, and glass) in a specific configuration. Twenty-four hours later, in the test phase, one of the outer objects was displaced (randomized between experimental groups). All trials were six minutes long and video recorded for analysis at a later time point. Between habituation phase and the familiarization phase, the mice were placed in their home cage and the arena was cleaned with 30% ethanol. All trials were manually scored for exploratory behavior with the program ELINE. The discrimination index for NOR was calculated by dividing the exploration time of the novel object by the total exploration time multiplied with 100. In contrast to NOR, in the SOR test it was not possible to keep all objects in the same distance from the walls and from each other in both familiarization and test phases, because the limitation of three objects in a rectangular space. Therefore, a correction was made for possible a priori preference for location or object. This correction was done by calculating the mean time spent on exploring each object in the three training sessions of the first day. These cumulative exploration times were set as a percentage of the overall exploration time. The difference between exploration time in the test trial and mean exploration time in the three training sessions was calculated and expressed as a percentage difference. A positive value depicts an increased preference of the individual towards the object in retrospect of possible innate preference at for hand Forced swim test The original FST 21 used to test antidepressants, consists of a pre-test session of fifteen minutes and a test session of five minutes twenty-four hours later. The now commonly used FST to assess depressive-like behavior without intervention consists of one trial 22 and was used in this study. Mice were placed in a 5,000 ml cylinder containing 2,000 ml of 27 C tap water for six minutes. All trials were recorded with a video camera positioned in front of the cylinder. These videos were analyzed with ELINE and scored for the behaviors swimming, struggling, and floating. Swimming was defined as a controlled motion in which the animals used all four paws. Floating behavior was defined as an immobile state with only small movements of one paw to keep balance. Finally, struggling was defined as any other 29

31 Chapter 2 behavior to keep the head above water. This was often done with hasty movements of all four paws towards the side of the cylinder with a vertical body position. Literature shows that an increase in floating behavior is consistent with a more depression-like phenotype 21, Home-cage activity To examine spontaneous activity in a familiar environment, individuals were monitored with a passive infrared (PIR) detector in their home cage. After seven days of individual housing, which included three days of handling, the animals were subjected to five consecutive days of PIR registration starting at ZT1. Starting at midnight, four days (96 hours) were analyzed with ACTOVIEW 23. This program calculates the average activity from the files produced by the activity management system (CAMS) collected from the PIR. 2.6 Balance beam The BB task was used to assess coordination and balance as previously described in Mazzola et al In short, animals were trained over three distances (10, 40, and 75 cm) to cross a square wooden beam (length 1 m, width 5 mm, height 10 mm, horizontally positioned 50 cm above the underlying surface) to their home cage. This is followed by a read-out trial of 100 cm. At the start and between trials, the animals were left in their home cage for one minute. In this study, the performance of the animal was measured by the number of correct steps as a percentage of the total steps necessary to cross the beam. A correct step was defined by the full placement of the hind paw on the beam from the initiation of the step to replacing it on the beam in a forward motion Elevated Plus Maze The second measurement of anxiety-like behavior is the EPM 25,26. This paradigm is also based on the innate exploratory behavior of the mice but this exploratory drive is counteracted by the reluctance of the mice to explore open and raised areas. The apparatus used for this paradigm was a plus-shaped maze with two open and two closed arms connected in the middle with an open center zone. The closed arms were surrounded by walls of 16 cm that were open at the top. The open arms solely had a small ridge (2 mm) surrounding the arms. The length of the arms was 29.5 cm and the center zone was 5x5 cm. The whole apparatus was positioned 50 cm above the ground. Low lighting conditions were used (30 lux open arms). At the start of the eight minutes trial, the animals were placed with their head in the middle of the center zone, pointing towards an open arm. The trials were manually scored for the following parameters: number of open arm entries and closed arm entries, and the time spent in open arms and closed arms. The maze was divided into three zones: open arms, closed arms and center. Entering into a new zone with four paws was seen as an entry and time recordings were taken from this point. A percentual score for each zone was calculated by dividing the time spent in a certain area by the total time. 30

32 Chapter Statistical analysis A one-way ANOVA was used in the statistical analysis of the data. When the one-way ANOVA resulted in a p-value 0.05, a Bonferroni posthoc test was performed. As the research question in the introduction did not focus on the comparison between PKU individuals and the opposing WT group, the statistical outcome of these comparisons (BTBR WT vs B6 PKU and B6 WT vs BTBR PKU) will not be discussed. To further investigate the learning and memory tasks, the discrimination index of each group was tested with a paired Student s t-test. A statistical significant difference was defined as p Values two standard deviations outside the mean were viewed as outliers and discarded from the analysis. 3. RESULTS Phe measurements in blood and brain. To examine if the identical point mutation in two different strains resulted in similar amino acid levels under the same food regimes, amino acid measurements were performed in blood and brain (figure 1A-B). Phe levels differed between the groups in blood (F(3,17)= , p<0.001) and brain (F(3,18)= , p<0.001). In blood, BTBR PKU showed Phe concentrations of ±82.3 µmol/l, a 6.3-fold increase compared to the BTBR WT (Phe 200.8±82.3 µmol/l, p<0.001). In B6 PKU, a Phe concentration of ±313.2 µmol/l was observed, which was a 26.2-fold rise compared to WT littermates (Phe 62.3±7.7 µmol/l, p<0.001). In respect to the blood Phe concentrations, brain Phe content in BTBR PKU mice were 699.5±46.9 nmol/g wet weight and in B6 PKU 666.8±65.4 nmol/g wet weight, resulting in respectively a 3.4-fold and 5.1-fold increase compared to WT littermates (BTBR WT 205.6±13.3 nmol/g wet weight, B6 WT 128.5±12.8 nmol/g wet weight, p<0.001 for both). A comparison between strains showed that B6 PKU individuals had higher levels of Phe in blood than BTBR (p=0.036), what was also found in the brain Phe levels between WT individuals (p=0.034). Additional amino acids measurements in blood and brain are provided as supplementary material. In blood of the B6 strain, additional genotype differences were found in blood valine (p=0.007), isoleucine (p=0.001) and leucine (p=0.002). In the BTBR strain, only differences were found in tyrosine (p<0.001). The altered blood amino acid concentrations did not translate to similar changes in brain amino acid concentration. In brain, only tyrosine levels were reduced in B6 PKU compared to B6 WT (p=0.001). 31

33 Chapter 2 Figure 2 Phenylalanine concentrations in blood and brain. (A) Phenylalanine concentrations in blood (µmol/l) (n=4-6), and (B) Phenylalanine content in the brain (nmol/g) (n=5-6) of BTBR and C57Bl/6 (B6) wild-type (WT) and phenylketonuria (PKU) mice. * p 0.05; mean±sem 3.2 Monoaminergic neurotransmitters in brain To further investigate the consequence on the biochemical level of the identical point mutation in both PKU strains, monoaminergic neurotransmitters together with the associated metabolite content were examined in the brain (Fig. 3A-C). First, in figure 3A, the catecholamine dopamine did not significantly differ between groups (F(3,19)=1.094, p=0.376). Second, further downstream the catecholamine pathway, norepinephrine (NE) did show differences between the groups (figure 3B; F(3,19)=20.384, p<0.001). Norepinephrine levels were reduced to 53% in B6 PKU compared to WT littermates (p<0.001) and a trend towards a reduction of 63% was observed in BTBR PKU (p=0.061). B6 WT showed a higher norepinephrine content compared to BTBR WT (p=0.002). Finally, in figure 2C, serotonin levels showed differences between the groups (F(3,19)=14.053, p<0.001) in which BTBR PKU showed a 57% reduction (p=0.012) and B6 PKU a 50% reduction (p<0.001). No significant differences were observed between the WTs and PKUs of both strains (both: p=1.000). 3.3 Activity The behavioral phenotype was examined in four domains, starting with locomotor activity that was assessed in three experimental test conditions. First, home-cage activity measurements were taken to assess baseline locomotor activity (figure 4A). A significant difference was observed between the WT of both strains (F(3,33)=10.185, p<0.001, BTBR WT vs B6 WT, p<0.001) and the PKU and WT individuals of the BTBR (p=0.017). Such a difference was not observed in the B6 strain (p=1.000). Second, the distance moved in an open field was measured to assess novelty-induced locomotion (figure 4B). In this novel environment, a significant difference was found between strains (F(3,34)= , p<0.001, BTBR WT vs BTBR B6 p=0.006, BTBR PKU vs B6 PKU, p<0.001) in which the B6 moved a greater distance. Furthermore, in contrast to the home-cage activity measurements, no significant differences were found 32

34 Chapter 2 in genotype (BTBR: p=1.000, B6: p=1.000). Finally, entries in the elevated plus maze were used to examine novelty-induced locomotion in a different arena. The results did not show significant differences between the groups (F(3,36)=0.916, p=0.443). 2 Figure 3 Neurotransmitter analyses. (A) Dopamine (B) Norepinephrine (C) Serotonin (n=6 for all groups except for BTBR PKU n=5). * p 0.05; mean±sem Figure 4 Activity. (A) Distance moved in an open field (n=9-10). (B) Home-cage activity measured by PIR (n=9-10). * p 0.05; mean±sem 33

35 Chapter Motor performance In figure 5 it is clear that the motor performance, assessed by the percentage of correct steps in the read-out trial, differed significantly (F(3,36)=16.479, p<0.001 respectively). In both strains, PKU individuals showed a lower percentage of correct steps compared to WT individuals (BTBR: p=0.001, B6 p<0.001). No significant differences were observed between strains (WTs p=1.000, PKUs p=1.000). Figure 5 Motor performance. The number of correct steps in the probe trail is depicted as a percentage of the total steps necessary to cross the beam (n=9-10). * p 0.05; mean±sem 3.5 Anxiety and depressive-like behavior Anxiety and depressive-like behavior were examined by the use of the OF, the EPM, and the FST (Fig.6A-D). In the OF, time spent in the corners significantly differed between groups (F(3,34)=9.164, p<0.001). B6 PKU significantly spent more time in the corners compared to the BTBR PKU (p<0.001). Particularly, the B6 PKU individuals spent approximately 50% of the time in the corners which was more than their WT littermates (p=0.024). In addition, the EPM showed a difference in the percentage of time spent in the closed arms (F(3,34)=10.302, p<0.001). This significant difference was found between the strains (WTs p=0.001, PKU s p=0.004) but not between the PKU and WT individuals of each strain (BTBR: p=1.000, B6: p=1.000). Finally, the FST used to examine depressive-like behavior, showed significant differences in the amount of time spent on floating (F(3,34)=6.155, p=0.002), struggling (F(3,34)=11.092; p<0.001), and swimming (F(3,34)=3.634, p=0.022). Floating and struggling differed only between B6 PKU and B6 WT but not between BTBR PKU and BTBR WT (floating: p=0.003, p=1.000, struggling; p=0.002, p=0.140, respectively). In swimming, only a significant reduction of swimming behavior was observed in BTBR PKU individuals compared to BTBR WT (BTBR: p=0.047, B6 p=0.221). 34

36 Chapter 2 2 Figure 6 Anxiety- and depressive-like behavior. (A) The time spent in the corners of the open field, the most sheltered areas of the arena. The habituation phase of the NOR was used as open field (n=9-10). (B) Time spent in the closed arms of the elevated plus maze (n=10). (C) Floating behavior in the forced swim test (n=9-10) (D) Struggling behavior in the forced swim test (n=9-10). * p 0.05; mean±sem 3.6 Learning and memory Learning and memory were assessed by the NOR and SOR tests (Fig7A,B). The performance on the discrimination index of the novel object and the displaced object showed a trend between groups (NOR: F(3,36)=2.818, p=0.053, SOR: (F(3,34)=2.775, p=0.056). To test whether each group learned the task, a paired t-test was used to examine if either the novel object differed from the same object or the displaced object from the non-displaced object. In NOR, the B6 WT (t(9)=-2.265, p=0.050) and B6 PKU (t(8)=2.762, p=0.030) learned the task. Within the BTBR WT a trend was observed (t(9)=-2.153, p=0.060). The PKU 35

37 Chapter 2 BTBR did not learn the task (t(9)=0.986, df=9, p=0.350). In SOR, BTBR WT (t(9)=2.335, p=0.044), B6 WT (t(8)=3.076, p=0.015), and B6 PKU (t(8)=2.762, p=0.025) learned the task. Again, PKU BTBR mice did not learn the task (t=-0.314, df=9, p=0.761). No significant differences were found in the exploration of the objects in the NOR test (F(3,36)=1.093, p=0.364). In the test session of the SOR, the BTBR mice explored the objects overall more than the B6 mice (F(3,36)=5.463, p=0.003, BTBR compared to B6 p<0.001). Figure 7 Learning and memory. (A) Discrimination index of NOR (n=10) (B) Discrimination index of SOR (n=8-10). * p 0.05, ~ p=0.06; mean±sem 4. DISCUSSION Here, we directly compared the two strains of the PKU mouse model in behavioral domains previously described for at least one of the two strains in literature (activity levels, motor performance, anxiety and/or depression-like behavior, and learning and memory) and PKUrelated biochemical parameters in the same laboratory using the same experimental settings. Distinct differences in behavioral outcome between the two strains were found in all four above-mentioned domains, regardless of comparable biochemical changes in amino acid and neurotransmitter content. The result of the mutation in the PAH enzyme on behavior was most pronounced in the PKU BTBR mice, revealing changes in home-cage activity, reduced motor performance, and learning and memory deficits. In contrast, compared to B6 WT mice, PKU B6 mice only showed reduced motor performance and indications of differences in anxiety-like behavior. Therefore, differences in the phenotypical outcome of the BTBR and B6 PKU mouse model seem to be primarily due to factors inherent to the genetic background of the mouse and much less to differences in biochemical parameters in blood and brain that are typically described for PKU pathology. 36

38 Chapter Learning and memory are intact in B6 PKU On a behavioral level, the cognitive outcome is consistently used to assess the severity of the PKU pathology and effectiveness of (new) treatments. This study confirms the cognitive deficits described for BTBR PKU mice in the literature. However, we show for the first time that B6 PKU mice can master a learning and memory paradigm despite severe disruptions in amino acid and neurotransmitter content in the brain. The differences in phenotypical behavior could lay in a different ability to understand the cognitive demands of a learning and memory paradigm of the genetic background. In the literature, a direct comparison between BTBR WT and B6 WT shows mixed results. As mentioned in the introduction, some articles show an intact ability to master a short-term or long-term memory task by both strains 11,13. In contrast, memory deficits in short-term novel object memory in the B6 13, reversal learning in the BTBR 11, and cued and contextual fear conditioning in BTBR 9,15 have been reported. The deficits found in BTBR could be restored with an increase in training 15 and cage enrichment 9. In our study, difficulties to master the learning and memory paradigms by BTBR WT were also found for the NOR but not for the SOR. This outcome could have been influenced by the order of testing (the NOR always preceded the SOR), or the number of training sessions (a single training phase in NOR, versus three training phases in SOR). Our results together with literature suggest that BTBR WT mice have more difficulties mastering a learning and memory paradigm compared to B6 WT mice. It is up for discussion whether and to what degree the PKU behavioral phenotype in the BTBR PKU is a result of weakening a poor learner, creating a deficit, or that the strong learner (B6 WT) is able to compensate Can PKU-related biochemical changes affect both strains differently? Both strains showed a vast increase in Phe levels and disrupted serotonin and norepinephrine levels in the brain (for norepinephrine a trend was observed for BTBR statistically) which is in accordance with previous studies 17, How these PKU-related changes can result in a different functional outcome is not clear. Concerning raised Phe concentrations, in in vitro models of PKU, increased Phe concentrations seems to affect post- and presynaptic markers, proteins involved in cytoskeleton organization, and neuronal morphology In vivo, although both strains show changes in different markers related to synaptic functioning, both BTBR PKU and B6 PKU show affected synaptic plasticity and overall neuronal functioning 29,31,36,37. Differences are found between BTBR WT and B6 WT in adult neurogenesis and neurodevelopmental markers but not in the synaptic markers synaptophysin and postsynaptic density protein (PSD-95) that are discussed in PKU literature 38. The reduced neurogenesis together with altered neurodevelopmental markers in BTBR WT indicates that BTBR could have a different brain development compared to B6 WT 38. As changes in Phe and neurotransmitters are chronically present at a very early age, we assume that neurodevelopment differs between BTBR PKU and B6 PKU. An indication that this is the case is given by the work of Andolina et al In their study they could improve dendritic spine maturation and performance in a short-term version of the NOR and 37

39 Chapter 2 SOR tests with a seven-day treatment (PND-14-21) with 5-hydroxytryptophan, a precursor of serotonin 29. It is not clear if B6 mice only have a different neonatal development or that they also have different susceptibility towards neurotransmitter depletion in life. Some indications present in literature suggest that B6 can differ from BTBR in their response to neurotransmitter manipulations. For instance, a comparison of acute tryptophan depletion, a method to deplete serotonin, both B6 and BTBR WT s show a decrease in serotonin but the tryptophan depletion only altered social interaction and social novelty behaviors in B6 and not in BTBR 39. Furthermore, administration of the serotonin agonist 3-chlorophenylpiperazine (CCP) in WT B6 has not affected locomotor activity nor performance of the B6 in learning and memory paradigm 40. Finally, in a conditional knock-out mouse maintained on a B6 genetic background, a significant decrease in serotonin and norepinephrine concentrations did not affect locomotor activity, motor performance and anxiety- and depression-related behaviors 41. These examples highlight that in addition to the differences found between BTBR WT and B6 WT in neurodevelopment, the consequence of neurotransmitter depletion in later life (without differences in development) could be different between BTBR and B6. Therefore, we hypothesize that the PKU-related changes in Phe and neurotransmitters affect both strains differently during neurodevelopment and later in life resulting in a difference in phenotypical behavior. 4.3 Considerations in testing behavior In this study, we identified differences between PKU of the B6 background and the PKU of the BTBR background. As all domains seem to be affected to some extent, an important consideration is the possibility that altered behavior in one domain can influence the outcome nonspecifically in another behavioral paradigm. For example, deficits found in motor performance in the BB of both genetic backgrounds could influence the outcome in the OF, EPM, SOR, NOR, and FST. However, we did not clearly find an influence of motor performance in behavioral paradigms with a low level of required motor performance (OF, EPM, NOR, SOR) as both strains did not reduce activity in the OF and the EPM, and no differences were found in exploration of the objects in the NOR. Therefore, the mice were not hampered to fulfill the key feature of the task. However, the task that required most motor skills, i.e. the FST, could be affected by the motor problems found in PKU. Accordingly, we observed that the PKU mice from both backgrounds showed difficulties in maintaining a floating position and correct swimming behavior during the FST. As a result, we believe that the changes found in the FST are primarily attributed to deficits in motor performance. Therefore, we conclude that the FST in PKU mice is not well suited for examining depressive-like behavior in PKU mice and conclusions in this domain should, therefore, be drawn with caution. 38

40 Chapter The translational value of the PKU mouse model The differences in behavioral outcome between PKU mice of both strains emphasizes that the consequences of the PAH mutation are influenced by other factors than Phe levels alone. Although the underlying mechanisms may be different, B6 PKU mice may resemble the human situation where some specific untreated PKU patients with high blood Phe concentrations have clearly escaped from the severe symptoms of PKU 42. To conclude, this study showed clear differences in PKU behavioral phenotype between BTBR and B6 mice despite similar biochemical phenotype. It contributes to a better translational insight in the use of the PKU mouse model. Future research should consider these differences when choosing one of the genetic strains to investigate the underlying mechanisms of PKU and/or new treatment targets. As the origin of BTBR and B6 PKU strains between labs may differ (or the frequency of backcrossing), our results also stress the need for a better genetic understanding of these strains used by research groups worldwide. Nevertheless, we would like to emphasize that both PKU strains have their own translational value for studying PKU and developing novel interventional strategies to battle the burden of the disease, as they may represent different patient populations. 2 SUPPLEMENTAL DATA Table 1. Amino acid concentrations in blood (µmol/l). Each column depicts the mean concentration of the amino acids ± standard deviation. The greek symbols are used to highlight significant differences; α= a difference between BTBR WT and BTBR PKU, β= a difference between B6 WT and B6 PKU, γ= a difference between BTBR PKU and B6 PKU (n=4-6). BTBR B6 WT PKU WT PKU Phenylalanine ± ± 87.5 α 58.3 ± ± β,γ Tyrosine 65.4 ± ± 10.9 α 53.0 ± ± 19.8 Valine ± ± ± ± 70.2 β Isoleucine 89.8 ± ± ± ± 15.4 β Leucine ± ± ± ± 27.5 β Histidine 62.8 ± ± ± ± 19.4 Threonine ± ± ± ±

41 Chapter 2 Table 2. Amino acid concentrations in brain (nmol/g). Each column depicts the mean concentration of the amino acids ± standard deviation. The greek symbols are used to highlight significant differences; α= a difference between BTBR WT and BTBR PKU, β= a difference between B6 WT and B6 PKU, γ= a difference between BTBR PKU and B6 PKU, δ= a difference between BTBR WT and B6 WT (n=5-6). BTBR B6 WT PKU WT PKU Phenylalanine ± ± 46.9 β ± 12.8 δ ± 65.4 β Tyrosine ± ± ± ± 1.5 β Valine 84.4 ± ± ± ± 10.5 γ Isoleucine 53.0 ± ± ± ± 5.1 Leucine ± ± ± ± 8.0 Histidine 90.4 ± ± ± ± 5.3 γ Threonine ± ± ± ± 32.2 Methionine 80.0 ± ± ± ± 8.2 γ Tryptophan 10.2 ± ± ± ±

42 Chapter 2 5. REFERENCES 1 Holmes A, Wrenn CC, Harris AP, Thayer KE, Crawley JN. Behavioral profiles of inbred strains on novel olfactory, spatial and emotional tests for reference memory in mice. Genes Brain Behav 2002; 1: Sittig LJ, Carbonetto P, Engel KA, Krauss KS, Barrios-Camacho CM, Palmer AA. Genetic Background Limits Generalizability of Genotype-Phenotype Relationships. Neuron 2016; 91: Alam I, McQueen AK, Acton D, Reilly AM, Gerard-O Riley RL, Oakes DK et al. Phenotypic severity of autosomal dominant osteopetrosis type II (ADO2) mice on different genetic backgrounds recapitulates the features of human disease. Bone 2017; 94: Doetschman T. Influence of Genetic Background on Genetically Engineered Mouse Phenotypes. 2009, pp Blau N, van Spronsen FJ, Levy HL. Phenylketonuria. Lancet 2010; 376: Shedlovsky A, McDonald JD, Symula D, Dove WF. Mouse models of human phenylketonuria. Genetics 1993; 134: Wahlsten D, Metten P, Crabbe JC. Survey of 21 inbred mouse strains in two laboratories reveals that BTBR T/+ tf/tf has severely reduced hippocampal commissure and absent corpus callosum. Brain Res 2003; 971: Jones-Davis DM, Yang M, Rider E, Osbun NC, da Gente GJ, Li J et al. Quantitative trait loci for interhemispheric commissure development and social behaviors in the BTBR T + tf/j mouse model of autism. PLoS One 2013; 8: e MacPherson P, McGaffigan R, Wahlsten D, Nguyen P V. Impaired fear memory, altered object memory and modified hippocampal synaptic plasticity in split-brain mice. Brain Res 2008; 1210: Ding Z, Georgiev P, Thöny B. Administration-route and genderindependent long-term therapeutic correction of phenylketonuria (PKU) in a mouse model by recombinant adenoassociated virus 8 pseudotyped vectormediated gene transfer. Gene Ther 2006; 13: Molenhuis RT, de Visser L, Bruining H, Kas MJ. Enhancing the value of psychiatric mouse models; differential expression of developmental behavioral and cognitive profiles in four inbred strains of mice. Eur Neuropsychopharmacol 2014; 24: Moy SS, Nadler JJ, Young NB, Perez A, Holloway LP, Barbaro RP et al. Mouse behavioral tasks relevant to autism: phenotypes of 10 inbred strains. Behav Brain Res 2007; 176: Cabib S, Pascucci T, Ventura R, Romano V, Puglisi-Allegra S. The behavioral profile of severe mental retardation in a genetic mouse model of phenylketonuria. Behav Genet 2003; 33: Nadler JJ, Zou F, Huang H, Moy SS, Lauder J, Crawley JN et al. Large-scale gene expression differences across brain regions and inbred strains correlate with a behavioral phenotype. Genetics 2006; 174: Stapley NW, Guariglia SR, Chadman KK. Cued and contextual fear conditioning in BTBR mice is improved with training or atomoxetine. Neurosci Lett 2013; 549:

43 Chapter 2 16 Puglisi-Allegra S, Cabib S, Pascucci T, Ventura R, Cali F, Romano V. Dramatic brain aminergic deficit in a genetic mouse model of phenylketonuria. Neuroreport 2000; 11: van Vliet D, Bruinenberg VM, Mazzola PN, van Faassen MHJR, de Blaauw P, Kema IP et al. Large Neutral Amino Acid Supplementation Exerts Its Effect through Three Synergistic Mechanisms: Proof of Principle in Phenylketonuria Mice. PLoS One 2015; 10: e Hovens IB, Schoemaker RG, van der Zee EA, Absalom AR, Heineman E, van Leeuwen BL. Postoperative cognitive dysfunction: Involvement of neuroinflammation and neuronal functioning. Brain Behav Immun 2014; 38: Ennaceur A, Delacour J. A new one-trial test for neurobiological studies of memory in rats. 1: Behavioral data. Behav Brain Res 1988; 31: Antunes M, Biala G. The novel object recognition memory: neurobiology, test procedure, and its modifications. Cogn Process 2012; 13: Porsolt RD, Anton G, Blavet N, Jalfre M. Behavioural despair in rats: a new model sensitive to antidepressant treatments. Eur J Pharmacol 1978; 47: Bogdanova O V, Kanekar S, D Anci KE, Renshaw PF. Factors influencing behavior in the forced swim test. Physiol Behav 2013; 118: Mulder C, Van Der Zee EA, Hut RA, Gerkema MP. Time-Place Learning and Memory Persist in Mice Lacking Functional Per1 and Per2 Clock Genes. J Biol Rhythms 2013; 28: Mazzola PN, Bruinenberg V, Anjema K, van Vliet D, Dutra-Filho CS, van Spronsen FJ et al. Voluntary Exercise Prevents Oxidative Stress in the Brain of Phenylketonuria Mice. JIMD Rep doi: /8904_2015_ Calabrese EJ. An assessment of anxiolytic drug screening tests: hormetic dose responses predominate. Crit Rev Toxicol 2008; 38: Belzung C, Griebel G. Measuring normal and pathological anxiety-like behaviour in mice: a review. Behav Brain Res 2001; 125: Sawin EA, Murali SG, Ney DM. Differential effects of low-phenylalanine protein sources on brain neurotransmitters and behavior in C57Bl/6-Pahenu2 mice. Mol Genet Metab 2014; 111: Pascucci T, Giacovazzo G, Andolina D, Accoto A, Fiori E, Ventura R et al. Behavioral and neurochemical characterization of new mouse model of hyperphenylalaninemia. PLoS One 2013; 8: e Andolina D, Conversi D, Cabib S, Trabalza A, Ventura R, Puglisi-Allegra S et al. 5-Hydroxytryptophan during critical postnatal period improves cognitive performances and promotes dendritic spine maturation in genetic mouse model of phenylketonuria. Int J Neuropsychopharmacol 2011; 14: Hörster F, Schwab MA, Sauer SW, Pietz J, Hoffmann GF, Okun JG et al. Phenylalanine reduces synaptic density in mixed cortical cultures from mice. Pediatr Res 2006; 59: Horling K, Schlegel G, Schulz S, Vierk R, Ullrich K, Santer R et al. Hippocampal synaptic connectivity in phenylketonuria. Hum Mol Genet doi: /hmg/ ddu

44 Chapter 2 32 Zhang Y, Zhang H, Yuan X, Gu X. Differential effects of phenylalanine on Rac1, Cdc42, and RhoA expression and activity in cultured cortical neurons. Pediatr Res 2007; 62: Schlegel G, Scholz R, Ullrich K, Santer R, Rune GM. Phenylketonuria: Direct and indirect effects of phenylalanine. Exp Neurol 2016; 281: Zhang H, Gu XF. A study of gene expression profiles of cultured embryonic rat neurons induced by phenylalanine. Metab Brain Dis 2005; 20: Li D, Gu X, Lu L, Liang L. Effects of phenylalanine on the survival and neurite outgrowth of rat cortical neurons in primary cultures: possible involvement of brain-derived neurotrophic factor. Mol Cell Biochem 2010; 339: Liang L, Gu X, Lu L, Li D, Zhang X. Phenylketonuria-related synaptic changes in a BTBR-Pah(enu2) mouse model. Neuroreport 2011; 22: Bruinenberg VM, van Vliet D, Attali A, de Wilde MC, Kuhn M, van Spronsen FJ et al. A Specific Nutrient Combination Attenuates the Reduced Expression of PSD-95 in the Proximal Dendrites of Hippocampal Cell Body Layers in a Mouse Model of Phenylketonuria. Nutrients 2016; 8. doi: /nu Stephenson DT, O Neill SM, Narayan S, Tiwari A, Arnold E, Samaroo HD et al. Histopathologic characterization of the BTBR mouse model of autistic-like behavior reveals selective changes in neurodevelopmental proteins and adult hippocampal neurogenesis. Mol Autism 2011; 2: Zhang WQ, Smolik CM, Barba-Escobedo PA, Gamez M, Sanchez JJ, Javors MA et al. Acute dietary tryptophan manipulation differentially alters social behavior, brain serotonin and plasma corticosterone in three inbred mouse strains. Neuropharmacology 2015; 90: Vetulani J, Sansone M, Bednarczyk B, Hano J. Different effects of 3-chlorophenylpiperazine on locomotor activity and acquisition of conditioned avoidance response in different strains of mice. Naunyn Schmiedebergs Arch Pharmacol 1982; 319: Isingrini E, Perret L, Rainer Q, Sagueby S, Moquin L, Gratton A et al. Selective genetic disruption of dopaminergic, serotonergic and noradrenergic neurotransmission: insights into motor, emotional and addictive behaviour. J Psychiatry Neurosci 2016; 41: Ramus SJ, Forrest SM, Pitt DB, Saleeba JA, Cotton RG. Comparison of genotype and intellectual phenotype in untreated PKU patients. J Med Genet 1993; 30:

45

46 CHAPTER 3 The behavioral phenotype of female phenylketonuria mice differs partially from male phenylketonuria mice Bruinenberg VM 1, van der Goot E 1, van Vliet D 2, de Groot M 2, Mazzola PN 1,2, van Spronsen FJ 2, van der Zee EA 1 1 Molecular Neurobiology, Groningen Institute for Evolutionary Life Sciences (GELIFES), University of Groningen, Groningen, the Netherlands, 2 Beatrix Children s Hospital, University Medical Center Groningen, Groningen, the Netherlands.

47 Chapter 3 ABSTRACT Introduction The use of both sexes in preclinical and clinical research is actively advocated in science. However, the implementation of this recommendation is often lacking. In the preclinical research of phenylketonuria (PKU) the inclusion of both sexes is limited. The heritable metabolic disorder PKU is caused by a mutation in the gene for phenylalanine hydroxylase, a key enzyme in the conversion of phenylalanine (Phe) to tyrosine. The consequent rise in Phe concentrations in blood and brain has detrimental effects in the brain, for example in neurotransmitter metabolism. Our recent study showed differences in behavioral outcome between the two genetic backgrounds of the PKU mouse model used in preclinical research 1. As that study only included male mice, the aim of the present study was to assess the behavioral phenotype of female PKU mice of both strains in the same behavioral paradigms; 1) activity, 2) motor performance, 3) anxiety-like behavior, and 4) learning and memory. Materials & Methods Female wild type (WT) and PKU mice of each strain were tested in two cohorts (BTBR WT, BTBR PKU, B6 WT, and B6 PKU, n=10 per group/per cohort). In the first cohort, the behavior of the mice was assessed in the open field (OF), long-term novel object recognition (NOR), and long-term spatial object recognition (SOR). In the second cohort, the mice were tested in home-cage activity, the elevated plus maze (EPM), and the balance beam (BB). Results 1) For activity, a PKU phenotype was seen in the BTBR strain only in the OF (p<0.001). A PKU phenotype was lacking in other activity measurements. 2) In motor performance, a similar deficit was found in the BTBR and B6 strain (both p<0.001). 3) Only in the OF, a preference for sheltered areas was found for the PKU mice of both backgrounds (BTBR: p=0.014, B6: p=0.047). 4) Both the BTBR PKU and the B6 PKU mice did not master the learning and memory paradigms (NOR: BTBR p=0.317, B6=0.516, SOR: BTBR p=0.565, B6 p=0.310) Conclusion/Discussion In contrast to the previously described differences in behavioral phenotype of male PKU mice of both genetic backgrounds, the PKU females of both strains showed a similar PKU phenotype. Together these studies highlight that the outcome of the PKU pathophysiology is influenced not only by genetic background, but also by sex. By including both males and females in PKU research, PKU research into the underlying mechanism of brain dysfunction and new treatment strategies will benefit the PKU population. 46

48 Chapter 3 1. INTRODUCTION Despite an equivalent number of males and females in the human population, neuroscience and biochemical research usually neglects this ratio between the sexes. Most of the literature focused on understanding underlying mechanisms of the healthy and the diseased state is biased towards males 2. Accordingly, the male-biased preclinical research has resulted in skewed data that is unable to be implemented in females 3. This imbalance has, for example, resulted in reports of adverse effects in females and withdrawal of drug treatments from the market 4. To prevent these adverse effects in females and the annulment of years of research, the use of both sexes is increasingly advocated 5. However, in many research fields these recommendations are not yet implemented, including the field of phenylketonuria (PKU). In PKU, the first step of the catabolic pathway of phenylalanine (Phe) is disrupted. The inborn error existing in the converting step of Phe to tyrosine causes the characterizing increase in Phe concentrations in blood and brain. When no restrictions are made in protein intake, patients exhibit severe mental disabilities and developmental delays 6. To understand the underlying mechanism of these detrimental effects, a PKU mouse model has been introduced 7. This PKU mouse model originated from mutagenesis in the black and tan brachyuryblack (BTBR) mouse strain, resulting in a mouse model that had a point mutation in the enzyme Phe hydroxylase (Pah). As a consequence, no catalytic activity of the enzyme is present and Phe concentrations rise in blood and brain. Clear limitations of the BTBR strain in the brain morphology and breeding capacity were the main reason to start back crossing of the original model onto the C57Bl/6 (B6). Recent research has shown a clear difference in the behavioral phenotype in male BTBR Pah enu2 and B6 Pah enu mice 1. However, it is not clear what the behavioral phenotype of the female mice of both strains is. 3 In studies concerning the PKU mouse model, limited studies investigate both sexes or females In some of these studies sex differences are described for behavioral response, biochemical consequence and/or treatment response. For example, while no sex differences were found in the marble burying test, vertical activity in the open field had a clear PKU phenotype in female B6 but not in male B6 9. In addition, the same study showed sex differences in the neurotransmitters (e.g. norepinephrine, epinephrine, the metabolite of dopamine 3,4-dihydroxyphenylacetic acid (DOPAC), and serotonin) and response to glycomacropeptide or amino acid supplementation 9. Furthermore, one of the first studies on gene therapy has shown sex difference in dose response and duration of adeno-associated virus-treatment in the ability to restore PAH functioning 8. These studies give a first indication that sex can influence the phenotypical outcome of PKU. However, a study examining an array of PKU-related behavioral paradigms in female PKU mice of both genetic backgrounds is lacking. Therefore, the aim of this study was to assess the behavioral phenotype of female PKU mice. From our own work, it is clear that the genetic background (BTBR or B6) of 47

49 Chapter 3 the PKU mouse model can influence the behavioral phenotype 1. Therefore, female mice of both genetic backgrounds are assed in the identical behavioral paradigms of this study 1. The behavioral paradigms can be subdivide in four behavioral domains: 1) activity, 2) motor performance, 3) anxiety-like behavior, and 4) learning and memory. 2. MATERIAL AND METHODS 2.1 Animal and experimental design The experimental design of this study was identical to the study described in Bruinenberg et al. (2016) 1. The fundamental steps of this design will be discussed briefly. Two cohorts of female wild-type (WT) and PKU mice of each strain (BTBR WT, BTBR PKU, B6 WT, and B6 PKU, n=10 per group/per cohort) were derived from heterozygous mating pairs. The offspring was weaned at post-natal 28 and group housed until the start of the experiment at the age of 4-5 months. The mice were subjected to a battery of behavioral paradigms. In the first cohort, the mice were tested in the open field (OF), long-term novel object recognition (NOR), and long-term spatial object recognition (SOR). In the second cohort, home-cage activity, the elevated plus maze (EPM), and the balance beam (BB) were performed. Seven days before the start of each cohort, the mice were individually housed and handled three times in the days preceding the start of the first test. Animals were kept on a 12/12 light/ dark cycle with unlimited access to food (RMH-B 2181, AB Diets, Phe: 8.7 g/kg) and water. The behavioral tests were performed between zeitgeber (ZT) 1 and 6. All proceedings were approved by the Institutional Animal Care and Use Committee of the University of Groningen. 2.2 Behavioral paradigms of cohort 1 The first cohort was subjected to the OF, NOR, and SOR. For these tests an arena (50x50x35 cm) with grey sides and a white floor was used. On the first day, the animals were introduced to this arena for 10 minutes to explore. This phase used as the OF test. The second day, two identical glass objects were placed in the arena. Again, the mice could explore the apparatus for 10 minutes. The third day, one of these objects was replaced by a novel glass object (the replaced object was randomized over the mice). Hereafter, the mice were not tested for five days. On day nine, the mice had four sessions of six minutes in the arena for the SOR: (1) habituation phase, (2-4) acquisition phases wherein a specific configuration of three different objects was presented (configuration was randomized over the mice). On day ten, this configuration was changed. All behavioral paradigms were video recorded. In the OF test, distance moved through the maze and time spend in certain locations of the maze (center zone, border zone, and corner zone (Figure 3A) 14 ) were analyzed with Ethovision v.11. The NOR and the SOR were analyzed by hand with the program ELINE (made in house). With this program, the percentage of time spending exploring an object can be registered. For the NOR, the difference in time between the time exploring the novel object to familiar object 48

50 Chapter 3 was divided by the total time exploring the objects. For the SOR, the exploration time of the three habituation phases were set as zero by dividing the time exploring the located object by the percentual exploring time of the same object in the habituation phase. 2.3 Behavioral paradigms of cohort 2 In the second cohort, the mice were subjected to three measurements, in chronological order; home-cage activity, elevated plus maze, and balance beam. Similar to cohort 1, the mice were individually housed seven days before the start of the experiment, including three days of handling, in cages that where placed underneath a passive infrared detector. After these seven days, a five-day recording period was started. In de final analysis, the 96 hours started from midnight on the first day was used for analysis with ACTOVIEW (made in house 15 ). After home-cage activity measurements, the animals were subjected to an elevated plus maze. The mice were placed in this plus-shaped maze, consisting of two open and two closed arms, for eight minutes. With ELINE, the number of open arm entries and closed arm entries, and the time spent in open arms and closed arms was investigated. Finally, the mice were tested on motor performance on the balance beam. In this task, the mice were trained to cross a one-meter long wooden beam by placing the mice further away from the safe house at the end of the beam (10 cm, 40 cm, and 75 cm). After this training, they were able to cross the one-meter beam in the read-out trail. The number of slips compared to the total number of steps in this read-out trail was used as read-out parameter Statistical analysis Normally distributed data, established with a Shapiro-Wilk test, was tested in a one- Way ANOVA (owanovo). Non-normally distributed data was analyzed with the nonparametric Kruskal-Wallis test. As in Bruinenberg et al. 2016, the statistical analysis was used to answer if differences were present between WT and PKU mice of each strain and between genotype of specific strains. When the p-value of the owanova was 0.05, a Bonforroni post-hoc test was used to examine the differences among these specific groups (BTBR WT vs. BTBR PKU, B6 WT vs. B6 PKU, BTBR WT vs. B6 WT, and BTBR PKU vs. B6 PKU). When the Kruskal-Wallis test was 0.05, a Mann-Whitney U test was used to examine these differences. If not specified differently, the data is represented as mean ± standard error of the mean. Tukey s hinges are used to assess the interquartile range (Q3- Q1). In the normally distributed data, values outside 1.5 x interquartile range were seen as outliers and excluded from the statistical analysis. 3. RESULTS 3.1 Activity Activity was explored in different paradigms within the battery of behavioral paradigms (Figure 1A/B). In Figure 1A, the average daily activity under home-cage conditions is 49

