originl rticle MT Open Inhiition of Glycolte Oxidse With Dicer-sustrte sirna Reduces Clcium Oxlte Deposition in Mouse Model of Primry Hyperoxluri Type 1 Chitli Dutt 1, Nicole Avithl-Curtis 1, Ntlie Pursell 1, Mrit Lrsson Cohen 1, Benjmin Holmes 1, Rohn Diwnji 1, Wei Zhou 1, Lucino Apponi 1, Mrtin Koser 1, Bo Ying 1, Dongyu Chen 1, Xue Shui 1, Utsv Sxen 1, Wendy A Cyr 1, Anee Shh 1, Nim Nzef 1, Weimin Wng 1, Mrc Arms 1, Henryk Dudek 1, Edurdo Slido 2, Bo D Brown 1 nd Chengjung Li 1 1 Dicern Phrmceuticls, Cmridge, Msschusetts, USA; 2 Instituto de Tecnologís Biomédics (ITB), Centro de Investigciones Biomédics de Cnris (CIBICAN), Centre for Biomedicl Reserch on Rre Diseses (CIBERER), Universidd L Lgun, Tenerife, Spin Primry hyperoxluri type 1 (PH1) is n utosoml recessive, metolic disorder cused y muttions of lnine-glyoxylte minotrnsferse (AGT), key heptic enzyme in the detoxifiction of glyoxylte rising from multiple norml metolic pthwys to glycine. Accumultion of glyoxylte, precursor of oxlte, leds to the overproduction of oxlte in the liver, which ccumultes to high levels in kidneys nd urine. Crystliztion of clcium oxlte (COx) in the kidney ultimtely results in renl filure. Currently, the only tretment effective in reduction of oxlte production in ptients who do not respond to high-dose vitmin B 6 therpy is comined liver/kidney trnsplnt. We explored n lterntive pproch to prevent glyoxylte production using Dicer-sustrte smll interfering RNAs (DsiRNAs) trgeting hydroxycid oxidse 1 (HAO1) mrna which encodes glycolte oxidse (), to reduce the heptic conversion of glycolte to glyoxylte. This pproch efficiently reduces mrna nd protein in the livers of mice nd nonhumn primtes. Reduction of heptic leds to normliztion of urine oxlte levels nd reduces COx deposition in preclinicl mouse model of PH1. Our results support the use of DsiRNA to reduce liver levels s potentil therpeutic pproch to tret PH1. Received 3 Octoer 215; ccepted 31 Decemer 215; dvnce online puliction 9 Ferury 216. doi:1.138/mt.216.4 INTRODUCTION The primry hyperoxluris (PH) re group of utosoml recessive disorders cused y the overproduction of oxlte. 1 4 Oxlte is n end product of glyoxylte metolism tht is produced in the liver nd excreted y the kidney. Overproduction of oxlte cuses precipittion of highly insolule clcium oxlte (COx) crystls in the kidney, resulting in urolithisis, nephroclcinosis, nd eventully renl filure. As PH ptients progress to end-stge renl disese (ESRD), systemic oxlosis (oxlte deposited in ll tissues) often occurs s the comintion of overproduction nd limited excretion of oxlte. 5,6 There re three forms of PH in which the underlying defects hve een identified, designted s PH types 1 (PH1), 2 (PH2), nd 3 (PH3). PH1, the most common (round 8% of ll PH ptients) nd severe sutype, is cused y muttions in the AGXT gene leding to impired ctivity of the enzyme AGT. 7 12 AGT plys centrl role in metolizing the intermedite glyoxylte, which is generted y metolism of glycolte, glycine, nd hydroxyproline in the liver (Figure 1). 7 Glyoxylte is converted into either glycine y AGT or into oxlte y lctte dehydrogense (encoded y the LDHA gene in the liver). 2 Reduced AGT ctivity leds to incresed ccumultion of glyoxylte nd susequent production of oxlte, the ltter of which is predominntly excreted in urine nd to lesser extent vi the gstrointestinl trct. Current tretments for PH1 focus on preserving renl function y decresing oxlte production nd crystlliztion through incresed fluid intke, intensive dilysis, orl dministrtion of clcium oxlte crystl inhiitors, such s citrte, nd high doses of pyridoxine (5 8 mg/kg/dy). 