51 Chapter 3 depicted (F(3,35)=3.909, p=0.017). No significant differences were found between the PKUs and WTs of both genetic backgrounds. However, BTBR PKU mice did show a higher activity than the B6 PKU mice (p=0.018). In Figure 2B, the distance moved through the OF is shown. This distance is indicative of novelty-induced activity. Here, the BTBR PKU mice covered less distance compared to the BTBR WT mice (p<0.001), showing a PKU phenotype effect within the BTBR strain. Furthermore, the BTBR PKU mice moved less than the B6 PKU mice (p<0.001). No effect of the PKU phenotype was seen in the B6 strain (p=0.097), and no differences were found between the WTs (p=0.109). Figure 1 Activity. (A) Passive infrared registration of home-cage activity (arbitrary units) (n=9-10). (B) The total distance moved in the open field (n=9-10). * p 0.05; mean±sem 3.2 Motor performance To assess motor performance in the mice, the number of correct steps relative to the total number of steps from the probe trail (100 cm) was calculated. In Figure 2, a clear difference in this score was shown between PKUs and WTs in both genetic strains (BTBR PKU vs. WT p<0.001, B6 PKU vs. WT p<0.001). No differences were observed between WTs or PKUs of both strains (WTs p=0.286, PKUs p=0.475). Figure 2 Motor performance. The number of correct steps in the balance beam test s probe trail (100 cm) is expressed as a percentage of the total number of steps necessary to cross the beam (n=9-10). * p 0.05; mean±sem 50

52 Chapter Anxiety-like behavior Anxiety-like behavior was examined in two behavioral paradigms, namely the OF and the EPM (Figure 3A-C). Both mazes used in these behavioral paradigms were designed to offer the mice sheltered and open areas, i.e. the corners compared to the center in the OF (Figure 3A) and closed arms versus open arms in the EPM. The preference of the animal to seek out these sheltered locations is thought to represent anxiety-like behavior. In the OF, the groups differed in time spent in the corners of the arena (F(3,35)=8.511, p<0.001). Both in the BTBR and B6 background, the PKU mice spent more time in the corners than the WT mice (BTBR: p=0.014, B6: 0.047). No differences were found between WTs (p=0.199) and PKUs (p=0.828). In the EPM, the PKU animals of both strains did not seek out the sheltered areas more than the WT animals, while the B6 PKU mice did spent more time in the closed arms compared with the BTBR PKU (p=0.003). 3 Figure 3 Anxiety-like behavior. (A) For Ethovion analysis, the open field was divided in three areas; 1) corner, 2) border, and 3) center. (B) The time spent in the corners (sum of area 1) of the open field are depicted (n=9-10). (C) Percentage of time spent in the closed arms of the elevated plus maze (n=9-10). * p 0.05; mean±sem 3.4 Learning and memory Two learning and memory paradigms were used to examine cognitive performance, namely NOR and SOR. In both tasks, the ability to identify a novel stimulus by increasing the exploration above change level was seen as intact ability to master the task. In the NOR, the BTBR WT mastered the task (t(9)=2.708, p=0.024) but the BTBR PKU did not (t(9)=1.059, p=0.317). In the B6 WT a trend was observed (t(9)=2.179, p=0.057) but this 51

53 Chapter 3 tendency was not present in the B6 PKU (t(9)=-0.677, p=0.516). In the SOR, a distinct PKU phenotype effect was seen in both stains wherein the WTs mastered the task (BTBR WT: t(9)=3.016, p=0.015, B6 WT: t(9)=4.381, p=0.002) and the PKUs did not (BTBR PKU: t(9)=0.565, p=0.586, B6 PKU: t(8)=1.084, p=0.310). Overall, trends were observed between differences between groups (NOR: F(3.36)=2.263, p=0.098, F(3.35)=2.706, p=0.060). No differences were found in total exploration time in both test phases of the paradigms (NOR: F(3,36)=1.873, p=0.152, SOR: F(3,33)=2.557, p=0.072). Figure 4 Learning and memory. (A) Discrimination index of the long-term novel object recognition (NOR) (n=9-10). A score of zero indicates an equal amount of time spent exploring both objects (B) The discrimination index of the long-term spatial object recognition (SOR) was corrected for any preference in location of object at for hand (n=9-10). * p 0.05, ~ p=0.057 above bars depict statistical analysis concerning change level; mean±sem 4. DISCUSSION In this study, the behavioral phenotype of female PKU mice was assessed in the four domains; 1) activity, 2) motor performance, 3) anxiety-like behavior, and 4) learning and memory. In contrast to male PKU mice of both genetic backgrounds 1, the effect of the PKU phenotype of the PKU females of both strains showed a similar outcome (Figure 5). The PKU phenotype in females was found in motor performance, anxiety-related behavior, and the learning and memory paradigms. Therefore, this study highlights that the PKU phenotype of the PKU mouse model is not only influenced by the genetic background 1, but can also be different between males and females. As numerous studies show that males and females can differ in (brain) morphology, behavior, and physiology, the discussion of this study will only concentrate on sex differences found in PKU patients. Two hypotheses related to steroid hormones and neurotransmitter metabolism, and differences in gene expression profiles between sexes will be considered. Before these topics are discussed, the limitations of the study will be discussed. 52

54 Chapter Limitations of the study In female mice, the fluctuating concentrations of ovarian hormones can influence behavior in for instance learning and memory 16,17. We did not monitor the cycle of the females during the behavioral paradigms in this study. We assume that mice would be at a random point in their cycle, as we have tested the females in small groups over a long period of time. However, we cannot be certain of this. It would be interesting to eliminate sex hormones to investigate the direct influence of sex hormones on the phenotypical outcome in future research. Furthermore, including biochemical parameters such as Phe and neurotransmitter concentration in brain, could give a better understanding of the underlying mechanism of the gender difference in the phenotypical outcome. 4.2 Differences in the manifestation of PKU between genders. As in preclinical research, clinical studies investigating gender differences in PKU are limited. Some of these studies describe gender differences in the outcome parameter, others do not. For example, no differences were found between PKU males and PKU females in IQ 18 or in gray matter volume of different brain regions 19. In contrast, differences were described in the occurrence of psychiatric disorders, visual attention, dietary control, quality of life, personality characteristics and behavior 18, The study of Stemerink et al. (2000) showed that the association between Phe concentrations in plasma and behavioral outcome is different between males and females (Age range: 8-20 years old) 24. In PKU males, a correlation was found between Phe concentrations measured over a two-year period prior to evaluation and the clusters introversion and positive-task orientation. In PKU females, the Phe concentrations of the first two years of life correlated with these clusters 24. This finding could indicate a gender difference in the response to elevated Phe concentrations 24. However, as pointed out by the authors, this result could also be influenced by differences in coping with the disorder, e.g. differences in social stress associated to PKU or maintaining the dietary treatment 24. In the PKU mouse model, differences are found between sexes in the B6 background but not in the BTBR background. For example, in the BTBR background, both the male and female BTBR PKU mice showed a PKU phenotype in the learning and memory paradigms. In the B6 background, only the female PKU mouse model showed behavioral deficit 1. Taken together, our present and previous 7 results suggest that (genetic) factors can influence the phenotypical outcome of PKU. 3 53

55 Chapter 3 Activity OF Home-cage activity Male Female BTBR B6 BTBR B6 Motor performance BB Anxiety-like behaviour OF EPM Learning and memory NOR SOR No significant difference Trend Significant difference Figure 5 The comparison between the PKU phenotype of male and female PKU mice. In our previous study, a PKU phenotype was observed in male BTBR mice in home-cage activity, balance beam (BB), novel object recognition (NOR), and spatial object recognition (SOR) 1. In the male B6, a PKU phenotype was only observed in the BB and the time spent in the corners of the open field (OF). In this current study, similar profile was seen in PKU phenotype in female BTBR and B6. Except for the distance moved trough OF that was only significant different between BTBR WT and BTBR PKU. (EPM=elevated plus maze) 4.2 Hypothesis 1: Sex hormones and neurotransmitter metabolism In PKU pathophysiology, disrupted neurotransmitter metabolism plays a key role in the development of brain dysfunction. In the PKU mouse model (B6) differences are observed in serotonin, epinephrine, norepinephrine and a metabolite of dopamine between female mice and male mice, regardless of genotype 9. It is not clear if these differences are also found in the BTBR PKU mouse model. Numerous studies have shown that neurotransmitter metabolism can be directly influenced by sex hormones. For example, estrogens can affect the neurotransmission of serotonin and dopamine by influencing concentrations, turnover, and number and function of receptors Furthermore, exposure patterns of sex hormones during development play a key role in the sexual dimorphic development of the brain which, in turn, can also influence the direct effect of sex hormones later in life 28,29. Therefore, the hypothesis could be raised that different hormonal exposure in (brain) development in male and female could predispose the sensitivity towards PKU-related changes in males 54

56 Chapter 3 and females differently. However, it is clear from the data we obtained in this study and the current literature that the gender differences are found in interaction with genetic background, indicating that the most likely hypothesis includes genetic factors. 4.3 Hypothesis II: Sex differences in gene expression. Despite similar genome sequences, sexual dimorphism is found in gene expression 30. Approximately 14% of the gene expression in the whole brain of mice is different between males and females 30. This estimation can differ between organs, brain regions and/or during development Although sexual dimorphic gene expression does not equal sex differences in protein levels 33, pinpointing gene expression differences could help identify modifying genes that reduce the impact of PKU-related biochemical changes in B6 PKU male mice. Especially, future research could investigate the overlap between differences in gene expression profiles between B6 and BTBR males and between B6 PKU male mice and B6 PKU female mice could narrow the number of genes of interest Conclusion This study, together with our previous report 1, emphasizes that the PKU phenotype is influenced by many factors including sex and genetic background. Investigating the underlying mechanisms in which these factors contribute to the PKU phenotype will increase our understanding of the detrimental effect of PKU. Furthermore, by including females within studies and as variable in statistical analysis more often, research into the underlying mechanism and new treatment strategies could further benefit the PKU population. 55

57 Chapter 3 5. REFERENCES 1 Bruinenberg VM, van der Goot E, van Vliet D, de Groot MJ, Mazzola PN, Heiner-Fokkema MR et al. The Behavioral Consequence of Phenylketonuria in Mice Depends on the Genetic Background. Front Behav Neurosci 2016; 10: Beery AK, Zucker I. Sex bias in neuroscience and biomedical research. Neurosci Biobehav Rev 2011; 35: Miller LR, Marks C, Becker JB, Hurn PD, Chen W-J, Woodruff T et al. Considering sex as a biological variable in preclinical research. FASEB J 2017; 31: Soldin OP, Mattison DR. Sex differences in pharmacokinetics and pharmacodynamics. Clin Pharmacokinet 2009; 48: Sandberg K, Umans JG, Georgetown Consensus Conference Work Group the GCCW. Recommendations concerning the new U.S. National Institutes of Health initiative to balance the sex of cells and animals in preclinical research. FASEB J 2015; 29: Blau N, van Spronsen FJ, Levy HL. Phenylketonuria. Lancet 2010; 376: Shedlovsky A, McDonald JD, Symula D, Dove WF. Mouse models of human phenylketonuria. Genetics 1993; 134: Mochizuki S, Mizukami H, Ogura T, Kure S, Ichinohe A, Kojima K et al. Long-term correction of hyperphenylalaninemia by AAV-mediated gene transfer leads to behavioral recovery in phenylketonuria mice. Gene Ther 2004; 11: Sawin EA, Murali SG, Ney DM. Differential effects of low-phenylalanine protein sources on brain neurotransmitters and behavior in C57Bl/6-Pahenu2 mice. Mol Genet Metab 2014; 111: Zagreda L, Goodman J, Druin DP, McDonald D, Diamond A. Cognitive deficits in a genetic mouse model of the most common biochemical cause of human mental retardation. J Neurosci 1999; 19: Mazzola PN, Bruinenberg V, Anjema K, van Vliet D, Dutra-Filho CS, van Spronsen FJ et al. Voluntary Exercise Prevents Oxidative Stress in the Brain of Phenylketonuria Mice. JIMD Rep doi: /8904_2015_ van Vliet D, Bruinenberg VM, Mazzola PN, van Faassen MHJR, de Blaauw P, Kema IP et al. Large Neutral Amino Acid Supplementation Exerts Its Effect through Three Synergistic Mechanisms: Proof of Principle in Phenylketonuria Mice. PLoS One 2015; 10: e van Vliet D, Bruinenberg VM, Mazzola PN, van Faassen MH, de Blaauw P, Pascucci T et al. Therapeutic brain modulation with targeted large neutral amino acid supplements in the Pah-enu2 phenylketonuria mouse model. Am J Clin Nutr 2016; 104: Hovens IB, Schoemaker RG, van der Zee EA, Absalom AR, Heineman E, van Leeuwen BL. Postoperative cognitive dysfunction: Involvement of neuroinflammation and neuronal functioning. Brain Behav Immun 2014; 38: Mulder C, Van Der Zee EA, Hut RA, Gerkema MP. Time-Place Learning and Memory Persist in Mice Lacking Functional Per1 and Per2 Clock Genes. J Biol Rhythms 2013; 28:

58 Chapter 3 16 Sabaliauskas N, Shen H, Molla J, Gong QH, Kuver A, Aoki C et al. Neurosteroid effects at α4βδ GABAA receptors alter spatial learning and synaptic plasticity in CA1 hippocampus across the estrous cycle of the mouse. Brain Res 2015; 1621: Hamson DK, Roes MM, Galea LAM. Sex Hormones and Cognition: Neuroendocrine Influences on Memory and Learning. Compr Physiol 2016; 6: Pietz J, Fätkenheuer B, Burgard P, Armbruster M, Esser G, Schmidt H. Psychiatric disorders in adult patients with early-treated phenylketonuria. Pediatrics 1997; 99: Christ SE, Price MH, Bodner KE, Saville C, Moffitt AJ, Peck D. Morphometric analysis of gray matter integrity in individuals with early-treated phenylketonuria. Mol Genet Metab 2016; 118: Fisch RO, Sines LK, Chang P. Personality characteristics of nonretarded phenylketonurics and their family members. J Clin Psychiatry 1981; 42: Cazzorla C, Cegolon L, Burlina AP, Celato A, Massa P, Giordano L et al. Quality of Life (QoL) assessment in a cohort of patients with phenylketonuria. BMC Public Health 2014; 14: Smith I, Beasley MG, Wolff OH, Ades AE. Behavior disturbance in 8-year-old children with early treated phenylketonuria. Report from the MRC/DHSS Phenylketonuria Register. J Pediatr 1988; 112: Craft S, Gourovitch ML, Dowton SB, Swanson JM, Bonforte S. Lateralized deficits in visual attention in males with developmental dopamine depletion. Neuropsychologia 1992; 30: Stemerdink BA, Kalverboer AF, van der Meere JJ, van der Molen MW, Huisman J, de Jong LW et al. Behaviour and school achievement in patients with early and continuously treated phenylketonuria. J Inherit Metab Dis 2000; 23: Pecins-Thompson M, Brown NA, Bethea CL. Regulation of serotonin re-uptake transporter mrna expression by ovarian steroids in rhesus macaques. Brain Res Mol Brain Res 1998; 53: Bethea CL, Phu K, Belikova Y, Bethea SC. Localization and regulation of reproductive steroid receptors in the raphe serotonin system of male macaques. J Chem Neuroanat 2015; 66 67: Fink G, Sumner BE, Rosie R, Grace O, Quinn JP. Estrogen control of central neurotransmission: effect on mood, mental state, and memory. Cell Mol Neurobiol 1996; 16: Knoll J, Miklya I, Knoll B, Dalló J. Sexual hormones terminate in the rat The significantly enhanced catecholaminergic/ serotoninergic tone in the brain characteristic to the post-weaning period. Life Sci 2000; 67: Espinosa P, Silva RA, Sanguinetti NK, Venegas FC, Riquelme R, González LF et al. Programming of Dopaminergic Neurons by Neonatal Sex Hormone Exposure: Effects on Dopamine Content and Tyrosine Hydroxylase Expression in Adult Male Rats. Neural Plast 2016; 2016: Yang X, Schadt EE, Wang S, Wang H, Arnold AP, Ingram-Drake L et al. Tissuespecific expression and regulation of sexually dimorphic genes in mice. Genome Res 2006; 16: Dewing P, Shi T, Horvath S, Vilain E. Sexually dimorphic gene expression in mouse brain precedes gonadal differentiation. Brain Res Mol Brain Res 2003; 118:

59 Chapter 3 32 Vawter MP, Evans S, Choudary P, Tomita H, Meador-Woodruff J, Molnar M et al. Gender-specific gene expression in postmortem human brain: localization to sex chromosomes. Neuropsychopharmacology 2004; 29: Xu J, Watkins R, Arnold AP. Sexually dimorphic expression of the X-linked gene Eif2s3x mrna but not protein in mouse brain. Gene Expr Patterns 2006; 6:

60 Chapter

61

62 CHAPTER 4 Sleep disturbances in Phenylketonuria: an explorative study in men and mice Vibeke M. Bruinenberg 1, Marijke C.M. Gordijn 2, Anita MacDonald 3, Francjan J. van Spronsen 4, Eddy A. Van der Zee 1 1Molecular Neurobiology, Groningen Institute for Evolutionary Life Sciences (GELIFES), University of Groningen, Groningen, the Netherlands, 2Chrono@work B.V. and Chronobiology, Groningen Institute for Evolutionary Life Sciences (GELIFES), University of Groningen, Groningen, the Netherlands, 3Birmingham Children s Hospital, Birmingham, United Kingdom, 4Beatrix Children s Hospital, University Medical Center Groningen, Groningen, the Netherlands. Front Neurol Apr 26;8:167. doi: /fneur ecollection 2017.

63 Chapter 4 SUMMARY Sleep problems have not been directly reported in Phenylketonuria (PKU). In PKU, the metabolic pathway of phenylalanine is disrupted, which, among others, causes deficits in the neurotransmitters and sleep modulators dopamine, norepinephrine and serotonin. Understanding sleep problems in PKU patients may help explain the pathophysiology of brain dysfunction in PKU patients. In this explorative study we investigated possible sleep problems in adult treated PKU patients and untreated PKU mice. In the PKU patients, sleep characteristics were compared to healthy first degree relatives by assessment of sleep disturbances, sleep-wake patterns, and sleepiness with the help of four questionnaires: Holland sleep disorder questionnaire, Pittsburgh sleep quality index, Epworth sleepiness scale, and Munich Chronotype Questionnaire. The results obtained with the questionnaires show that PKU individuals suffer more from sleep disorders, a reduced sleep quality, an increased latency to fall asleep, and experience more sleepiness during the day. In the PKU mice, activity patterns were recorded with passive infrared recorders. PKU mice switched more often between active and non-active behavior and shifted a part of their resting behavior into the active period, confirming that sleep quality is affected as a consequence of PKU. Together, these results give the first indication that sleep problems are present in PKU. More detailed future research will give a better understanding of these problems which could ultimately result in the improvement of treatment strategies by including sleep quality as an additional treatment target. Keywords (max 6): inherited metabolic disorder; Pah enu2 mice; PKU patients; PKU mice; neurotransmitters; sleep disorders 62

64 Chapter 4 1. INTRODUCTION In 2014, Gadoth and Oksenberg reviewed the incidence of sleep and sleep disorders in patients with inherited metabolic disease (IMD). Although their review focused on sleeprelated breathing disorders among severely affected subjects with IMD, the general title suggested that sleep problems could be missed or underestimated in these conditions. A metabolic disease in which brain modulators of sleep are severely affected but attention for sleep research is very limited is phenylketonuria (PKU). PKU is caused by an inborn error in the metabolic pathway of phenylalanine (Phe) that disrupts the conversion of Phe to tyrosine. As a result, Phe concentrations build up in blood and brain and the ability to intrinsically produce the dopamine precursor tyrosine is lost. These changes do not solely affect the metabolism of dopamine. Also, reduced concentrations of noradrenaline and serotonin are found in PKU patients 1,2 and in the PKU mouse model 3. These neurotransmitters are known to be important regulators of sleep, wakefulness, and switches between these states 4,5. Nevertheless, these abnormalities in neurotransmitter availability are not specifically linked to possible sleep problems in PKU research. In PKU research, a few studies have indirectly investigated sleep regulators or sleep. First, in treated and untreated PKU patients, sleep- EEG measurements indicate differences in the number of sleep spindles despite similar REM and non-rem distribution compared to healthy controls 6. Second, in early treated PKU infants (4-18 weeks old), EEG measurements show differences in the development of sleep compared to healthy controls 7. Finally, in the PKU mouse model high levels of orexin A (hypocretin 1) were reported, a neuropeptide that is associated with wakefulness 8,9. This made the authors suggest hyperactivity in PKU, however, the exact consequence of these increased levels are not clear while hyperactivity is not consistently described in PKU mice Currently, PKU treatment remains suboptimal in which disturbances in executive functions, mood, social cognition, and in internalizing problems such as depression and anxiety are described in early-treated PKU patients 11,12. As it is well established that altered sleep negatively influences cognitive performance, most notably in the domains of executive functioning 13 15, and mood by impacting feelings of depression, anxiety and stress 16,17, sleep related issues could very well serve as an explanation of the PKU brain dysfunction despite diet and drug treatment. Understanding the presence and severity of sleep problems in PKU patients and its pathophysiology could ultimately result in the improvement of treatment strategies by including sleep quality as an additional treatment target. Therefore, the aim of this explorative study was to investigate the presence of sleep disturbances in PKU patients with questionnaires together with analyses of rest/wake patterns in PKU mice, indirectly reflecting sleep characteristics which could confirm the PKU-specific nature of putative sleep issues in PKU patients. As sleep is influenced by, among others, genetic factors 18 21, first-degree 63

65 Chapter 4 relatives (FDR) of PKU patients and wild-type (WT) littermates of each genetic strain of the PKU mouse model were used as controls. 2. MATERIALS & METHODS 2.1. PKU patients Subjects In the summer of 2016, participants for this study were recruited by distributing a link to an electronic survey to subjects associated with the Dutch PKU patient organization. PKU patients and FDR, who did not do shift work in the past three months, were asked to fill out four questionnaires with ten additional questions (date of birth, zip code, gender, height, bodyweight, PKU or control, treatment of PKU, other health issues, smoking, and the use of sleep promoting drugs). All participants were informed about the scientific purpose of the study and agreed to participate. They completed the questionnaires completely anonymous. To ensure that the questionnaires were not completed by the same individuals more than once, the submissions were checked for uniqueness, focusing on date of birth and IP address. In total, 47 subjects of which 25 PKU patients and 23 controls participated. The Medical Ethics Review Committee (METC) of the University of Groningen concluded that the Medical Research Involving Subjects Act did not apply to this study Sleep questionnaires Four validated questionnaires were included in the survey: (1) Holland Sleep Disorders Questionnaire (HSDQ) 22 ; 40 items, (2) Pittsburgh Sleep Quality Index (PSQI) 23 ; 19 items, (3) Epworth Sleepiness Questionnaire (ESS) 24 ; 8 items. (4) Munich Chronotype Questionnaire (MCTQ) 25 ; 16 items. Firstly, the HSDQ gives a general score to identify the possible occurrence of a sleep disorder and can differentiate between six main categories of sleep disorders (insomnia, parasomnia, circadian rhythm sleep disorders (CRSD), hypersomnia, sleep-related movement disorders such as for instance restless legs syndrome, and sleeprelated breathing disorder) 22. Secondly, for the PSQI, seven component scores can be derived (subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction) and computed to a global score 23. Thirdly, the ESS is used to examine sleepiness during the day (for instance during reading) 24. Finally, the MCTQ was used to identify chronotype (the preferred timing of sleep) calculated by taking the mid-point of sleep on free days corrected for the sleep debt acquired during working days PKU mouse study To investigate the rest/wake pattern in PKU mice, the home-cage activity of adult (4-7 months) WT and PKU mice of BTBR and C57Bl/6 (B6) background was monitored. Both genetic strains of the PKU mouse model have a point mutation in the gene encoding for 64

66 Chapter 4 phenylalanine hydroxylase (PAH) causing Phe to rise in blood and brain 10. The PKU model was originally described for the BTBR background but the model was latter crossed back on to the C57Bl/6J background. Currently, both genetic strains are used in PKU research. In-house heterozygous mating pairs were used to breed the following groups of mice: BTBT WT, BTBR PKU, B6 WT, and B6 PKU. These mice were weaned on post-natal day 28 and genotype was established with quantitative PCR 10. The experiment consisted of a habituation phase and a data acquiring phase. After a seven-day habituation phase, the activity of individual housed mice (cage: 33x15x14cm with nesting material and paper role) were monitored for seven days with passive infrared detectors (PIR). During the whole experiment, animals were on a 12/12 light/dark cycle and had ad libitum access to water and normal chow (RMH-B 2181, ABdiets, Phe: 8.7 g/kg). Data was analyzed with ACTOVIEW (made in-house, described in 27 ). This program calculated the average daily activity, diurnality ((Sum activity light phase- sum activity dark phase)/ total activity) and fragmentation 28 from the files produced by the activity management system (CAMS) collected from the PIR. All proceedings were carried out in accordance to the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (The ARRIVE Guidelines Checklist) approved by the Institutional Animal Care and Use Committee of the University of Groningen Statistics The statistical analysis was executed with the statistical software IBM SPSS Statistics for Windows, Version 22.0 (Armonk, NY: IBM Corp.). Within this program the Shapiro-Wilk was used to test normality of the data. The activity of the mice was normally distributed and a multi-variate ANOVA was used to test the factors group and gender of overall activity, fragmentation and diurnality. Differences in group characteristics in the patient study were not normally distributed and therefore tested with the non-parametric Mann-Whitney U test. The chronotype score of the MCTQ, normally distributed, was tested with a univariate ANOVA for group. Age and gender were included as cofactors. The ordinal nature of the HSDQ, PSQI and ESS scores, made it not possible to test confounding factors parametrically. Therefore, gender and age were explored with generalized linear models, using an ordinal model. Differences in the frequency of occurrence of sleep disorders was also tested with a generalized linear model, using a binary model. Non-parametric Levene s test was used to test the homogeneity of the data. 3. RESULTS 3.1 PKU patients Subjects This study is a pilot study serving as a proof-of-concept for sleep-related issues due to PKU. For this reason, individuals were included when they correctly filled out at least one of the questionnaires. Not all questionnaires were correctly filled out by all subjects which resulted 65

67 Chapter 4 in different group characteristics for each questionnaire (Table 1). For all questionnaires responses, the FDR controls were significantly older than the PKU subjects (HSDQ: p<0.05, PSQI: p<0.05, ESS: p<0.05, MCTQ: p<0.05) but no differences were found in BMI (HSDQ: p=0.68, PSQI: p=0.92, ESS: p=0.99, MCTQ: p=0.90). Furthermore, no significant differences were found in gender distribution of the groups (HSDQ: t(44)=-0.603, p=0.55, PSQI: t(35)=0.172, p=0.87, ESS: t(37)=0.143, p=0.89, MCTQ: t(39)=-0.260, p=0.80). One FDR control was excluded because she used several sleep-promoting drugs (Trazodon and zolpidem tartrate) Frequency of sleep disorders. The global score of the HSDQ is used to identify the presence of a sleep disorder with an overall accuracy of 88% (κ: 0.75) 22. In PKU patients, 48 % had a global score above the cut-off of 2.02, indicative of a sleep disorder, compared to 19 % of FDR controls (Fig. 1A: b=1.367, Wald χ 2 (1,N=46)=3.983, p<0.05). Especially, among the six main sleep disorder categories, PKU patients had a higher score for both insomnia and CRSD (Fig 1B: b=-1.527, Wald χ 2 (1,N=46)=7.626, p<0.05, Fig 1C: b=-1.593, Wald χ 2 (1,N=46)=8.115, p<0.05, respectively), but not for the other four categories (parasomnia; b=-0.985, Wald χ 2 (1,N=46)=2.941, p=0.086, hypersomnia; b=-0.985, Wald χ 2 (1,N=46)=3.287, p=0.070, restless legs syndrome; b=-0.696, Wald χ 2 (1,N=46)=1.757, p=0.185 and sleep-related breathing disorder; b=-0.693, Wald χ 2 (1,N=46)=1.688, p=0.194). Age and gender did not significantly contribute to these models. These results reveal a higher frequency of sleep disorders, more specifically insomnia and CRSD, in PKU patients Sleep quality The global PSQI, comprised of seven sleep related components, is used to classify poor (>5) and good sleepers ( 5). This global score was significantly higher in the PKU patients of which more people were classified as poor sleepers (57%) compared to the FDR controls of which 25% were classified as poor sleepers (Fig. 2A: b=-1.674, Wald χ 2 (1,N=37)=7.102, p<0.05). Within the global score, two component scores differed between PKU patients and FDR controls. The first was the component score for latency to fall asleep which was significantly increased in PKU patients (FDR: 0.44±0.63, PKU: 1.43±0.98, b=-2.306, Wald χ 2 (1,N=37)=9.733, p<0.05). The second one was the component score for subjective sleep quality. This score was computed from the question in which subjects could indicate their quality of sleep during the past month on a scale of very good (score 0) to very bad (score 3). This score was significantly higher in PKU patients than in the FDR controls (FDR: 0.69±0.60, PKU: 1.38±0.92, b=-1.655, Wald χ 2 (1,N=37)=5.516, p<0.05). Age and gender did not significantly contribute to these models. These results suggest that sleep quality is reduced in PKU patients compared to FDR controls. 66

68 Chapter 4 Table 1. Subject characteristics. Data is presented as mean ± standard deviation. Abbreviations: F=female, M=male, HSDQ= Holland Sleep Disorders Questionnaire, PSQI=Pittsburgh Sleep Quality Index, ESS=Epworth Sleepiness Questionnaire, MCTQ=Munich Chronotype Questionnaire, PKU= phenylketonuria patients, FDR-control= first degree relatives-control HSDQ PSQI ESS MCTQ PKU FDR-control PKU FDR-control PKU FDR-control PKU FDR-control Number Age 29.2 ± ± ± ± ± ± ± ± 10.1 BMI 24.1 ± ± ± ± ± ± ± ± 4.2 Gender 17F/ 8M 16F/ 5M 15F/ 6M 11F/ 4M 16 F/6 M 12 F/ 5M 16F/ 8M 12 F/ 5M Smoking Heath issues Sleep-promoting drugs KUVAN KUVAN+protein restricted Protein restricted

69 Chapter Sleepiness during the day In the ESS, the participant subjectively rate the chance of dozing off during eight situations from none (score 0) to high (score 3). This score is significantly higher in PKU patients than FDR controls (Fig 2B: (b=-1.608, Wald χ 2 (1,N=39)=6.597, p<0.05). A main effect of age and gender did not contribute significantly to the model. These results suggest that PKU patients experience more sleepiness during daytime Chronotype. In the MCTQ, the sleep schedules of the participants on working and non-working days were asked. From this data, we could calculate chronotype, defined as the mid-sleep on free days corrected for the potential sleep debt acquired during the working days 26. Chronotype is dependent on age and gender 26, therefore, statistical analysis included age and gender as a cofactor. No significant differences were found between chronotype scores in PKU patients and FDR controls or for any of the cofactors in the complete model (Group: F(1,37)=1.287, p=0.26, Gender: F(1,37)=0.954, p=0.34, Age: F(1,37)=1.941, p=0.17) PKU mice Rest/wake patterns. No main or interaction effects were observed for sex in all parameters, therefore, data of males and female were grouped. The fragmentation score is indicative of the frequency that active behavior is switched to non-active behavior and vice versa. In PKU mice, in both strains, the fragmentation score was increased compared to WT (Fig. 3A; F(3,56)=10.803, p<0.05, BTBR p<0.05, p<0.05). Although differences were found in overall activity between the WT s of each strain (BTBR WT: ± , B6 WT: ± (mean±sd) F(3,56)=3.858, p<0.05, BTBR WT vs B6 WT p=0.019), the increase in fragmentation score found in PKU mice did not coincide with a change in overall activity (BTBR PKU: ±605.23, B6 PKU: ±860.06, BTBR p=0.67, B6 p=1.000). However, a shift did occur in the timing of the rest/active behavior. The negative diurnality score, reflecting night activity in animals, became less negative in PKU mice (Fig. 3B; F(3,56)=8.235, p<0.001, BTBR p<0.05, B6 p<0.05). These results reveal that PKU mice have increased fragmentation and a shift in diurnality (more inactive in active phase). 68

70 Chapter 4 4 Figure 1. HSDQ (A) Results from the global score of HSDQ indicate that 48% of the PKU patients have a sleep disorder compared to 19% of the FDR controls. (B) PKU patients have a significant higher insomnia score than FDR controls. Six PKU patients are above the cut-off score compared to zero FDR controls (C). Although only two PKU patients are above the cut-off score, PKU patients have significant higher CRSD score compared to FDR controls. Data represents individual scores with median. Dotted line represents cut-off score between having sleep problem or not. ** p<0.01 Figure 2. PSQI and ESS. (A) The global score of the Pittsburgh Sleep Quality Index (PSQI) was significantly higher in PKU patients compared to FDR control. The dotted line represents the cut-off between good and poor sleepers. 57% of the PKU patients are above this cut-off and categorized as poor sleepers. (B) Epworth Sleepiness Questionnaire (ESS). Patients experience more sleepiness during the day than FDR controls. Data represents individual scores with median. * p<0.05, ** p<

71 Chapter 4 Figure 3. Characteristics of the rest/wake pattern in mice (A) In both genetic strains of the PKU mouse model, an increase in fragmentation is seen in the PKU mice compared to WT littermates. Furthermore, a significant difference is found between the fragmentation score of PKU mice of each strain. (B) In the graph negative diurnality scores are observed. This indicates that we are investigating animals which are active in the dark. PKU mice have a less distinct negative score suggesting that they shift part of their resting behavior into the light phase. Data are depicted as mean ± standard error of the mean. * p<0.05, ** p< DISCUSSION In this explorative study, we investigated sleep characteristics in PKU patients with questionnaires and analyzed the rest/wake patterns in PKU mice. In the PKU patients study, we showed that PKU patients compared to FDR controls have more sleep disorders, a reduced sleep quality, an increased latency to fall asleep, and experience more sleepiness during the day. In the PKU mice, we found an increase in fragmentation and a shift in diurnality. The increase in fragmentation indicated that the PKU mice switch more often between active and non-active behavior. This score did not coincide with changes in overall activity. In addition, PKU mice shift a part from their resting behavior into the active phase (a shift in diurnality). Both experiments strongly support the hypothesis that sleep is affected in PKU. This seems to be directly related to the disorders as the deficits were found in both PKU mouse strains despite their genetic differences and cognitive sensitivity to the PKU condition Study limitations Sleep is influenced by, among others, genetic factors, age, and gender For this reason, this study recruited FDR of PKU patients as control group. No differences were found in gender distributions between the groups, but differences were found in age between FDR controls and PKU patients. Although we recognize these limitations, we believe that it does not hamper our results obtained in the first three questionnaires (HSDQ, PSQI, ESS) for the following reason. Literature shows either no effect or a deterioration of sleep with aging. For 70

72 Chapter 4 example, the PSQI is not affected by age 29. The ESS score is influenced by an age x gender interaction, wherein females tend to have higher scores between 0-39 age group compared to males 30. Around the ages 40-49, the ESS score of males deteriorates reaching the same ESS score as females. Therefore, the ESS score is either higher or stays constant with increasing age 30. This implies that the higher scores found in the younger PKU patients compared to the older FDR controls are contrary to what would be expected and likely reflect a real indication of sleep problems in PKU patients. Moreover, no significant effect of age was found in our study. 4.2 Sleep characteristics are altered in PKU The different measurements of sleep used in this study reveal a variety of sleep problems and show some specifically affected characteristics of sleep. In the HSDQ, PKU patients show a higher incidence in sleep disorders, but only in the main categories insomnia and CRSD scores were higher in PKU patients than in FDR controls. CRSD were changed to circadian rhythm sleep-wake disorders (CRSWD) in the third edition of the International classification of sleep disorders 31. CRSWD are sleep disorders grouped under dyssomnias, a group of sleep disorders which show insomnia, excessive sleepiness, or difficulty initiating or maintaining sleep 31. In the current study, we found an increased score for insomnia and an increased sleepiness during the day in the ESS in PKU patients, supporting the idea that PKU patients experience problems specific for this cluster of sleep disorders. CRSWD are disorders related to the timing of sleep 31. Some are a consequence of external circumstances, e.g. shift work or jet lag, others have potentially a more internal, neurological basis, e.g. delayed sleep phase syndrome (DSPS). In DSPS, the latency to fall asleep opposed to the desired time to fall asleep is delayed. DSPS patients experience difficulties to shift their sleep/ wake pattern to an earlier time point in response to environmental time cues, for example traveling from Europe to Asia, and do not experience sleepiness when they are able to sleep at their desired time, as is possible for instance during a holiday or vacation period. In this study, we showed that PKU patients report an increased latency to fall asleep in the PSQI. However, we did not see a difference in the midsleep on non-working days when age was a cofactor. This could be because midsleep is strongly influenced by age and the age was not distributed evenly over the full width of both groups 26. Therefore, an important future direction is to compare age-matched controls to PKU patients. Further research in PKU should focus on 1) more objective measurements of sleep such as polysomnography, or sleepwake rhythm analysis with 2) phase shift experiments in PKU mice to identify problems in shifting sleep/wake patterns (and neurological substrates), 3) monitoring sleepiness during the day specifically during holidays, and 4) core body temperature and dim-light melatonin rhythm monitoring to investigate if PKU patients experience a blunted or delayed internal rhythm of physiological markers. 4 71

73 Chapter The switch between sleep and wakefulness is defective in PKU The HSDQ identifies sleep disorders and attribute certain scores to six symptom clusters. These clusters may be due to different sleep disorders, possibly with different pathophysiological mechanism. For instance, the comorbidity of insomnia with other sleep disorders is very high 22. The PKU mouse study did identify a more specific aspect of sleep, namely increased fragmentation. In general, an increased fragmentation score indicates an increase in switching behavior. Switching between sleep and wakefulness is thought to be regulated by a flip-flop switch that results from mutual inhibition of sleep-promoting pathways and wakepromoting pathways 32. Several cholinergic and monoaminergic projections are important in these pathways, such as serotonin, norepinephrine, and dopamine 32. As these latter neurotransmitters are affected in PKU mice (and in untreated PKU patients), it could be that the increased fragmentation score is a consequence of disruptions in this switch. In earlytreated patients on diet, neurotransmitter deficiencies in dopamine and serotonin are still present 2. These deficiencies could possibly affect the switch between sleep and wakefulness and cause fragmentation of the sleep/wake rhythm in PKU patients, more difficulty falling asleep and as a consequence daytime sleepiness Conclusion This explorative study is the first to investigate sleep disturbances both in PKU patients and PKU mice. In PKU patients, we demonstrate more sleep disorders, a reduced sleep quality, an increased latency to fall asleep, and more sleepiness during the day. We show in PKU mice an increased fragmentation and a shift in diurnality. These results produce the first evidence to suggest that sleep problems occur in PKU. The resulting complaints associated with altered sleep are comparable to the cognitive symptoms described for early and continuously treated PKU patients. More detailed future research will give a better understanding and further identify sleep problems in PKU which could ultimately result in the improvement of treatment strategies by including sleep quality as an additional treatment target. ACKNOWLEDGEMENTS The authors would like to thank Dr. Marina C. Giménez for her help with the electronic survey. 72

74 Chapter 4 5. REFERENCES 1 McKean CM. The effects of high phenylalanine concentrations on serotonin and catecholamine metabolism in the human brain. Brain Res 1972; 47: Burlina AB, Bonafé L, Ferrari V, Suppiej A, Zacchello F, Burlina AP. Measurement of neurotransmitter metabolites in the cerebrospinal fluid of phenylketonuric patients under dietary treatment. J Inherit Metab Dis 2000; 23: van Vliet D, Bruinenberg VM, Mazzola PN, van Faassen MH, de Blaauw P, Pascucci T et al. Therapeutic brain modulation with targeted large neutral amino acid supplements in the Pah-enu2 phenylketonuria mouse model. Am J Clin Nutr 2016; 104: Holst SC, Valomon A, Landolt H-P. Sleep Pharmacogenetics: Personalized Sleep-Wake Therapy. Annu Rev Pharmacol Toxicol 2016; 56: Eban-Rothschild A, Rothschild G, Giardino WJ, Jones JR, de Lecea L. VTA dopaminergic neurons regulate ethologically relevant sleep wake behaviors. Nat Neurosci 2016; 19: Schulte FJ, Kaiser HJ, Engelbart S, Bell EF, Castell R, Lenard HG. Sleep patterns in hyperphenylalaninemia: a lesson on serotonin to be learned from phenylketonuria. Pediatr Res 1973; 7: De Giorgis GF, Nonnis E, Crocioni F, Gregori P, Rosini MP, Leuzzi V et al. Evolution of daytime quiet sleep components in early treated phenylketonuric infants. Brain Dev; 18: Surendran S, Rady PL, Szucs S, Michals- Matalon K, Tyring SK, Matalon R. High level of orexin A observed in the phenylketonuria mouse brain is due to the abnormal expression of prepro-orexin. Biochem Biophys Res Commun 2004; 317: Surendran S, Campbell GA, Tyring SK, Matalon K, McDonald JD, Matalon R. High levels of orexin A in the brain of the mouse model for phenylketonuria: possible role of orexin A in hyperactivity seen in children with PKU. Neurochem Res 2003; 28: Bruinenberg VM, van der Goot E, van Vliet D, de Groot MJ, Mazzola PN, Heiner-Fokkema MR et al. The Behavioral Consequence of Phenylketonuria in Mice Depends on the Genetic Background. Front Behav Neurosci 2016; 10: Jahja R, van Spronsen FJ, de Sonneville LMJ, van der Meere JJ, Bosch AM, Hollak CEM et al. Social-cognitive functioning and social skills in patients with early treated phenylketonuria: a PKU-COBESO study. J Inherit Metab Dis 2016; 39: Waisbren SE, Noel K, Fahrbach K, Cella C, Frame D, Dorenbaum A et al. Phenylalanine blood levels and clinical outcomes in phenylketonuria: a systematic literature review and meta-analysis. Mol Genet Metab 2007; 92: Banks S, Dinges DF. Behavioral and physiological consequences of sleep restriction. J Clin Sleep Med 2007; 3: Couyoumdjian A, Sdoia S, Tempesta D, Curcio G, Rastellini E, DE Gennaro L et al. The effects of sleep and sleep deprivation on task-switching performance. J Sleep Res 2010; 19:

75 Chapter 4 15 Goel N, Rao H, Durmer JS, Dinges DF. Neurocognitive consequences of sleep deprivation. Semin Neurol 2009; 29: Short MA, Louca M. Sleep deprivation leads to mood deficits in healthy adolescents. Sleep Med 2015; 16: Meerlo P, Havekes R, Steiger A. Chronically restricted or disrupted sleep as a causal factor in the development of depression. Curr Top Behav Neurosci 2015; 25: Lind MJ, Aggen SH, Kirkpatrick RM, Kendler KS, Amstadter AB. A Longitudinal Twin Study of Insomnia Symptoms in Adults. Sleep 2015; 38: Barclay NL, Gregory AM. Quantitative genetic research on sleep: A review of normal sleep, sleep disturbances and associated emotional, behavioural, and health-related difficulties. Sleep Med Rev 2013; 17: Gottlieb DJ, O Connor GT, Wilk JB. Genome-wide association of sleep and circadian phenotypes. BMC Med Genet 2007; 8: S9. 21 Paine S-J, Gander PH, Harris R, Reid P. Who reports insomnia? Relationships with age, sex, ethnicity, and socioeconomic deprivation. Sleep 2004; 27: Kerkhof GA, Geuke MEH, Brouwer A, Rijsman RM, Schimsheimer RJ, Van Kasteel V. Holland Sleep Disorders Questionnaire: a new sleep disorders questionnaire based on the International Classification of Sleep Disorders-2. J Sleep Res 2013; 22: Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res 1989; 28: Johns MW. Reliability and factor analysis of the Epworth Sleepiness Scale. Sleep 1992; 15: Roenneberg T, Kuehnle T, Juda M, Kantermann T, Allebrandt K, Gordijn M et al. Epidemiology of the human circadian clock. Sleep Med Rev 2007; 11: Roenneberg T, Kuehnle T, Pramstaller PP, Ricken J, Havel M, Guth A et al. A marker for the end of adolescence. Curr Biol 2004; 14: R1038 R Mulder C, Van Der Zee EA, Hut RA, Gerkema MP. Time-Place Learning and Memory Persist in Mice Lacking Functional Per1 and Per2 Clock Genes. J Biol Rhythms 2013; 28: van Someren EJ, Hagebeuk EE, Lijzenga C, Scheltens P, de Rooij SE, Jonker C et al. Circadian rest-activity rhythm disturbances in Alzheimer s disease. Biol Psychiatry 1996; 40: Grandner MA, Kripke DF, Yoon I-Y, Youngstedt SD. Criterion validity of the Pittsburgh Sleep Quality Index: Investigation in a non-clinical sample. Sleep Biol Rhythms 2006; 4: Boyes J, Drakatos P, Jarrold I, Smith J, Steier J. The use of an online Epworth Sleepiness Scale to assess excessive daytime sleepiness. Sleep Breath doi: /s x. 31 American Academy of Sleep Medicine. International Classification of Sleep Disorders Third Edition (ICSD-3) Saper CB, Fuller PM, Pedersen NP, Lu J, Scammell TE. Sleep State Switching. Neuron 2010; 68:

76 Chapter

77

78 CHAPTER 5 A novel treatment strategy for phenylketonuria: exploring the possibilities of nutrients to improve brain function Vibeke M.Bruinenberg 1, Priscila N. Mazzola 1,2, Danique van Vliet 2, Danielle S. Counotte 3, Maryam Rakhshandehroo 3, Mirjam Kuhn 3, Francjan J. van Spronsen 2, Eddy A. van der Zee 1 1 Molecular Neurobiology, Groningen Institute for Evolutionary Life Sciences (GELIFES), University of Groningen, Groningen, the Netherlands, 2 Beatrix Children s Hospital, University Medical Center Groningen, Groningen, the Netherlands, 3 Nutricia Research, Nutricia Advanced Medical Nutrition, Utrecht, the Netherlands

79 Chapter 5 Abbreviations AA Arachidonic acid BDNF Brain-derived neurotrophic factor EPA Eicosapentaenoic acid GABA Gamma-aminobutyric acid HMGR 3-hydroxy-3-methylglutaryl coenzyme A reductase LIMK1 LIM kinase 1 LNAA Large neutral amino acids PAH Phenyalanine hydroxylase Phe Phenylalanine PKU Phenylketonuria Rac 1 Ras-related C3 botulinum toxin substrate 1 ROS Reactive oxygen species TPH-2 Tryptophan hydroxylase 2 Trp Tryptophan Tyr Tyrosine UMP Uridine-5 -monophosphate 78

80 Chapter 5 ABSTRACT Phenylketonuria (PKU) is an inherited metabolic disorder where disrupted conversion of Phe into tyrosine leads to accumulation of phenylalanine (Phe) in blood and brain. Increased Phe impairs brain functioning by negatively affecting four domains: neurotransmitter metabolism, white matter integrity, oxidative balance, and synapse functioning. Traditionally, PKU treatment aims to lower blood Phe concentrations, however, because treatment is often difficult to maintain, blood Phe concentrations can rise and fluctuate in PKU patients. Therefore, a treatment strategy focusing on relieving the Phe effects on brain functioning by means of specific nutrients could yield new treatment possibilities. In this review, the possibility of an additional nutritional intervention for PKU is presented through reviewing the beneficial effects of nutritional components in the healthy or diseased brain in these four domains. Based on the described positive effects in the non-pku-literature, the hypothesized alternative nutritional treatment for PKU should include omega-3 fatty acids, B-vitamins, and antioxidants. 5 79

81 Chapter 5 1. INTRODUCTION Hippocrates stated, Let food be thy medicine and medicine be thy food. This statement illustrates the age-old belief that food can be used to treat a disease. In recent years, the concept that nutrition can be used to aid cognition and cognitive disorders, has regained more attention 1,2. The rapid increase in research and thereby knowledge has raised awareness into the widespread possibilities of different nutritional components as a treatment strategy for disorders affecting neurocognition. If any disease should be addressed in which the importance of nutrition has been recognized at an early stage, it is phenylketonuria (PKU; OMIM # ). This inherited metabolic disease is caused by mutations in the gene encoding for the hepatic enzyme phenylalanine hydroxylase (PAH) responsible for converting phenylalanine (Phe) into tyrosine (Tyr). The loss of functionality of the enzyme results in high concentrations of Phe in blood and brain. These increased concentrations of Phe can affect four domains of brain functioning through 1) disrupting the neurotransmitter metabolism of especially serotonin and dopamine 3, 2) altering the white matter integrity 4, increasing oxidative stress 5, and affecting synapse functioning 6. In untreated patients, this results in mental disablement, problems with movement, and seizures 7. To minimize these consequences, current dietary treatment aims to severely restrict Phe intake as early as possible with a protein restricted diet plus an artificial amino acid mixture. Despite this treatment, subtle and specific deficits in cognitive functioning, such as processing speed, attention, and working memory are found in earlytreated PKU patients 8,9. The association between Phe concentrations and symptomology in patients, treated and untreated, is well established 7. Therefore, treatment strategies in PKU traditionally have focused on lowering Phe concentrations. However, as current dietary treatment is often difficult to maintain, Phe concentrations can rise and fluctuate in PKU patients 10. An alternative intervention strategy focusing on reducing the functional consequences of high Phe on the four above mentioned domains by means of specific nutrients could be of great interest. Applying nutrients to aid brain functioning was examined, among others, by the Wurtman lab. This resulted in a specific formulation of nutrients registered as Fortasyn Connect (FC) 11. FC includes the omega-3 PUFAs DHA and EPA, but also choline, UMP, phospholipids, folic acid, vitamins B6, B12, C, E, and selenium 12. These nutrients constitute the precursors and cofactors for the formation of membranes through the biosynthetic Kennedy pathway 13, and dietary supplementation with these nutrients has been shown to improve brain function (for review 12 ). In this review, the affected four functional domains neurotransmitter metabolism, white matter integrity, oxidative stress, and synaptic functioning will be discussed. Hereafter, these 80

82 Chapter 5 domains will be discussed in light of possible targets of nutrients present in FC to battle PKU-specific damage to them. Literature concerning these nutrients in PKU patients and/or models of the disease is limited. Therefore, particularly nutrients that have beneficial effects in the healthy or diseased brain other than PKU will also be discussed as potential candidates to relieve cognitive deficits in PKU. 2. THE AFFECTED DOMAINS: 2.1 Neurotransmitter metabolism The deficiency in the conversion of Phe to Tyr causes problems with neurotransmitter metabolism via different routes. Besides the poor intrinsic synthesis of Tyr, high blood Phe levels interfere with the brain s availability of other large neutral amino acids (LNAAs), including Tyr and Tryptophan (Trp), via interfering with transport over the blood-brain barrier 14,15. In a direct matter, reduced Tyr availability in the brain leads to impaired dopamine synthesis, as Tyr is the amino acid precursor for dopamine. Indirectly, Phe impairs the availability of other LNAAs such as Trp, therefore diminishing the synthesis of serotonin, which has Trp as its precursor. Moreover, high brain Phe concentrations inhibit the activity of the enzymes Tyr- and Trp hydroxylases that are important for the rate-limiting steps in the conversion of Tyr and Trp to dopamine and serotonin, respectively 16. Indeed, lower than normal levels of dopamine and serotonin metabolites have been observed in the CSF of PKU patients3. In addition, reduced levels of monoamines (serotonin, dopamine and norepinephrine) have been shown in the brain of mouse models of PKU White matter integrity Myelin alterations in PKU, first reported in , have become a consistent feature of PKU in vitro and in vivo models, untreated, and treated PKU patients 4,23. In patients, metabolic control expressed as plasma Phe, brain Phe, and the fluctuations in Phe during the life time is associated with the degree of white matter abnormalities 4, Nevertheless, the exact influence of Phe on myelin is not clear yet. For instance, in cell culture, oligodendrocytes subjected to high concentrations of Phe switch phenotype from a myelinated to a nonmyelinated form 29. However, Phe and/or the metabolites of Phe (phenylpyruvate, and phenylacetate) do not affect oligodendrocyte progenitor cell proliferation (development), oligodendrocyte migration (function) nor induce mortality of oligodendrocytes (survival) 30. This indirect effect could be via the effect of Phe on cholesterol and protein synthesis. Both processes are necessary for the production and maintenance of myelin sheaths. Cholesterol synthesis is affected through the negative effect of Phe on the rate-controlling enzyme, 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGR). Both in PKU mice as well as in Phe-exposed oligodendrocyte cultures, the activity of this enzyme is reduced 31. The negative effect of Phe on protein synthesis is confirmed in vitro and in vivo in the PKU mouse model and in PKU patients High concentrations of Phe, initiated by the PAH inhibitor alpha- 81

83 Chapter 5 methylphenylalanine, are suggested to influence protein synthesis through inhibiting protein synthesis initiation and protein elongation 34,35. Taken together, the myelin deficits found in PKU are more likely to be caused by the effect of Phe on cholesterol and protein synthesis rather than by a direct effect of Phe on oligodendrocytes. This does not exclude other direct effects of Phe on myelin. 2.3 Oxidative stress Oxidative stress plays a role in the pathogenesis of several neurocognitive disorders and has been studied in PKU (reviewed by Ribas et al. 5 ). Oxidative stress is an imbalance between (increased) production and/or (decreased) removal of reactive species. This imbalance can be caused by accumulation of reactive species and/or decreased antioxidant defense. Multiple arguments support the potential role of oxidative stress in the pathophysiology of PKU. Firstly, the accumulation of Phe together with its metabolites can increase the generation of reactive species mostly derived from oxygen (reactive oxygen species, ROS), that in turn can react with biomolecules causing oxidative damage to proteins, lipids, carbohydrates, and DNA. As a consequence, cell injury or even cell death can occur. Secondly, in general, compared with other organs, the brain is especially susceptible to oxidative stress due to its relatively poor antioxidant defense and its high oxygen consumption 36. Thirdly, the Pherestricted diet may hamper the intake of antioxidants and/or their precursors in treated patients, leading to oxidative stress by reducing antioxidant defenses 5,37. Finally, metabolites of Phe inhibit antioxidant enzymes 38,39. These arguments are supported by several findings confirming the presence of oxidative stress within PKU, shown in vitro 37,40,41, in the brains of animal models of PKU (chemically-induced hyperphenylalaninemia rat model: ; genetic PKU mouse model BTBR: 44,45 ; C57BL/6: 46 ), and in plasma of patients Furthermore, antioxidant defenses are lower in PKU patients by means of lower activity of antioxidant enzymes, higher biomolecule damage and lower availability of antioxidants in blood 49,50, Synapse functioning Together with the previously discussed myelination problems, Bauman and Kemper 56 (1982) reported morphological changes in dendritic arborizations and reduced numbers of synaptic spines in untreated PKU patients. Likewise, in treated PKU patients, reduced synaptic or neuronal density is suggested because of the reduced grey matter found in specific regions of the brain 57. Generally, the altered synaptic functioning in PKU is demonstrated by 1) changes in synaptic morphology, 2) reduced expression of proteins involved in synaptic functioning, and 3) functional outcome changes (synaptic efficacy). The first aspect is supported with research, in vivo and in vitro, showing a decreased number of synapses, reduced width of the synaptic cleft, reduced thickness of the post-synaptic density, altered arborization of dendritic trees, and reduced neurite length 44, A causal role of Phe on these structural changes was found through the effect of Phe on the pathway important for dendritic elongation and arborization, F-actin remodeling of the cytoskeleton. High 82

84 Chapter 5 concentrations of Phe in vitro affects F-actin remodeling through reducing the expression of Cofilin, phosphorylation status of Cofilin and the activity of Ras-related C3 botulinum toxin substrate 1 (Rac1), GTPase being important for phosphorylation of Cofilin via LIM kinase 1 (LIMK1) 61,62. The morphological changes are shown by altered expression of the presynaptic markers synaptophysin and synaptosomal-associated protein-25, postsynaptic markers synaptopodin and spinopodin, synaptic functioning related proteins synapsin 2 and dihydropyrimidinase-related protein 2 6,60,63. Taken together, these changes result in functional outcome changes in long-term potentiation, a form of synaptic strengthening relevant to learning and memory, and the ability for the PKU BTBR mouse to master a learning and memory task 64. In summary, a nutritional intervention in PKU should preferably have the following features: a) increase neurotransmitter metabolism for serotonin and dopamine, b) improve white matter integrity, c) increase oxidative defenses, and d) improve synaptic functioning. In the next section, nutrients affecting these four domains will be discussed. 3. NUTRIENTS AFFECTING BRAIN STRUCTURE AND FUNCTION 3.1 Can nutritional intervention improve neurotransmitter metabolism for serotonin and dopamine? In PKU, the precursors of the neurotransmitter pathways and enzymes of the rate-limiting steps are affected. Although this is only a part of the synthesis pathway, all possible targets that can positively influence synthesis will be discussed (the effect of nutrients on serotonin is reviewed 65,66 ). Various nutrients can intervene in the synthesis pathway and release of neurotransmitters, for example vitamin D, vitamin B6, and omega-3 fatty acids. This is first highlighted by in vivo experiments that revealed that vitamin D can increase the expression of tryptophan hydroxylase 2 (TPH-2), the rate-limiting enzyme within serotonin synthesis 67. Second, the co-enzymatic form of vitamin B6, pyridoxal phosphate, is an important cofactor in the neurotransmitter pathways of serotonin but also in others such as melatonin, dopamine, norepinephrine, and gamma-aminobutyric acid (GABA) 68. Third, serotonin release from the pre-synapse can be inhibited by E2 series prostaglandins 69. These prostaglandins are produced from the unsaturated omega-6 fatty acid, arachidonic acid (AA), while the E3 series prostaglandins are produced from the omega-3 fatty acid eicosapentaenoic acid (EPA). The production of these prostaglandins is affected by the competition of EPA and AA at the cyclooxygenase pathway. A ratio in favor of EPA inhibits the production of E2 series prostaglandins from AA 70. This finding suggests that EPA supplementation, via inhibition of E2 series prostaglandins, could promote the release of serotonin from pre-synapses. Finally, fatty acids can affect cell membrane fluidity and consequently the receptor availability of e.g. serotonin and dopamine 71,72. In addition, DHA supplementation in rats, via micro-emulsions with linseed oil, resulted in an increase in serotonin and dopamine concentrations in the 5 83

85 Chapter 5 brain 73. Omega-3 fatty acid supplementation also had a positive effect on dopamine in a rat model of Parkinson s disease 74 and prevented dopamine-deficits induced by food allergies in mice 75. In the first study, supplementation of DHA and/or uridine-5 -monophosphate (UMP) increased dopamine concentrations and the activity of Tyr hydroxylase, the rate-limiting step in dopamine synthesis. Interestingly, the combination of DHA and UMP increased the beneficial effects even more, which indicates that the combination of certain nutrients could increase the positive effects of a single nutrient. In conclusion, nutrients, such as vitamin B6, vitamin D, fatty acids, and UMP can mediate neurotransmitter metabolism by regulating neurotransmitter uptake, synthesis and release. However, research with these nutrients relating to neurotransmitter concentrations in PKU is very limited. Merely one study examined the effect of B6 treatment on serotonin in PKU patients 76. In this study, a seven days B6 treatment in non-treated PKU patients did not significantly change blood serotonin concentration. Although this study did not show positive results, more PKU specific studies are needed for definite conclusions. In light of the clear beneficial effects of nutrients in other research fields, a combination of different nutrients could yield positive effect(s) in neurotransmitter metabolism in PKU. 3.2 Can a nutritional intervention improve white matter integrity? In general, various nutrients are important for myelination and myelin recovery. For example, iron, B6, and B12 deficiencies can cause problems with myelination 77 79, omega-3 fatty acid supplementation supported myelination in a model of traumatic brain injury 80, vitamin D3 increased myelination in a model of nerve injury 81, and in a chronic cerebral hypoperfusion model, L-carnitine led to increased myelination 82. These findings highlight the possible positive effect of several nutrients on myelination. However, the implications of these studies for PKU are not yet clear, as in these studies improved myelination is often expressed by myelin sheath thickness or myelin markers 82. In PKU, these parameters have not been measured. A measure that is affected in PKU is cholesterol and protein synthesis, which is considered essential in myelination. Therefore, selectively increasing e.g. cholesterol biosynthesis by increasing the activity the rate-controlling enzyme, HMGR, could yield possible treatment strategies for PKU. Possible candidates are omega-3 fatty acids as treatment of one and two weeks in glial cells increased HMGR activity in primary glial cell cultures 83. To conclude, various nutrients are important for myelination. However, the significance of these nutrients for this specific domain in PKU patients is not clear yet. 3.3 Can a nutritional intervention restore oxidative balance or increase oxidative defenses? Restoring the balance between reactive species and antioxidants by supplementing with antioxidants or their precursors could have beneficial effects for PKU patients. Several studies examined antioxidant treatments in PKU models. Moraes and coworkers extensively examined the positive effects of lipoic acid in PKU in in vitro and in vivo experiments 40,84. They showed in a chemically induced hyperphenylalaninemia rat model (repetitive injections of Phe 84

86 Chapter 5 and alpha-methyl-phe as PAH inhibitor) as well as in Phe-treated brain homogenates of rats, that lipoic acid administration had a positive effect on antioxidant enzyme activity and nonenzymatic antioxidant systems that were impaired by Phe 40,84. Furthermore, in another acute hyperphenylalaninemia model in which rodents received intracerebroventricular injections of Phe before training in a behavioral paradigm, Phe changed the behavior and increased oxidative stress. These effects could be counteracted by pretreatment with creatine and/or pyruvate intraperitoneally 85,86. Finally, in a rat model of maternal hyperphenylalaninemia, vitamin C, vitamin E, as well as melatonin showed positive effects on parameters of oxidative stress caused by Phe 87,88. In PKU patients, L-carnitine and selenium had a beneficial effect on the oxidative stress parameters 89. In conclusion, increasing the availability of antioxidants may improve patients oxidative defenses, thereby reducing oxidative stress. 3.4 Can a nutritional intervention improve synaptic functioning? The use of a nutritional intervention to improve synaptic functioning is often discussed in relation to early life interventions by either replenishing deficiencies if these occur - or supplementing certain nutrients during pregnancy up to early adolescence 90. For instance, supplementation of vitamin B12 during pregnancy up to three months increased brain-derived neurotrophic factor (BDNF) gene expression and protein concentrations in the hippocampus of male Wistar rats 91. When adding omega 3-fatty acids to the same supplementation period, the BDNF gene expression and protein concentration were significantly higher in the hippocampus compared to the vitamin B12 supplementation group, and it increased the BDNF concentrations also in the cortex 91. Furthermore, supplementation of DHA and/or uridine during gestation and early-life positively affected pre- and postsynaptic markers as well as hippocampal dendritic spine density 92. Nutrient supplementation studies also show benefits later in life. This finding is illustrated by research showing that supplementation of DHA increases dendritic spine density in the hippocampus in adult gerbils 93. Supplementation of DHA, EPA, UMP or a combination of DHA with UMP and EPA with UMP also increased pre- and postsynaptic protein expression in adult gerbils 94. On a functional level, in vitro experiments in hippocampal primary cultures have shown that DHA supplementation could increase synaptic activity 95. In conclusion, supplementation of especially DHA, EPA, and UMP solely or in combination can have a positive effect on synaptic functioning REVIEWING EXISTING LITERATURE IN PKU At the beginning of this review, four domains were introduced that are thought to contribute to the cognitive deficits observed in PKU patients. However, these domains are not separate entities but are highly interrelated, each contributing to the overall cognitive functioning of an individual. Apart from oxidative stress, research specifically focused on the effect of nutrients in PKU is very limited. Some studies examined the effects of nutrient supplementation on the overall cognitive functioning of PKU patients Most of these studies have focused on DHA 85

87 Chapter 5 and/or EPA supplementation and find inconclusive results. Some studies report beneficial effects after supplementation on coordination, fine motor skills, and EEG measurements of processing visual stimuli Others do not find beneficial effects of DHA supplementation on cognitive processing speed and executive functioning 99. Several arguments could explain why these results are contradicting. Firstly, the duration of treatment is of great importance. Recuperation from abnormalities as a result from omega-3 fatty acid deprivation requires weeks to months for different cell types and organelles of the rodent brain 100. The possible transcending positive effect of these fatty acids in PKU could take even longer. Secondly, the targeted patient group is always key in the design of the study. In healthy subjects, omega-3 fatty acid supplementation increases cognitive performance in infants but not in other age groups 101. In PKU, different patient groups can be identified, e.g. early treated, compliant, non-compliant, and recently a new group is developing: early-treated PKU patients in latelife. This group is new because screening and early dietary treatment of patients was initiated approximately years ago, depending on the country. Therefore, age-differences and different treatment-histories in PKU patients should be taken into consideration when designing an interventional study with a combination of specific nutrients. Finally, providing a combination of specific nutrients or multiple nutrients in one pathway could elicit an effect that cannot be seen when only a single nutrient is offered 91, PERSPECTIVES OF A NUTRITIONAL INTERVENTION FOR PKU The beneficial effects of the various nutrients previously described could also be discussed in light of the effect on neuronal membrane, one of the sites at which the four identified domains in PKU converge. The neuronal membranes are the principal site of action for many neuronal activities, such as synaptic functioning (including receptor and ion channel activity), neurotransmitter release and optimal exchange of nutrients and other molecules 102,103. In addition, myelin is largely made up of lipids that can be derived from the same precursors as those that make up the neuronal membrane. Finally, antioxidants protect the neuronal membrane from oxidative stress, thus maintaining its integrity, stability, and ultimately its function. Neuronal membranes are mainly composed of phospholipids, the most abundant of which are generated primarily via the well-characterized metabolic pathway known as the Kennedy pathway 11. The rate of phospholipid synthesis through the Kennedy pathway depends on the availability of three dietary precursors; UMP, choline and the omega-3 polyunsaturated fatty acid DHA 11. Each precursor, on its own, is rate limiting. Furthermore, in addition to numerous other supportive roles of brain function, certain B-vitamins, more specifically folic acid and vitamins B6 and B12, act as cofactors by directly affecting liver metabolism of DHA and choline, thereby increasing their availability for membrane phospholipid synthesis 12. Furthermore, vitamin C, E, and selenium play critical roles in protecting the neuronal membrane through their antioxidant properties. Finally, 86

88 Chapter 5 phospholipids exert their supportive role through increasing precursor availability, by acting as direct precursors to choline 104 and by increasing intestinal absorption of DHA and EPA 105. This particular combination of nutrients named Fortasyn Connect, containing DHA, EPA, UMP, choline, folate, vitamin B12, vitamin B6, phospholipids, vitamin C, vitamin E, and selenium has been shown to positively affect brain phospholipid synthesis 11,93,106,107, neurite outgrowth 108,109, the number of dendritic spines 92,93, neurotransmitter release and signaling 109,110, levels of pre- or post-synaptic protein 11,93,94,109, and levels of phospholipid also found in myelin 109. Moreover, this combination of nutrients has improved white matter integrity in an animal model of Alzheimer s disease 111. In human subjects, different measures can be used as a proxy to study functioning of the brain, and more specifically functioning of synapses. Clinical studies in the context of Alzheimer s disease have shown that Fortasyn Connect leads to improved brain function, and more importantly improvements in memory performance 112,113. An EEG study in humans substantiates improvements in synaptic connectivity through demonstrated preservation of functional brain connectivity and brain network organization 114. All these studies together suggest that Fortasyn Connect could be of great interest to address cognitive deficits in PKU patients. Testing this combination of nutrients in a PKU mouse model on the level of the four different domains and cognitive performance is now warranted. The first proof-of-concept study has indicated that Fortasyn Connect could be beneficial for the expression of a postsynaptic marker in a specific region in the hippocampus in the PKU mouse model 115. This indicates that Fortasyn Connect could be beneficial for at least of the domains, namely synaptic functioning. These results warrant additional exploration of the effects of this specific nutrient combination on outcome parameters in PKU. 5 Conflict of Interest (COI) Statement: This manuscript was written by researchers employed by the University of Groningen and the Beatrix Children s hospital, University Medical Center Groningen. Both institutions received funding from Nutricia Advanced Medical Nutrition, Utrecht, the Netherlands. Three of the co-authors are currently working at this company (DSC, MR, and MK). 87

89 Chapter 5 Table 1: Summary of single nutrient interventions. Limited studies examined single nutrient interventions in PKU (depicted in the first row). Therefore, single nutrients that have beneficial effects in the healthy or diseased brain other than PKU are included in the table (divided in three domains). As various studies did examine the effect of antioxidants on oxidative stress in PKU, other literature concerning antioxidants and oxidative stress are not included. PKU related literature Omega-3 fatty acids UMP Iron Vitamin B6 Vitamin B12 Vitamin D L-carnitine Antioxidants Did not significantly change blood serotonin concentration 76 Beneficial effect for oxidative stress parameters 89 Beneficial effect for oxidative stress parameters 40,84 89 Neurotransmitter metabolism Show positive effects on dopamine and serotonin Shows a positive effect on dopamine 75 Co-enzymatic form of vitamin B6 is an important cofactor in the synthesis pathway of serotonin 68 Influences synthesis pathway 67 White matter integrity Prevented myelination Deficiencies problems 80 result in complications 77 Deficiencies result in complications 78 Deficiencies result in complications 79 Increases Increased myelination 81 myelination 82 Synaptic functioning Increased BDNF, pre-and postsynaptic markers, hippocampal dendritic spine density, and synaptic activity Increased pre-and postsynaptic markers, and hippocampal dendritic spine density 11,94 Increased BDNF 91 88

90 Chapter 5 6. REFERENCES 1 Weiser MJ, Butt CM, Mohajeri MH. Docosahexaenoic Acid and Cognition throughout the Lifespan. Nutrients 2016; 8. doi: /nu Forbes SC, Holroyd-Leduc JM, Poulin MJ, Hogan DB. Effect of Nutrients, Dietary Supplements and Vitamins on Cognition: a Systematic Review and Meta-Analysis of Randomized Controlled Trials. Can Geriatr J 2015; 18: Burlina AB, Bonafé L, Ferrari V, Suppiej A, Zacchello F, Burlina AP. Measurement of neurotransmitter metabolites in the cerebrospinal fluid of phenylketonuric patients under dietary treatment. J Inherit Metab Dis 2000; 23: Anderson PJ, Leuzzi V. White matter pathology in phenylketonuria. Mol Genet Metab 2010; 99 Suppl 1: S Ribas GS, Sitta A, Wajner M, Vargas CR. Oxidative stress in phenylketonuria: what is the evidence? Cell Mol Neurobiol 2011; 31: Horling K, Schlegel G, Schulz S, Vierk R, Ullrich K, Santer R et al. Hippocampal synaptic connectivity in phenylketonuria. Hum Mol Genet doi: /hmg/ ddu Waisbren SE, Noel K, Fahrbach K, Cella C, Frame D, Dorenbaum A et al. Phenylalanine blood levels and clinical outcomes in phenylketonuria: a systematic literature review and meta-analysis. Mol Genet Metab 2007; 92: Moyle JJ, Fox AM, Arthur M, Bynevelt M, Burnett JR. Meta-analysis of neuropsychological symptoms of adolescents and adults with PKU. Neuropsychol Rev 2007; 17: Enns GM, Koch R, Brumm V, Blakely E, Suter R, Jurecki E. Suboptimal outcomes in patients with PKU treated early with diet alone: revisiting the evidence. Mol Genet Metab; 101: MacDonald A, Gokmen-Ozel H, van Rijn M, Burgard P. The reality of dietary compliance in the management of phenylketonuria. J Inherit Metab Dis 2010; 33: Wurtman RJ, Cansev M, Sakamoto T, Ulus IH. Administration of docosahexaenoic acid, uridine and choline increases levels of synaptic membranes and dendritic spines in rodent brain. World Rev Nutr Diet 2009; 99: van Wijk N, Watkins CJ, Hageman RJJ, Sijben JCW, Kamphuis PGHJ, Wurtman RJ et al. Combined dietary folate, vitamin B-12, and vitamin B-6 intake influences plasma docosahexaenoic acid concentration in rats. Nutr Metab (Lond) 2012; 9: Kennedy EP WS. The function of cytidine coenzymes in the biosynthesis of phospholipides. J Biol Chem 1956; 222: van Spronsen FJ, Hoeksma M, Reijngoud D-J. Brain dysfunction in phenylketonuria: is phenylalanine toxicity the only possible cause? J Inherit Metab Dis 2009; 32: de Groot MJ, Sijens PE, Reijngoud D-J, Paans AM, van Spronsen FJ. Phenylketonuria: brain phenylalanine concentrations relate inversely to cerebral protein synthesis. J Cereb Blood Flow Metab 2015; 35: Ogawa S, Ichinose H. Effect of metals and phenylalanine on the activity of human tryptophan hydroxylase-2: comparison with that on tyrosine hydroxylase activity. Neurosci Lett 2006; 401:

91 Chapter 5 17 Pascucci T, Andolina D, Mela I La, Conversi D, Latagliata C, Ventura R et al. 5-Hydroxytryptophan rescues serotonin response to stress in prefrontal cortex of hyperphenylalaninaemic mice. Int J Neuropsychopharmacol 2009; 12: Pascucci T, Andolina D, Ventura R, Puglisi- Allegra S, Cabib S. Reduced availability of brain amines during critical phases of postnatal development in a genetic mouse model of cognitive delay. Brain Res 2008; 1217: Pascucci T, Ventura R, Puglisi-Allegra S, Cabib S. Deficits in brain serotonin synthesis in a genetic mouse model of phenylketonuria. Neuroreport 2002; 13: Puglisi-Allegra S, Cabib S, Pascucci T, Ventura R, Cali F, Romano V. Dramatic brain aminergic deficit in a genetic mouse model of phenylketonuria. Neuroreport 2000; 11: van Vliet D, Bruinenberg VM, Mazzola PN, van Faassen MHJR, de Blaauw P, Kema IP et al. Large Neutral Amino Acid Supplementation Exerts Its Effect through Three Synergistic Mechanisms: Proof of Principle in Phenylketonuria Mice. PLoS One 2015; 10: e ALVORD EC, STEVENSON LD, VOGEL FS, ENGLE RL. Neuropathological findings in phenyl-pyruvic oligophrenia (phenylketonuria). J Neuropathol Exp Neurol 1950; 9: Huttenlocher PR. The neuropathology of phenylketonuria: human and animal studies. Eur J Pediatr 2000; 159 Suppl: S Hood A, Antenor-Dorsey JA V, Hershey T, Rutlin J, Shimony JS, McKinstry RC et al. White matter integrity and executive abilities in individuals with phenylketonuria. Mol Genet Metab 2013; 109: Anderson PJ, Wood SJ, Francis DE, Coleman L, Warwick L, Casanelia S et al. Neuropsychological functioning in children with early-treated phenylketonuria: impact of white matter abnormalities. Dev Med Child Neurol 2004; 46: Bick U, Ullrich K, Stöber U, Möller H, Schuierer G, Ludolph AC et al. White matter abnormalities in patients with treated hyperphenylalaninaemia: magnetic resonance relaxometry and proton spectroscopy findings. Eur J Pediatr 1993; 152: Leuzzi V, Tosetti M, Montanaro D, Carducci C, Artiola C, Antonozzi I et al. The pathogenesis of the white matter abnormalities in phenylketonuria. A multimodal 3.0 tesla MRI and magnetic resonance spectroscopy (1H MRS) study. J Inherit Metab Dis 2007; 30: Mastrangelo M, Chiarotti F, Berillo L, Caputi C, Carducci C, Di Biasi C et al. The outcome of white matter abnormalities in early treated phenylketonuric patients: A retrospective longitudinal long-term study. Mol Genet Metab doi: /j. ymgme Dyer CA, Kendler A, Philibotte T, Gardiner P, Cruz J, Levy HL. Evidence for central nervous system glial cell plasticity in phenylketonuria. J Neuropathol Exp Neurol 1996; 55: Schoemans R, Aigrot M-S, Wu C, Marée R, Hong P, Belachew S et al. Oligodendrocyte development and myelinogenesis are not impaired by high concentrations of phenylalanine or its metabolites. J Inherit Metab Dis 2010; 33: Shefer S, Tint GS, Jean-Guillaume D, Daikhin E, Kendler A, Nguyen LB et al. Is there a relationship between 3-hydroxy- 3-methylglutaryl coenzyme a reductase activity and forebrain pathology in the PKU mouse? J Neurosci Res 2000; 61:

92 Chapter 5 32 Smith CB, Kang J. Cerebral protein synthesis in a genetic mouse model of phenylketonuria. Proc Natl Acad Sci U S A 2000; 97: Hoeksma M, Reijngoud D-J, Pruim J, de Valk HW, Paans AMJ, van Spronsen FJ. Phenylketonuria: High plasma phenylalanine decreases cerebral protein synthesis. Mol Genet Metab 2009; 96: Hughes J V, Johnson TC. The effects of phenylalanine on amino acid metabolism and protein synthesis in brain cells in vitro. J Neurochem 1976; 26: Binek-Singer P, Johnson TC. The effects of chronic hyperphenylalaninaemia on mouse brain protein synthesis can be prevented by other amino acids. Biochem J 1982; 206: Halliwell B GJ. Free radicals in biology and medicine Fernandes CG, Leipnitz G, Seminotti B, Amaral AU, Zanatta A, Vargas CR et al. Experimental evidence that phenylalanine provokes oxidative stress in hippocampus and cerebral cortex of developing rats. Cell Mol Neurobiol 2010; 30: Mazumder MK, Paul R, Borah A. β-phenethylamine--a phenylalanine derivative in brain--contributes to oxidative stress by inhibiting mitochondrial complexes and DT-diaphorase: an in silico study. CNS Neurosci Ther 2013; 19: Rosa AP, Jacques CED, Moraes TB, Wannmacher CMD, Dutra A de M, Dutra- Filho CS. Phenylpyruvic acid decreases glucose-6-phosphate dehydrogenase activity in rat brain. Cell Mol Neurobiol 2012; 32: Moraes TB, Zanin F, da Rosa A, de Oliveira A, Coelho J, Petrillo F et al. Lipoic acid prevents oxidative stress in vitro and in vivo by an acute hyperphenylalaninemia chemically-induced in rat brain. J Neurol Sci 2010; 292: Kienzle Hagen ME, Pederzolli CD, Sgaravatti AM, Bridi R, Wajner M, Wannmacher CMD et al. Experimental hyperphenylalaninemia provokes oxidative stress in rat brain. Biochim Biophys Acta 2002; 1586: Simon KR, Dos Santos RM, Scaini G, Leffa DD, Damiani AP, Furlanetto CB et al. DNA damage induced by phenylalanine and its analogue p-chlorophenylalanine in blood and brain of rats subjected to a model of hyperphenylalaninemia. Biochem Cell Biol 2013; 91: Mazzola PN, Terra M, Rosa AP, Mescka CP, Moraes TB, Piccoli B et al. Regular exercise prevents oxidative stress in the brain of hyperphenylalaninemic rats. Metab Brain Dis 2011; 26: Liang L, Gu X, Lu L, Li D, Zhang X. Phenylketonuria-related synaptic changes in a BTBR-Pah(enu2) mouse model. Neuroreport 2011; 22: Ercal N, Aykin-Burns N, Gurer-Orhan H, McDonald JD. Oxidative stress in a phenylketonuria animal model. Free Radic Biol Med 2002; 32: Solverson P, Murali SG, Brinkman AS, Nelson DW, Clayton MK, Yen C-LE et al. Glycomacropeptide, a low-phenylalanine protein isolated from cheese whey, supports growth and attenuates metabolic stress in the murine model of phenylketonuria. Am J Physiol Endocrinol Metab 2012; 302: E