13,14 Indequte response to these therpies cn e followed y removl of ostructing stones or comined liver/kidney trnsplnt. However, these tretments re insufficient, s progressive nephroclcinosis leds to ESRD during the first two to three decdes in most ptients. 13 Severl novel moleculr pproches to reduce toxic oxlte lod hve een proposed nd re currently under investigtion, including oxlte degrding enzyme therpy, correction of AGT mis-trgeting nd virlly medited AGXT gene therpies. 15 23 An lternte pproch of sustrte reduction, y which the production of glyoxylte is reduced through inhiition of glycolte oxidse (), cn e potentilly effective therpeutic option. 24 27 Glycolte is highly solule smll molecule nd cn e eliminted in the urine t high concentrtions without inducing kidney dmge or other toxicity. Inhiition of reduces the production of glyoxylte nd susequently oxlte, while incresing glycolte levels. In the ccompnying pper, 28 reduction of urinry oxlte levels ws oserved Correspondence: Chengjung Li, Dicern Phrmceuticls, 87 Cmridgeprk Drive, Cmridge, Msschusetts 214, USA. E-mil: cli@dicern.com nd Edurdo Slido, Deprtment of Pthology, Universidd de L Lgun School Medicine, Tenerife 3832, Spin. E-mil: eslido@ull.es 77 www.moleculrtherpy.org vol. 24 no. 4, 77 778 pr. 216
DsiRNA Trgeting HAO1 Prevents Hyperoxluri Ethylene glycol Glucose Food Alcohol dehydrogense Glyoxl ADH Glycoldehyde GULO Gulonolctone oxidse ** GLO1 GLO2 Glyoxylse 1 Glyoxylse 2 ADH Alcohol dehydrogense Glycine Glycolte Hydroxyproline ALDH DAO * AGT HAO1 GRHPR HYPDH Aldehyde dehydrogense PH1 PH2 HOG HOGA PH3 LDH Non enzymtic? Oxlte * Rodents do not express DAO in the liver ** lcking in humns, pes nd monkeys Figure 1 Oxlte metolism pthwy overview. ADH, lcohol dehydrogense; AGT, lnine-glyoxylte minotrnsferse; ALDH, ldehyde dehydrogense; GLO, glyoxylse; GRHPR, glyoxylte reductse hydroxypyruvte reductse; GULO, gulonolctone oxidse; HAO1, hydroxycid oxidse 1; HOG, 4-hydroxy-2-oxoglutrte; HOGA, 4-hydroxy-2-oxoglutrte ldolse; HYPDH (PRODH2), hydroxyproline dehydrogense; LDH, lctte dehydrogense. when n Agxt-deficient (Agxt -/- ) mouse model, which exhiits key fetures of PH1 due to completely lcking of Agxt mrna nd protein, ws crossed with Ho1-deficient mouse model. 28 Moreover, mice with deficiencies in oth AGT (Agxt-deficient) nd (Ho1-deficient) displyed norml growth nd development when compred to their heterozygous littermtes. Tken together, these studies suggest inhiition could e potentilly effective nd sfe therpeutic option for PH1 ptients. In this study, we investigte RNA interference (RNAi) medited enzyme reduction s novel potentil sustrte reduction strtegy for PH1 ptients. Dicer-sustrte smll interfering sir- NAs (DsiRNAs) hve een demonstrted to specificlly nd effectively reduce the expression of the trget mrna nd protein. 29 We hve identified DsiRNAs tht trget the HAO1 mrna encoding protein. Here we show tht dministrtion of HAO1 DsiRNA chieves potent nd durle knockdown of HAO1 mrna nd protein in mice nd nonhumn primtes (NHPs). In ddition, we demonstrte tht dministrtion of HAO1 DsiRNA significntly reduces urine oxlte nd prevents kidney dmge in the hyperoxluric Agxt -/- mouse model of PH1. These results support further explortion of reduction through HAO1-trgeting DsiRNA s potentil novel pproch for the tretment of PH1. RESULTS exhiits potent nd specific ctivity in wildtype mice We identified our led HAO1 DsiRNA y sequentil screening in vitro followed y testing in vivo (Supplementry Figure S1). The led HAO1 DsiRNA formulted in lipid nnoprticle, designted s, showed potent, dose-dependent reduction of Ho1 mrna (ED5:.22 mg/kg, Figure 2) nd protein (Figure 2) in mouse liver. The homogeneity of Ho1 mrna nd protein knockdown in mouse heptocytes were confirmed y in situ hyridiztion nd immunohistochemistry, respectively. As shown in Figure 2c,d, tretment with.1 or 1 