93 Chapter 5 47 Sitta A, Barschak AG, Deon M, Barden AT, Biancini GB, Vargas PR et al. Effect of short- and long-term exposition to high phenylalanine blood levels on oxidative damage in phenylketonuric patients. Int J Dev Neurosci 2009; 27: Gassió R, Artuch R, Vilaseca MA, Fusté E, Colome R, Campistol J. Cognitive functions and the antioxidant system in phenylketonuric patients. Neuropsychology 2008; 22: Schulpis KH, Tsakiris S, Traeger-Synodinos J, Papassotiriou I. Low total antioxidant status is implicated with high 8-hydroxy- 2-deoxyguanosine serum concentrations in phenylketonuria. Clin Biochem 2005; 38: Artuch R, Colomé C, Sierra C, Brandi N, Lambruschini N, Campistol J et al. A longitudinal study of antioxidant status in phenylketonuric patients. Clin Biochem 2004; 37: Sierra C, Vilaseca MA, Moyano D, Brandi N, Campistol J, Lambruschini N et al. Antioxidant status in hyperphenylalaninemia. Clin Chim Acta 1998; 276: Wilke BC, Vidailhet M, Favier A, Guillemin C, Ducros V, Arnaud J et al. Selenium, glutathione peroxidase (GSH-Px) and lipid peroxidation products before and after selenium supplementation. Clin Chim Acta 1992; 207: Schulpis KH, Papastamataki M, Stamou H, Papassotiriou I MA. The effect of diet on total antioxidant status, ceruloplasmin, transferrin and ferritin serum levels in phenylketonuric children. - PubMed - NCBI pubmed/ (accessed 3 Sep2015). 54 Colomé C, Artuch R, Vilaseca M-A, Sierra C, Brandi N, Lambruschini N et al. Lipophilic antioxidants in patients with phenylketonuria. Am J Clin Nutr 2003; 77: van Bakel MM, Printzen G, Wermuth B, Wiesmann UN. Antioxidant and thyroid hormone status in selenium-deficient phenylketonuric and hyperphenylalaninemic patients. Am J Clin Nutr 2000; 72: Bauman ML, Kemper TL. Morphologic and histoanatomic observations of the brain in untreated human phenylketonuria. Acta Neuropathol 1982; 58: Christ SE, Price MH, Bodner KE, Saville C, Moffitt AJ, Peck D. Morphometric analysis of gray matter integrity in individuals with early-treated phenylketonuria. Mol Genet Metab 2016; 118: Andolina D, Conversi D, Cabib S, Trabalza A, Ventura R, Puglisi-Allegra S et al. 5-Hydroxytryptophan during critical postnatal period improves cognitive performances and promotes dendritic spine maturation in genetic mouse model of phenylketonuria. Int J Neuropsychopharmacol 2011; 14: Cordero ME, Trejo M, Colombo M, Aranda V. Histological maturation of the neocortex in phenylketonuric rats. Early Hum Dev 1983; 8: Hörster F, Schwab MA, Sauer SW, Pietz J, Hoffmann GF, Okun JG et al. Phenylalanine reduces synaptic density in mixed cortical cultures from mice. Pediatr Res 2006; 59: Schlegel G, Scholz R, Ullrich K, Santer R, Rune GM. Phenylketonuria: Direct and indirect effects of phenylalanine. Exp Neurol 2016; 281:

94 Chapter 5 62 Zhang Y, Zhang H, Yuan X, Gu X. Differential effects of phenylalanine on Rac1, Cdc42, and RhoA expression and activity in cultured cortical neurons. Pediatr Res 2007; 62: Imperlini E, Orrù S, Corbo C, Daniele A, Salvatore F. Altered brain protein expression profiles are associated with molecular neurological dysfunction in the PKU mouse model. J Neurochem 2014; 129: Cabib S, Pascucci T, Ventura R, Romano V, Puglisi-Allegra S. The behavioral profile of severe mental retardation in a genetic mouse model of phenylketonuria. Behav Genet 2003; 33: Patrick RP, Ames BN. Vitamin D and the omega-3 fatty acids control serotonin synthesis and action, part 2: relevance for ADHD, bipolar disorder, schizophrenia, and impulsive behavior. FASEB J 2015; 29: Patrick RP, Ames BN. Vitamin D hormone regulates serotonin synthesis. Part 1: relevance for autism. FASEB J 2014; 28: Kaneko I, Sabir MS, Dussik CM, Whitfield GK, Karrys A, Hsieh J-C et al. 1,25-Dihydroxyvitamin D regulates expression of the tryptophan hydroxylase 2 and leptin genes: implication for behavioral influences of vitamin D. FASEB J 2015; 29: Shabbir F, Patel A, Mattison C, Bose S, Krishnamohan R, Sweeney E et al. Effect of diet on serotonergic neurotransmission in depression. Neurochem Int 2013; 62: Schlicker E, Fink K, Göthert M. Influence of eicosanoids on serotonin release in the rat brain: inhibition by prostaglandins E1 and E2. Naunyn Schmiedebergs Arch Pharmacol 1987; 335: Wada M, DeLong CJ, Hong YH, Rieke CJ, Song I, Sidhu RS et al. Enzymes and receptors of prostaglandin pathways with arachidonic acid-derived versus eicosapentaenoic acid-derived substrates and products. J Biol Chem 2007; 282: Heron DS, Shinitzky M, Hershkowitz M, Samuel D. Lipid fluidity markedly modulates the binding of serotonin to mouse brain membranes. Proc Natl Acad Sci U S A 1980; 77: Heinrichs SC. Dietary omega-3 fatty acid supplementation for optimizing neuronal structure and function. Mol Nutr Food Res 2010; 54: Sugasini D, Lokesh BR. Rats given linseed oil in microemulsion forms enriches the brain synaptic membrane with docosahexaenoic acid and enhances the neurotransmitter levels in the brain. Nutr Neurosci 2015; 18: Cansev M, Ulus IH, Wang L, Maher TJ, Wurtman RJ. Restorative effects of uridine plus docosahexaenoic acid in a rat model of Parkinson s disease. Neurosci Res 2008; 62: de Theije CGM, van den Elsen LWJ, Willemsen LEM, Milosevic V, Korte-Bouws GAH, Lopes da Silva S et al. Dietary long chain n-3 polyunsaturated fatty acids prevent impaired social behaviour and normalize brain dopamine levels in food allergic mice. Neuropharmacology 2015; 90: Berman JL, Justice P, Hsia DY. Effect of vitamin B 6 on blood 5-hydroxytryptamine concentration. Ann N Y Acad Sci 1969; 166: Badaracco ME, Siri MVR, Pasquini JM. Oligodendrogenesis: the role of iron. Biofactors 2010; 36:

95 Chapter 5 78 Kirksey A, Morré DM, Wasynczuk AZ. Neuronal development in vitamin B6 deficiency. Ann N Y Acad Sci 1990; 585: Agamanolis DP, Chester EM, Victor M, Kark JA, Hines JD, Harris JW. Neuropathology of experimental vitamin B12 deficiency in monkeys. Neurology 1976; 26: Pu H, Guo Y, Zhang W, Huang L, Wang G, Liou AK et al. Omega-3 polyunsaturated fatty acid supplementation improves neurologic recovery and attenuates white matter injury after experimental traumatic brain injury. J Cereb Blood Flow Metab 2013; 33: Chabas J-F, Stephan D, Marqueste T, Garcia S, Lavaut M-N, Nguyen C et al. Cholecalciferol (vitamin D₃) improves myelination and recovery after nerve injury. PLoS One 2013; 8: e Ueno Y, Koike M, Shimada Y, Shimura H, Hira K, Tanaka R et al. L-carnitine enhances axonal plasticity and improves white-matter lesions after chronic hypoperfusion in rat brain. J Cereb Blood Flow Metab 2015; 35: Vignikin R, Grundt IK, Scotto J, Wolfrom C, Raulin J, Gautier M. Effects of polyunsaturated fatty acids on human and rat cells. In vitro versus in vivo experiments. In Vivo 1989; 3: Moraes TB, Jacques CED, Rosa AP, Dalazen GR, Terra M, Coelho JG et al. Role of catalase and superoxide dismutase activities on oxidative stress in the brain of a phenylketonuria animal model and the effect of lipoic acid. Cell Mol Neurobiol 2013; 33: Berti SL, Nasi GM, Garcia C, Castro FL de, Nunes ML, Rojas DB et al. Pyruvate and creatine prevent oxidative stress and behavioral alterations caused by phenylalanine administration into hippocampus of rats. Metab Brain Dis 2012; 27: Dos Reis EA, Rieger E, de Souza SS, Rasia- Filho AA, Wannmacher CMD. Effects of a co-treatment with pyruvate and creatine on dendritic spines in rat hippocampus and posterodorsal medial amygdala in a phenylketonuria animal model. Metab Brain Dis 2013; 28: Martínez-Cruz F, Osuna C, Guerrero JM. Mitochondrial damage induced by fetal hyperphenylalaninemia in the rat brain and liver: its prevention by melatonin, Vitamin E, and Vitamin C. Neurosci Lett 2006; 392: Martinez-Cruz F, Pozo D, Osuna C, Espinar A, Marchante C, Guerrero JM. Oxidative stress induced by phenylketonuria in the rat: Prevention by melatonin, vitamin E, and vitamin C. J Neurosci Res 2002; 69: Sitta a, Vanzin CS, Biancini GB, Manfredini V, de Oliveira a B, Wayhs C a Y et al. Evidence that L-carnitine and selenium supplementation reduces oxidative stress in phenylketonuric patients. Cell Mol Neurobiol 2011; 31: Prado EL, Dewey KG. Nutrition and brain development in early life. Nutr Rev 2014; 72: Rathod R, Khaire A, Kemse N, Kale A, Joshi S. Maternal omega-3 fatty acid supplementation on vitamin B12 rich diet improves brain omega-3 fatty acids, neurotrophins and cognition in the Wistar rat offspring. Brain Dev 2014; 36:

96 Chapter 5 92 Cansev M, Marzloff G, Sakamoto T, Ulus IH, Wurtman RJ. Giving uridine and/or docosahexaenoic acid orally to rat dams during gestation and nursing increases synaptic elements in brains of weanling pups. Dev Neurosci 2009; 31: Sakamoto T, Cansev M, Wurtman RJ. Oral supplementation with docosahexaenoic acid and uridine-5 -monophosphate increases dendritic spine density in adult gerbil hippocampus. Brain Res 2007; 1182: Cansev M, Wurtman RJ. Chronic administration of docosahexaenoic acid or eicosapentaenoic acid, but not arachidonic acid, alone or in combination with uridine, increases brain phosphatide and synaptic protein levels in gerbils. Neuroscience 2007; 148: Cao D, Kevala K, Kim J, Moon H-S, Jun SB, Lovinger D et al. Docosahexaenoic acid promotes hippocampal neuronal development and synaptic function. J Neurochem 2009; 111: Koletzko B, Beblo S, Demmelmair H, Hanebutt FL. Omega-3 LC-PUFA supply and neurological outcomes in children with phenylketonuria (PKU). J Pediatr Gastroenterol Nutr 2009; 48 Suppl 1: S Beblo S, Reinhardt H, Demmelmair H, Muntau AC, Koletzko B. Effect of fish oil supplementation on fatty acid status, coordination, and fine motor skills in children with phenylketonuria. J Pediatr 2007; 150: Agostoni C, Verduci E, Massetto N, Fiori L, Radaelli G, Riva E et al. Long term effects of long chain polyunsaturated fats in hyperphenylalaninemic children. Arch Dis Child 2003; 88: Yi SHL, Kable JA, Evatt ML, Singh RH. A randomized, placebo-controlled, doubleblind trial of supplemental docosahexaenoic acid on cognitive processing speed and executive function in females of reproductive age with phenylketonuria: A pilot study. Prostaglandins Leukot Essent Fatty Acids 2011; 85: Bourre JM, Durand G, Pascal G, Youyou A. Brain cell and tissue recovery in rats made deficient in n-3 fatty acids by alteration of dietary fat. J Nutr 1989; 119: Jiao J, Li Q, Chu J, Zeng W, Yang M, Zhu S. Effect of n-3 PUFA supplementation on cognitive function throughout the life span from infancy to old age: a systematic review and meta-analysis of randomized controlled trials. Am J Clin Nutr 2014; 100: Yehuda S, Rabinovitz S, Carasso RL, Mostofsky DI. The role of polyunsaturated fatty acids in restoring the aging neuronal membrane. Neurobiol Aging; 23: Vigh L, Escribá P V, Sonnleitner A, Sonnleitner M, Piotto S, Maresca B et al. The significance of lipid composition for membrane activity: new concepts and ways of assessing function. Prog Lipid Res 2005; 44: Wurtman RJ, Hirsch MJ, Growdon JH. Lecithin consumption raises serum-freecholine levels. Lancet (London, England) 1977; 2: O Doherty PJ, Kakis G, Kuksis A. Role of luminal lecithin in intestinal fat absorption. Lipids 1973; 8: Cansev M, Watkins CJ, van der Beek EM, Wurtman RJ. Oral uridine-5 - monophosphate (UMP) increases brain CDP-choline levels in gerbils. Brain Res 2005; 1058:

97 Chapter Holguin S, Martinez J, Chow C, Wurtman R. Dietary uridine enhances the improvement in learning and memory produced by administering DHA to gerbils. FASEB J 2008; 22: Pooler AM, Guez DH, Benedictus R, Wurtman RJ. Uridine enhances neurite outgrowth in nerve growth factor-differentiated PC12 [corrected]. Neuroscience 2005; 134: Cansev M, van Wijk N, Turkyilmaz M, Orhan F, Sijben JWC, Broersen LM. A specific multi-nutrient enriched diet enhances hippocampal cholinergic transmission in aged rats. Neurobiol Aging 2015; 36: Savelkoul PJM, Janickova H, Kuipers AAM, Hageman RJJ, Kamphuis PJ, Dolezal V et al. A specific multi-nutrient formulation enhances M1 muscarinic acetylcholine receptor responses in vitro. J Neurochem 2012; 120: Zerbi V, Jansen D, Wiesmann M, Fang X, Broersen LM, Veltien A et al. Multinutrient diets improve cerebral perfusion and neuroprotection in a murine model of Alzheimer s disease. Neurobiol Aging 2014; 35: Scheltens P, Kamphuis PJGH, Verhey FRJ, Olde Rikkert MGM, Wurtman RJ, Wilkinson D et al. Efficacy of a medical food in mild Alzheimer s disease: A randomized, controlled trial. Alzheimers Dement 2010; 6: 1 10.e Scheltens P, Twisk JWR, Blesa R, Scarpini E, von Arnim CAF, Bongers A et al. Efficacy of Souvenaid in mild Alzheimer s disease: results from a randomized, controlled trial. J Alzheimers Dis 2012; 31: e Waal H, Stam CJ, Lansbergen MM, Wieggers RL, Kamphuis PJGH, Scheltens P et al. The effect of souvenaid on functional brain network organisation in patients with mild Alzheimer s disease: a randomised controlled study. PLoS One 2014; 9: e Bruinenberg VM, van Vliet D, Attali A, de Wilde MC, Kuhn M, van Spronsen FJ et al. A Specific Nutrient Combination Attenuates the Reduced Expression of PSD-95 in the Proximal Dendrites of Hippocampal Cell Body Layers in a Mouse Model of Phenylketonuria. Nutrients 2016; 8. doi: /nu

98 Chapter

99

100 CHAPTER 6 A specific nutrient combination attenuates the reduced expression of PSD-95 in the proximal dendrites of hippocampal cell body layers in a mouse model of phenylketonuria Vibeke M. Bruinenberg 1, Danique van Vliet 2, Amos Attali 3, Martijn C. de Wilde 3, Mirjam Kuhn 3, Francjan J. van Spronsen 2, Eddy A. van der Zee 1, * 1 Molecular Neurobiology, University of Groningen, Groningen, the Netherlands, 2 Division of Metabolic Diseases, Beatrix Children s Hospital, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands 3 Nutricia Research, Nutricia Advanced Medical Nutrition, Utrecht, the Netherlands *Correspondence: e.a.van.der.zee@rug.nl; Tel.:

101 Chapter 6 ABSTRACT The inherited metabolic disease phenylketonuria (PKU) is characterized by increased concentrations of phenylalanine in the blood and brain, and as a consequence neurotransmitter metabolism, white matter, and synapse functioning are affected. A specific nutrient combination (SNC) has been shown to improve synapse formation, morphology and function. This could become an interesting new nutritional approach for PKU. To assess whether treatment with SNC can affect synapses, we treated PKU mice with SNC or an isocaloric control diet for 12 weeks, starting at postnatal day 31. Immunostaining for postsynaptic density protein 95 (PSD-95), a postsynaptic density marker, was carried out in the hippocampus, striatum and prefrontal cortex. Compared to WT mice on normal chow without SNC, PKU mice on the isocaloric control showed a significant reduction in PSD-95 expression in the hippocampus, specifically in the granular cell layer of the dentate gyrus, with a similar trend seen in the CA1 and CA3 pyramidal cell layer. No differences were found in the striatum or prefrontal cortex. PKU mice on a diet supplemented with SNC showed improved expression of PSD-95 in the hippocampus. This study gives the first indication that SNC supplementation has a positive effect on hippocampal synaptic deficits in PKU mice. Keywords: Synaptic proteins; Hippocampus; PSD-95; nutrient combination; phenylketonuria 100

102 Chapter 6 1. INTRODUCTION The primary defect in the inherited metabolic disease phenylketonuria (PKU) is the disrupted phenylalanine (Phe) metabolism, caused by mutations in the gene encoding for the hepatic enzyme phenylalanine hydroxylase, which normally converts Phe to tyrosine. When no dietary Phe restriction is applied, this causes an increase in blood and brain Phe concentration compared to healthy controls. Although many questions still remain to be answered, a clear correlation between Phe concentration in blood and brain and the cognitive symptoms of PKU has been shown 1 3. Increased Phe concentrations disrupt neurotransmitter metabolism, white matter integrity and affect synapse functioning in PKU patients and in models of PKU Concerning the latter, a disruption of neuronal connectivity and synaptic morphology became evident in Golgi analyses of both PKU patients and PKU mice, showing a decreased number of spines, width of the synaptic cleft and thickness of the post-synaptic density, indicative of reduced synaptic function 5,6,12. These observed morphological abnormalities are corroborated by a decrease in proteins associated with synaptic functioning 9,13,14. As different markers of synaptic functioning have been examined in relation to PKU, the postsynaptic marker postsynaptic density-95 (PSD-95) has not. This protein is of interest since it is highly associated with growth and functioning of dendritic spines and modulates long-term potentiation, a process important for learning and memory ). Phe-induced neuro-morphological changes are reversible, providing a window of opportunity for interventions even after the expression of symptoms due to high Phe exposure 6,18. Our study targets these synaptic deficits with a specific nutrient combination (SNC) monitored via the expression of PSD-95. This SNC contains uridine monophosphate (UMP), docosahexaenoic acid (DHA), eicosapentaenoic acid (EPA), choline, phospholipids, folic acid, vitamins B12, B6, C, and E, and selenium. This combination provides rate-limiting nutrients critical for the synthesis of phospholipids, a major component of (synaptic) membranes, and has shown beneficial effects on synapse formation, morphology and function in mouse models of Alzheimer s disease 19. Due to the multiple pathways and precursors involved in membrane formation, intervention with single components of this SNC would, very likely, lead to sub-optimal synthesis of phospholipids and therefore limited beneficial effects. Indeed, limited and inconsistent evidence is available for the effect of supplementation with single components of this SNC in PKU patients To investigate if SNC can overcome synaptic deficits in PKU, we study here the effect of SNC on the expression of PSD-95 in the hippocampus, striatum and prefrontal cortex in the C57BL/6 PKU mouse model

103 Chapter 6 2. MATERIALS AND METHODS 2.1 Animals and dietary intervention In this study, 31 day-old male and female homozygous C57BL/6 Pah enu2 (PKU; gender balanced groups, n=60) mice and their wild-type (WT; n=10) littermates were fed for 12 weeks with different diets. Mice were bred at the University of Groningen, The Netherlands. At the start of the experiment all mice were housed singly. The genotype of the animals was established by quantitative PCR analysis from DNA extracts from tail tissue 23. PKU mice were randomly assigned to the following groups: a high-phe, mid-phe, or low-phe diet (Phe contents of 8.8, 6.4, or 4.4 g/kg diet respectively) either with or without SNC. The Phecontents of 8.8 and 6.4 are in the normal range of standard chow. The 4.4 g/kg of Phe in the food is a slight reduction compared to commercial available standard chows but still contains the minimal nutritional requirement for laboratory animals 24. The different Phe concentrations in food resulted in the following Phe concentrations in blood: WT control; 49.6±5.9, PKU 8.8; ±305.4, PKU 6.4; ±199.8, and PKU 4.4; 1065±150.4 (mean ± standard deviation). WT control mice received a high-phe diet without SNC. A WT control group on this same diet with SNC was not included because potential beneficial effects are considered irrelevant to elucidate the hypothesized mode of action of SNC in the PKU model. All components were in accordance to the minimal nutritional requirement for laboratory animals 24. This study was approved by the ethical committee of the University of Groningen, The Netherlands. 2.2 Tissue preparation After 12 weeks of dietary treatment, all animals were euthanized via a single intraperitoneal injection of pentobarbital. Blood was collected via heart punction and animals were transcardially perfused with 4% paraformaldehyde (PFA; in 0.1 M phosphate buffer (PB); ph7.4). Brains were post-fixed for 24 hours and subsequently rinsed with 0.01 M PB. After exposing the brains to 30% buffered sucrose, they were snap-frozen with liquid nitrogen and stored at -80 C. 2.3 Immunohistochemistry Coronal brain sections (20 µm thick) were processed for immunohistochemical analysis of the postsynaptic marker PSD-95 with a free-floating technique according to the following steps: 1) incubation with 0.3% H2O2 for 30 minutes, 2) incubation with 1:1000 monoclonal mouse anti-psd-95, Millipore, MABN68, 1% normal goat serum (NGS) and 0.5% Triton-X for 2 hours in a water bath at 37⁰C, 24 hours at room temperature and subsequent storage for 48 hours at 4⁰C, 3) incubation with secondary antibody solution (1:500 Biotin-SPconjugated affinipure Goat-anti-Mouse, Jackson, code: Lot# , 1% NGS and 0.5% Triton-X) for 2 hours at room temperature (RT), 4) incubation with 1:400 AB complex, Vectastain PK-6100 standard in TBS for 2 hours at RT, 5) color development 102

104 Chapter 6 was initiated by the introduction of 100 µl 0.1% H2O2 to the 3,3 -diaminobenzidine (DAB; 7 mg/15 ml) solution. The sections were rinsed multiple times with Tris buffered Saline (ph 7.4) between the above described steps. The specificity of the primary antibody was tested with omitting the primary antibody in the protocol of the staining, which resulted in the absence of detectable immunostaining, and western blot, which showed a band at 95 kda. The optical density (OD) of the staining was measured with a Quantimet 550 image analysis system (Leica, Cambridge, UK) as has been used before 25,26. In the hippocampus, 10 regions of interest were measured between bregma coordinates and mm (see Fig 1 for delineation and abbreviations). In the striatum, the mean OD of three fixed areas located in the caudoputamen were calculated (between bregma coordinates 1.18 and 0.14 mm). In addition, the infralimbic and prelimbic area of the prefrontal contex were measured between the bregma coordinates 1.98 and 1.54 mm. The OD of these regions was corrected for background staining by subtracting the OD of the corpus callosum. Both hemispheres of three sections of each individual were measured. Due to freeze artifacts three animals were discarded: one animal from the high-phe without SNC group and two animals from the low-phe without SNC group. 2.4 Statistical analysis The distribution of all parameters was checked with the Shapiro-Wilk normality test. Normally distributed data were tested with a One-way ANOVA with the Bonferroni test as post-hoc test. The Kruskal-Wallis test was used for non-parametric data with the Mann- Whitney U test as post-hoc test. All statistical analyses were performed with the Statistical Package for Social Sciences (SPSS) (V 16.0)

105 Chapter 6 Figure 1. PSD-95 expression in the C57BL/6 PKU mouse. (A) An overview picture of the PSD-95 immunostaining. The following 10 areas of interest were measured: (B) CA3: Stratum lucidum (SL), CA3 pyramidal cell layer (CA3 pyr) (C) DG: outer and middle molecular layer (OML/MML), the inner molecular layer (IML), inner (DG-IB) and outer blade (DG-OB) of the dentate gyrus, hilus, (D) CA1: Stratum oriens (SO), CA1 pyramidal cell layer (CA1 pyr) and the Stratum radiatum (SR) (E) a detailed picture of the granular layers present in the DG. (F) a detailed picture of the CA1 pyramidal cell layer. Arrows indicate clear staining within the granular layer. The size bar indicates 100 µm. The detail pictures within E and F are digitally enlarged. 104

106 Chapter 6 3. RESULTS Body weight and blood Phe concentrations were measured to monitor the dietary treatment. The absolute Phe intake, corrected for body weight, food intake, and the measured Phe in the food, confirmed that Phe intake was significantly different between the different Phe content groups but not between the different Phe content groups and their corresponding SNC groups (One-way ANOVA df=6 F=64.181, post hoc Bonferroni analysis: High-Phe- Mid-Phe; p= High-Phe-Low-Phe; p=0.000, Mid-Phe-Low-Phe; p=0.000, High- Phe with and without; p=0.298, Mid-Phe with and without; p=1.000, Low-Phe with and without; p=1.000). Furthermore, as clinically relevant, both males and females were used in this study. It is known from the literature that hippocampal spine density is affected by the estrous cycle and hormones associated with this cycle 27. However, we did not find significant differences between males and females in the PSD-95 OD, and therefore pooled values from males and females. Non-parametric tests were used to examine the PSD-95 immunoreactivity differences. No significant differences were found between the Phe-content groups of PKU mice for all brain regions studied. Therefore, all PKU mice from the different Phe-groups were pooled into two groups; those with and without SNC supplementation (final groups: WT, PKU with and PKU without SNC). In the hippocampus, a significant difference was found within the inner blade and outer blade of the dentate gyrus (Kruskal-Wallis: DG-IB; p=0.009, DG- OB; p=0.008). Post hoc testing revealed that PSD-95 OD in the DG-IB and DG-OB was significantly decreased in PKU mice fed with diets without SNC compared to WT by 54% and 64%, respectively (Mann-Whitney U test DG-IB; p=0.005, DG-OB; p=0.005). Although these significant differences were still present in DG-IB and DG-OB between the WT and PKU mice fed with diets with SNC, the decrease was considerably less: 29% and 27%, respectively (Mann-Whitney U test DG-IB; p=0.031, DG-OB; p=0.026). A clear trend was observed in the CA1 and CA3 cell layers (Kruskal-Wallis: CA1; p=0,053, CA3; p=0,064), where PKU mice on diets without SNC showed a reduction in PSD-95 OD of 44% in the CA1 and 51% in the CA3 compared to WT mice. In the CA1, this difference was attenuated for the PKU mice fed with diets with SNC, as the difference between this group and WT mice was only 8%. In the striatum and prefrontal cortex, no significant differences were found between the groups (Kruskal-Wallis: p=0,830, p=0.930, respectively); no PKU PSD- 95 expression deficit was present in PKU mice, and no change was induced by SNC

107 Chapter 6 4. DISCUSSION To the best of our knowledge, this is the first report of reduced hippocampal PSD-95 expression in the PKU mouse model. The most important finding of our study is that SNC supplementation seems to dampen PKU PSD-95 OD deficits towards WT values in specific areas of the hippocampus. Therefore, this study indicates that SNC supplementation could have a positive effect on synaptic functioning by lessening the reduced expression of PSD- 95 in the hippocampus of C57BL/6 PKU mice in the regions affected in PKU. This positive effect was most clear in the dendrites of the DG granular cells at the level of the cell layer, and to a lesser extent in CA1 pyramidal cell dendrites at the level of the cell layer. PSD-95 is a scaffolding protein which is postsynaptically present in glutamatergic and serotonergic synapses 15,28. In general, PSD-95 is implicated in postsynaptic plasticity and maturation of excitatory synapses via the interchange of AMPA and NMDA receptors in the postsynaptic compartment 15, The most marked reduction in PSD-95 in the PKU hippocampus was found in the DG granular layer, and somewhat less prominent in the CA1 and CA3 regions. In all cases, the reduction was limited to the most proximal part of the dendrites. At present, it is unclear which hippocampal input circuitry anatomically matches best the affected terminal fields. The hippocampus receives input from various sources, which deviate in their terminal fields. For example, part of the input from the entorhinal cortex terminates on the proximal dendrites of DG granular cells and CA3 pyramidal cells, and the Schaffer collaterals originating from CA3 pyramidal neurons project on the CA1 proximal dendrites 32,33. A reduction of PSD-95 in PKU mice, specifically within the proximal dendrites, could suggest weakening of synaptic connectivity in these neuronal circuits which could negatively impact learning and memory. The supplementation with SNC attenuates the PKU-specific reduction in hippocampal PSD-95 expression towards WT levels. Although the effect was not statistically significant for all regions, SNC treatment seems to particularly affect the CA1, DG-IB, and DG-OB. Apparently, certain cellular properties of these regions are more susceptible to the effect of SNC. Alternatively, the duration of the treatment was not sufficient to have a strong positive effect in all hippocampal regions. The relation between PSD-95 and the glutamatergic AMPA and NMDA receptors suggests that the PKU-specific reduction in PSD-95 could specifically affect these receptors. However, the literature concerning this topic is somewhat contradictory. Martynyuk et al. 34 found a significant increase in the Glu1 and Glu2/3 subunits of AMPA receptors and a total increase in NMDA receptor densities, a suggested compensatory mechanism for the acute suppressive effect of high Phe concentrations on glutamatergic synaptic transmission 32. Despite the upregulation in postsynaptic glutamate receptors, Martynyuk et al. 34 also report preliminary data indicating that the functional activity of glutamatergic synaptic transmission in the PKU brain is still reduced. Although they do not specify to what degree, it is possible to 106

108 Chapter 6 envisage that the reduction in function is associated with the reduction in PSD-95 found in our study. In consensus with literature, this study shows reduced levels of a protein associated with synaptic functioning in PKU mice 9,13,14. In contrast, recent literature showed an increase in synapse number and an increase in the postsynaptic markers synaptopodin and spinophilin in specific regions of the hippocampus (striatum radiatum CA1 and stratum lucidum CA3) which could suggest that this results in an increase in PSD-95 in the same regions 14. However, our study did not show significant differences between PKU and WT individuals in the same hippocampal subregions. Horling and colleagues 14 found a specific upregulation of thin and branched spines, and the postsynaptic markers associated with the cytoskeleton, suggesting that the less mature spines are affected. These immature spines, which are not fully engaged in synaptic activity, could contain less PSD-95 compared to the types that were not affected, e.g. mushroom type. Hence, our data suggest that the increase in the number of spines and postsynaptic markers found by Horling and colleagues does not come with an overall increase in PSD-95 expression. To conclude, this study shows that SNC supplementation could have a positive effect on PSD-95 expression in specific hippocampal subregions affected in C57BL/6 PKU mice. The examination of additional pre-and postsynaptic markers and functional outcomes, e.g. executive functioning, will be key to the subsequent more extensive investigation of synaptic dysfunction in PKU mice and the beneficial effects of SNC supplementation. 6 Acknowledgments: The authors thank Jan Keijser, Kunja Slopsema, and Els van der Goot for their skillful assistance during the studies. The research leading to these results has received funding from Nutricia Advanced Medical Nutrition. Author contribution: All authors agreed to be listed and approved the submitted version of the publication. VMB, DVV, AA, MCDW, MK, FJVS, and EAVZ conceived and designed the experiments; VMB performed the experiments; VMB analyzed the data; AA, MCDW, and MK contributed reagents/materials/analysis tools; VMB, DVV, AA, MCDW, MK, FJVS, and EAVZ wrote the paper. Conflict of interest: AA, MCDW, and MK are employed by Nutricia Advanced Medical Nutrition. The content is solely the responsibility of the authors and does not necessarily represent the official views of Nutricia Research. 107

109 Chapter 6 Figure 2. PSD-95 expression is reduced specifically in the hippocampus of the PKU mouse model. (A) No significant differences are found between the three groups in striatum (Kruskal-Wallis: p=0,830) (B) No significant differences are found between the three groups in the prefrontal cortex (Kruskal-Wallis: p=0.930) (C) Compared to WT mice on normal chow without SNC, PKU mice on the isocaloric control showed a significant reduction in PSD-95 expression in the hippocampus, specifically in the granular cell layer of the dentate gyrus (Kruskal-Wallis: DG-IB p=0,009, DG-OB p=0,008), with a similar trend in the CA1 and CA3 pyramidal cell layer (Kruskal-Wallis: CA1 pyr p=0,053, CA3 pyr p=0,064). A significant difference was found between the WT group and both PKU groups for the DG-IB and DG-OB (WT compared to PKU without SNC: Mann-Whitney U test DG-IB; p=0.005, DG-OB; p= WT compared to PKU with SNC: Mann-Whitney U test DG-IB; p=0.031, DG-OB; p=0.026) (arrow bars depict SEM). 108

110 Chapter 6 5. REFERENCES 1 Blau N, van Spronsen FJ, Levy HL. Phenylketonuria. Lancet 2010; 376: Waisbren SE, Noel K, Fahrbach K, Cella C, Frame D, Dorenbaum A et al. Phenylalanine blood levels and clinical outcomes in phenylketonuria: a systematic literature review and meta-analysis. Mol Genet Metab 2007; 92: Jahja R, Huijbregts SCJ, de Sonneville LMJ, van der Meere JJ, van Spronsen FJ. Neurocognitive evidence for revision of treatment targets and guidelines for phenylketonuria. J Pediatr 2014; 164: e2. 4 Joseph B, Dyer CA. Relationship between myelin production and dopamine synthesis in the PKU mouse brain. J Neurochem 2003; 86: Liang L, Gu X, Lu L, Li D, Zhang X. Phenylketonuria-related synaptic changes in a BTBR-Pah(enu2) mouse model. Neuroreport 2011; 22: Andolina D, Conversi D, Cabib S, Trabalza A, Ventura R, Puglisi-Allegra S et al. 5-Hydroxytryptophan during critical postnatal period improves cognitive performances and promotes dendritic spine maturation in genetic mouse model of phenylketonuria. Int J Neuropsychopharmacol 2011; 14: Cordero ME, Trejo M, Colombo M, Aranda V. Histological maturation of the neocortex in phenylketonuric rats. Early Hum Dev 1983; 8: Loo YH, Fulton T, Miller K, Wisniewski HM. Phenylacetate and brain dysfunction in experimental phenylketonuria: synaptic development. Life Sci 1980; 27: Hörster F, Schwab MA, Sauer SW, Pietz J, Hoffmann GF, Okun JG et al. Phenylalanine reduces synaptic density in mixed cortical cultures from mice. Pediatr Res 2006; 59: Adler-Abramovich L, Vaks L, Carny O, Trudler D, Magno A, Caflisch A et al. Phenylalanine assembly into toxic fibrils suggests amyloid etiology in phenylketonuria. Nat Chem Biol 2012; 8: Hood A, Antenor-Dorsey JA V, Hershey T, Rutlin J, Shimony JS, McKinstry RC et al. White matter integrity and executive abilities in individuals with phenylketonuria. Mol Genet Metab 2013; 109: Bauman ML, Kemper TL. Morphologic and histoanatomic observations of the brain in untreated human phenylketonuria. Acta Neuropathol 1982; 58: Imperlini E, Orrù S, Corbo C, Daniele A, Salvatore F. Altered brain protein expression profiles are associated with molecular neurological dysfunction in the PKU mouse model. J Neurochem 2014; 129: Horling K, Schlegel G, Schulz S, Vierk R, Ullrich K, Santer R et al. Hippocampal synaptic connectivity in phenylketonuria. Hum Mol Genet doi: /hmg/ ddu El-Husseini AE, Craven SE, Chetkovich DM, Firestein BL, Schnell E, Aoki C et al. Dual palmitoylation of PSD-95 mediates its vesiculotubular sorting, postsynaptic targeting, and ion channel clustering. J Cell Biol 2000; 148:

111 Chapter 6 16 Nagura H, Ishikawa Y, Kobayashi K, Takao K, Tanaka T, Nishikawa K et al. Impaired synaptic clustering of postsynaptic density proteins and altered signal transmission in hippocampal neurons, and disrupted learning behavior in PDZ1 and PDZ2 ligand binding-deficient PSD-95 knockin mice. Mol Brain 2012; 5: van der Zee EA. Synapses, spines and kinases in mammalian learning and memory, and the impact of aging. Neurosci Biobehav Rev 2015; 50: Embury JE, Charron CE, Martynyuk A, Zori AG, Liu B, Ali SF et al. PKU is a reversible neurodegenerative process within the nigrostriatum that begins as early as 4 weeks of age in Pah(enu2) mice. Brain Res 2007; 1127: van Wijk N, Broersen LM, de Wilde MC, Hageman RJJ, Groenendijk M, Sijben JWC et al. Targeting synaptic dysfunction in Alzheimer s disease by administering a specific nutrient combination. J Alzheimers Dis 2014; 38: Beblo S, Reinhardt H, Demmelmair H, Muntau AC, Koletzko B. Effect of fish oil supplementation on fatty acid status, coordination, and fine motor skills in children with phenylketonuria. J Pediatr 2007; 150: Koletzko B, Beblo S, Demmelmair H, Hanebutt FL. Omega-3 LC-PUFA supply and neurological outcomes in children with phenylketonuria (PKU). J Pediatr Gastroenterol Nutr 2009; 48 Suppl 1: S Yi SHL, Kable JA, Evatt ML, Singh RH. A randomized, placebo-controlled, doubleblind trial of supplemental docosahexaenoic acid on cognitive processing speed and executive function in females of reproductive age with phenylketonuria: A pilot study. Prostaglandins Leukot Essent Fatty Acids 2011; 85: Mazzola PN, Bruinenberg V, Anjema K, van Vliet D, Dutra-Filho CS, van Spronsen FJ et al. Voluntary Exercise Prevents Oxidative Stress in the Brain of Phenylketonuria Mice. JIMD Rep doi: /8904_2015_ APA National Research Council. Nutrient Requirements of Laboratory AnimalsNo Title. Fourth rev. The National Academies Press: Washington DC, Hovens IB, Schoemaker RG, van der Zee EA, Absalom AR, Heineman E, van Leeuwen BL. Postoperative cognitive dysfunction: Involvement of neuroinflammation and neuronal functioning. Brain Behav Immun 2014; 38: Van der Zee EA, Keijser JN. Localization of pre- and postsynaptic cholinergic markers in rodent forebrain: a brief history and comparison of rat and mouse. Behav Brain Res 2011; 221: Shors TJ, Falduto J, Leuner B. The opposite effects of stress on dendritic spines in male vs. female rats are NMDA receptordependent. Eur J Neurosci 2004; 19: Allen JA, Yadav PN, Roth BL. Insights into the regulation of 5-HT2A serotonin receptors by scaffolding proteins and kinases. Neuropharmacology 2008; 55: Zhang Y, Matt L, Patriarchi T, Malik ZA, Chowdhury D, Park DK et al. Capping of the N-terminus of PSD-95 by calmodulin triggers its postsynaptic release. EMBO J 2014; 33: Yudowski GA, Olsen O, Adesnik H, Marek KW, Bredt DS. Acute inactivation of PSD-95 destabilizes AMPA receptors at hippocampal synapses. PLoS One 2013; 8: e

112 Chapter 6 31 MacGillavry HD, Song Y, Raghavachari S, Blanpied TA. Nanoscale scaffolding domains within the postsynaptic density concentrate synaptic AMPA receptors. Neuron 2013; 78: Xu J-Y, Zhang J, Chen C. Long-lasting potentiation of hippocampal synaptic transmission by direct cortical input is mediated via endocannabinoids. J Physiol 2012; 590: Turner DA, Buhl EH, Hailer NP, Nitsch R. Morphological features of the entorhinalhippocampal connection. Prog Neurobiol 1998; 55: Martynyuk AE, Glushakov A V, Sumners C, Laipis PJ, Dennis DM, Seubert CN. Impaired glutamatergic synaptic transmission in the PKU brain. Mol Genet Metab 2005; 86 Suppl 1: S

113

114 CHAPTER 7 Long-term treatment with a specific nutrient combination in phenylketonuria mice improves recognition memory Vibeke M. Bruinenberg, Danique van Vliet 2, Els van der Goot 1, Minke L. de Vries 1, Danielle S. Counotte 3, Mirjam Kühn 3, Francjan J. van Spronsen 2, Eddy A. van der Zee 1 1 GELIFES, University of Groningen, Groningen, The Netherlands: 2 Child Hosp, Univ Med Cent, Groningen, The Netherlands; 3 Nutricia Research, Nutricia Advanced Medical Nutrition, Utrecht, The Netherlands