mg/kg resulted in significntly reduced Ho1 mrna nd protein expression in nerly ll heptocytes. These results demonstrte tht is potent gent tht efficiently suppresses protein expression in murine heptocytes. chieves rpid nd durle trget knockdown In order to determine the sensitivity nd durility of trget knockdown, we monitored the expression of Ho1 mrna nd protein for up to 6 dys following single dose of HAO1-1. rpidly suppressed Ho1 mrna expression, chieving ~9% reduction within 12 hours fter injection (Figure 3). Notly, t dy 45, the remining suppression of Ho1 mrna oserved in mice treted with.3 nd 1 mg/kg of ws ~2 nd 4%, respectively. We lso oserved tht protein ws reduced y more thn 9% y 5 dys post-dose, fter which the levels were reduced to elow detectle limits nd remined undetectle until dy 29 (Figure 3). protein levels recovered nd were ck to seline levels etween dys 45 nd 6 (Figure 3). In conclusion, single tretment of t dose of.3 or 1 mg/kg is le to rpidly produce sustined inhiition of protein expression in vivo. reduces urine oxlte in the Agxt -/- PH1 mouse model To further evlute the ctivity of s potentil therpeutic for PH1, we employed murine model of PH1. Agxt -/- mice hve germline muttion in the gene encoding AGT, resulting in high seline urinry oxlte levels compred to their wild-type Moleculr Therpy vol. 24 no. 4 pr. 216 771
DsiRNA Trgeting HAO1 Prevents Hyperoxluri % Ho1 mrna remining 1 ED5:.22 mg/kg 75 5 c.3.1 Dose (mg/kg).3.1.3 1. mg/kg Dose (mg/kg) d Liver Spleen 2 6 μm 6 μm 6 μm 6 μm 4 3 μm 3 μm 3 μm 3 μm.1 mg/kg.3 mg/kg Figure 2 displys potent in vivo ctivity in mice. The led HAO1 DsiRNA formulted in LNP () ws i.v. injected into wild-type mice with vrying doses s indicted. Liver smples were collected for nlysis of mrna fter 24 hours nd for protein on dy 5 fter single-dose injection. () Results of rel-time reverse trnscription PCR (men ± SD) indicte tht inhiits Ho1 mrna expression with high potency (ED5:.22 mg/kg). () Results of western nlysis show tht displyed potent inhiition of protein expression 5 dys fter injection. (c) Ho1 mrna ws detected in liver tissues y in situ hyridiztion nd showed tht smples isolted from nimls injected t.1 or 1 mg/kg hve lower hyridiztion signls compred to injected nimls. Br = 1 μm. (d) protein levels were mesured y immunohistochemistry. Results showed tht, either t.1 or.3 mg/kg, effectively nd homogenously reduced liver protein in mouse heptocytes, compred to injected controls. Spleen sections were used s negtive controls. Br = 3 or 6 μm. LNP, lipid nnoprticle. counterprts. 23 Mle Agxt -/- nimls recevied single i.v. dose of t.3 mg/kg, nd urine smples were nlyzed for oxlte, glycolte nd cretinine. Glycolte ws investigted s potentil iomrker of inhiition, sed on the ssumption tht suppression of protein would inhiit the conversion of glycolte to glyoxylte, resulting in incresed uriniry glycolte concentrtions. Urinry cretinine levels were used s control to normlize for vritions in urine concentrtion. The urinry oxlte levels normlized to cretinine were significntly reduced 5 dys fter tretment nd remined low until dy 15, efore slowly returning to seline levels y dy 28 (Figure 4). Simultneously, we oserved tht tretment resulted in incresed production of urinry glycolte (Figure 4). These results indicte tht tretment reduced urine oxlte through n on-trget suppression of expression, locking the conversion of glycolte into glyoxylte nd resulting in susequent elevtion of urinry glycolte. Our results lso indicte tht urinry glycolte could e used s phrmcodynmic mrker to monitor the response to. Furthermore, the results tht urinry oxlte levels instntly nd inversely mirrored glycolte levels suggest tht heptic glycolte is min source of glyoxylte nd oxlte overproduction