115 Chapter 7 ABSTRACT Introduction In phenylketonuria (PKU), a gene mutation in the phenylalanine metabolic pathway causes accumulation of phenylalanine (Phe) in blood and brain. Although early introduction of a Phe-restricted diet can prevent severe symptoms from developing, patients who are diagnosed and treated early still experience deficits in cognitive functioning indicating shortcomings of current treatment. In the search for new and/or additional treatment strategies, a combination of specific nutrient combination (SNC) was postulated to improve brain function in PKU. In this study, we examined the effect of SNC on memory and motor function. Material & Methods 48 homozygous wild-types (WT, +/+) and 96 PKU BTBR Pah2 (-/-) male and female mice received dietary interventions from postnatal day 31 till 10 months of age. These mice were subdivided in the following 6 groups: high Phe diet (WT C-HP, PKU C-HP), high Phe plus specific nutrient combination (WT SNC-HP, PKU SNC-HP), PKU low-phe diet (PKU C-LP), and PKU low-phe diet plus specific nutrient combination (PKU SNC- LP). Memory and motor function in mice was tested at 3,6, and 9 months after treatment initiation in the open field (OF), novel object recognition test (NOR), spatial object recognition test (SOR), and the balance beam (BB). Results In the NOR, we found that PKU mice despite being subjected to high Phe conditions could master the task on all three time points when supplemented with SNC. Under low Phe conditions, the PKU mice on control diet can master the NOR on all time points and mice on the supplemented diet can master the task at time point 6 and 9. SNC supplementation did not consistently influence the performance in the OF, SOR or BB. Conclusion This study is the first long-term intervention study in BTBR PKU mice that shows that SNC supplementation can specifically benefit novel object recognition. Future research is necessary to identify the mode of action of SNC supplementation. 114

116 Chapter 7 1. INTRODUCTION The detrimental effects of increased phenylalanine (Phe) concentrations on the brain are clearly visible in the metabolic disorder, Phenylketonuria (PKU, OMIM ). In this disorder, a mutation in the gene encoding for the hepatic enzyme phenylalanine hydroxylase causes a disruption in the conversion of Phe to tyrosine. Consequently, when no restriction is made in natural protein intake, Phe accumulates in blood and brain. Neonatal screening facilitates early introduction of treatment preventing symptoms such as severe cognitive disabilities and epilepsy. Nonetheless, even early-continuously-treated patients experience deficits in cognitive functioning, in for instance processing speed, attention, working memory and social-cognitive functioning 1 3, highlighting that current treatment is not optimal. In the search for new and/or additional treatment strategies, a combination of specific nutrients (SNC) was postulated to relieve the functional and neurobiological effects of increased Phe 4. The specific nutrients selected were originally combined to facilitate the Kennedy pathway as they are important precursors, cofactors and antioxidants for this pathway. This pathway is important in phospholipid synthesis, a key feature of the neuronal membrane. Functional neuronal membranes are important for synaptic functioning and neurotransmitter release 5,6, domains affected in PKU. Furthermore, other PKU-related problems in, for instance, white matter integrity and oxidative stress could benefit from these nutrients as antioxidants present in the SNC could relieve oxidative stress 7 9 and the phospholipids present in the SNC are an important part of myelin 10. Therefore, the specific nutrient combination (SNC) aims to detain the effect of Phe on neurotransmitter metabolism, white matter integrity, oxidative stress and synaptic functioning, in the end, positively affecting functional outcome. 7 The period in which SNC supplementation could have an added effect on the current recommended life-long treatment with a Phe restricted diet in PKU patients is not clear. Although life-long treatment is recommended, at present, little is known about the progression of the disorder under this treatment during aging as it is implemented approximately years ago. Therefore, the aim of this study is twofold: 1) to investigate a phe-restricted treatment in an aging PKU mouse model, 2) to examine the effect of SNC on the behavioral performance of PKU mice under high Phe and low Phe conditions. 2. METHODS 2.1 Animals A breeding colony of heterozygous (+/-) mating pairs generated 48 wild-types (WT, +/+) and 96 PKU BTBR Pah2 (-/-) male and female mice. Original breeding pairs were kindly provided by prof. Puglisi-Allegra from the Sapienza, University of Roma, Rome, Italy. On postnatal day (PND) 28 the animals were weaned and the genetic status of the animals was established 115

117 Chapter 7 via quantitative PCR analysis on DNA extracted from ear tissue (as previously described 11 ). After weaning, all littermates were kept in the breeding cage (26x42x15 cm) without the mother until PND 31. On PND 31, the animals were group housed in sex-matched pairs of two in cages of the same dimensions (26x42x15 cm) with sawdust bedding and cage enrichment in the form of nesting material, paper rolls, and a small wooden stick. To prevent fighting among the males, a red transparent house was added in their cages in the adult stage. The condition of the housing facility was kept constant at a temperature of 21±1 C with a 12/12 light/dark cycle. The animals received fresh food every day (between zeitgeber time (ZT) 8-10) and had ad libitum access to water. Leftover food was collected every day before given fresh food. Together with a thorough search of bedding after cage cleaning, the difference with the offered food gave an estimation of weekly food intake per pair. Furthermore, body weight of the mice was measured during these weekly cage cleaning (ZT 8-10). The dietary intervention started at PND 31 until 10 months of age. The length of the experiment required clear humane endpoints. These were set on a decrease in body weight of 15% together with other sickness behavior e.g. inactive behavior, arched back. If one of the pair of mice was excluded from the experiment, females were placed in pairs of three. Males were kept solitary. All experimental procedures were approved by an independent ethics committee for animal experimentation (6504E, Groningen, the Netherlands) and complied with the principles of good laboratory animal care following the European Directive for the protection of animals used for scientific purposes Dietary intervention The dietary intervention started on PND 31. At this time point, pairs of mice were assigned to one of the following six groups: WT control high Phe diet (WT C-HP), WT high Phe diet plus SNC (WT SNC-HP), PKU high Phe diet (PKU C-HP), PKU high Phe diet plus SNC (PKU SNC-HP), PKU low-phe diet (PKU C-LP), and PKU low-phe diet plus SNC (PKU SNC-LP). The high Phe diet is a normal diet for WT animals. Each group consisted of 12 males and 12 females. As the mice were at the pre-adolescence stage at the beginning of the experiments, the diet was based on the growth diet AIN-93G. In the adult stage (13 weeks), the mice were switched to diets based on the maintenance diet AIN-93M manufactured by the same supplier (Research Diet Services BV, Wijk bij Duurstede, The Netherlands). The key characteristics of the diets were kept the same; normal diet with or without SNC had Phe concentrations of 6.2 g/kg and tyrosine concentrations of 15 g/kg and low-phe diet (based on previous literature 12 ) had Phe concentrations of 2.0 g/kg and tyrosine concentrations of 15 g/kg. The specifics of the diets are depicted in Table

118 Chapter 7 Table 1. Nutritional content of experimental diets. * The corresponding mineral and vitamin premixes to the growth (G: AIN-93G) or maintenance (M: AIN-93M) diets were used. g /kg diet C-HP SNC-HP C-LP SNC-HP G M G M G M G M Carbohydrates 666,5 718,4 638,6 690,5 670,7 722,6 642,8 694,7 Fat 50,0 50,0 50,0 50,0 50,0 50,0 50,0 50,0 Dietary fibre 50,0 50,0 50,0 50,0 50,0 50,0 50,0 50,0 Protein 186,0 134,1 186,0 134,1 181,8 129,9 181,8 129,9 Amino acids Alanine 4,6 3,3 4,6 3,3 4,6 3,3 4,6 3,3 Arginine 6,4 4,5 6,4 4,5 6,4 4,5 6,4 4,5 Aspartic acid 12,2 8,0 12,2 8,0 12,2 8,0 12,2 8,0 Cystine 3,7 2,4 3,7 2,4 3,7 2,4 3,7 2,4 Glutamic acid 36,3 25,5 36,3 25,5 36,3 25,5 36,3 25,5 Glycine 3,2 2,3 3,2 2,3 3,2 2,3 3,2 2,3 Histidine 4,6 3,3 4,6 3,3 4,6 3,3 4,6 3,3 Isoleucine 8,2 5,9 8,2 5,9 8,2 5,9 8,2 5,9 Leucine 15,7 10,9 15,7 10,9 15,7 10,9 15,7 10,9 Lysine 16,3 9,2 16,3 9,2 16,3 9,2 16,3 9,2 Methionine 4,6 3,3 4,6 3,3 4,6 3,3 4,6 3,3 Phenylalanine 6,2 6,2 6,2 6,2 2,0 2,0 2,0 2,0 Proline 20,5 14,3 20,5 14,3 20,5 14,3 20,5 14,3 Serine 9,7 6,7 9,7 6,7 9,7 6,7 9,7 6,7 Threonine 6,7 4,7 6,7 4,7 6,7 4,7 6,7 4,7 Tryptophan 2,1 1,6 2,1 1,6 2,1 1,6 2,1 1,6 Tyrosine 15,0 15,0 15,0 15,0 15,0 15,0 15,0 15,0 Valine 10,0 7,0 10,0 7,0 10,0 7,0 10,0 7,0 7 Mineral premix* 35,0 35,0 35,0 35,0 35,0 35,0 35,0 35,0 Vitamin premix* 10,0 10,0 10,0 10,0 10,0 10,0 10,0 10,0 Additives 2,5 2,5 2,5 2,5 2,5 2,5 2,5 2,5 SNC 27,5 27,5 27,5 27,5 Totaal (g) Energie (kcal/kg diet) 3859,9 3859,9 3748,4 3748,4 3859,9 3859,9 3748,4 3748,4 117

119 Chapter Behavioral paradigms During the dietary intervention, the animals were behaviorally assessed every 12 weeks starting at four months of age. Five days before each test session, a blood spot was taken. Each test session consisted of an open field (OF), novel object recognition (NOR), spatial object recognition (SOR) and a balance beam (BB). The animals were tested between ZT1-6. All procedures and experimental setup were described in our previous study 13. In short, the habituation phase of the NOR was used as OF. On day 1 of the testing session, the animals were placed in the middle of a square arena (50x50x35 cm) to explore the arena freely for ten minutes. The subsequent day the animals could explore two identical objects for ten minutes in the familiarization phase of the NOR. Again 24 hours later, one of the objects was replaced with a novel object wherein the mouse could explore this new setting for 10 minutes. After a period of 5 days, the animals were tested in the SOR. The first day the animals were exposed to four sessions of 6 minutes. The first session was similar to the habituation phase of the NOR. In the second to the fourth session, the animals could freely explore three different objects (in shape, color, and texture) in a specific configuration. Between sessions, the animals were placed back in the home cage for 2 minutes. The second day, one of the objects was moved to a different location. The objects, the starting condition, and the displaced object were randomized over trails. Both the NOR and SOR were performed in a separate room recorded with a camera (Panasonic WVCP500) connected to a computer outside the room with Media recorder (Noldus, The Netherlands). The balance beam was performed in the housing facility 24 hours later. During this task, the animals had to cross a square wooden beam (length 1 m, width 5 mm, height 10 mm, horizontally positioned 50 cm above the underlying surface) over four distances (10, 40, 75, and 100 cm). The final distance was used as read-out trail. In this trail, the number of correct steps and total steps necessary to cross the beam were manually scored and calculated to a percentage. A step was considered correct if the hind paw had a full placement on the beam at the initiation and end of the forward movement. The open field was analyzed with Ethovision v.11 (Noldus, The Netherlands). In this analysis, the arena was divided into a center zone, four border zones, and four corner zones 14. Activity was quantified by the distance moved and anxiety-like behavior was examined by the preference of the animal to seek out the more sheltered zones, the corners. In the NOR and SOR, the exploration time of each object was manually scored with the program ELINE (made in house). For the NOR, the discrimination index (DI) was calculated by the time spent exploring the novel object minus the time exploring the same object divided by the total exploration time of both objects 15. For the SOR, the exploration time of the first three training sessions was compared to the time exploring in the test session. The mice mastered these learning paradigms when they explored either the novel object or the relocated object above change level. 118

120 Chapter Statistics All statistical analysis was performed with SPSS All data were checked for normality (Shapiro-Wilk test) and homogeneity of variance (Levene s test). Food intake was tested non-parametrically with a Kruskal-Wallis and post hoc analysis was done with a Mann- Whitney U test. A fixed effects linear mixed model was used to test repeated measurements of body weight, and the behavioral paradigms. As the groups were not fully balance between the genotypes (WT mice are not able to receive low Phe diet), two models were tested; 1) a factorial analysis of the factors: time, genotype, and specific nutrient combination was used to investigate differences between WT and PKU mice on high Phe diet and the influence of SNC supplementation within these groups, 2) a factorial analysis of the factors: time, specific nutrient combination, and Phe condition was used to examine differences between the four different diets in PKU mice (C-HP, SNC-HP, C-LP, and SNC-LP). For body weight, we assume that the body weight measurements taken near each other are more related to each other than the measurements taken with a larger time interval (for example we expect that the body weight measurement of week 13 is more similar to week 14 than week to 41). Therefore, the repeated covariance type was set to first order autoregressive. For the behavioral this assumption was not made, therefore a diagonal covariance type was selected. Finally, the ability to master the task was investigated by comparing the DI to change level (0) with a t-test. No corrections were made for multiplicity. A p-value equal to or less than 0.05 was considered significant. If not differently specified, data are expressed as mean ± standard error of the mean. 3. RESULTS General health, body weight, and food intake In the course of the experiment, 21 animals were excluded from the experiment. The excluded animals were not in a specific treatment group (Kruskal-Wallis test, groups p=0.081) and dropout was not skewed by genotype or sex (Kruskal-Wallis test, genotype p=0.134, sex p=0.480). The general health of the animals was, among others, monitored by body weight and food intake. Both parameters were split for growth diet and maintenance diet, the first 9 weeks and starting from 3 months respectively. Furthermore, male and female were analyzed separately as food intake and body weight was different between the sexes. In addition, graphs and analysis was split for all groups on a high Phe diet (model 1) and all PKU groups (model 2). In figure 1A-B, an increase in body weight over time is found in males and females (Female; time p<0.001, Male; time: p<0.001). In females, the progression of the body weight differed between WT and PKU mice (genotype x time p<0.001) and an interaction was found between genotype and SNC supplementation (genotype x time x SNC p=0.005). In male, similar results were found but no interaction was found with SNC supplementation (time p<0.001, genotype x time p<0.001, genotype x time x SNC p=0.871). In figure 1C-D 119

121 Chapter 7 the curves of all PKU groups are depicted. Again, an increase in body weight was observed for both sexes (Female; time p<0.001, Male; time: p<0.001). In both sexes, this increase of body weight over time was different between high Phe and low Phe conditions (Female: Phe condition x time p=0.002, Male: Phe condition x time p=0.009). SNC supplementation did not significantly contribute to this model. In figure 1E-F, the average daily food intake is depicted. A mere indication of the differences in food intake can be drawn from these data, as only group housed individuals of the same genotype were included. Nevertheless, no differences were found in food intake between the groups in female mice (p=0.173), and in male mice only between WT C-HP and PKU SNC-HP (p=0.036). In maintenance diet (Figure 2A-D), the growth of the animals persisted (Fig 2A Female; time p<0.001, Fig 2B Male; time: p<0.001, Figure 2C Female; time p<0.001, Figure 2D Male; time: p<0.001). In male mice, the differences in the increase of body weight over time between WT and PKU on high Phe diet were still present (genotype x time p=0.002) but in females this interaction was no longer significant (genotype x time p=0.754). In PKU mice (figure 2C-D), similar results were found (Female: Phe condition x time p<0.001, Male Phe condition x time p=0.002). Daily food intake was also similar to the results found in growth diet (Male: WT C-HP and PKU SNC-HP p=0.036). 3.2 Open field The distance covered in the open field was used to explore differences in novelty-induced exploration. Gender differences were observed, and therefore, male and female mice were analyzed separately. In figure 3A trough 3D, it is evident that all groups move less through the maze (Fig 3A Female; time p<0.001, Fig 3B Male; time: p<0.001, Figure 3C Female; time p<0.001, Figure 3D Male; time: p<0.001). In female mice, the WT mice covered more distance in the maze compared to PKU high Phe mice (p<0.001). In male mice, the progression over time was different between WT and PKU high Phe mice (p=0.013) but no main effect was found (p=0.663). The distance moved was differently influenced by SNC supplementation in WT and PKU mice (p=0.001). In female PKU mice (Figure 3C), PKU mice on low Phe diet covered more distance through the maze compared to PKU mice on high Phe diet (p=0.022). Furthermore, the progression over time was different between these groups (p=0.017). In male PKU mice (Figure 3D), the difference between high Phe and low Phe diet was inversed. The mice in the high Phe condition moved more through the maze compared to the low Phe condition (p=0.006). The progress over time was not different between the conditions (p=0.254). In addition to the distance moved through the arena, the preference of the mice to explore more sheltered areas of the arena was investigated in the open field. This was done by examining time spent in the corners. Over time, the time spent in the corners was not constant (Fig 4A Female; time p<0.001, Fig 4B Male; time: p<0.001, Figure 4C Female; time p<0.001, Figure 4D Male; time: p<0.001). In the females, the PKU high Phe mice spent less time in 120

122 Chapter 7 the corners compared WT mice (Fig 4A, genotype p<0.001) in which the progress in time was also different (p=0.040). In male mice, no difference was observed between WT and PKU high Phe mice or the progression in time (genotype=0.906, time x genotype=0.055). In the PKU mice, a trend was observed between low and high Phe conditions in female mice (p=0.064) but not in male mice (p=0.586). 7 Figure 1 Growth diet. Results are separated for females and male (graphs A,C,E and graphs B,D,F, respectively). In figure A and B, the bodyweight curves of the first eight weeks of treatment, starting on postnatal day 31, are depicted for all groups on high Phe diet (WT C-HP, WT SNC-HP, PKU C-HP, PKU SNC-HP). In figure C and D, the body weight curves for all PKU mice groups are depicted. Mean daily food intake is depicted in graph E and F (median depicted). Graphs A-D: mean± standard error of the mean (SEM), x-axis depict days. Graphs E-F: median. 121

123 Chapter 7 Figure 2 Maintenance diet. Graphs are identically organized as figure 2. In graphs A-D, the bodyweight curves of the last 28 weeks of dietary treatment are depicted, starting at week 13. In graphs E-F, mean daily food intake is depicted. Graphs A-D: mean± SEM, x-axis depict weeks. Graphs E-F: median 122

124 Chapter 7 Figure 3 Distance moved in open field. (A) Female mice on high Phe diet, (B) male mice on high Phe diet, (C) female PKU mice, and (D) male PKU mice. (mean± SEM) 7 Figure 4 Time spent in corners of the maze. The time spent in the corners of the maze is thought to represent anxiety-like behavior as the mice seek out the more sheltered areas of the arena. (A) Female mice on high Phe diet, (B) male mice on high Phe diet, (C) female PKU mice, and (D) male PKU mice. (mean± SEM) 123

125 Chapter Learning and memory paradigms: NOR and SOR No differences were observed in sex, and therefore, males and females were analyzed together. In the model, a significant difference was found between WT and PKU high Phe diet in DI (p<0.001). SNC supplementation improved the performance of PKU mice (p<0.001) whereas no differences were found between Phe conditions (p=0.324) or a different response of Phe conditions to the SNC supplementation (p=0.236). Although this statistical analysis highlight the difference between the groups, it doesn t give insight in the ability of the mice to master the learning and memory paradigm. To investigate this, DI was compared to change level. From figure 5A, it is clear that the mice of the WT-groups (C-HP t(23)=3.407, p=0.002, SNC-HP t(23)=3.715, p=0.001), PKU SNC-HP (t(20)=2.915, p=0.009), and PKU C-LP (t(23)=3.646 p=0.001) are able to master the task after three months of treatment. However, the PKU C-HP (t(22)=0.070 p=0.944) and PKU SNC-LP (t(23)=1.707 p=0.103) did not. After six months of treatment (Figure 5B), all groups except for the PKU C-HP learned the task (WT C-HP t(23)=7.789 p<0.001, WT SNC-HP t(23)=6.924 p<0.001, PKU C-HP t(20)=0.826 p=0.418, PKU SNC-HP t(18)=3.573 p=0.002, PKU C-LP t(23)=2.137, p=0.043, PKU SNC-LP t(22)=2.648 p=0.015). After the nine month of treatment, for a second time, all groups mastered the NOR task, with the exception of PKU C-HP (WT C-HP t(21)=3.482 p=0.002, WT SNC-HP t(20)=3.081 p= 0.006, PKU C-HP t(21)=1.729 p=0.098, PKU SNC-HP t(15)=6.037 p<0.001, PKU C-LP t(22)=2.230 p=0.036, PKU SNC-LP t(20)=5.461, p<0.001). In PKU mice, the SNC supplementation did not affect the time spent on exploring the objects ( p=0.294). PKU on high Phe diet did spent more time exploring the objects compared to the low Phe diet groups (p=0.012). In WT mice, SNC supplementation increased the exploration (p=0.033). The analysis of the SOR data, did not reveal a PKU phenotype of the PKU control high Phe group compared to the WT group (p=0.768). Upon a closer look, by comparing the DI to change level and eliminating values two standard deviations outside the mean, similar results were found after three months of treatment as previously described in the NOR. The mice of the WT-groups (C-HP t(23)=2.450, p=0.022, SNC-HP t(22)=2.123, p=0.045), PKU SNC- HP (t(19)=2.228, p=0.038), and PKU C-LP (t(23)=2.193 p=0.037) are able to master the task after three months of treatment. The PKU C-HP (t(19)=-.069 p=0.946) and PKU SNC- LP (t(23)=1.385 p=0.179) did not. However, the WT mice did no longer master the SOR after six months of treatment (C-HP t(22)=2.450, p=0.074, SNC-HP t(23)=1.719, p=0.099). 124

126 Chapter 7 Figure 5 Novel object recognition. Discrimination index ((exploration novel object- exploration same object)/total exploration time) is tested against change level (0). * represent a significant difference from change level. (mean± SEM) 3.4 Balance beam The relative number of correct steps was used to investigate the motor performance. In both females and males (Figure 6A-B), a clear difference was observed between WT and PKU mice on high Phe diet (Fig 6A Female; p<0.001, Fig 4B Male; p<0.001). In female PKU mice (Figure 6A), SNC supplementation reduced the relative number of correct steps (p=0.037). In male PKU mice, this effect was not observed (p=0.785). When comparing low and high Phe conditions (Figure 6C-D), only female mice performed better in the low Phe groups (Figure 6C Female; fit of model AIC= , p<0.001, Figure 6D Male; fit of model AIC= , p=0.863). SNC supplementation did not influence the performance of the PKU mice groups (female: p=0.587, male: p=0.671)

127 Chapter 7 Figure 6 Motor performance. The relative number of correct steps made in the probe trail (100 cm) is depicted. (A) Female mice on high Phe diet, (B) male mice on high Phe diet, (C) female PKU mice, and (D) male PKU mice. (mean± SEM). 4. DISCUSSION In this study, a long-term dietary intervention with SNC supplementation, a low-phe diet or a combination of these diets was investigated in male and female BTBR PKU mice. These groups of mice were compared to WT controls with or without supplementation. During the 9 months of intervention, the animals were tested three time (after 3, 6, and 9 months of treatment) in the open field, NOR, SOR and the balance beam. In the NOR, we found that PKU mice despite being subjected to high Phe conditions could master the NOR at all three time points when supplemented with SNC. Under low Phe conditions, the PKU mice on control diet could master the NOR on all time points and mice on the supplemented diet could master the task at time point 6 and 9. SNC supplementation did not consistently influence the performance in the open field, SOR or the balance beam. This indicates that SNC supplementation specifically influences the performance in the NOR. 4.1 A long-term study: Opportunities and limitations This study was the first to investigate the long-term effect on behavior of different treatments in the BTBR PKU mouse model. As a consequence, we were unprepared for the loss in animals during the experiment due to early aging. This led to unbalanced groups and smaller numbers of animals to be tested. However so, we believe that the exclusion of animals from the experiment was not skewed towards any experimental group and that the remaining animals were in good condition, enabling us to draw firm conclusions. Nevertheless, this study shows that future studies should be aware of this early drop-out of BTBR mice. 126

128 Chapter 7 In this study, we performed multiple rounds of behavioral testing in the same animal to examine changes over time. In the open field, we found a reduction in locomotor activity and an increase in time spent in the corners even in the WT mice. Although this is not described in BTBR mice, our results are in line with other mouse studies that show reduced locomotor activity in aging mice in the open field 16,17. The preference for a certain location in the maze is thought to depend on the properties of the maze and lighting conditions that could explain the mixed results found for this parameter between studies 16,17. In the NOR and the balance beam, no clear differences were found over time that suggest that either the multiple testing rounds or the aging effect of a period of 6 months does not influence the performance in these tasks. In contrast, the SOR was affected by either one or both of these factors. At time point 3, the WT groups showed similar abilities to master the task compared to the NOR. However, at latter time points, the WT individuals were not able to master this task. Although it is possible that multiple rounds of behavioral testing can influence this outcome, impairments of spatial memory are reported in normal aging mice and BTBR mice Timing of treatment Early intervention in PKU patients can prevent the irreversible cognitive disabilities found in untreated or late-diagnosed PKU patients 22, suggesting a specific window of treatment. Although guidelines are given, this window is not that exactly defined 22. For example, a Pherestricted diet can still have positive effects in late-diagnosed PKU patients 23. In our current study, we have introduced the SNC supplementation and low-phe conditions at PND 31. At this time point, the maturation of the brain and the characteristic behavior in mice is thought to represent the adolescence stage in humans 24. The introduction of our treatment would, therefore, surpass the early intervention window of PKU treatment to prevent cognitive disabilities. Nevertheless, the PKU mice on a control low Phe diet did master the NOR paradigm. Suggesting, at least in PKU mice, that the low Phe conditions introduced later in life can be beneficial for object recognition memory. Furthermore, no early cognitive decline in novel object recognition was observed under these conditions. 7 The timing of diet was also important in the SNC supplementation. Under high Phe conditions, PKU mice that received SNC supplementation were able to master the NOR task at all three time points. In contrast to the SNC supplementation in low Phe conditions, where the animals only mastered the NOR task at time point 6 and 9. The original basis SNC supplementation in Alzheimer s disease was the idea that Alzheimer s patients had a greater need for renewal of synapses than healthy aged-matched controls 25. By supplementing specific nutrients, this nutritional need could be met and consequently improve synaptic functioning 25. The positive results of SNC in the high Phe condition and the low Phe condition on latter time points could, therefore, be a consequence of meeting a nutritional need for the renewal of synapses. A need that perhaps was not present at time point

129 Chapter Specificity of SNC supplementation Besides the beneficial effect of the SNC nutrients on the Kennedy pathway, the nutrients (separately) of the SNC combination are likely to positively affect synaptic functioning, neurotransmitter metabolism, white matter integrity, and oxidative stress. In this study, we were not able to explore the underlying molecular mechanism in which SNC supplementation exerted its positive effect in the brain. However, by comparing the different outcome parameters, we could hypothesize certain brain regions and domains to be affected. In the PKU mice, SNC supplementation (only in females) and low Phe conditions improved body weight. An outcome parameter, at times, used as an indication of efficacy of treatment in PKU mice 26,27. Therefore, the improvements found in body weight could indicate efficacy of treatment by positively effecting growth and/or metabolic stress in PKU 26. Future research is needed to establish the effects of SNC on body composition, energy balance, and the possible difference between male and female mice in this response. In the behavioral paradigms, SNC supplementation specifically improved the NOR. The long-term object recognition protocol used in this study is thought to involve the hippocampal system 15,28. This brain region receives input from the perirhinal cortex, which in turn collects information from other brain regions involved in visual, olfactory, and somatosensory perception. All these brain regions are important in object recognition, and the consolidation, acquisition and retrieval of the memory necessary for mastering the NOR paradigm 15,29. The positive effect of SNC supplementation on the performance in the NOR could possibly come from the effect of SNC on synaptic plasticity within the hippocampus, important in the novel object recognition 30. Although neurotransmitter release and signaling are positively affected by SNC supplementation 10,31, it is not clear if one or more of the these domains (white matter integrity, neurotransmitter metabolism and 4.4 CONCLUSION This study is the first long-term intervention study in BTBR PKU mice that shows that SNC supplementation can specifically benefit novel object recognition. Future studies are necessary to identify the mode of action of SNC supplementation. 128

130 Chapter 7 5. REFERENCES 1 Jahja R, van Spronsen FJ, de Sonneville LMJ, van der Meere JJ, Bosch AM, Hollak CEM et al. Social-cognitive functioning and social skills in patients with early treated phenylketonuria: a PKU-COBESO study. J Inherit Metab Dis 2016; 39: Moyle JJ, Fox AM, Arthur M, Bynevelt M, Burnett JR. Meta-analysis of neuropsychological symptoms of adolescents and adults with PKU. Neuropsychol Rev 2007; 17: Enns GM, Koch R, Brumm V, Blakely E, Suter R, Jurecki E. Suboptimal outcomes in patients with PKU treated early with diet alone: revisiting the evidence. Mol Genet Metab; 101: Bruinenberg VM, van Vliet D, Attali A, de Wilde MC, Kuhn M, van Spronsen FJ et al. A Specific Nutrient Combination Attenuates the Reduced Expression of PSD-95 in the Proximal Dendrites of Hippocampal Cell Body Layers in a Mouse Model of Phenylketonuria. Nutrients 2016; 8. doi: /nu Yehuda S, Rabinovitz S, Carasso RL, Mostofsky DI. The role of polyunsaturated fatty acids in restoring the aging neuronal membrane. Neurobiol Aging; 23: Vigh L, Escribá P V, Sonnleitner A, Sonnleitner M, Piotto S, Maresca B et al. The significance of lipid composition for membrane activity: new concepts and ways of assessing function. Prog Lipid Res 2005; 44: Martínez-Cruz F, Osuna C, Guerrero JM. Mitochondrial damage induced by fetal hyperphenylalaninemia in the rat brain and liver: its prevention by melatonin, Vitamin E, and Vitamin C. Neurosci Lett 2006; 392: Martinez-Cruz F, Pozo D, Osuna C, Espinar A, Marchante C, Guerrero JM. Oxidative stress induced by phenylketonuria in the rat: Prevention by melatonin, vitamin E, and vitamin C. J Neurosci Res 2002; 69: Sitta a, Vanzin CS, Biancini GB, Manfredini V, de Oliveira a B, Wayhs C a Y et al. Evidence that L-carnitine and selenium supplementation reduces oxidative stress in phenylketonuric patients. Cell Mol Neurobiol 2011; 31: Cansev M, van Wijk N, Turkyilmaz M, Orhan F, Sijben JWC, Broersen LM. A specific multi-nutrient enriched diet enhances hippocampal cholinergic transmission in aged rats. Neurobiol Aging 2015; 36: Mazzola PN, Bruinenberg V, Anjema K, van Vliet D, Dutra-Filho CS, van Spronsen FJ et al. Voluntary Exercise Prevents Oxidative Stress in the Brain of Phenylketonuria Mice. JIMD Rep doi: /8904_2015_ Pascucci T, Giacovazzo G, Andolina D, Accoto A, Fiori E, Ventura R et al. Behavioral and neurochemical characterization of new mouse model of hyperphenylalaninemia. PLoS One 2013; 8: e Bruinenberg VM, van der Goot E, van Vliet D, de Groot MJ, Mazzola PN, Heiner-Fokkema MR et al. The Behavioral Consequence of Phenylketonuria in Mice Depends on the Genetic Background. Front Behav Neurosci 2016; 10: Hovens IB, Schoemaker RG, van der Zee EA, Absalom AR, Heineman E, van Leeuwen BL. Postoperative cognitive dysfunction: Involvement of neuroinflammation and neuronal functioning. Brain Behav Immun 2014; 38:

131 Chapter 7 15 Antunes M, Biala G. The novel object recognition memory: neurobiology, test procedure, and its modifications. Cogn Process 2012; 13: Shoji H, Takao K, Hattori S, Miyakawa T. Age-related changes in behavior in C57BL/6J mice from young adulthood to middle age. Mol Brain 2016; 9: Brouwer-Brolsma EM, Schuurman T, de Groot LCPMG, Feskens EJM, Lute C, Naninck EFG et al. No role for vitamin D or a moderate fat diet in aging induced cognitive decline and emotional reactivity in C57BL/6 mice. Behav Brain Res 2014; 267: Weber M, Wu T, Hanson JE, Alam NM, Solanoy H, Ngu H et al. Cognitive Deficits, Changes in Synaptic Function, and Brain Pathology in a Mouse Model of Normal Aging. eneuro 2015; 2. doi: / ENEURO Saroja SR, Kim E-J, Shanmugasundaram B, Höger H, Lubec G. Hippocampal monoamine receptor complex levels linked to spatial memory decline in the aging C57BL/6J. Behav Brain Res 2014; 264: MacPherson P, McGaffigan R, Wahlsten D, Nguyen P V. Impaired fear memory, altered object memory and modified hippocampal synaptic plasticity in split-brain mice. Brain Res 2008; 1210: Stapley NW, Guariglia SR, Chadman KK. Cued and contextual fear conditioning in BTBR mice is improved with training or atomoxetine. Neurosci Lett 2013; 549: van Spronsen FJ, van Wegberg AM, Ahring K, Bélanger-Quintana A, Blau N, Bosch AM et al. Key European guidelines for the diagnosis and management of patients with phenylketonuria. Lancet Diabetes Endocrinol doi: /s (16) Koch R, Moseley K, Ning J, Romstad A, Guldberg P, Guttler F. Long-term beneficial effects of the phenylalanine-restricted diet in late-diagnosed individuals with phenylketonuria. Mol Genet Metab 1999; 67: Semple BD, Blomgren K, Gimlin K, Ferriero DM, Noble-Haeusslein LJ. Brain development in rodents and humans: Identifying benchmarks of maturation and vulnerability to injury across species. Prog Neurobiol 2013; : van Wijk N, Broersen LM, de Wilde MC, Hageman RJJ, Groenendijk M, Sijben JWC et al. Targeting synaptic dysfunction in Alzheimer s disease by administering a specific nutrient combination. J Alzheimers Dis 2014; 38: Solverson P, Murali SG, Brinkman AS, Nelson DW, Clayton MK, Yen C-LE et al. Glycomacropeptide, a low-phenylalanine protein isolated from cheese whey, supports growth and attenuates metabolic stress in the murine model of phenylketonuria. Am J Physiol Endocrinol Metab 2012; 302: E Ney DM, Hull AK, van Calcar SC, Liu X, Etzel MR. Dietary glycomacropeptide supports growth and reduces the concentrations of phenylalanine in plasma and brain in a murine model of phenylketonuria. J Nutr 2008; 138: Reger ML, Hovda DA, Giza CC. Ontogeny of Rat Recognition Memory measured by the novel object recognition task. Dev Psychobiol 2009; 51: Oliveira AMM, Hawk JD, Abel T, Havekes R. Post-training reversible inactivation of the hippocampus enhances novel object recognition memory. Learn Mem 2010; 17:

132 Chapter 7 30 Sarkisyan G, Hedlund PB. The 5-HT7 receptor is involved in allocentric spatial memory information processing. Behav Brain Res 2009; 202: Savelkoul PJM, Janickova H, Kuipers AAM, Hageman RJJ, Kamphuis PJ, Dolezal V et al. A specific multi-nutrient formulation enhances M1 muscarinic acetylcholine receptor responses in vitro. J Neurochem 2012; 120:

133

134 CHAPTER 8 Large neutral amino acid supplementation exerts its effect through three synergistic mechanisms: proof of principle in phenylketonuria mice Danique van Vliet 1,2, Vibeke M. Bruinenberg 2, Priscila N. Mazzola 1,2, Martijn H.J.R. van Faassen 3, Pim de Blaauw 3, Ido P. Kema 3, M. Rebecca Heiner-Fokkema 3, Rogier D. van Anholt 4, Eddy A. van der Zee 2, Francjan J. van Spronsen 1 * 1 University of Groningen, University Medical Center Groningen, Beatrix Children s Hospital, Groningen, The Netherlands. 2 University of Groningen, Center of Behavior and Neurosciences, Department of Molecular Neurobiology, Groningen, The Netherlands. 3 University of Groningen, University Medical Center Groningen, Department of Laboratory Medicine, Groningen, The Netherlands. 4 Independent Researcher, Deventer, The Netherlands. PLoS One Dec 1;10(12):e doi: /journal.pone ecollection 2015.

135 Chapter 8 ABSTRACT Background Phenylketonuria (PKU) was the first disorder in which severe neurocognitive dysfunction could be prevented by dietary treatment. However, despite this effect, neuropsychological outcome in PKU still remains suboptimal and the phenylalanine-restricted diet is very demanding. To improve neuropsychological outcome and relieve the dietary restrictions for PKU patients, supplementation of large neutral amino acids (LNAA) is suggested as alternative treatment strategy that might correct all brain biochemical disturbances caused by high blood phenylalanine, and thereby improve neurocognitive functioning. Objective As a proof-of-principle, this study aimed to investigate all hypothesized biochemical treatment objectives of LNAA supplementation (normalizing brain phenylalanine, nonphenylalanine LNAA, and monoaminergic neurotransmitter concentrations) in PKU mice. Methods: C57Bl/6 Pah-enu2 (PKU) mice and wild-type mice received a LNAA supplemented diet, an isonitrogenic/isocaloric high-protein control diet, or normal chow. After six weeks of dietary treatment, blood and brain amino acid and monoaminergic neurotransmitter concentrations were assessed. Results In PKU mice, the investigated LNAA supplementation regimen significantly reduced blood and brain phenylalanine concentrations by 33% and 26%, respectively, compared to normal chow (p<0.01), while alleviating brain deficiencies of some but not all supplemented LNAA. Moreover, LNAA supplementation in PKU mice significantly increased brain serotonin and norepinephrine concentrations from 35% to 71% and from 57% to 86% of wild-type concentrations (p<0.01), respectively, but not brain dopamine concentrations (p=0.307). Conclusions This study shows that LNAA supplementation without dietary phenylalanine restriction in PKU mice improves brain biochemistry through all three hypothesized biochemical mechanisms. Thereby, these data provide proof-of-concept for LNAA supplementation as a valuable alternative dietary treatment strategy in PKU. Based on these results, LNAA treatment should be further optimized for clinical application with regard to the composition and dose of the LNAA supplement, taking into account all three working mechanisms of LNAA treatment. 134

136 Chapter 8 1. INTRODUCTION Phenylketonuria (PKU; OMIM ) is the first disorder in which severe neurocognitive dysfunction could be prevented by dietary treatment. It is caused by a deficient activity of the hepatic enzyme phenylalanine hydroxylase (PAH; EC ), normally converting phenylalanine (Phe) to tyrosine. If left untreated, classical PKU symptomatology is almost exclusively restricted to the brain, including severe neurocognitive dysfunction, seizures, and psychiatric problems, correlating with high blood Phe concentrations 1. This selective brain vulnerability to high blood Phe concentrations is hypothesized to be related to the transport characteristics for Phe at the blood-brain barrier (BBB) 2,3. At the BBB, the large neutral amino acid transporter 1 (LAT1) is the predominant transport system for all large neutral amino acids (LNAA), and is saturated for >95% 4. Combined with the fact that LAT1 shows a high affinity to Phe, increased blood Phe concentrations strongly increase brain Phe influx, outcompeting the transport of other LNAA 5-8. Based on both the increased Phe and the decreased non-phe LNAA transport across the BBB, different brain biochemical disturbances underlie brain dysfunction in PKU 3,9. High brain Phe concentrations have been found to be neurotoxic and to affect brain metabolism 10-14, while reduced brain availability of non-phe LNAA has been related to impaired cerebral protein synthesis 6,15. In addition, impaired cerebral monoaminergic neurotransmitter synthesis may result from outcompeted brain uptake of their amino acid precursors tyrosine and tryptophan 16, and/or from an inhibitory effect of high brain Phe concentrations on tyrosine hydroxylase (TH) and tryptophan hydroxylase (TPH) 17. Thus far, blood Phe reduction has been the primary target of treatment in PKU. This can be accomplished by a severe Phe-restricted diet and, in some patients, by pharmacological treatment with tetrahydrobiopterin. However, early- and continuously treated PKU patients still show impaired executive functioning and are prone to develop anxiety and depressive symptoms 18,19. Moreover, the Phe-restricted diet is socially demanding and hard to comply with 20. Therefore, an alternative pathophysiology-based treatment strategy directly targeting the brain is required. One such possible treatment strategy includes supplementation of LNAA (without Phe) that aims to restore the disturbed LNAA transport across the BBB without dietary Phe restriction. Based on aforementioned hypotheses on PKU pathophysiology, LNAA supplementation could serve to: 1) decrease brain Phe, 2) increase brain non-phe LNAA, and/or 3) increase brain monoaminergic neurotransmitter concentrations Previous research on LNAA treatment in PKU has primarily focused on brain and blood Phe reduction as treatment objective In addition, LNAA supplementation in PKU patients was recently found to increase blood and urine melatonin concentrations, which might reflect increased brain serotonin synthesis 29. However, not all brain biochemical treatment 135

137 Chapter 8 objectives have been investigated. In consequence, optimal composition and dosing of LNAA treatment is currently unknown, limiting its clinical application. As a first step to develop an optimal LNAA treatment regimen, the aim of the present study was to assess all abovementioned hypothesized biochemical treatment objectives of LNAA supplementation in a PKU mouse model. 2. MATERIAL & METHODS 2.1 Animals To establish a new breeding colony, breeding pairs of C57Bl/6 Pah-enu2 mice had been kindly provided by Prof. B. Thöny from the division of Clinical Chemistry and Biochemistry of the University Children s Hospital in Zurich in Switzerland. From heterozygous (+/-) mating, wild-type (WT, +/+), heterozygous, and PKU (-/-) mice of both sexes were obtained. After weaning at four weeks of age, genetic characterization was performed by quantitative PCR analysis on DNA extracted from ear tissue. Animals were housed individually at 21±1ºC on a 12-hr light-dark cycle (7:00 am-7:00 pm), and water and AM-II food pellets (Arie Block BV, Woerden, The Netherlands) were offered ad libitum. In total, 42 WT (21 male, 21 female) and 46 PKU (23 male, 23 female) mice were used. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (S1 File; ARRIVE guidelines checklist). The protocol was approved by the Institutional Animal Care and Use Committee of the University of Groningen (Permit Number: 6504A). 2.2 Experimental design At postnatal day 37, animals were assigned to one of three treatment groups based on genotype and sex, receiving either normal chow, a high-protein diet, or a LNAA supplemented diet. Dietary treatment was continued for six weeks. During the first week of treatment, body weight and food intake were measured daily, after which body weight and food intake were determined at weekly intervals. Food intake was manually assessed by the difference between food given and left on the cages tops using a scale. Spilled food was not measured because it has been shown to represent less than 0.1 g/day/mouse 30. After six weeks of dietary treatment, animals were euthanized by combined heart puncture and decapitation under inhalation-anesthetics with isoflurane 2-3 hours after the beginning of the light phase. 2.3 Experimental diets The basal diet during the experiment was based on the composition of AIN-93M 31, which was also administered in unadjusted form to the untreated control group (normal chow). The experimental LNAA diet was based on the LNAA regimen as used in the study by Pietz et al. 26. It was produced by adding LNAA to the basal diet at the expense of cornstarch on a weight-for-weight basis. The added amount of LNAA was equal to the amount of 136

138 Chapter 8 (natural) protein in the basal diet, in effect doubling the amount of protein/amino acids in the LNAA supplemented diet. The LNAA added in the LNAA supplemented diet included equal amounts of l-tyrosine, l-tryptophan, l-valine, l-isoleucine, l-leucine, l-methionine, and l-histidine. To control for the extra amount of protein in the LNAA supplemented diet, a high-protein control diet was included. This high-protein diet was produced by adding extra casein to the AIN-93M diet at the expense of cornstarch on a weight-for-weight basis, and was calculated to result in an isonitrogenic and isocaloric control diet for the experimental LNAA diet. Diets were prepared by Research Diet Services B.V. (Wijk bij Duurstede, The Netherlands). 2.4 Biochemical analyses To obtain plasma and brain material for biochemical analyses, blood was collected by heart puncture and whole brains were removed. Blood was centrifuged at 1500 g for 10 min and plasma was collected and stored at -80 ºC until further analysis. The cerebrum was snap frozen in liquid nitrogen and stored at -80 ºC until further preparation. Frozen cerebrum was crushed in liquid nitrogen and divided into aliquots. Frozen brain powder for amino acid measurements was processed to 20% (weight: volume (w:v)) homogenates in phosphatebuffered saline (ph 7.4), and for tryptophan, indole and catecholamine measurements to 2% (w:v) homogenates in acetic acid (0.08 M). Brain homogenates were sonified on ice at W. Next, samples were centrifuged at 800 rcf for 10 min (4 ºC), and the supernatant/ internatant was put on ice to be used for further analysis. For brain amino acid (except for tryptophan) measurements, norleucine in sulfosalicylic acid was added as an internal standard to the 20% brain homogenate (1:1, v:v). Samples were vortexed and centrifuged at rcf for 4 min. Plasma amino acid measurements were performed according to the same method, using 50 µl plasma instead of 20% brain homogenate. Amino acid concentrations were measured with a method based on ion exchange chromatography with post column derivatization with Ninhydrin on a Biochrom 30+ analyser (Pharmacia Biotech, Cambridge, UK). 8 For tryptophan and monoaminergic neurotransmitter measurements, an anti-oxidative solution was prepared in demineralised water (0.4 g/l ascorbic acid and g/l ethylenediaminetetraacetic acid). For tryptophan and indole measurements, 25 µl of the anti-oxidative solution was added to 25 µl of the 2% brain homogenate. For catecholamine measurements, 40 µl of the anti-oxidative solution was added to 10 µl of the 2% brain homogenate. Plasma tryptophan measurements were performed using 25 µl plasma instead of 2% brain homogenate. Analysis of tryptophan and monoaminergic neurotransmitter concentrations was performed using liquid chromatography in combination with isotope dilution mass spectrometry, essentially as described by Van de Merbel et al

139 Chapter Statistical analyses Statistical analyses were performed using the software IBM SPSS Statistics for Windows, Version (Armonk, NY: IBM Corp.). Data of the one animal that was euthanized preterm, were excluded from analyses. All tests were performed two-sided at a significance level of α=0.05. Analyses on brain and blood biochemistry as well as on weekly food intake (per g body weight) were performed by two-way ANOVA with genotype and diet as independent variables. In case of not normally distributed data (assessed by Shapiro-Wilk test) or unequal variances (assessed by Levene s test), analyses were performed on log-transformed data. If the interaction between genotype and diet and/or a main effect of diet was found to be significant, data were further analyzed by one-way ANOVA and Tukey s post-hoc tests for PKU and WT mice separately. The effect of dietary treatment on body weight was analyzed for WT and PKU male and female mice separately by repeated-measures ANOVA with one between factor (diet, three levels: normal chow, high-protein diet, and LNAA diet) and one within factor (time, 7 levels: 0, 1, 2, 3, 4, 5, and 6 weeks) and Tukey s post-hoc analysis. To investigate whether brain Phe concentrations in PKU mice were primarily determined by blood Phe concentrations or by dietary treatment, multiple linear regression analysis was performed with blood Phe concentrations and dietary treatment as independent variables. Data are expressed as mean ± standard deviation (SD). 3. RESULTS 3.1 Food and LNAA intake Amino acid contents of the different diets are presented in Table 1. Weekly food intake (expressed as g food/g body weight/week), as shown in Fig 1, decreased during the first weeks for all treatment groups, stabilizing in the later weeks of dietary treatment. In both the fifth and sixth week of dietary treatment, two-way ANOVA analyses showed a significant main effect of genotype on food intake (p<0.01). Moreover, in both weeks, a significant main effect of dietary treatment (p<0.01), and a significant interaction between genotype and dietary treatment on food intake was observed (p<0.01 for week 5, p<0.05 for week 6). In PKU mice, food intake in both weeks was lower on high-protein diet than on normal chow (p<0.05), while being even lower on LNAA supplemented diet (p<0.01). In WT mice, food intake in both weeks did not significantly differ for any of the dietary treatments (p=0.376 and p=0.322). Based on the weekly food intakes and amino acid contents of the different diets, mean daily intakes of individual LNAA during the six week 138

140 Chapter 8 dietary treatment were calculated for all experimental groups as shown in the Supplemental Table 1 (S1 Table). Table 1. Nutritional content of the experimental diets (g/kg diet). Nutritional content* normal chow LNAA diet high-protein diet Carbohydrates Fat Dietary fibre Protein Amino acids** LNAA L-Phenylalanine (0.1) 12.2 (6.2) L-Tyrosine (15.3) 12.1 (7.3) L-Valine (17.6) 15.9 (8.5) L-Isoleucine (17.0) 12.4 (6.4) L-Leucine (17.1) 23.1 (12.2) L-Methionine (16.7) 6.9 (3.9) L-Histidine (15.8) 6.8 (3.5) L-Threonine (0.3) 11.3 (5.9) non-lnaa L-Aspartic acid (0.4) 19.4 (10.1) L-Serine (0.2) 15.8 (8.1) L-Glutamic acid (1.7) 59.6 (31.4) Glycine (0.0) 5.2 (2.5) L-Alanine (0.3) 8.2 (4.3) L-Lysine (0.8) 19.5 (10.4) L-Arginine (0.5) 9.5 (5.3) Contents are not shown for L-Tryptophan, L-Proline, and L-cyst(e)ine, as these were not measured due to technical limitations. Differences with LNAA contents of normal chow are stated in brackets. LNAA, large neutral amino acid. *Mineral and vitamin premixes were also included in accordance with the composition of the AIN93M diet 30. **Dietary contents as measured in samples of prepared food pellets

141 Chapter 8 Fig 1. Weekly food intake for WT (A) and PKU (B) mice on different diets Numbers of mice on normal chow, LNAA supplemented diet, and high-protein diet were n=14, n=14, and n=14 for WT mice respectively, while being n=15, n=14, and n=15 for PKU mice. Error bars represent SEM. 3.2 Body weight and general health At initiation of treatment, body weight of PKU mice (male 13.7 ± 3.6 g; female: 12.8 ± 2.1 g) was lower than of WT mice (male 18.9 ± 1.5 g; female: 15.6 ± 1.8 g), but did not significantly differ between dietary groups in PKU nor WT mice. Body weight curves during treatment were significantly lower for WT female mice on LNAA supplemented diet compared to either normal chow (p<0.05) or high-protein diet (p<0.01), but did not significantly differ between dietary treatments for WT male (p=0.485) and PKU male or PKU female mice (p=0.397 and p=0.343) (Fig 2). During the experiment, one animal (WT female on LNAA supplemented diet) was euthanized on the 19 th day after inclusion, because of too much weight loss. Pathological examination showed hydrocephalus, which is sometimes found in C57Bl/6 (both WT and PKU) mice. Fig 2. Body weights during the experiment Mean body weights for A) male and B) female WT (dashed lines) and PKU (solid lines) mice on different diets. Numbers of mice on normal chow, LNAA supplemented diet, and high-protein diet were n=14, n=14, and n=14 for WT mice respectively, while being n=15, n=14, and n=15 for PKU mice. Error bars represent SEM. 140

142 Chapter Plasma amino acids Plasma concentrations of individual LNAA in PKU and WT mice on control and LNAA supplemented diets are depicted in Fig 3. Two-way ANOVA analyses showed a significant main effect of genotype on plasma Phe, tyrosine, tryptophan (p<0.01), and threonine concentrations (p<0.05). In PKU compared to WT mice on normal chow, plasma Phe concentrations were increased by almost 30-fold. Plasma tyrosine, tryptophan, and threonine concentrations in PKU mice were reduced to 55%, 77%, and 79%, respectively, of concentrations in WT mice on normal chow. For other LNAA, no significant main effect of genotype was observed on plasma concentrations. A significant main effect of dietary treatment was observed on plasma concentrations of all LNAA except for threonine (p=0.073 for threonine; p<0.05 for tyrosine; p<0.01 for all other LNAA). Moreover, two-way ANOVA analyses showed a significant interaction between genotype and dietary treatment on plasma Phe and methionine concentrations (p<0.01 for both). In PKU mice, plasma Phe concentrations on LNAA supplemented diet were significantly reduced to 67% of concentrations on normal chow (p<0.01), while plasma Phe concentrations on high-protein diet were significantly higher than on normal chow (p<0.01). In WT mice on LNAA, plasma Phe concentrations were reduced compared to control diets (p<0.05), but concentrations on high-protein diet did not significantly differ from those on normal chow (p=0.929). Plasma concentrations of supplemented LNAA were higher on LNAA supplementation compared to control diets, both in PKU and WT mice, although this did not reach statistical significance for all LNAA. Plasma concentrations of non-lnaa amino acids in PKU and WT mice on control and LNAA supplemented diets are presented in Supplemental table 2 (S2 Table). In both PKU and WT mice, plasma glycine and lysine concentrations on LNAA supplemented diet were lower compared to control diets (p<0.05), just not reaching statistical significance for lysine in PKU mice (p=0.051). In addition, plasma serine concentrations were lower on LNAA supplementation compared to control diets in WT mice (p<0.05), but not in PKU mice (p=0.407)

143 Chapter 8 Fig 3. Plasma LNAA concentrations Plasma concentrations of A) phenylalanine, B) tyrosine, C) tryptophan, D) valine, E) isoleucine, F) leucine, G) methionine, H) histidine, and I) threonine in WT and PKU mice after six weeks of receiving different diets. Numbers of mice on normal chow, LNAA supplemented diet, and high-protein diet were n=14, n=12, and n=14 for WT mice respectively, while being n=14, n=12, and n=15 for PKU mice. Untransformed data are expressed as mean ± SEM. * p<0.05; ** p<0.01; p<0.05 and p<0.01 compared to WT mice on normal chow. 3.4 Brain amino acids Brain concentrations of individual LNAA in PKU and WT mice on control and LNAA supplemented diets are depicted in Fig 4. Two-way ANOVA analyses showed a significant main effect of genotype on brain concentrations of all LNAA except for methionine (p<0.01 for all LNAA but methionine, p=0.523 for methionine). In PKU compared to WT mice on normal chow, brain Phe concentrations were increased by 8.3-fold. Also, brain histidine concentrations in PKU mice were elevated to 139% of WT concentrations on normal chow. Brain concentrations of all other non-phe LNAA but methionine were reduced in PKU mice on normal chow, ranging from 52% of concentrations in corresponding WT mice for tyrosine to 77% for leucine. A significant main effect of dietary treatment was observed on brain Phe, tryptophan, methionine, and threonine concentrations (p<0.01 for all). Moreover, two-way ANOVA analyses showed a significant interaction between genotype and dietary treatment on brain Phe, methionine, and histidine concentrations (p<0.01 for Phe and methionine, p<0.05 for histidine). Brain Phe concentrations in PKU mice on LNAA supplementation were reduced to 142

144 Chapter 8 74% of concentrations on normal chow (p<0.01), still being 6.1-fold higher when compared to WT concentrations on normal chow. Brain Phe concentrations in PKU mice on highprotein diet were higher than in PKU mice on normal chow (p<0.05). Brain tryptophan, histidine, and methionine concentrations in PKU mice on LNAA supplementation were significantly higher when compared to PKU mice on normal chow (p<0.05), resulting in concentrations of 100% and 166% of WT concentrations on normal chow for tryptophan and histidine, and an elevation by 3.6-fold for methionine. In contrast, brain threonine concentrations in PKU mice on LNAA supplementation were lower when compared to PKU mice on control diets (p<0.01). In WT mice, brain Phe concentrations on LNAA supplementation tended to be lower when compared to normal chow, although this just did not reach statistical significance (p=0.053). Similar to PKU mice, brain methionine concentrations were higher, whereas brain threonine concentrations were lower on LNAA supplementation as compared to control diets (p<0.01 for all). Brain concentrations of non-lnaa amino acids in PKU and WT mice on control and LNAA supplemented diets are presented in Supplemental Table 3 (S3 Table). In both PKU and WT mice, brain serine and glycine concentrations on LNAA supplemented diet were lower compared to control diets (p<0.05), although this did not reach statistical significance for glycine in PKU mice on LNAA supplementation compared to high-protein diet (p=0.064). In addition, in WT mice only, brain lysine concentrations were lower (p<0.01), whereas brain taurine concentrations were higher on LNAA supplementation compared to control diets (p<0.05)

145 Chapter 8 Fig 4 Brain LNAA concentrations Brain concentrations of A) phenylalanine, B) tyrosine, C) tryptophan, D) valine, E) isoleucine, F) leucine, G) methionine, H) histidine, and I) threonine in WT and PKU mice after six weeks of receiving different diets. Numbers of mice on normal chow, LNAA supplemented diet, and high-protein diet were n=13, n=12, and n=14 for WT mice respectively, while being n=14, n=12, and n=14 for PKU mice. Untransformed data are expressed as mean ± SEM. * p<0.05; ** p<0.01; p<0.05 and p<0.01 compared to WT mice on normal chow. 3.5 Brain monoaminergic neurotransmitters Brain monoaminergic neurotransmitter and associated metabolite concentrations in PKU and WT mice are depicted in Fig 5. In the catecholamine pathway, two-way ANOVA analyses showed significant main effects of genotype on brain dopamine and norepinephrine concentrations (p<0.01 for both). Brain dopamine and norepinephrine concentrations in PKU mice on normal chow were reduced to 76% and 57%, respectively, of concentrations in WT mice on the same diet. In the serotonergic pathway, significant main effects of genotype were observed on brain concentrations of serotonin and its associated metabolite 5-hydroxyindoleacetic acid (5-HIAA) (p<0.01 for both). Brain serotonin and 5-HIAA concentrations in PKU mice on normal chow were decreased to 35% and 26%, respectively, of concentrations in corresponding WT mice. Dietary treatment had no significant effect on brain catecholamine or serotonin concentrations in WT mice, while in PKU mice it did. A significant main effect of dietary treatment was 144

146 Chapter 8 observed on brain catecholamine, serotonin, and 5-HIAA concentrations (p<0.01 for all but dopamine, p<0.05 for dopamine). Moreover, two-way ANOVA analyses showed a significant interaction between genotype and dietary treatment on brain norepinephrine, serotonin, and 5-HIAA (p<0.01 for all), but not dopamine concentration (p=0.766). In the catecholamine pathway, LNAA supplementation in PKU mice resulted in significantly higher brain norepinephrine concentrations compared to normal chow (p<0.01), and partially restored its deficit to an average of 86% of concentrations in WT mice on normal chow. In contrast, brain dopamine concentrations did not significantly differ in PKU mice between the dietary treatments (p=0.307). In WT mice, no significant differences were observed between any of the dietary treatments for neither dopamine (p=0.104) nor norepinephrine (p=0.283). In the serotonergic pathway, in PKU mice on LNAA supplementation, both brain serotonin and 5-HIAA concentrations were significantly increased when compared to control diets (p<0.01), partially restoring their concentrations to an average of 71% and 67%, respectively, of concentrations in WT mice on normal chow. In WT mice on LNAA supplementation, brain serotonin concentrations tended to be higher compared to normal chow (p=0.051), and brain 5-HIAA concentrations tended to be higher compared to high-protein diet (p=0.097), but both did not reach statistical significance. 8 Fig 5 Brain monoaminergic neurotransmitter concentrations Brain concentrations of A) dopamine, B) norepinephrine, C) serotonin in WT and PKU mice after six weeks of receiving different dietary treatments. Numbers of mice on normal chow, LNAA supplemented diet, and high-protein diet were n=13, n=12, and n=14 for WT mice respectively, while being n=15, n=13, and n=15 for PKU mice. Untransformed data are expressed as mean ± SEM. **p<0.01; p<0.01 compared to WT mice on normal chow. 3.6 Relation between plasma and brain Phe To investigate whether the reduction of brain Phe concentrations on LNAA supplementation in PKU mice was primarily related to an effect at the BBB or especially related to reduced blood Phe concentrations, the relationship between blood and brain Phe concentrations was assessed in PKU mice on LNAA supplementation and control diets (Fig 6). Multiple linear regression analysis showed that brain Phe concentrations in PKU mice were significantly 145

147 Chapter 8 predicted (adjusted R 2 =0.801, F=75.525, p<0.01) by both blood Phe concentrations (B=0.127, SE B=0.026, β=0.521) and LNAA supplemented diet (B= , SE B=39.006, β=-0.451). Brain non-phe LNAA concentrations in PKU mice on normal chow, LNAA supplementation, and high-protein diet did not show a clear relationship with either their respective blood concentrations nor with blood Phe concentrations (data not shown). Fig 6 Plasma versus brain Phe concentrations in PKU mice on different dietary treatments Relationship between plasma Phe and brain Phe concentrations in PKU mice on normal chow (n=13), LNAA supplemented diet (n=12), and high-protein diet (n=14). 3.7 Relation between brain monoaminergic neurotransmitters and their precursors To investigate whether the increase of brain serotonin and norepinephrine on LNAA supplementation in PKU mice was primarily due to (1) increased brain availability of their precursors, or to (2) enhanced conversion of their precursors, brain monoaminergic neurotransmitters were assessed in relation to their respective (amino acid) precursors (Fig 7). Two-way ANOVA analyses showed a significant main effect of genotype on ratios of brain dopamine/tyrosine, norepinephrine/dopamine, and serotonin/tryptophan (p<0.01 for all), but not norepinephrine/tyrosine concentrations (p=0.184). Ratios of brain dopamine/ tyrosine were increased, while rations of brain norepinephrine/dopamine and serotonin/ tryptophan concentrations were reduced in PKU compared to WT mice on normal chow. A significant main effect of dietary treatment was observed on ratios of brain norepinephrine/ dopamine concentrations only (p<0.01). In addition, two-way ANOVA analyses showed 146

148 Chapter 8 a significant interaction between genotype and dietary treatment on ratios of brain norepinephrine/dopamine and serotonin/tryptophan concentrations (p<0.01 for both). In PKU mice on LNAA supplementation, ratios of brain norepinephrine/dopamine were increased compared to control diets (p<0.01). Also, ratios of brain serotonin/tryptophan concentrations in PKU mice on LNAA supplementation were higher as compared to control diets (p<0.05). Fig 7 Ratios of brain monoaminergic neurotransmitters to precursors Ratios of brain A) dopamine/tyrosine, B) norepinephrine/tyrosine, C) norepinephrine/dopamine, and D) serotonin/tryptophan concentrations in WT and PKU mice after six weeks of receiving different dietary treatments. Numbers of mice on normal chow, LNAA supplemented diet, and high-protein diet were n=13, n=12, and n=14 for WT mice respectively, while being n=14, n=12, and n=14 for PKU mice. Untransformed data are expressed as mean ± SEM. *p<0.05; **p<0.01; and p<0.05; p<0.01 compared to WT mice on normal chow DISCUSSION This study is the first to investigate all hypothesized biochemical treatment effects of LNAA supplementation using one LNAA supplementation regimen and a single experimental design. Besides reducing blood Phe concentrations, the present study showed that LNAA supplementation without dietary Phe restriction in PKU mice could directly improve brain biochemistry through three mechanisms: 1) reducing brain Phe concentrations, 2) attenuating the brain deficiencies of some, but not all, LNAA, and 3) increasing brain serotonin and norepinephrine, but not dopamine concentrations. Before discussing these results in more detail, we will first address some methodological issues. 147

149 Chapter 8 Firstly, the present experiment was performed in Pah-enu2 (PKU) mice, a well-established model that resembles the genetics, biochemistry and neurobiology of classical PKU in humans. Biochemically, PKU mice show high blood and brain Phe concentrations in combination with brain non-phe LNAA and monoaminergic neurotransmitter deficits that also characterize human PKU biochemistry. K m -values for LNAA transport at the BBB have not been determined for (PKU) mice 33. However, K m -values for individual LNAA as determined in vivo in rats and in human brain capillaries showed a significant correlation 34, while being approximately 8-40 times lower for humans than for rats. This could imply that competition between Phe and non-phe LNAA for transport across the BBB takes place at lower plasma concentrations in humans than in rats (and maybe mice), so that LNAA supplementation might be even more effective in PKU patients. One of the important advantages of this PKU mouse model over clinical studies is the possibility to measure not only Phe (that can be determined in humans by magnetic resonance spectroscopy (MRS)) but also non-phe LNAA and monoaminergic neurotransmitter concentrations in brain (that at present cannot be measured by MRS). Secondly, the LNAA supplemented diet used was based on the study by Pietz et al. (1999) that investigated the effect of concomitant LNAA supplementation during an oral Phe challenge on brain Phe uptake and EEG activity in PKU patients 26. The LNAA supplementation regimen used by Pietz et al. consisting of equal amounts (150 mg/kg body weight) of all non- Phe LNAA except for threonine, in total, approximated the daily dietary protein intake for adults. To translate this acute regimen for PKU patients to chronic treatment in PKU mice, the total amount of added LNAA in the present study was equal to the amount of natural protein in the basal diet. In full accordance with the study by Pietz et al., threonine was not supplemented, even though Sanjurjo et al. (2003) have shown threonine supplementation alone (50 mg/kg/d) reduced blood Phe concentrations by 36% in PKU patients LNAA supplementation reduces blood Phe concentrations Although LNAA supplementation is suggested to improve brain metabolism primarily by restoring the unbalanced LNAA transport at the BBB, LNAA supplementation in PKU mice was also found to significantly reduce blood Phe concentrations to 67% of concentrations on normal chow. This is in accordance with previous studies on LNAA supplementation in PKU mice showing plasma Phe reductions to % of concentrations in untreated PKU controls 23,24. LNAA supplementation has been hypothesized to exert this effect through competition with Phe for uptake at the gut-blood barrier 21, or through increased Phe utilization due to higher net protein synthesis 21. In support of this last hypothesis, food intake of PKU mice during the final weeks of the experiment was significantly lower on LNAA supplemented diet than on control diets, while body weight did not significantly differ. This may suggest that LNAA supplementation indeed induced anabolism in PKU mice, thereby demanding a lower dietary protein (and thus food) intake, and by that 148

150 Chapter 8 contributing to the observed plasma Phe reduction. As expected, plasma concentrations of supplemented LNAA in PKU mice were all increased. Surprisingly, however, plasma tyrosine concentrations in PKU mice remained low-normal on the currently applied LNAA supplemented diet. Even when correcting for the 20% of the supplemented tyrosine that is assumed not to be absorbed in the gastrointestinal tract 21, tyrosine intake still was 3.5 times higher in the LNAA supplemented diet compared to normal chow. Similar to plasma tyrosine concentrations, brain tyrosine concentrations were not significantly increased on LNAA supplemented diet either, while brain norepinephrine - the end product of tyrosinederived neurotransmitter metabolism - was significantly increased on LNAA supplemented diet. Therefore, a possible explanation might be that all supplemented tyrosine was used to partly restore the profound brain norepinephrine deficiency, and thereby did not result in increased plasma and brain tyrosine concentrations. 4.2 LNAA supplementation reduces brain Phe concentrations, while attenuating brain deficiencies of some but not all non-phe LNAA In brain, Phe concentrations on LNAA supplementation in PKU mice were significantly reduced by 26% compared to concentrations on normal chow. This is in good agreement with previous studies on LNAA supplementation in both PKU patients 25 and mice 24, showing brain Phe reductions of 20-46%. Moreover, it may well support the finding of a clear competitive effect on Phe transport across the BBB in PKU patients by Pietz et al. (1999) using a comparable LNAA supplement 26. As besides brain Phe, blood Phe concentrations were also reduced in PKU mice on LNAA supplementation, the question arises whether the reduced brain Phe concentrations might be due to the reduced plasma Phe concentrations rather than a direct effect at the BBB level. Multiple linear regression analysis suggests, however, that LNAA supplementation reduced brain Phe concentrations in PKU mice through a combined effect of both plasma Phe reduction and enhanced competition at the BBB. 8 This is the first time that brain non-phe LNAA concentrations have been reported in response to LNAA supplementation in PKU. The LNAA supplementation regimen that was used, changed most brain non-phe LNAA concentrations, restoring brain tryptophan in PKU mice to WT concentrations. Brain methionine concentrations were even significantly increased by 5.5-fold in PKU mice on LNAA supplementation compared to normal chow, corresponding with the similarly strong increases in blood. Although the cerebral and systemic effects and possible toxicity due to these strongly elevated methionine concentrations are not fully understood 36, at least these results warrant against indiscriminate supplementation of methionine in PKU. Also, brain histidine concentrations in PKU mice were even further increased on LNAA supplementation. From the fact that histidinaemia, in which brain histidine concentrations are much more increased, is not associated with any brain dysfunction, we may conclude that this elevation of brain histidine as observed in PKU mice 149

151 Chapter 8 probably does not have clinical significance 37. On the other hand, brain concentrations of threonine, which was not included in the LNAA supplement, were significantly reduced in both PKU and WT mice on LNAA supplemented diet. This further supports the idea that highly unbalanced LNAA intake may induce brain deficiencies of some LNAA. It can be concluded from these results that LNAA supplementation in PKU mice indeed attenuates brain Phe concentrations and attenuates brain deficiencies of (at least some) non- Phe LNAA. At the same time, results suggest that the relationships between brain non-phe LNAA concentrations and their respective plasma concentrations as well as plasma Phe concentrations are complex and differ for each non-phe LNAA, given the amount of LNAA supplementation used in this study. Development of the optimal LNAA supplementation regimen that can both effectively reduce brain Phe concentrations and improve brain concentration of all non-phe LNAA therefore clearly deserves further research. 4.3 LNAA treatment improves brain serotonin and norepinephrine, but not dopamine, concentrations Besides its effect on brain LNAA concentrations, LNAA supplementation in PKU mice significantly increased brain serotonin from 35% to 71% of concentrations in WT mice. Also, brain norepinephrine in PKU mice on LNAA supplementation increased from 57% to 86% of concentrations in WT mice, whereas brain dopamine concentrations remained unchanged. Although brain monoaminergic neurotransmitter concentrations in response to LNAA supplementation have not been reported previously, a recent study on LNAA supplementation in PKU patients showed increased melatonin (a serotonin metabolite) concentrations in plasma and urine, which - according to the authors- could be a possible new marker for brain serotonin synthesis in PKU patients 29. As C57Bl/6 is one of many mouse strains being deficient in melatonin 38, unfortunately, we were unable to correlate brain serotonin and plasma melatonin concentrations. Regarding the clinical importance of brain monoaminergic neurotransmitters in PKU, traditionally, especially brain dopamine deficiency has been associated with cognitive and mood disturbances in PKU 39,40. However, brain norepinephrine impairments may have been underestimated, while cerebral norepinephrine abnormalities have been associated with many (neuro)psychiatric disorders 41. Both insufficient precursor availability and impaired TH and TPH activity by inhibition of excessive brain Phe concentrations have been hypothesized to account for the brain monoaminergic neurotransmitter deficits in PKU. The present results suggest that the relative contribution of each of these mechanisms may be different for the dopaminergic and serotoninergic pathways in PKU. Regarding the brain catecholamine deficiencies, the increased ratio of brain dopamine/tyrosine and unaffected ratio of brain norepinephrine/ tyrosine in PKU mice could be explained in two ways, each supporting one of the two aforementioned main theories on brain monoaminergic neurotransmitter deficiencies in 150

152 Chapter 8 PKU. Firstly, it may indicate that insufficient brain tyrosine availability rather than inhibition of TH by high Phe would be responsible for the brain catecholamine deficiencies observed in PKU. This would support the report by Fernstrom et al. (2007) concluding that increased brain Phe is not likely to impair catecholamine synthesis in PKU, whereas low brain tyrosine does 42. In consequence, this would implicate that even higher blood tyrosine concentrations may be needed to restore brain dopamine concentrations. Secondly, the increased ratio of brain dopamine/tyrosine and unaffected ratio of brain norepinephrine/tyrosine in PKU compared to WT mice on normal chow may be explained by the fact that brain dopamine is not exclusively derived from brain tyrosine. This would support the report by Ikeda et al. (1967) showing that TH, at least in vitro, could synthesize catecholamines also from Phe, which is abundant in the PKU brain. This would suggest that insufficient precursor amino acid availability in brain would not be the primary mechanism underlying reduced dopamine concentrations, but inhibition of TH by high Phe is 43. In consequence, this would imply that further reduction of brain Phe concentrations would probably be most effective to increase brain dopamine concentrations in PKU. Regarding the observed serotonin deficits in PKU, the reduced ratios of brain serotonin/tryptophan in PKU mice suggest that inhibition of TPH by high Phe does play an important role in the cerebral serotonin impairments characterizing PKU. This is in good agreement with the in vitro observation that Phe inhibits TPH more strongly than TH 44. To further discriminate between the importance of both hypothesized mechanisms underlying brain monoaminergic neurotransmitter impairments in PKU, future studies need to investigate the effects of selective brain tyrosine and tryptophan increase and selective brain Phe reduction on monoaminergic neurotransmitter concentrations in PKU mice. To conclude, this study was the first to investigate all hypothesized biochemical treatment objectives of LNAA supplementation in PKU. Results in PKU mice showed that LNAA supplementation improves brain biochemistry in PKU by three synergistic mechanisms. Thereby, this study provides proof-of-concept for LNAA supplementation as a possible alternative treatment strategy for PKU that improves brain biochemistry by targeting the unbalanced LNAA transport across the BBB. Before clinical application should be considered, however, further optimization of LNAA treatment with regard to the LNAA being supplemented and their dose is required, taking into account all three brain biochemical treatment objectives. 8 ACKNOWLEDGEMENTS The authors express their gratitude to Mrs. H.A. Kingma, Mrs. E.Z. Jonkers, Mrs. K. Boer, and Mrs. H. Adema for their analytical support. 151

153 Chapter 8 REFERENCES 1 Blau N, van Spronsen FJ, Levy HL. Phenylketonuria. Lancet 2010 Oct 23;376(9750): Pardridge WM. Blood-brain barrier carriermediated transport and brain metabolism of amino acids. Neurochem Res 1998 May;23(5): van Spronsen FJ, Hoeksma M, Reijngoud DJ. Brain dysfunction in phenylketonuria: is phenylalanine toxicity the only possible cause? J Inherit Metab Dis 2009 Feb;32(1): Smith QR. Transport of glutamate and other amino acids at the blood-brain barrier. J Nutr 2000 Apr;130(4S Suppl):1016S-22S. 5 Knudsen GM, Hasselbalch S, Toft PB, Christensen E, Paulson OB, Lou H. Bloodbrain barrier transport of amino acids in healthy controls and in patients with phenylketonuria. J Inherit Metab Dis 1995;18(6): de Groot MJ, Hoeksma M, Reijngoud DJ, de Valk HW, Paans AM, Sauer PJ, et al. Phenylketonuria: reduced tyrosine brain influx relates to reduced cerebral protein synthesis. Orphanet J Rare Dis 2013 Sep 4;8(1): Shulkin BL, Betz AL, Koeppe RA, Agranoff BW. Inhibition of neutral amino acid transport across the human blood-brain barrier by phenylalanine. J Neurochem 1995 Mar;64(3): Oldendorf WH, Sisson BW, Silverstein A. Brain uptake of selenomethionine Se 75. II. Reduced brain uptake of selenomethionine Se 75 in phenylketonuria. Arch Neurol 1971 Jun;24(6): de Groot MJ, Hoeksma M, Blau N, Reijngoud DJ, van Spronsen FJ. Pathogenesis of cognitive dysfunction in phenylketonuria: review of hypotheses. Mol Genet Metab 2010;99 Suppl 1:S de Groot MJ, Sijens PE, Reijngoud DJ, Paans AM, van Spronsen FJ. Phenylketonuria: brain phenylalanine concentrations relate inversely to cerebral protein synthesis. J Cereb Blood Flow Metab 2014 Oct Glushakov AV, Glushakova O, Varshney M, Bajpai LK, Sumners C, Laipis PJ, et al. Long-term changes in glutamatergic synaptic transmission in phenylketonuria. Brain 2005 Feb;128(Pt 2): Martynyuk AE, Glushakov AV, Sumners C, Laipis PJ, Dennis DM, Seubert CN. Impaired glutamatergic synaptic transmission in the PKU brain. Mol Genet Metab 2005 Dec;86 Suppl 1:S Shefer S, Tint GS, Jean-Guillaume D, Daikhin E, Kendler A, Nguyen LB, et al. Is there a relationship between 3-hydroxy-3- methylglutaryl coenzyme a reductase activity and forebrain pathology in the PKU mouse? J Neurosci Res 2000 Sep 1;61(5): Horster F, Schwab MA, Sauer SW, Pietz J, Hoffmann GF, Okun JG, et al. Phenylalanine reduces synaptic density in mixed cortical cultures from mice. Pediatr Res 2006 Apr;59(4 Pt 1): Smith CB, Kang J. Cerebral protein synthesis in a genetic mouse model of phenylketonuria. Proc Natl Acad Sci U S A 2000 Sep 26;97(20): Hommes FA, Lee JS. The control of 5-hydroxytryptamine and dopamine synthesis in the brain: a theoretical approach. J Inherit Metab Dis 1990;13(1):

154 Chapter 8 17 Pascucci T, Andolina D, Mela IL, Conversi D, Latagliata C, Ventura R, et al. 5-Hydroxytryptophan rescues serotonin response to stress in prefrontal cortex of hyperphenylalaninaemic mice. Int J Neuropsychopharmacol 2009 Sep;12(8): Smith I, Knowles J. Behaviour in early treated phenylketonuria: a systematic review. Eur J Pediatr 2000 Oct;159 Suppl 2:S Jahja R, Huijbregts SC, de Sonneville LM, van der Meere JJ, van Spronsen FJ. Neurocognitive evidence for revision of treatment targets and guidelines for phenylketonuria. J Pediatr 2014 Apr;164(4): e2. 20 MacDonald A. Diet and compliance in phenylketonuria. Eur J Pediatr 2000 Oct;159 Suppl 2:S van Spronsen FJ, de Groot MJ, Hoeksma M, Reijngoud DJ, van Rijn M. Large neutral amino acids in the treatment of PKU: from theory to practice. J Inherit Metab Dis 2010 Dec;33(6): Matalon R, Michals-Matalon K, Bhatia G, Burlina AB, Burlina AP, Braga C, et al. Double blind placebo control trial of large neutral amino acids in treatment of PKU: effect on blood phenylalanine. J Inherit Metab Dis 2007 Apr;30(2): Matalon R, Michals-Matalon K, Bhatia G, Grechanina E, Novikov P, McDonald JD, et al. Large neutral amino acids in the treatment of phenylketonuria (PKU). J Inherit Metab Dis 2006 Dec;29(6): Matalon R, Surendran S, Matalon KM, Tyring S, Quast M, Jinga W, et al. Future role of large neutral amino acids in transport of phenylalanine into the brain. Pediatrics 2003 Dec;112(6 Pt 2): Moats RA, Moseley KD, Koch R, Nelson M,Jr. Brain phenylalanine concentrations in phenylketonuria: research and treatment of adults. Pediatrics 2003 Dec;112(6 Pt 2): Pietz J, Kreis R, Rupp A, Mayatepek E, Rating D, Boesch C, et al. Large neutral amino acids block phenylalanine transport into brain tissue in patients with phenylketonuria. J Clin Invest 1999 Apr;103(8): Schindeler S, Ghosh-Jerath S, Thompson S, Rocca A, Joy P, Kemp A, et al. The effects of large neutral amino acid supplements in PKU: an MRS and neuropsychological study. Mol Genet Metab 2007 May;91(1): Koch R, Moseley KD, Yano S, Nelson M,Jr, Moats RA. Large neutral amino acid therapy and phenylketonuria: a promising approach to treatment. Mol Genet Metab 2003 Jun;79(2): Yano S, Moseley K, Azen C. Large neutral amino acid supplementation increases melatonin synthesis in phenylketonuria: a new biomarker. J Pediatr 2013 May;162(5): Bachmanov AA, Reed DR, Tordoff MG, Price RA, Beauchamp GK. Nutrient preference and diet-induced adiposity in C57BL/6ByJ and 129P3/J mice. Physiol Behav 2001 Mar;72(4): Reeves PG, Nielsen FH, Fahey GC,Jr. AIN- 93 purified diets for laboratory rodents: final report of the American Institute of Nutrition ad hoc writing committee on the reformulation of the AIN-76A rodent diet. J Nutr 1993 Nov;123(11): van de Merbel NC, Hendriks G, Imbos R, Tuunainen J, Rouru J, Nikkanen H. Quantitative determination of free and total dopamine in human plasma by LC-MS/ MS: the importance of sample preparation. Bioanalysis 2011 Sep;3(17):