in the Agxt /- hyperoxluric nimls (Supplementry Figure S2). The results tht normlized oxlte reduction nd glycolte elevtion correlted well t different time points lso indicte tht is remrkly effective trget for oxlte reduction in the PH1 disese mouse model (Supplementry Figure S2). Urinry oxlte nd glycolte levels were llowed to return to seline levels, fter which ws re-dministered t.3 mg/kg (on dy 33 nd gin on dy 4). Reproducily, oxlte reduction nd glycolte elevtion were detected with tretment, suggesting tht the oserved iologicl response is specificlly triggered y. Repeted doses of control DsiRNA showed no effect on urinry oxlte in the Agxt -/- mice (Supplementry Figure S3). We lso confirmed tht tretment reduced not only the oxlte concentrtion in the urine smples, ut lso the mount of excreted 772 www.moleculrtherpy.org vol. 24 no. 4 pr. 216
DsiRNA Trgeting HAO1 Prevents Hyperoxluri % Ho1 mrna remining 1 5.3 mg/kg 1 mg/kg 1 5 1 15 2 3 35 4 45 Dose 6 hours 12 hours 1 18 hours Dys 2 3 4 5 1 15 2 29 4 45 6 Dys Figure 3 chieves rpid nd durle trget knockdown in mice. Wild-type mice were injected with single i.v. dose of t.3 or 1 mg/kg. () At the indicted time points, five nimls were scrificed, nd liver smples were evluted for mrna levels y rel-time reverse trnscription PCR. The results (men ± SD) demonstrted tht Ho1 mrna levels were quickly diminished nd the inhiitory effect of single dose of lsted for more thn month. () Representtive results of western nlysis from ech time point of mice injected with single dose of or t.3 mg/kg. Results showed tht protein grdully decresed strting 6 hours postdose, reching levels elow the detection limit y dy 1 nd reemerging etween dys 29 nd 4. oxlte over 24 hours in the treted nimls (Figure 4c). reduced protein expression in Agxt -/- mice to the sme extent s ws oserved in wild-type mice (Supplementry Figure S3). Consistent with the results reported in the Ho1 nd Agxt doule knockout mice, 28 our results further suggest tht HAO1 could e potentil therpeutic trget for PH1. protects ginst kidney dmge elicited y ethylene glycol in Agxt -/- mice Mle Agxt -/- mice develop vrile degrees of kidney clcium oxlte deposits when fed ethylene glycol (EG), precursor of glycolte. 23 To determine whether reduction of oxlte y HAO1 tretment is sufficient to prevent kidney dmge, mle Agxt -/- mice were provided wter contining.7% EG while eing treted with either or once per week for 3 weeks. Additionlly, one group of nimls underwent delyed tretment, receiving one dose of following 2 weeks of EG dministrtion (Figure 5). Body weight nd wter consumption were mesured during the 3 weeks of study period (Supplementry Figure S4). At the end of 3 weeks, nimls tht received only EG showed high oxlte levels (Figure 5) nd vrile degrees of COx deposits in their kidneys (Figure 5c,d,e). Animls tht strted tretment with simultneously with EG dministrtion displyed no chnge in urinry oxlte levels (Figure 5) nd showed no kidney COx deposition (Figure 5d,e). These results indicte tht reduction of urinry oxlte levels is sufficient to prevent COx deposits in the kidney. Additionlly, nimls tht received delyed tretment showed miniml COx deposits, despite hving high oxlte levels for 2 weeks prior to tretment with (Figure 5,d,e). Formtion of kidney crystls in this model usully occurs fter 2 weeks of EG dministrtion, suggesting tht not only the level of urinry oxlte ut lso the durtion of oxlte elevtion re criticl for kidney COx depositions. Our dt showed tht kidney dmge cn e prevented y reducing the urine oxlte levels efore certin threshold point tht triggers the deposition of COx in kidney. Tken together, our results demonstrte tht utilizing -medited inhiition of to reduce the level nd/or durtion of elevted