155 Chapter 8 33 Martynyuk AE, van Spronsen FJ, Van der Zee EA. Animal models of brain dysfunction in phenylketonuria. Mol Genet Metab 2010;99 Suppl 1:S Hargreaves KM, Pardridge WM. Neutral amino acid transport at the human bloodbrain barrier. J Biol Chem 1988 Dec 25;263(36): Sanjurjo P, Aldamiz L, Georgi G, Jelinek J, Ruiz JI, Boehm G. Dietary threonine reduces plasma phenylalanine levels in patients with hyperphenylalaninemia. J Pediatr Gastroenterol Nutr 2003 Jan;36(1): de Rezende MM, D Almeida V. Central and systemic responses to methionine-induced hyperhomocysteinemia in mice. PLoS One 2014 Aug 25;9(8):e Coulombe JT, Kammerer BL, Levy HL, Hirsch BZ, Scriver CR. Histidinaemia. Part III: Impact; a prospective study. J Inherit Metab Dis 1983;6(2): Roseboom PH, Namboodiri MA, Zimonjic DB, Popescu NC, Rodriguez IR, Gastel JA, et al. Natural melatonin knockdown in C57BL/6J mice: rare mechanism truncates serotonin N-acetyltransferase. Brain Res Mol Brain Res 1998 Dec 10;63(1): Christ SE, Huijbregts SC, de Sonneville LM, White DA. Executive function in early-treated phenylketonuria: profile and underlying mechanisms. Mol Genet Metab 2010;99 Suppl 1:S Surtees R, Blau N. The neurochemistry of phenylketonuria. Eur J Pediatr 2000 Oct;159 Suppl 2:S Sara SJ. The locus coeruleus and noradrenergic modulation of cognition. Nat Rev Neurosci 2009 Mar;10(3): Fernstrom JD, Fernstrom MH. Tyrosine, phenylalanine, and catecholamine synthesis and function in the brain. J Nutr 2007 Jun;137(6 Suppl 1):1539S-1547S; discussion 1548S. 43 Ikeda M, Levitt M, Udenfriend S. Phenylalanine as substrate and inhibitor of tyrosine hydroxylase. Arch Biochem Biophys 1967 May;120(2): Ogawa S, Ichinose H. Effect of metals and phenylalanine on the activity of human tryptophan hydroxylase-2: comparison with that on tyrosine hydroxylase activity. Neurosci Lett 2006 Jul 3;401(3):

156 Chapter 8 SUPPORTING INFORMATION S1 Table. Average LNAA intakes of the different experimental groups (mg/g body weight/day) WT PKU high-protein diet LNAA diet normal chow high-protein diet LNAA diet normal chow Phenylalanine 0.92 ± ± ± ± ± ± 0.13 Tyrosine 0.74 ± ± ± ± ± ± 0.13 Valine 1.14 ± ± ± ± ± ± 0.17 Isoleucine 0.91 ± ± ± ± ± ± 0.13 Leucine 1.68 ± ± ± ± ± ± 0.25 Methionine 0.46 ± ± ± ± ± ± 0.07 Histidine 0.49 ± ± ± ± ± ± 0.07 Threonine 0.82 ± ± ± ± ± ± 0.12 LNAA intake is given in mg/g body weight/day, expressed as mean ± SD. Dietary intake is not shown for Tryptophan, as this could not be measured in the food pellets. Numbers of mice on normal chow, LNAA supplemented diet, and high-protein diet were n=13, n=13, and n=14 for WT mice respectively, while being n=15, n=14, and n=15 for PKU mice

157 Chapter 8 S2 Table. Plasma non-lnaa amino acid concentrations after six weeks of receiving different diets WT PKU Normal chow LNAA diet High-protein diet Normal chow LNAA diet High-protein diet Taurine 425 ± ± ± ± 120 a 348 ± 72 # 256 ± 66 # Aspartic acid 19 ± 7 16 ± 7 16 ± 8 17 ± 8 16 ± 9 12 ± 4 Serine 203 ± 46** 144 ± 36** # 182 ± 39 # 189 ± ± ± 46 Asparagine 71 ± ± ± ± ± ± 19 Glutamate 48 ± ± ± ± 16 a 32 ± ± 10 Glutamine 585 ± ± ± ± 157* 506 ± ± 70* Proline 191 ± ± ± ± ± ± 72 Glycine 249 ± 30** 130 ± 50** ## 222 ± 26 ## 232 ± 89** 144 ± 52** # 196 ± 45 # Alanine 749 ± ± ± ± 142 a 695 ± 178 ## 482 ± 160 ## Citrulline 77 ± ± ± ± 17 a ** 88 ± ± 16* Ornithine 67 ± ± ± ± ± ± 16 Lysine 550 ± 160** 300 ± 64** ## 552 ± 140 ## 426 ± 130 a 322 ± ± 97 Arginine 113 ± ± ± ± 24 aa 68 ± ± 19 Plasma concentrations are expressed in µmol/l (mean ± SD). Concentrations in PKU mice on normal chow are compared to WT mice on normal chow (a <0.05 and aa<0.01). Within the groups of WT and PKU mice, concentrations that differ between dietary treatment groups are indicated (*<0.05; **<0.01; #<0.05; and ##<0.01). 156

158 Chapter 8 S3 Table. Brain non-lnaa amino acid concentrations after six weeks of receiving different diets WT PKU High-protein diet LNAA diet Normal chow High-protein diet LNAA diet Normal chow Taurine 9913 ± 702* ± 488* ## 9839 ± 534 ## ± ± ± 684 Aspartic acid 4833 ± ± ± ± 459 a 4501 ± ± 317 Serine 1012 ± 96** 617 ± 102** ## 1012 ± 93 ## 1137 ± 98 aa ** 872 ± 174** ## 1121 ± 146 ## Asparagine 80 ± 9 72 ± ± ± ± ± 8 Glutamate ± ± ± ± 636 aa 9627 ± ± 679 Glutamine 4702 ± ± 677 # 4846 ± 992 # 4264 ± ± ± 743 Proline 131 ± ± ± ± ± ± 70 Glycine 1496 ± 152** 1153 ± 143** ## 1467 ± 157 ## 1817 ± 373 aa * 1466 ± 285* 1724 ± 265 Alanine 874 ± ± ± ± 84 aa 745 ± ± 95 GABA 3876 ± ± ± ± ± ± 525 Lysine 328 ± 37** 282 ± 28** ## 329 ± 42 ## 304 ± ± ± 40 Arginine 260 ± ± ± ± ± ± 73 Brain concentrations are expressed in nmol/g wet weight (mean ± SD). Concentrations in PKU mice on normal chow are compared to WT mice on normal chow (a <0.05 and aa<0.01). Within the groups of WT and PKU mice, concentrations that differ between dietary treatment groups are indicated (*<0.05; **<0.01; #<0.05; and ##<0.01)

159

160 CHAPTER 9 Therapeutic brain modulation with targeted Large Neutral Amino Acid supplements in the Pah-enu2 phenylketonuria mouse model D. van Vliet, V.M. Bruinenberg, P.N. Mazzola, H.J.R. van Faassen, P. de Blaauw, T. Pascucci, S. Puglisi-Allegra, I.P. Kema, M.R. Heiner-Fokkema, E.A. van der Zee, F.J. van Spronsen 1 University of Groningen, University Medical Center Groningen, Beatrix Children s Hospital, Groningen, The Netherlands (DvV, PNM, FJvS). 2 University of Groningen, Groningen Institute for Evolutionary Life Sciences (GELIFES), Department of Molecular Neurobiology, Groningen, The Netherlands (VMB, PNM, EAvdZ). 3 University of Groningen, University Medical Center Groningen, Department of Laboratory Medicine, Groningen, The Netherlands (HJRvF, PdB, IPK, MRHF). 4 Sapienza University, Fondazione Santa Lucia, Department of Psychology and Centro Daniel Bovet, Rome, Italy (TP, SPA). 5Fondazione Santa Lucia, IRCCS, Rome, Italy (TP, SPA) Am J Clin Nutr Nov;104(5): Epub 2016 Sep 21.

161 Chapter 9 Sources of support: This research was funded by a research grant from the National PKU Alliance (USA) and the University Medical Center of Groningen. Experimental diets were provided by Nutricia Research. Running head: Towards LNAA treatment for PKU Keywords: phenylketonuria, inborn error of metabolism, large neutral amino acids Abbreviations BBB blood-brain barrier 5-HIAA 5-hydroxyindoleacetic acid His histidine Ile isoleucine LAT1 large neutral amino acid transporter type 1 Leuleucine LNAA large neutral amino acids Met methionine PAH phenylalanine hydroxylase Phe phenylalanine PKU phenylketonuria Thr threonine Tyr tyrosine Trp tryptophan Val valine VIL valine, isoleucine, and leucine WT wild-type 160

162 Chapter 9 ABSTRACT Background: Phenylketonuria treatment consists mainly of a phenylalanine-restricted diet, which leads to suboptimal neurocognitive and psychosocial outcomes. Supplementation of large neutral amino acids (LNAAs) has been suggested as an alternative dietary treatment strategy to optimize neurocognitive outcome in phenylketonuria and has been shown to influence 3 brain pathobiochemical mechanisms in phenylketonuria, but its optimal composition has not been established. Objective: In order to provide additional pathobiochemical insight and develop optimal LNAA treatment, several targeted LNAA supplements were investigated with respect to all 3 biochemical disturbances underlying brain dysfunction in phenylketonuria. Design: Pah-enu2 (PKU) mice received 1 of 5 different LNAA supplemented diets beginning at postnatal day 45. Control groups included phenylketonuria mice receiving an isonitrogenic and isocaloric high-protein diet or AIN-93M diet, and wild-type mice receiving the AIN- 93M diet. After 6 weeks, brain and plasma amino acid profiles and brain monoaminergic neurotransmitter concentrations were measured. Results: Brain Phe concentrations were most effectively reduced by supplementation of LNAAs, such as Leu and Ile, with a strong affinity for the LNAA transporter type 1. Brain non- Phe LNAAs could be restored on supplementation, but unbalanced LNAA supplementation further reduced brain concentrations of those LNAAs that were not (sufficiently) included in the LNAA supplement. To optimally ameliorate brain monoaminergic neurotransmitter concentrations, LNAA supplementation should include Tyr and Trp together with those LNAAs that effectively reduce brain Phe concentrations. The requirement for Tyr supplementation is higher than it is for Trp, and the relative effect of brain Phe reduction is higher for serotonin than for dopamine and norepinephrine. 9 Conclusions: The study shows that all 3 biochemical disturbances underlying brain dysfunction in phenylketonuria can be targeted by specific LNAA supplements. The study thus provides essential information for the development of optimal LNAA supplementation as an alternative dietary treatment strategy to optimize neurocognitive outcome in patients with phenylketonuria. 161

163 Chapter 9 1. INTRODUCTION Phenylketonuria (PKU; Online Mendelian Inheritance in Man no ) is an inborn error of phenylalanine (Phe) metabolism, characterized by an impaired conversion of Phe to tyrosine (Tyr). Left untreated, high blood Phe concentrations correlate with classical PKU symptoms, including severe neurocognitive dysfunction, seizures, and psychiatric problems. Based on this correlation, blood Phe reduction by a Phe-restricted diet is the treatment of choice 1. Although the symptoms of PKU are almost exclusively restricted to brain dysfunction and this is known to correlate with blood Phe concentrations, with the current blood Phe-lowering treatment strategies, neurocognitive and psychosocial outcomes remain suboptimal 2,3 and treatment adherence is known to decline with age 4. An alternative pathophysiology-based treatment directly targeting brain biochemistry to improve brain function of patients with PKU is therefore highly anticipated. In the pathophysiology of PKU, disturbed amino acid transport across the blood-brain barrier (BBB) plays a central role, with increased brain Phe influx outcompeting the brain influx of other large neutral amino acids (LNAAs) that share the same transport system 5. Three main hypotheses have been postulated to explain the mechanism(s) by which disturbed BBB transport of LNAAs impairs brain function in PKU 6. The first hypothesis states that brain dysfunction is primarily the consequence of neurotoxic high brain Phe concentrations. The second hypothesis, substantiated by impaired cerebral protein synthesis in PKU 7, postulates that insufficient availability of non-phe LNAAs to the brain is an additional mechanism. The third hypothesis presumes that especially impaired cerebral monoaminergic neurotransmitter synthesis is of specific importance. Founded on these hypotheses, dietary supplementation of LNAAs has been suggested as a possible alternative treatment strategy for PKU that could target all the pathophysiological mechanisms causing brain dysfunction in PKU. Previous studies have investigated various LNAA supplements in patients with PKU and Black and Tan Brachyury (BTBR) Pahenu2 (PKU) mice In theory, the various LNAA supplements may have served different biochemical pathways inside the brain, but most studies have focused primarily on the effects on blood and brain Phe concentrations and showed inconsistent results 24. In a proof-of-principle study, we investigated all of the hypothesized biochemical treatment objectives, and showed that LNAA supplementation 1) reduced brain Phe concentrations, 2) attenuated the brain deficiencies of some but not all supplemented LNAAs, and 3) increased concentrations of brain neurotransmitters in PKU mice 25. LNAA supplementation could, therefore, serve different brain biochemical treatment goals; however, thus far, insufficient 162

164 Chapter 9 understanding of the effects of various LNAA supplements and optimal LNAA composition hampers clinical application. Our study attempted to correlate biochemical brain treatment objectives with specific requirements for the composition of LNAA supplementation by therapeutic modulation of blood and brain biochemistry with different LNAA supplements in BTBR Pah-enu2 PKU mice. 2. MATERIAL AND METHODS 2.1 Animals This study was performed in BTBR Pah-enu2 mice. The Pah-enu2 mouse is a wellestablished PKU mouse model that resembles the genetics, biochemistry, and neurobiology of PKU in humans 26. As previously described, one of the important advances of this mouse model over the human PKU model is that brain concentrations of LNAAs other than Phe and monoaminergic neurotransmitters can be measured directly 25. From heterozygous (+/-) mating, we obtained wild-type (WT, +/+), heterozygous, and PKU (-/-) mice of both sexes. After weaning at 4 wk of age, genetic characterization was performed by quantitative polymerase chain reaction analysis on DNA extracted from ear tissue. Water and RMH-B food pellets (Arie Block BV) were offered ad libitum. From the start of the experiment, animals were individually housed and kept at 21±1ºC on a 12-h light-dark cycle ( ). In total, we used 16 WT mice (8 male, 8 female) and 112 PKU (56 male, 56 female) mice. All procedures and treatments were carried out in strict accordance with the National Research Council guidelines. The experimental protocol was approved by the Institutional Animal Care and Use Committee of the University of Groningen (Permit Number: 6504D). 2.2 Experimental design At postnatal day 45, animals were included in the experiment, and PKU mice were assigned to 1 of 7 different dietary treatment groups. Following the order in which the animals were born, PKU mice were successively assigned to each of the 7 experimental groups. This procedure was carried out for male and female mice separately and avoided assigning littermates to the same experimental group as much as possible (Supplemental Figure 1). During the first week of dietary treatment, body weight and food intake were measured daily. Afterward, body weight and food intake were determined weekly. Dietary treatment was continued for 6 wk. At the end of the experiment, animals were killed by combined heart puncture and decapitation under inhalation anesthetics with isoflurane Experimental diets The basic diet was AIN-93M 27, which was administered in unadjusted form to the PKU and WT control groups. The compositions of the investigated LNAA-supplemented diets 163

165 Chapter 9 were principally based on theoretical considerations regarding optimal regimens to achieve each of the biochemical treatment objectives of LNAA treatment in PKU 24. To reduce brain Phe concentrations, the reported effects of Val, Ile, and Leu 13,14 and that especially Leu and Ile but not valine (Val) have relatively low Michaelis-Menten constant values for transport across the BBB 5,28 made us hypothesize that a Leu+Ile-supplemented diet may be especially effective. To stimulate cerebral protein synthesis, however, supplementation of all LNAAs competing with the high Phe concentrations in blood for transport across the BBB may be required 24. Moreover, to explicitly increase brain availability of amino acid precursors for monoaminergic neurotransmitter synthesis, a Tyr+Trp supplemented diet was suggested as being more effective. In addition, aside from theoretical considerations, a Thr-supplemented diet was investigated because Thr supplementation was shown to be beneficial in reducing plasma Phe concentrations in patients with PKU 23. The experimental LNAA diets were produced by adding LNAAs to the AIN-93M diet at the expense of cornstarch on a weight-for-weight basis. The added amount of Tyr and Trp in the Tyr+Trp-supplemented diet was calculated to be 20% and 10%, respectively, of the amount of protein in the AIN-93M diet ( g/kg diet), which corresponds to ~200 and 100 g/kg body weight in humans, if compared with a mean natural protein intake in humans of 1 g kg body weight -1 d -1. Similarly, the added amount of Leu and Ile in the Leu+Ilesupplemented diet was calculated to be 20% and 15%, respectively. The added amount of Thr in the Thr-supplemented diet was calculated to be 5% of the amount of protein in the AIN-93M diet. The total amount of added LNAAs in both diets containing additional non- Phe LNAAs with or without Thr [LNAA(+Thr) and LNAA(-Thr) diets] was equal to the amount of protein in the AIN-93M diet ( g/kg diet). The LNAA mixtures included equal amounts of 17.7 mg/kg [in LNAA(-Thr)] or 15.5 mg/kg diet [in LNAA(+Thr)] of l-tyr, l-trp, l-val, l-ile, l-leu, l-methionine (Met), and l-histidine (His) [and l-thr in the LNAA(+Thr) diet only]. The high-protein diet was produced by adding extra casein to the AIN-93M diet at the expense of cornstarch on a weight-for-weight basis. The added amount of extra casein was calculated to result in an isonitrogenic and isocaloric control diet for the diets posing the highest amino acid loads [the LNAA(-Thr) and LNAA(+Thr) diets]. Diets were prepared by Research Diet Services B.V. Amino acid analyses in the different diets are presented in Supplemental Table Brain collection and biochemical analyses To obtain brain material for biochemical analyses, whole brains were removed. Brains were divided into cerebellum, brain stem, and cerebrum, the latter being further divided into 2 hemispheres. The collected brain samples were individually snap frozen in liquid nitrogen and stored at -80ºC until further preparation. One-half cerebrum and blood samples were further processed for the analyses of brain and plasma amino acid and monoaminergic neurotransmitter concentrations, as described previously 25. Monoaminergic 164

166 Chapter 9 neurotransmitters (monoamines) and associated metabolites for which concentrations in the brains were assessed included dopamine, norepinephrine, and normetanephrine in the catecholamine pathway, and serotonin and 5-hydroxyindoleacetic acid (5-HIAA) in the serotonergic pathway. 2.5 Statistical analyses Statistical analyses have been performed on the data collected on all of the animals except for one mouse that died prematurely. Statistical analyses were performed with the use of IBM SPSS Statistics for Windows, Version All of the tests were performed 2-sided at a significance level of α=0.05. Data were expressed as means ± SDs unless otherwise indicated. Brain and plasma biochemistry was analyzed in 2 steps. If data were not normally distributed (assessed by Shapiro-Wilk test) or in case of unequal variances (assessed by Levene s test), then analyses were performed on log-transformed data. First, each experimental group was individually compared with PKU mice receiving the AIN-93M diet through use of Student s t tests and Bonferroni correction for multiple testing. Second, all of the experimental groups that were significantly different from PKU mice on AIN-93M diet were compared by 1-factor ANOVA and Tukey s post hoc analysis. The effect of dietary treatment on body weight was analyzed on log-transformed data by repeated-measures ANOVA and Tukey s post hoc analysis, with one between-subjects factor (treatment group, 8 levels), one within-subjects factor (time, 7 levels: 0,1, 2, 3, 4, 5, and 6 wk), and sex as a covariate. Weekly food intake (per gram of body weight) was analyzed by 1-factor ANOVA with Tukey s post hoc analysis. To assess which parameters of brain Phe, Tyr, and Trp concentrations correlated with brain monoaminergic neurotransmitter concentrations in PKU mice, we performed multiple linear regression analyses using backward selection. These analyses were performed for brain serotonin, dopamine, and norepinephrine concentrations as dependent variables, with genotype and brain Phe, Tyr, and Trp concentrations as independent variables RESULTS 3.1 General health and dietary intake All of the experimental diets were tolerated well by the mice. Of the 128 mice included in the experiment, 1 PKU male mouse receiving the Leu+Ile-supplemented diet died unexpectedly 31 d after inclusion. Postmortem macroscopic pathological examination showed no pathology. A wound in the neck of one PKU male mouse receiving the Thr-supplemented diet was found during the final week of the experiment caused by excessive grooming. Such excessive grooming is sometimes observed in mice, especially male BTBR (both WT and 165

167 Chapter 9 PKU) mice. Body weight curves during 6-wk-long dietary treatment were different for male and female mice (p<0.001), but they did not significantly differ between treatment groups (p=0.402) (Supplemental Figure 2). Weekly food intake (expressed as g food g body weight -1 wk -1 ) decreased after the first week for all experimental groups, and remained relatively stable in the later treatment weeks (data not shown). Based on the weekly food intakes and amino acid contents of the different diets, mean daily intakes of individual LNAAs during the last week of the 6-wk dietary treatment were calculated for all of the experimental groups and presented in Supplemental Table Plasma LNAAs Plasma LNAA concentrations in PKU mice receiving the AIN-93M diet and experimental diets and in WT and PKU control animals are shown in Figure 1. In PKU mice receiving the AIN-93M diet, plasma Phe concentrations were 490% higher than the concentrations in WT animals receiving the AIN-93M diet (p<0.001; Figure 1A), whereas plasma concentrations of non-phe LNAAs were similar or lower than in WT mice (Figure 1B-I). When assessing the LNAA-supplemented diets, plasma Phe concentrations in PKU mice receiving LNAA(+Thr) and LNAA(-Thr) diets were 23% lower than the concentrations observed in the AIN-93M diet (p<0.01 for both; Figure 1A). On selective supplementation of Tyr+Trp, Leu+Ile, or Thr, however, plasma Phe concentrations did not significantly differ from those in the AIN-93M diet, whereas with the high-protein diet, plasma Phe concentrations were higher than they were with the AIN-93M diet (p<0.001; Figure 1A). Plasma non-phe LNAA concentrations were significantly higher in mice receiving diets in which these particular non-phe LNAAs had been supplemented than they were in PKU mice receiving the AIN-93M diet; the exception was plasma Tyr and His concentrations in LNAA(+Thr), which were not statistically significantly different from concentrations in PKU mice receiving the AIN-93M diet (p=0.062 and p=0.190, respectively; Figure 1B-I). In PKU mice receiving the high protein diet, plasma Tyr, Val, Ile, and Leu concentrations were higher than in PKU mice receiving the AIN-93M diet (p<0.01), but these elevations were significantly less than those observed in mice receiving the experimental diets (Figure 1B-I). 3.3 Brain LNAAs Brain concentrations of individual LNAAs in PKU mice receiving the experimental diets and in WT as well as in PKU control animals are shown in Figure 2. Brain Phe concentrations in PKU mice receiving the AIN-93M diet were 310% higher than concentrations in WT mice receiving the AIN-93M diet (p<0.001; Figure 2A). Brain His concentrations in PKU mice receiving the AIN-93M diet were 49% higher than in WT mice receiving the AIN-93M diet (p<0.001; Figure 2H). Brain concentrations of all other non-phe LNAAs were lower in PKU 166

168 Chapter 9 mice receiving the AIN-93M diet, ranging from 66% of concentrations in corresponding WT mice for Tyr to 85% for Met (Figure 2B-G,I). Figure 1. Plasma concentrations of A) phenylalanine, B) tyrosine, C) tryptophan, D) valine, E) isoleucine, F) leucine, G) methionine, H) histidine, and I) threonine in WT and PKU mice after 6 wk of receiving different diets. Numbers of mice are n=15 or n=16 for all treatment groups. Untransformed data are expressed as means ± SEMs. * p<0.05; ** p<0.01; *** p<0.001 (2-sided) compared with PKU mice receiving the AIN-93M diet unless otherwise indicated. Statistical analyses were performed in 2 steps: 1) each experimental group was individually compared with PKU mice receiving the AIN-93M diet by using Student s t tests and Bonferroni correction for multiple testing, and 2) all experimental groups that were significantly different from PKU mice receiving the AIN-93M diet were compared by 1-factor ANOVA and Tukey s post hoc analysis. If data were not normally distributed (assessed by Shapiro- Wilk test), or showed unequal variances (assessed by Levene s test), analyses were performed on logtransformed data. LNAA, large neutral amino acid; PKU, phenylketonuria; WT, wild-type 9 167

169 Chapter 9 Brain Phe concentrations on LNAA(-Thr)-, LNAA(+Thr)-, and Leu+Ile-supplemented diets were comparable, being 25-29% lower than concentrations in PKU mice receiving the AIN- 93M diet (p<0.001; Figure 2A). On selective supplementation of Tyr+Trp or Thr, however, brain Phe concentrations in PKU mice did not differ (p=1.000 for both), whereas brain Phe concentrations in the high-protein diet were even higher than they were in PKU mice receiving the AIN-93M diet (p<0.05). All brain non-phe LNAA disturbances could be improved by either one or more of the experimental diets. As may be observed in Figure 2B-I, almost all non-phe LNAA concentrations were higher if being supplemented in the diet while in general being lower than concentrations in PKU mice receiving the AIN-93M diet if not being supplemented, specifically for brain Met, His, and Thr concentrations. Compared with PKU mice receiving the AIN-93M diet, the LNAA(-Thr) diet resulted in higher brain concentrations of all supplemented non-phe LNAAs (p<0.05 for Tyr and p<0.001 for others), except for Ile, Leu, and Met (p=0.093; p=0.236; and p=0.115, respectively). Brain concentrations of Thr, which was not included in the LNAA(-Thr) diet, were even lower than they were in PKU mice receiving the AIN-93M diet (p<0.001; Figure 2I). The LNAA(+Thr) diet showed similar results, but brain non-phe LNAA concentrations were lower than they were in the LNAA(-Thr) diet, and in contrast to the LNAA(-Thr) diet, brain Thr concentrations were higher than they were in PKU mice receiving the AIN-93M diet (p<0.001). Selective Tyr+Trp supplementation restored brain Tyr and Trp concentrations to WT levels, but further impaired brain Thr concentrations compared with PKU mice receiving the AIN- 93M diet (p<0.001). Similarly, on selective Leu+Ile supplementation, only brain Ile and Leu concentrations were higher than in PKU mice receiving the AIN-93M diet (p<0.05 for both). Besides reducing brain Phe concentrations, Leu+Ile supplementation also resulted in lower brain Met, His, and Thr but not Tyr and Trp concentrations than the concentrations in PKU mice receiving the AIN-93M diet (p<0.001 for all). Selective Thr supplementation provided higher brain Thr concentrations than those in PKU mice receiving the AIN-93M diet (p<0.001) without affecting the brain concentrations of any other LNAAs, including Phe. The high-protein control diet resulted in even higher brain His concentrations than those in PKU mice receiving the AIN-93M diet (p<0.05). 3.4 Brain monoaminergic neurotransmitters Brain monoamine and associated metabolite concentrations in PKU mice receiving the experimental diets as well as in WT and PKU control animals are shown in Figure 3. In the catecholamine pathway, brain dopamine, norepinephrine, and normetanephrine concentrations in PKU mice receiving the AIN-93M diet were, respectively, 85% (p<0.01), 61% (p<0.001), and 82% (p<0.05) of the concentrations in WT mice receiving the AIN- 93M diet. Tyr+Trp, LNAA(-Thr), and LNAA(+Thr) diets similarly resulted in higher brain norepinephrine concentrations than those in PKU mice receiving the AIN-93M diet (p<

170 Chapter 9 for all), and partially restored its deficit to 79-85% of the concentrations in WT mice (Figure 3B). Brain normetanephrine concentrations were higher in these 3 LNAA-supplemented diets than those in PKU mice receiving the AIN-93M diet (p<0.01), and they no longer differed from WT levels (p=0.360; Figure 3C). For brain dopamine, exclusively the Tyr+Trp diet resulted in higher concentrations than those in PKU mice receiving the AIN-93M diet (p<0.05; Figure 3A). Selective Leu+Ile or Thr supplementation as well as the high-protein diet did not significantly change brain dopamine or norepinephrine concentrations. 9 Figure 2. Brain concentrations of A) phenylalanine, B) tyrosine, C) tryptophan, D) valine, E) isoleucine, F) leucine, G) methionine, H) histidine, and I) threonine in WT and PKU mice after 6 wk of receiving different diets. Numbers of mice are n=15 or n=16 for all treatment groups. Untransformed data are expressed as means ± SEMs. * p<0.05; ** p<0.01; *** p<0.001 (2-sided) compared with PKU mice receiving the AIN-93M diet unless otherwise indicated. Statistical analyses were performed in 2 steps: 1) each experimental group was individually compared with PKU mice receiving the AIN-93M diet by using Student s t tests and Bonferroni correction for multiple testing, and 2) all experimental groups that were significantly different from PKU mice receiving the AIN-93M diet were compared by 1-factor ANOVA and Tukey s post hoc analysis. If data were not normally distributed (assessed by Shapiro- Wilk test), or showed unequal variances (assessed by Levene s test), analyses were performed on logtransformed data. LNAA, large neutral amino acid; PKU, phenylketonuria; WT, wild-type 169

171 Chapter 9 In the serotonergic pathway, brain serotonin and 5-HIAA concentrations in PKU mice receiving the AIN-93M diet were 46% and 27%, respectively, of concentrations in WT mice receiving the AIN-93M diet (p<0.001 for both; Figure 3D,E). The Tyr+Trp diet resulted in higher brain serotonin and 5-HIAA concentrations than those in PKU mice receiving the AIN-93M diet (p<0.001 for both) and partially restored their deficiencies to, respectively, 64% and 46% of the concentrations in WT mice. In mice receiving the LNAA(-Thr) and LNAA(+Thr) diets, brain serotonin concentrations were further restored to 83%, and brain 5-HIAA concentrations were 71% and 65%, respectively, of WT concentrations (p<0.001 for all). Selective supplementation of Leu+Ile or Thr did not demonstrate a significant effect on either brain serotonin or 5-HIAA concentrations if compared with PKU mice receiving the AIN-93M diet (p=0.421 for Thr diet on 5-HIAA, and p=1.000 for other), whereas brain 5-HIAA concentrations were even lower in PKU mice receiving the high-protein diet than the AIN-93M diet (p<0.001). Figure 3. Brain concentrations of A) dopamine, B) norepinephrine, C) normetanephrine, D) serotonin, and E) 5-hydroxyindoleacetic acid (5-HIAA) in WT and PKU mice after 6 wk of receiving different diets. Numbers of mice are n=15 or n=16 for all treatment groups. Untransformed data are expressed as means ± SEMs. * p<0.05; ** p<0.01; *** p<0.001 (2-sided) compared with PKU mice receiving the AIN-93M diet unless otherwise indicated. Statistical analyses were performed in 2 steps: 1) each experimental group was individually compared with PKU mice receiving the AIN-93M diet by using Student s t tests and Bonferroni correction for multiple testing, and 2) all experimental groups that were significantly different from PKU mice receiving the AIN-93M diet were compared by 1-factor ANOVA and Tukey s post hoc analysis. If data were not normally distributed (assessed by Shapiro-Wilk test), or showed unequal variances (assessed by Levene s test), analyses were performed on log-transformed data. LNAA, large neutral amino acid; PKU, phenylketonuria; WT, wild-type 170

172 Chapter Relation between brain monoaminergic neurotransmitters and brain amino acid biochemistry To further investigate the correlation between brain amino acid and monoamine concentrations and the possible mechanism(s) by which LNAA supplementation in PKU mice restored brain monoamine concentrations, linear regression analyses were performed. Assessing the catecholamine pathway, multivariate linear regression analyses showed that both brain dopamine and norepinephrine concentrations could be predicted by a combination of brain Phe and Tyr concentrations (R 2 =0.127, p<0.001 for dopamine; R 2 =0.563, p<0.001 for norepinephrine; Figure 4A,B), showing a negative correlation with brain Phe concentrations and a positive correlation with brain Tyr concentrations. When considering brain Phe and Tyr concentrations, genotype no longer demonstrated a significant correlation (p=0.780 for dopamine; p=0.704 for norepinephrine). Moreover, no interaction was observed between brain Phe and Tyr concentrations (p=0.319 for dopamine; p=0.263 for norepinephrine). In the serotonergic pathway, brain serotonin concentrations could be predicted by brain Phe and Trp concentrations and genotype (R 2 =0.753, p<0.001; Figure 4C) without significant interactions between these variables. In this model, brain serotonin concentrations were negatively correlated with brain Phe levels and positively correlated with brain Trp concentrations. In addition, when corrected for both brain Phe and Trp concentrations, brain serotonin concentrations in PKU mice were 0.49 nmol/g wet weight lower than in WT mice. 9 Figure 4. Predicted compared with actual brain concentrations of A) dopamine, B) norepinephrine, and C) serotonin in WT and PKU mice after 6 wk of receiving different diets. Lines of equality x=y are added for comparative illustration. Numbers of mice are n=16 for WT and n=110 for PKU. To assess which parameters of brain phenylalanine (Phe), tyrosine (Tyr), and tryptophan (Trp) concentrations were correlated with brain monoaminergic neurotransmitter concentrations in PKU mice, multiple linear regression analyses by use of backward selection were performed. Both brain dopamine and norepinephrine concentrations could be predicted by a combination of brain Phe (negatively correlated) and Tyr (positively correlated) concentrations (R 2 =0.127, p<0.001 for dopamine; R 2 =0.563, p<0.001 for norepinephrine), with no significant correlation with genotype and no significant interaction between brain Phe and Tyr. Brain serotonin concentrations could be predicted by brain Phe (negatively correlated) and Trp (positively correlated) concentrations as well as genotype (0.49 nmol/g wet weight lower than in WT mice, when corrected for Phe and Trp) (R 2 =0.753, p<0.001) without significant interactions between these parameters. PKU, phenylketonuria; WT, wild-type. 171

173 Chapter 9 4. DISCUSSION Our study compared different LNAA supplements with respect to all biochemical brain and blood treatment objectives of LNAA treatment in PKU to provide further pathobiochemical insight and identify novel therapeutic options. After comparison of the response of (brain) biochemical impairments in PKU to different LNAA supplements, 5 main results were identified: 1) plasma Phe reduction could be accomplished by high-dose supplementation of all LNAAs (either with or without Thr) but not on selective supplementation of Leu+Ile, Tyr+Trp, or Thr; 2) brain Phe concentrations were similarly reduced on supplementation of all LNAAs (either with or without Thr) and on selective supplementation of Leu+Ile but not Tyr+Trp or Thr; 3) brain non-phe LNAA concentrations could be effectively restored on supplementation, but were more impaired if they were not included in selective LNAA supplements; and 4) brain monoaminergic neurotransmitter concentrations improved by increasing precursor amino acids (Tyr+Trp diet) rather than by selectively reducing brain Phe (Leu+Ile diet); the combination of both strategies [on LNAA(+Thr) and LNAA(-Thr) diets] was most effective, especially for the serotonergic pathway. Thus far, plasma Phe reduction has been the primary target of all of the treatment strategies in PKU. Although LNAA supplementation in PKU is suggested to improve brain metabolism primarily through its effect on the transport of Phe and other LNAAs at the BBB, it also reduces plasma Phe concentrations 24. In our study, plasma Phe concentrations in PKU mice were 23% lower with LNAA supplementation including all LNAAs (either with or without Thr) than they were in PKU mice receiving the AIN-93M diet. This finding is in accordance with previous results obtained in PKU mice, showing that LNAA supplementation including (nearly) all LNAAs could reduce plasma Phe concentrations to 47-67% of untreated PKU controls 21,22,25. Regarding the underlying mechanism, plasma Phe reduction has been assumed to be the result of supplementation of Thr in particular. In contrast to the effect of reduced plasma Phe concentrations on Thr supplementation (50 mg kg -1 d -1 ) in patients with PKU 23, we did not observe this effect in mice. In addition, both LNAA(-Thr) and LNAA(+Thr) diets showed comparable effects on plasma Phe concentrations. The data presented demonstrate that our previous hypothesis that plasma Phe reduction on LNAA treatment is the result of increased net protein synthesis rather than competition at the gut-blood barrier 24,25 is unlikely. Body weight was comparable between dietary treatment groups, and supplementation of Tyr (in the Tyr+Trp diet) the most limiting plasma LNAA in PKU did not affect plasma Phe concentrations. Alternatively, it has been suggested that plasma Phe reduction on LNAA supplementation is caused by competition of LNAAs with Phe for transport across the gut-blood barrier 11. The Leu+Ile diet, which showed a clear competitive effect on Phe transport across the BBB, did not result in plasma Phe reduction in this study. Thus, high-dose supplementation of all LNAAs (either with or without Thr) can 172

174 Chapter 9 effectively reduce plasma Phe concentrations in PKU, but our data do not allow for definite conclusions regarding the underlying mechanism(s). A second aim of LNAA supplementation in PKU is to reduce brain Phe concentrations by increasing the competition with Phe for transport across the BBB. In our study, brain Phe concentrations were reduced by 24-29% in all experimental diets but not in Tyr+Trp and Thr. The findings on LNAA(-Thr) and LNAA(+Thr) are in accordance with previous studies on LNAA supplementation that used similar regimens with equal amounts of all non- Phe LNAAs except for Thr in PKU mice 25 and patients with PKU 8. The fact that selective Leu+Ile supplementation was especially effective in reducing brain Phe concentrations in PKU mice confirms our hypothesis based on the high affinities of Leu and Ile to the large neutral amino acid transporter 1 (LAT1) in the rat 5. Moreover, this result provides further biochemical support for the findings on cerebrospinal fluid biochemistry as well as improved neuropsychological functioning and electroencephalogram activity on Val, Ile, and Leu supplementation in patient with PKU 13,14,29. The fact that brain Phe reduction on Leu+Ile was not accompanied by reduced Phe concentrations in plasma further supports the notion that brain Phe reduction on LNAA is at least in part the result of a direct effect at the BBB level 8,25. Thus, our results indicate that supplementation of LNAAs with a strong affinity to the LAT1 transporter, such as Leu and Ile, is especially effective in reducing brain Phe concentrations rather than reducing plasma Phe. A less well considered treatment objective for LNAA supplementation in PKU includes restoring brain non-phe LNAA concentrations. The present study confirms our previous data that supplementation of (nearly) all LNAAs improves brain non-phe LNAA concentrations 25. On the one hand, inclusion of any specific LNAA in the LNAA-supplemented diet almost invariably led to increased brain concentrations of that particular LNAA compared with concentrations in the AIN-93M diet. On the other hand, supplementation (selective) of LNAAs with a high affinity for the LAT1 transporter further impaired brain concentrations of unsupplemented LNAAs. Leu+Ile supplementation further reduced brain His and Met concentrations, and brain Thr concentrations were even lower in LNAA(-Thr) than in PKU mice receiving the AIN-93M diet. Selective supplementation of LNAAs with a lower affinity for LAT1, such as Tyr+Trp or Thr, did not impair brain concentrations of unsupplemented LNAAs. These findings stress the importance of well-balanced LNAA supplements and the need to evaluate the effects on brain concentrations of all LNAAs, especially if supplements include LNAAs that are known to have a strong affinity for LAT1. 9 Finally, LNAA supplementation aims to improve brain monoaminergic neurotransmitter concentrations in PKU, which are thought to be impaired because of insufficient brain availability of Tyr and Trp, or because of high brain Phe concentrations inhibiting the activity of Tyr and Trp hydroxylases 24. To our knowledge, our study is the first to directly compare 173

175 Chapter 9 the effects of 1) increased brain Tyr and Trp availability, 2) reduced brain Phe concentrations, and 3) the combination of both on brain monoamine concentrations in PKU mice. Effects were most prominent for brain norepinephrine and serotonin rather than dopamine concentrations, which is consistent with several studies in PKU mice and measurements in cerebrospinal fluid of patients with PKU 25, Selective Tyr+Trp supplementation, including 24.8 and 12.4 mg/kg diets of additional Tyr and Trp, ameliorated both brain norepinephrine and serotonin concentrations, whereas no such effects were observed for Leu+Ile. Both LNAA(-Thr)- and LNAA(+Thr)-supplemented diets were even more effective than Tyr+Trp in improving brain serotonin concentrations, and were similarly effective as Tyr+Trp in improving brain norepinephrine concentrations. On the one hand, these results imply that both insufficient brain Tyr and Trp availability and increased brain Phe concentrations contribute to the brain monoamine impairments in PKU, but on the other hand, again imply that the relative contribution of both mechanisms is different for the dopaminergic and serotoninergic pathways 25. Clearly, increasing cerebral Leu did decrease cerebral Phe, but did not increase serotonin, suggesting that a decrease of cerebral Phe was not important enough or the increase of Leu counteracted this effect. In addition, the requirement for Tyr supplementation seems higher than it is for Trp supplementation. Taken together, brain monoamine concentrations could be improved by increasing precursor amino acids rather than by selectively reducing brain Phe concentrations. The combination of both strategies through supplementation of all LNAAs was the most effective approach, especially for the serotonergic pathway. In conclusion, the present study compared different LNAA supplements with respect to all brain and blood biochemical treatment objectives of LNAA treatment in PKU to provide more pathobiochemical insight and identify novel therapeutic options. Our results show that targeted LNAA supplements can influence specific brain biochemical features in PKU and thereby suggest a valuable new opportunity to further elucidate the relative importance of the different brain biochemical disturbances in PKU brain dysfunction. Our results also provide essential knowledge for the development of the optimal LNAA treatment of patients with PKU. ACKNOWLEDGEMENTS The authors express their gratitude to Mrs. H.A. Kingma, Mrs. E.Z. Jonkers, Mrs. K. Boer, and Mrs. H. Adema for their analytical support, and to Dr. E.H. Schölvinck for English editing of the manuscript. Also, we are grateful to Nutricia Research for providing the experimental diets. DvV, EAvdZ, and FJvS designed the research; DvV, VMB, and PNM conducted the animal experiment; DvV, HJRvF, PdB, IPK, and MRHF were responsible for biochemical analyses; 174

176 Chapter 9 TP and SPA provided essential material; DvV analyzed data; DvV, EAvdZ, and FJvS wrote the paper; DvV had primary responsibility for the final content. All authors read and approved the final manuscript. COMPETING INTERESTS EAvdZ has received advisory board fees from Arla Foods. FJvS has received research grants, advisory board fees, and speaker s honoraria from Merck Serono and Nutricia Research, has received speaker s honoraria from Vitaflo, and has received advisory board fees from Arla Foods. All other authors have declared not to have conflicts of interest. All authors have read the journal s policy on disclosure of potential conflicts of interest. This research was funded by a grant from the National PKU Alliance (USA). Experimental diets were provided by Nutricia Research

177 Chapter 9 5. REFERENCES 1. Blau N, van Spronsen FJ, Levy HL. Phenylketonuria. Lancet 2010;376: Jahja R, Huijbregts SC, de Sonneville LM, van der Meere JJ, van Spronsen FJ. Neurocognitive evidence for revision of treatment targets and guidelines for phenylketonuria. J Pediatr 2014;164:895,899.e2. 3. Smith I, Knowles J. Behaviour in early treated phenylketonuria: a systematic review. Eur J Pediatr 2000;159 Suppl 2:S Walter JH, White FJ, Hall SK, MacDonald A, Rylance G, Boneh A, Francis DE, Shortland GJ, Schmidt M, Vail A. How practical are recommendations for dietary control in phenylketonuria? Lancet 2002;360: Smith QR. Transport of glutamate and other amino acids at the blood-brain barrier. J Nutr 2000;130:1016S-22S. 6. van Spronsen FJ, Hoeksma M, Reijngoud DJ. Brain dysfunction in phenylketonuria: is phenylalanine toxicity the only possible cause? J Inherit Metab Dis 2009;32: Hoeksma M, Reijngoud DJ, Pruim J, de Valk HW, Paans AM, van Spronsen FJ. Phenylketonuria: High plasma phenylalanine decreases cerebral protein synthesis. Mol Genet Metab 2009;96: Pietz J, Kreis R, Rupp A, Mayatepek E, Rating D, Boesch C, Bremer HJ. Large neutral amino acids block phenylalanine transport into brain tissue in patients with phenylketonuria. J Clin Invest 1999;103: Schindeler S, Ghosh-Jerath S, Thompson S, Rocca A, Joy P, Kemp A, Rae C, Green K, Wilcken B, Christodoulou J. The effects of large neutral amino acid supplements in PKU: an MRS and neuropsychological study. Mol Genet Metab 2007;91: Koch R, Moseley KD, Yano S, Nelson M,Jr, Moats RA. Large neutral amino acid therapy and phenylketonuria: a promising approach to treatment. Mol Genet Metab 2003;79: Matalon R, Michals-Matalon K, Bhatia G, Burlina AB, Burlina AP, Braga C, Fiori L, Giovannini M, Grechanina E, Novikov P, et al. Double blind placebo control trial of large neutral amino acids in treatment of PKU: effect on blood phenylalanine. J Inherit Metab Dis 2007;30: Moats RA, Moseley KD, Koch R, Nelson M,Jr. Brain phenylalanine concentrations in phenylketonuria: research and treatment of adults. Pediatrics 2003;112: Berry HK, Brunner RL, Hunt MM, White PP. Valine, isoleucine, and leucine. A new treatment for phenylketonuria. Am J Dis Child 1990;144: Jordan MK, Brunner RL, Hunt MM, Berry HK. Preliminary support for the oral administration of valine, isoleucine and leucine for phenylketonuria. Dev Med Child Neurol 1985;27: Batshaw ML, Valle D, Bessman SP. Unsuccessful treatment of phenylketonuria with tyrosine. J Pediatr 1981;99: Lou H. Large doses of tryptophan and tyrosine as potential therapeutic alternative to dietary phenylalanine restriction in phenylketonuria. Lancet 1985;2: Lou HC, Lykkelund C, Gerdes AM, Udesen H, Bruhn P. Increased vigilance and dopamine synthesis by large doses of tyrosine or phenylalanine restriction in phenylketonuria. Acta Paediatr Scand 1987;76:

178 Chapter Mazzocco MM, Yannicelli S, Nord AM, van Doorninck W, Davidson-Mundt AJ, Greene CL. Cognition and tyrosine supplementation among school-aged children with phenylketonuria. Am J Dis Child 1992;146: Pietz J, Landwehr R, Kutscha A, Schmidt H, de Sonneville L, Trefz FK. Effect of highdose tyrosine supplementation on brain function in adults with phenylketonuria. J Pediatr 1995;127: Lines D, Magarey A, Raymond J, Robertson E. Tyrosine supplementation in phenylketonuria. J Paediatr Child Health 1997;33: Matalon R, Surendran S, Matalon KM, Tyring S, Quast M, Jinga W, Ezell E, Szucs S. Future role of large neutral amino acids in transport of phenylalanine into the brain. Pediatrics 2003;112: Matalon R, Michals-Matalon K, Bhatia G, Grechanina E, Novikov P, McDonald JD, Grady J, Tyring SK, Guttler F. Large neutral amino acids in the treatment of phenylketonuria (PKU). J Inherit Metab Dis 2006;29: Sanjurjo P, Aldamiz L, Georgi G, Jelinek J, Ruiz JI, Boehm G. Dietary threonine reduces plasma phenylalanine levels in patients with hyperphenylalaninemia. J Pediatr Gastroenterol Nutr 2003;36: van Spronsen FJ, de Groot MJ, Hoeksma M, Reijngoud DJ, van Rijn M. Large neutral amino acids in the treatment of PKU: from theory to practice. J Inherit Metab Dis 2010;33: van Vliet D, Bruinenberg VM, Mazzola PN, van Faassen MH, de Blaauw P, Kema IP, Heiner-Fokkema MR, van Anholt RD, van der Zee EA, van Spronsen FJ. Large Neutral Amino Acid Supplementation Exerts Its Effect through Three Synergistic Mechanisms: Proof of Principle in Phenylketonuria Mice. PLoS One 2015;10:e Martynyuk AE, van Spronsen FJ, Van der Zee EA. Animal models of brain dysfunction in phenylketonuria. Mol Genet Metab 2010;99 Suppl 1:S Reeves PG, Nielsen FH, Fahey GC,Jr. AIN- 93 purified diets for laboratory rodents: final report of the American Institute of Nutrition ad hoc writing committee on the reformulation of the AIN-76A rodent diet. J Nutr 1993;123: Hargreaves KM, Pardridge WM. Neutral amino acid transport at the human bloodbrain barrier. J Biol Chem 1988;263: Berry HK, Bofinger MK, Hunt MM, Phillips PJ, Guilfoile MB. Reduction of cerebrospinal fluid phenylalanine after oral administration of valine, isoleucine, and leucine. Pediatr Res 1982;16: Harding CO, Winn SR, Gibson KM, Arning E, Bottiglieri T, Grompe M. Pharmacologic inhibition of L-tyrosine degradation ameliorates cerebral dopamine deficiency in murine phenylketonuria (PKU). J Inherit Metab Dis 2014; 31. Pascucci T, Giacovazzo G, Andolina D, Accoto A, Fiori E, Ventura R, Orsini C, Conversi D, Carducci C, Leuzzi V, et al. Behavioral and neurochemical characterization of new mouse model of hyperphenylalaninemia. PLoS One 2013;8:e Burlina AB, Bonafe L, Ferrari V, Suppiej A, Zacchello F, Burlina AP. Measurement of neurotransmitter metabolites in the cerebrospinal fluid of phenylketonuric patients under dietary treatment. J Inherit Metab Dis 2000;23:

179

180 CHAPTER 10 Summary and general conclusion

181 Chapter SUMMARY AND INTEGRATION OF RESULTS The main findings of the experimental chapters of this thesis are: Chapter 2: Distinct differences are found between the genetic background of the PKU mouse model despite similar biochemical phenotype. Chapter 3: In contrast to the findings of chapter 2, the PKU females of both strains show similar behavioral outcome. Chapter 4: Sleep problems are present in PKU mice and patients. Chapter 6: A specific nutrient combination (SNC) can improve a postsynaptic marker in specific subregions of the hippocampus. Chapter 7: SNC supplementation improves the outcome in novel object recognition (NOR) despite high Phe concentrations in the food. Chapter 8: Equal amounts of non-phe large neutral amino acids (LNAA s) (except for Threonine) can reduce brain Phe concentrations, improve the concentrations of some of the non-phe LNAA s, and improve brain levels of serotonin and norepinephrine. Chapter 9: Tailoring the LNAA composition to reduced Phe, improve brain non-phe LNAA s and neurotransmitter concentration, showed that improvements of these three biochemical aims were not always accompanied with a reduction of Phe in plasma. Overall, these experimental chapters emphasize that increased Phe concentrations in plasma, currently the gold standard in the clinic, is not the sole predictor of the phenotypical outcome of PKU. Below these results will be discussed in more depth by highlighting the importance to the PKU research field and integrating these results with findings in current literature. This is followed by future perspectives taking the studies of this dissertation as a starting point for future research. The translational value of the PKU mouse model The translational value of a model in general implies a benefit for the modeled situation. However, the true benefit is not always exact. A benefit could be a better understanding of the disease or developing a new treatment strategy. Yet, in the end, the model is only beneficial when results can be extrapolated or translated to the modeled condition. One of the models used in preclinical phenylketonuria (PKU) research, is the PKU mouse model. PKU mice have a mutation in both copies of the phenylalanine hydroxylase (PAH) gene (homozygous) causing a functional deficit, comparable to PKU patients. The mutation is bred in the black and tan, brachyury (BTBR) mouse and the C57Bl/6J (B6) mouse. In chapter 2, we show distinct differences in the behavioral outcome between these strains, most striking in learning and memory. These differences cannot be attributed to biochemical differences in Phe concentrations or neurotransmitter concentrations of norepinephrine and serotonin while these show a clear PKU phenotype in both genetic backgrounds. In chapter 3, we illustrate that the phenotypical outcome of PKU mice is not only different between genetic background but also between males and females. It is demonstrated that B6 female mice show deficits in learning and memory in contrast to the male counterparts. This suggests that the consequence of the PAH mutation is influenced by the sex and genetic background of 180

182 Chapter 10 mice. If the preclinical results can be extrapolated to the clinical situation, one would expect difference in the manifestation of PKU in PKU patients. Are there individual differences found in the manifestation of PKU in patients? Even though reports are sparse, some studies describe PKU patients that escaped the severe cognitive deficits despite high Phe concentrations (Table. 1). Although diagnosis and inclusion criteria of these atypical PKU patients can differ slightly, Table 1 lists eight families that all escaped severe cognitive deficits. This suggests a genetic influence. The predicted association between genotype and atypical phenotype was investigated by Ramus and colleguas 1,2. They showed that PAH activity, based on genotype, could not fully explain the atypical phenotype as untreated siblings with the same PAH genotype could show remarkable differences in cognitive outcome 2. This was also suggested by the reports described in Table 1, the third and the fourth subgroup. Here, studies where listed that showed differences in PKU outcome among siblings. Ramus et al. further investigated other modifying genes influencing phenotypical outcome. However, the examination of polymorphisms in tyrosine hydroxylase or markers closely linked to the PAH gene yielded no possible candidates. As the point mutation, and so the predicted PAH activity of the two genetic background, is the same, the two genetic backgrounds of the PKU mouse model could help facilitate the research in these modifying genes. In the PKU mice studies we found that genetic background and sex of the mice could influence the phenotypical outcome. In PKU research, studies concerning gender differences are limited Despite these limited reports, indications of sex differences are reported. First, in Table 1, 25 atypical female and 17 atypical male PKU are described. This skewed distribution is in contrast to our findings in B6 PKU mice where females are thought to be more affected in the cognitive domain than males. However, the larger number of reports in females could be an artifact of the identification process of atypical PKU patients. The atypical PKU patients are often identified when their siblings have typical PKU or when children of atypical PKU patients show mental disabilities, typical PKU or PKU identified with via newborn screening. Atypical PKU mothers will have a higher change to be identified via this later route, as even non-pku children can be affected in utero by the high concentrations of Phe of the mother (maternal PKU). This differs from atypical PKU fathers, where healthy children can be born when the child has a healthy mother. Second, in Table 1 a separation was made between genders of the affected siblings of atypical PKU patients. Eight reports discussed affected sisters, in the face of four brothers. However, this distribution is not statistically different (c 2 (1,N=13)=1.17,p= ). The outcome could be influenced by the selecting parameter, IQ, as other behavioral disturbances could be present in atypical 10 1 All statistical analysis was performed in IBM SPSS Statistics for Windows, Version 22.0 (Armonk, NY: IBM Corp.) 181

183 Chapter 10 PKU patients despite normal IQ (Hsia et al. 1970). Finally, indeed, in studies concerning treated PKU patients, sex differences are observed in the pathophysiology of PKU such as differences in the occurrence of psychiatric disorders, visual attention, dietary control, quality of life, and personality 10,11,13, Together, these studies highlight that genetic factors and sex can influence the outcome of PKU Considerations of the PKU mouse model The PKU mouse model has clear similarities with the modeled situation; 1) PKU mice have a point mutation in the PAH gene, 2) the mutation causes Phe concentration to rise in blood and brain, 3) biochemical consequences in amino acids and neurotransmitter concentration are similar to PKU patients, 4) cognitive deficits are found in BTBR PKU mice, 5) B6 PKU mice might resemble atypical PKU patients in the cognitive domain, and 6) problems in sleep are found in PKU mice and patients (Chapter 4). Utilizing the model to its full potential is not only identifying the strengths, but also considering the limitations. First, the original PKU mouse model was described in the black and tan, brachyury (BTBR) strain 18. This strain of mice are commonly used in N-nitrosoN-ethylurea (ENU) genetic screens because of the relative high forward mutation rates after a single dose of the mutagen in these mice compared to, for example, the B6 mouse 18,19. Yet, the BTBR wild-type (WT) mice show abnormalities in brain morphology and behavior Therefore, the BTBR PKU mouse model was crossed back on the B6 background. The question that arises is: Why are both genetic backgrounds still used in preclinical PKU research as these abnormalities are known in the BTRB background? Without underestimating the influence of these abnormalities on the PKU phenotype in BTBR, only BTBR PKU mice consistently mimics the typical PKU patients in behavioral outcome. Therefore, when behavioral outcome is included in the outcome parameters, the BTBR background of the PKU mouse model is preferred. Secondly, there are limitations in the behaviors mice can express and, therefore, model. For instance, in early-treated PKU patients deficits are still found in working memory, executive functioning and personality 26,27. Although specific behavioral paradigms are designed to investigate these cognitive functions, they are often dependent on the definition used and measure only certain components of the behavior. Finally, newborn screening makes it possible to identify PKU patients early-in-life and start Phe-restricted as soon as possible. One could question, what type of PKU patient the PKU mouse model is a model for in the future. When this is the early-treated PKU patient, practical problems could appear as intervening in the first days of the pups is very difficult. To conclude, the PKU mouse models the PKU patients adequately. If the use of either one is a conscious decision, the BTBR PKU and the B6 PKU mice have their translational value for studying PKU. 182

184 Chapter 10 Table 1 Case reports of atypical PKU patients. Reports were included when the gender of the PKU patients and PKU siblings were stated. First two columns show author and date. The third column depict the males (M) and females (F) with normal or borderline intelligence (IQ>70). When more than 1 individual is depicted in this column, the patients were siblings. The categorization of the fourth column is described at the bottom of the table. The dotted line distinguishes between four subgroups. The first group consist of single cases atypical PKU patients without siblings (with PKU) or siblings are not reported. The second subgroup, consists of atypical PKU patients related to each other without (more) siblings or typical PKU siblings. The third subgroup, consist of atypical PKU patients with a typical PKU brother. The final subgroup, consist of atypical PKU patients with a typical PKU sister. Authors Date Siblings Other Siblings Allen M A Coates M A Dyken and Culley (Case 4) M A Frankenburg (Case B) F A Hsia (Yannet) M A Perry F A Fisch F A Leonard F B Partington F B Woolf (Case 1-2) F A Colombo M; 1F B Dyken and Culley (Case 1-3) M; 1F B Frankenburg (Case C) M; 1F B Hsia F B Koch M B Perry M;3F B Zinger F B Blainey F C Kasim F C Perry (Case 1-2) M C Stevenson F C Brugger M D Coffelt F D Hsia F D Hsia (Yannet) F D Jervis M D Knox M D Mabry F D Tapia M D Total 17M;25F 10 A not certain or not reported B no siblings (with PKU) C typical PKU brother D typical PKU sister 183

185 Chapter 10 New treatment strategies in PKU Specific nutrient combination A specific nutrient combination (SNC) comprised of uridine monophosphate (UMP), docosahexaenoic acid (DHA), eicosapentaenoic acid (EPA), choline, phospholipids, folic acid, vitamins B12, B6, C, and E, and selenium was investigated for the first time as new (additional) treatment strategy for PKU. In the first proof-of-concept study, described in chapter 6, male and female B6 WT and PKU mice were offered a diet supplemented with SNC for post-natal day (PND) 31 up to PND 115. After this period, a positive effect of SNC supplementation was found in the post-synaptic marker, postsynaptic density protein 95 (PSD-95), in specific sub-regions of the hippocampus. As this protein is associated with growth and functioning of dendritic spines 28,29, the SNC supplementation is hypothesized to restore the affected synaptic function in PKU mice. The possible improvement in synaptic functioning could implicate an improvement in functional outcome. Therefore, in chapter 7, we investigated the behavioral outcome of SNC supplementation in the BTBR WT and PKU mice (male and female). In this long-term intervention study, we show that SNC supplementation can improve novel object recognition memory in high Phe and low Phe conditions. These two studies show that a nutrient combination that specifically targets consequences of Phe in the brain can improve functional outcome despite the high Phe concentrations in blood and brain and possibly other PKU-related biochemical disturbances. This is a novel view on treatment in PKU. Current treatment and research into new treatment strategies focusses on the biochemical outcome of the treatment, for example reducing Phe concentrations. Therefore, these two chapters highlight that (new) treatment strategies, perhaps, do not need to be exclusively confined to the original idea of reducing Phe concentrations. Identifying the mode of action of the SNC could help pinpoint the critical (biochemical) parameter(s) or brain networks involved in the functional outcome of PKU. One treatment that also moves beyond the idea of reducing Phe concentrations, is LNAA treatment LNAA The concept of treating PKU with LNAA supplementation is not a new approach to PKU However, despite this research, the optimal composition of the LNAA treatment is not yet identified. In addition, the integration of mulptiple biochemical treatment objectives is lacking (for example (1) reducing brain Phe concentrations, (2) improve protein synthesis by restoring brain non-phe LNAA concentrations, and (3) improve monoaminergic neurotransmitter synthesis by increasing amino acid precursors) 38. In chapter 8, LNAA treatment of equal amounts of the LNAA s tyrosine, tryptophan, valine, isoleucine, leucine, methionine, and histidine (based on Pietz et al. 1999) was added to the diet of B6 WT and PKU mice (both sexes). This study showed that such a regime can reduce brain Phe concentrations, improve the concentrations of some of the non-phe LNAA s, and improve brain levels of serotonin and norepinephrine. In chapter 9, the composition of the LNAA 184

186 Chapter 10 regime was refined and tailored to the three biochemical aims. This resulted in five specific LNAA regimes that were offered to B6 PKU (both sexes) mice for six weeks and compared to a normal diet in WT and PKU mice and a high-protein control diet in PKU mice. The outcome for the three biochemical aims were: First, brain Phe concentrations could be reduced by LNAA treatment with or without threonine and with supplementation of solely leucine and isoleucine. Secondly, non-phe LNAA concentrations could be restored if they were added to the supplementation. When they were not added to the regime, they were more impaired. Finally, monoaminergic neurotransmitter concentrations were improved when the precursors (tyrosine+ tryptophan) were added and with both LNAA treatments (with or without threonine). The improvements found for all three biochemical aims was not always accompanied with reduced Phe concentrations in plasma. Again, these studies highlight that the blood Phe concentrations alone, as currently used clinically, is not always indicative of the detrimental effects in the brain. Therefore, it is of importance to evaluate the consequence for all three biochemical aims, in future LNAA studies but also other PKU research. Although we recognize that investigating all three biochemical aims in PKU patient can be challenging. For example, non-invasive methods of measuring neurotransmitter concentrations repeatedly in the brain are currently not at hand. Finally, for the LNAA studies, we can conclude that the balance between all components of the LNAA s is of great importance. 2. FUTURE PERSPECTIVES 2.1. Future perspective I: the PKU mouse model Can a physiological challenge influence the BTBR and B6 differently? In Chapter 2, we show that male B6 mice are able to master the learning and memory paradigms despite similar biochemical changes in blood and brain. Therefore, we hypothesize in the discussion of that chapter that the biochemical changes could affect the B6 PKU differently than BTBR PKU during the neurodevelopment and adult life. In the discussion of this chapter, we make the assumption that the PKU-related biochemical differences are not influenced by the physiological challenge of the forced swim test two hours before sacrifice. However, it is possible that the physiological challenge (e.g. the massive release of stress hormones) elicits or masks changes of the naïve state that, in the end, could mask possible differences between the genetic strains. To explore this hypothesis, home-cage controls (age-matched, sex-matched, and housed under the same conditions) were compared to the tested BTBR WT, BTBR PKU, B6 WT, and B6 PKU of Chapter 2. Statistical analysis with multivariate ANOVA (factors; genetic background (BTBR/B6), state (naïve/tested), and genotype (WT/PKU)), was performed on amino acid concentrations in blood and brain and neurotransmitter concentrations in brain (Figure 1-3)

187 Chapter 10 Figure 1 Amino acids in plasma of both genetic backgrounds. Except for Phenylalanine (Phe: F(1,45)=1.038 p=0.315), in interaction effect between genotype was found for all LNAA s (Tyrosine: F(1,45)=4.542 p<0.05, Valine: F(1,45)= p<0.05, Leucine: F(1,45)= p<0.05, Histidine: F(1,45)=4.360 p<0.05, Threonine: F(1,45)= p<0.05, and Isoleucine: F(1,45)=8.861 p<0.05). Furthermore, an interaction effect of genetic background and genotype was observed for all LNAA s (Phe: F(1,45)= p<0.05, Tyrosine: F(1,45)= p<0.05,valine: F(45,1)= p<0.05, Leucine: F(1,45)=9.005 p<0.05, Histidine: F(1,45)=8.952 p<0.05, Threonine: F(1,45)= p<0.05, Isoleucine: F(1,45)=7.059 p<0.05). WT= wild type, PKU= phenylketonuria, (N)= naïve mice, (T)= tested mice (Chapter 2). 186

188 Chapter 10 For amino acids in the blood, a significant interaction effect was found between genotype and state for all LNAA depicted in Figure 1, except for Phe. This indicates that the response to a physiological challenge is different between WT s and PKU s but not between the genetic backgrounds. Genetic background did interact with genotype for all LNAA depicted in Figure 1. From the graphs it is evident that the WTs of each genetic background differ from each other. The results found in blood amino acids could not be directly extrapolated to brain amino acids (Figure 2). Only, Phe concentrations were affected by an interaction of genotype and state and background and state. The tested B6 PKU mice showed lower Phe concentrations than the naïve B6 PKU, a response to the challenge that is not observed in the WT s or the BTBR PKU. As in the plasma, genetic background and genotype did interact, except for tryptophan and threonine. However, in the brain the concentration of these amino acids were found to differ between the PKU s of each strain. Overall, BTBR PKU and B6 PKU seem to respond similar to physiological challenge in amino acid concentrations in blood and brain. From the graphs, only the response of histidine in the brain stands out. The BTBR PKU and B6 PKU seem to be affected in opposite direction, however, only a trend was observed (p=0.058). For the neurotransmitters, a clear PKU phenotype is found for dopamine, norepinephrine, serotonin and the turn-over of serotonin (Figure 3A-C) independently of genetic background. Overall, the genetic backgrounds differed in norepinephrine and the turn-over of serotonin. Furthermore, only the turn-over of serotonin was affect by the physiological challenge wherein a trend was observed in a difference between WTs and PKUs. These results suggest that there are no major differences in neurotransmitter concentrations between BTBR PKU and B6 PKU. Therefore, these results raise the hypothesis that neurotransmitter concentrations alone are not likely to explain the difference found in behavioral outcome. Interestingly, although it is a trend, it seems that PKU individuals of both genetic backgrounds cannot increase the turn-over of serotonin after a physiological challenge, a response that is also observed in WT littermates. Together, this data suggest that the response in amino acids and neurotransmitters to a physiological challenge is not different for the genetic backgrounds despite clear difference between WT and PKU

189 Chapter 10 Figure 2 Amino acids in brain of both genetic backgrounds. Phe concentrations are affected by an interaction of genotype and state (F(1,45)=4.461, p<0.05) and background and state (F(1,45)=4.694, p<0.05). Genetic background and genotype interacted, except for Tryptophan and Threonine (Phe: F(1,45)=7.573 p<0.05, Tyrosine: F(1,45)=5.281 p<0.05, Tryptophan: F(1,45)= p= 0.272, Valine: F(1,45)=8.357 p<0.05, Isoleucine: F(1,45)=5.460 p<0.05, Leucine: F(1,45)=6.349 p<0.05, Methionine; F(1,45)=6.789 p<0.05, Histidine: F(1,45)=6.863 p<0.05, Threonine: F(1,45)=3.798 p=0.059). A main effect of the physiological challenge was found for Histidine and Methionine (F(1,45)=4.934 p<0.05, F(1,45)=4.387, p<0.05, respectively). In Histidine, a trend was observed in the interaction of genotype and state (F(1,45)=3.831, p=0.058). WT= wild type, PKU= phenylketonuria, (N)= naïve mice, (T)= tested mice (Chapter 2) Identifying the difference between BTBR and B6 From Chapter 2 it is clear that genetic factors influence the PKU phenotype. To identify candidate genes, different genetic screening tools can be used such as RNA-seq. This approach would allow us to identify candidate genes that are upregulated, for example following a learning event. Likewise, it would help elucidate gene clusters that are distinctively expressed by the different genetic strains or sexes. Another possibility is to investigate m-rna profiles during neurodevelopment. The previous found similarities in PKU phenotype of both genetics background in amino acids and neurotransmitters (paragraph 2.1.1) withholds the hypothesis that the PKU mutation in different genetic backgrounds differently affects these PKU induced changes, presumably during the neurodevelopment and adult life. 188

190 Chapter 10 Research showed that a seven-day treatment in BTBR PKU pups (post-natal day 14-21) with 5-hydroxytryptophan, a precursor of serotonin, can improve dendritic spine maturation and performance in a short-term version of the NOR and SOR tests 39. Investigating differences between genetic backgrounds around this time window, could help identifying the processes protecting the PKU mutant mice with a B6 background. Figure 3 Neurotransmitters in blood of both genetic backgrounds. (A) PKU phenotype was observed for dopamine (F(1,45)=5.698 p<0.05), norepinephrine (F(1,45)= p<0.05), Serotonin (F(1,45)= p<0.05) and turn-over of serotonin (5-HIAA/Serotonin F(1,45)=58.272, p<0.05). Independently of the physiological challenge, differences are found between BTBR and B6 in norepinephrine (F(1,45)= p<0.05) and the turn-over of serotonin (F(1,45)= p<0.05). The turn-over of serotonin seems to be affected by the physiological challenge (F(1,45) p<0.05) wherein a trend was observed in a difference between WT and PKU s (F(1,45)=3.166, p=0.083). WT= wild type, PKU= phenylketonuria, (N)= naïve mice, (T)= tested mice (Chapter 2) Future perspective II: Sleep research in PKU In Chapter 4, we show that PKU patients have more sleep disorders, reduced sleep quality, increased latency to fall asleep, and experience more sleepiness during the day compared to first degree relatives. In the PKU mice, we found an increase in fragmentation, more switches between active and non-active behavior, and a shift in diurnality, shifting a part from their resting behavior into the active phase. Both experiments strongly support the hypothesis that 189

191 Chapter 10 sleep is affected in PKU, a starting point for sleep research in PKU. An important new avenue as sleep problems can negatively affect cognitive functioning, e.g. executive functioning 40,41 and mood, e.g. feelings of depression, anxiety and stress 42,43. These are disturbances similar to the described disturbances in early-and continuous treated PKU patients (e.g. in executive functioning, mood, social cognition, and in internalizing problems such as depression and anxiety 26,27 ) Improving the electronic survey. The proof-of-concept study in chapter 4 examined a small number of PKU patients and first-degree relatives (FDR) controls. In retrospect, small adjustments could yield more information in future cohorts. First, the distribution in age was different between PKU patients and FDR controls. In the study, the recruitment of healthy sibling of PKU patients was more difficult than expected at forehand. Besides investigating more time in this recruitment, an age-matched control group from the general population could be included to examine the differences between PKU patients and a healthy population. Second, the recruitment of subjects was done via a specific call concerning a sleep questionnaire via Dutch PKU society on a non-committal basis. This could have resulted in a responder bias as people of the Dutch PKU society are probably a subclass of PKU patients. Dispersing the questionnaire via health care professionals could possibly give a better representation of the PKU community. Third, in the questionnaire one question was added concerning the treatment of the PKU patients. However, it is not clear how well the patients were controlled. Including a question about recent Phe concentrations or involving the treating health care professionals, could possibly associate the Phe concentration or type of treatment with the sleep disturbances. Finally, the leading treatment for PKU patients is a Phe restricted diet supplemented with artificial amino acids, vitamins, and minerals. The timing of the artificial mix could influence the timing of sleep. Evaluation of this timing should be included in the future survey. Together, these adjustments would give more insights in the relation between sleep and PKU Moving beyond the proof-of-concept study The proof-of-concept study indicated that the timing of sleep and switching from active to non-active behavior could be affected in PKU. Corroboration of these findings should be done in future studies. Timing of sleep is influenced by sleep regulatory processes and circadian rhythm 44. The sleep regulatory processes could be influenced by neurotransmitter disturbances in PKU. Improving these neurotransmitter concentrations by either Pherestriction or LNAA supplementation could identify if this processes are at hand. To investigate if PKU influences the circadian rhythm, PKU mice could be exposed to continuous lighting regimes (constant light or dark) to identify disruptions in the internal free-running rhythm of the animal or phase-shift experiments to identify problems in shifting sleep/wake patterns. The use of PKU mice could help identify the neurobiological substrates. In PKU 190

192 Chapter 10 patients, core body temperature and dim-light melatonin rhythm could be monitored to investigate if PKU patients experience a blunted or delayed internal rhythm of physiological markers. These experiments would give the first indications concerning the underlying mechanism of sleep problems in PKU Possible treatment strategy: Exercise A possible treatment strategy to improve the altered sleep/wake pattern is exercise. This treatment strategy is, for instance, investigated in aged individuals where disturbed temporal regulation of rest/wake cycle, as found in PKU mice, is part of the altered rest/wake pattern. Voluntary exercise in these mice improved parameters associated with a strengthened rest/ wake pattern 45. On the basis of these findings, the hypothesis is that providing a running wheel to PKU mice would strengthen their rest/wake patterns. Although the study was not designed to examine the effect of exercise on rest/wake patterns, the study of Mazzola and colleagues (2015) did offer running wheels to female WT and PKU B6 mice (4 mo) 6. Reanalysis of the data showed reduced fragmentation and a shift in diurnality in PKU mice (Figure 4A,B one-tailed t-test: fragmentation t(17)=1.951 p<0.05, diurnality t(17)=2.071 p<0.05). The affected rest/wake pattern in these PKU mice suggest that voluntary exercise does not improve the circadian rhythm. However, as previously pointed out, this study was not designed to examine rest/wake patterns. The type of running wheel (with bars) could have affected the outcome as motor deficits could cause the PKU mice to have difficulties to run for long consecutive periods. A closed running wheel and passive infrared registration of all activity could reveal other results. Therefore, this data does not allow for definite conclusions, but is nevertheless in support of the data presented in chapter Figure 4 Fragmentation and diurnality score of exercising PKU mice. WT mice (n=9) and PKU mice (n=10) had excess to a running wheel for 53 days. Fragmentation score and diurnality was calculated over data collected from this period. In both scores, the WT and PKU mice differed from each other (one-tailed t-test: fragmentation t(17)=1.951 p<0.05, Diurnality t(17)=2.071 p<0.05). 191

193 Chapter Future perspective III: New treatment strategies in PKU SNC and LNAA In chapter 6 and 7, specific improvements are found after SNC supplementation. However, a definite mode of action of SNC was not clearly identified in these PKU studies, despite the clear hypotheses we had (Chapter 5). One hypothesis was synaptic functioning. Although chapter 6 have indicated improved synaptic functioning in the B6 PKU, the differences in age, exposure to behavioral paradigms, housing conditions and genetic backgrounds of the PKU mouse model yield caution when relating the improved synaptic functioning to the improved functional outcome. To examine a direct effect of SNC on neurons, one could investigate primary neuronal cell cultures with and without Phe. SNC supplementation under these conditions could help identify a direct effect of SNC on post-and pre-synaptic markers or specific changes in neuronal morphology. Another hypothesis was neurotransmitter metabolism. Neurotransmitters concentrations were measured in the whole brain of the BTBR mice of chapter 7 2. From figure 5 of the discussed neurotransmitter are affecting the functional outcome the PKU mice in the NOR, it is clear that SNC supplementation did not influence whole brain concentration of neurotransmitters or turnover of serotonin within these groups under these circumstances. With the caution that regional differences in the brain may occur, it is unlikely that the neurotransmitter metabolism R. To examine the mode of action of SNC, future research should, therefore, elaborate on the effect of SNC on the domains raised in chapter 5 (synaptic functioning, neurotransmitter metabolism oxidative stress, and white matter integrity). Within these studies, attention must be applied to regional differences of the brain as the SNC effect found in chapter 6 was very specific. Furthermore, when restoring neurotransmitter deficits in PKU is one of the biochemical aims of treatment, combining the strengths of SNC supplementation and LNAA treatment could be of great interest Placing SNC and LNAA on the scale of new treatment strategies in PKU The search of new treatment strategies for PKU is broad with diverse treatment objectives. The different treatment strategies could be placed along a figurative scale, on one end alternatives for amino acids mixture which could improve compliance and metabolic control and, at the other end, targeting the cause of the disease (Figure 6). The treatments will be discussed from the left side from this schematic representation to the right side (Figure 6). First, Glycomacropeptide (GMP) is a naturally occurring protein that is lacking Phe. This protein can be a protein source for PKU patients that could improve compliance and metabolic control. However, due to the process of getting GMP free from the whey protein, some Phe will be in the product offered to patients 46. Second, SNC, discussed in Chapter 5, 6, and 7, is designed to relieve the effects of high Phe on the brain. Although 2 Neurotransmitter measurements were performed according to the protocols described in chapter 8 and 9 192

194 Chapter 10 Figure 5 The effect of SNC supplementation on neurotransmitter concentrations in PKU mice. The statistical analysis with a multivariate ANOVA (factors sex and group) showed no interaction effect between sex and group. Therefore, the graphs include both male and female mice. For dopamine, no differences were observed between the groups. For norepinephrine, the PKU mice on high Phe (C-HP and SNC-HP) significantly differed from WTs (WT C-HP and WT SNC-HP) and PKUs on low Phe diet (C-LP and SNC-LP). These differences were also observed in Serotonin. The turnover of serotonin (5-HIAA/Serotonin) differed between WT C-HP and both high Phe PKU groups (C-HP p=0.013, SNC- HP p=0.011) and between WT SNC-HP with all PKU groups (PKU C-HP p<0.001, PKU SNC-HP p<0.001, PKU C-LP p=0.005, PKU SNC-LP p=0.041). * p<0.05, 5-HIAA= 5-Hydroxyindoleacetic acid (metabolite of serotonin), n=12, mean±standard error of the mean

195 Chapter 10 the first positive reports are included in this thesis, at this moment, SNC supplementation is more likely to be added to the current treatment strategy or LNAA treatment than be a treatment strategy on its own. Third, LNAA, discussed in Chapter 8 and 9, has been found to lower Phe concentrations, improve other non-phe LNAA s concentrations in brain and improve neurotransmitter concentrations. However, the optimal composition of LNAA s is yet to be determined. Fourth, Phe in high concentrations can self-assemble to toxic fibrils with an amyloid-like structure 47. Inhibiting this self-assemble by antibodies or small molecules could yield new treatment strategies Targeting the amyloid structure is not a novel approach in neuroscience, as this is also done in neurodegenerative diseases, such Alzheimer s disease. Fifth, tetrahydrobiopterin (BH4) is the essential in the conversion of Phe to tyrosine. Depending on the genotype in the PAH gene, some PKU patients benefit from supplementation of the synthetic form of BH4. Furthermore, a recent study showed that high doses of BH4 can also be beneficial in PKU conditions normally not responding to BH4, namely in PKU mice 50. Besides the role in Phe metabolism, BH4 is an important cofactor of tyrosine hydroxylase and tryptophan hydroxylase, key enzymes in the synthesis of dopamine and serotonin. High doses of BH4 can be beneficial for the turnover of these neurotransmitters in the brain 50. Fifth, Phenylalanine Ammonia Lyase (PAL) is an enzyme derived from yeast and fungi that can degrade Phe to small amounts of ammonia and to trans-cinnamic acid, a harmless metabolite. Several clinical trials have been performed with an injectable form of this enzyme, ravpal PEG 51. This treatment is successful in lowering Phe concentrations, but the multiple injections could cause irritation to the injection side and immune reactions 51. Research had been investing in oral administration of PAL, but, at this moment, this is not as successful as ravpal PEG Finally, gene therapy is a form of treatment in which, for PKU, the PAH gene is carried into body by a stable virus, for example an adeno associated virus. This virus can be directed towards the liver, however, the corrected PAH activity is not permanent and reintroduction of the virus is difficult as immune reaction towards the virus can occur Another possibility is muscle-directed gene therapy. Normally, the muscle has no PAH enzyme and the cofactor to convert Phe to tyrosine. Therefore, successful muscle-directed gene therapy doesn t solely contain the PAH gene but also contain the BH4-biosynthetic enzymes GTP cyclohydrolase I (GTPCH) and 6-pyruvoyl-tetrahydropterin synthase (PTPS) 61. To conclude, several new treatment strategies are explored in the PKU research field. When depicting these along a scale they can be ranked on the target of the treatment. However, to some extent, the scale can be replaced by non-invasive (left) to invasive (right) or even shortterm solutions (left) to long-term solutions (right). If the development indeed will follow this last subdivision is not clear. However, when it advances are made on the right hand of the scale, treatment strategies on the left hand of the scale become less relevant. 194

196 Chapter 10 Figure 6 Schematic representation of PKU treatment strategies. Abbreviations and symbols: GMP= Glycomacropeptide, SNC= Specific nutrient combination, LNAA= large neutral amino acid, # Phe= Interfering in the assemble of toxic Phe fibils, BH4=tetrahydrobiopterin, PAL= Phenylalanine ammonia lyase. 3. FINAL CONCLUSION Taken all results and consideration together the following implications can be identified. First, all experimental chapters (except for chapter 4) show that Phe concentrations in plasma are not always a good predictor of the pathophysiological outcome. Future research should focus on identifying predictors (e.g. genetic background) and markers that can help monitor biochemical changes in the brain. Second, chapter 4 recognizes for the first time sleep problems in PKU. Being aware of sleep problems in PKU, and, in the end, treating sleep problems in PKU can, hopefully, improve treatment. Finally, this thesis contained studies that showed new aspects of the disease (PKU strain differences and sleep problems) and a treatment strategy not investigated before (SNC supplementation). Hopefully, this will be the basis of follow-up studies in PKU that could, in the end, improve the treatment of PKU and release the burden of the strict dietary treatment PKU patients have to adhere to