urinry oxlte is sufficient to protect ginst kidney dmge cused y prolonged hyperoxluri in PH1 mouse model. Cumultive inhiitory effect through frequent dministrtion of low-dose Since PH1 is chronic disese nd requires regulr tretment, we tested whether recurring tretment of would yield ny cumultive effects. Wild-type mice were injected with.3,.1, or.3 mg/kg of i-monthly nd were susequently nlyzed for the expression of Ho1 mrna nd protein t different time points. We oserved dose-dependent decrese in oth Ho1 mrna nd protein during the first month of tretment (Supplementry Figure S5). Notly, oth Ho1 mrna nd protein levels hd trough points tht were lower during the second cycle thn the first cycle. Additionlly, protein levels were lmost indistinguishle mong ll three dose level groups fter three doses of (Supplementry Figure S5). These results demonstrte the cumultive effect of recurring doses of. We hve lso demonstrted tht, with monthly dministrtion of, there were no significnt dverse effects in liver nd the hemtopoietic system during the course of 15 dys (Supplementry Figure S6). Tken together, these results re consistent with the results of Agxt nd Ho1 doule knockout mice 28 nd further support tht could e sfe nd effective trget for PH1. Moleculr Therpy vol. 24 no. 4 pr. 216 773
DsiRNA Trgeting HAO1 Prevents Hyperoxluri Oxlte/cretinine (mg/g) treted versus group 1..5. 7 14 21 28 35 42 49 Dose Dys Dose Dose Glycolte/cretinine (mg/g) treted versus group 4 3 2 1 7 14 21 28 35 42 49 Dose Dys Dose Dose c Totl oxlte (μmol/24 hours) 2.5 2. 1.5 1..5 d Oxlte/cretinine (mg/g) 6 4 2 LC-MS Enzymtic. Dy 42 Dy 72 Dy Dy 2 Dy 5 Figure 4 reduces urinry oxlte nd elevtes urinry glycolte levels in mouse PH1 model. Mle Agxt -/- mice were injected i.v. with (N = 4) t.3 mg/kg or (N = 5) s control on dy, dy 33, nd dy 4. Urine smples were mnully collected on predose dy 2, s well s dys 2, 5, 8, 12, 14, 16, 2, 23, 26, 28, 33, 35, 36, 41, nd 48, nd nlyzed for oxlte, glycolte, nd cretinine concentrtions y LC/ MS. () The chnge in urine oxlte levels in the tretment group ws clculted s the urine oxlte concentrtion normlized to cretinine (milligrm of oxlte per grm of cretinine) reltive to the verge normlized urine oxlte concentrtion of the control group, where the urine oxlte (milligrm of oxlte per grm of cretinine) in the control group ws set t 1. LC/MS nlysis (men ± SD) reveled tht reduced oxlte levels rpidly, nd the effects lsted for up to 15 dys efore eginning to return to seline levels. () Urine glycolte levels were mesured, normlized to cretinine, nd expressed s vlue reltive to the verge of the control group (which ws set t 1, s in Figure 3). simultneously induced the elevtion of urinry glycolte levels nd the reduction of oxlte levels. Results of these experiments with vlues plotted without normliztion to tretment group nd expressed s individul niml re provided in Supplementry Figure S2. (c) Twenty-four hour urine smples were lso collected using metolic cges from dy 41 to 42 nd from dy 71 to 72. The results (men ± SD) indicted tht oxlte excretion ws sustntilly lower fter tretment compred to control (). (d) Oxlte levels were mesured y oth enzymtic ssy nd LC/MS methods. The normlized oxlte vlues (milligrm per grm of cretinine) determined y oth methods were compred using urine smples collected on dy, dy 2 nd dy 5. The dt (men ± SD) indicte good correltion etween oth methods nd further confirm tht effectively reduces oxlte production. The rrows indicte the dys on which the nimls were treted with or. HAO1 DsiRNA pproch displys potent, durle, nd specific ctivity in NHPs To further confirm tht our HAO1 DsiRNA pproch is effective cross species, we evluted HAO1 DsiRNA in NHP-specific formultion in mle nd femle cynomolgus monkeys. We demonstrted tht this HAO1 DsiRNA potently suppressed mrna nd protein expression in dose-dependent mnner with n ED5 of less thn.3 mg/kg (Figure 6,). Similrly, this HAO1 DsiRNA chieved durle inhiition of HAO1 mrna expression in NHPs t.3 mg/kg, with mrna levels remining significntly suppressed (4% suppression) up to dy 29 fter single i.v. injection (Figure 6c). Furthermore, s we oserved in mice, the urinry glycolte levels were elevted in dose-dependent mnner in the nimls tht received HAO1 DsiRNA (Figure 6d). Similr to urine glycolte, plsm glycolte levels displyed n equl degree of response to HAO1 DsiRNA (Figure 6e), further suggesting tht glycolte is consistent iomrker for HAO1 trget inhiition. In further support of this hypothesis tht knockdown of the HAO1 trget cn 774 www.moleculrtherpy.org vol. 24 no. 4 pr. 216
DsiRNA Trgeting HAO1 Prevents Hyperoxluri Group 1 Group 2 Group 3 Group 4 Dy Dy 7 Dy 14 Wter Wter/ 1,2 9 Ethylene glycol EG/ Ethylene glycol Ethylene glycol EG// EG/HAO-1 Oxlte/cretinine (mg/g) 1,5 6 3 Wter/ EG/ EG// EG/ 2 7 14 21 Dys c 4 4 2 2 d e Wter/ EG/ EG// EG/ Wter/ EG/ EG// EG/ Figure 5 protects ginst EG-induced kidney dmge in PH1 mouse model. () Mle Agxt -/- mice were divided into four groups contining seven mice ech nd were supplied with either.7% EG in their drinking wter (groups 2 to 4) or regulr drinking wter s control (group 1). Animls of groups 1 nd 2 were injected with three doses of on dys, 7, nd 14. Animls of group 3 were injected with two doses of on dys nd 7 s well s single dose of on dy 14. Group 4 received three doses of on dys, 7, nd 14. Both nd were given i.v.. Urine smples were collected weekly nd nlyzed for oxlte levels with enzymtic ssy. All nimls were scrificed, nd the kidneys were collected nd evluted for COx crystls on dy 22. () Animls drinking.7 % of EG (EG/) nd treted with displyed continuous elevtion of urinry oxlte levels. Animls drinking EG nd injected with (EG/) exhiited urinry oxlte levels tht remined t seline concentrtion. Mice drinking EG nd given delyed tretment of (EG//) showed n initil elevtion of urine oxlte levels (milligrm per grm of cretinine) during the first 2 weeks tht returned ck to seline levels following tretment. Dshed lines indicte estimted rnge of oxlte levels of wild-type C57BL/6 mice tht re not littermtes nd, therefore, not fully mtched to the Agxt -/- mice in terms of genetic ckground. (c) Histologicl nlysis of kidney tissue ws performed to detect COx crystls using Pizzlto stining. Representtive imges of kidney sections show nimls of the EG/ group developed COx crystls tht exhiited chrcteristic irefringence under polrized light (upper pnel) nd positive stining using the Pizzlto s method (lower pnel). (d) Histologicl nlysis indictes tht is le to prevent the formtion of COx crystls nd protect the kidneys from further dmge in the Agxt -/- mice tht received EG (compre EG/ nd EG/ groups). Delyed tretment of lso effectively reduced COx crystls in the kidneys, though to lesser extent. Ech picture shows representtive section from n individul niml of ech group. Br = 15 μm. (e) Whole kidney sections were imged, scnned, nd reconstituted with imging softwre to further demonstrte tht prevents the formtion of COx crystls in treted nimls. e predicted using iomrkers, we oserved notle correltion etween HAO1 mrna knockdown nd plsm glycolte elevtion t stedy stte fter multiple doses of HAO1 DsiRNA injection (Figure 6f). Collectively, our dt demonstrte tht the HAO1 DsiRNA pproch produces potent nd durle responses nd tht these responses cn e relily monitored not only in mice ut lso in NHPs. DISCUSSION The emergence of RNAi therpeutics provides gret potentil to develop tretments tht were previously considered impossile. The min chllenges for RNAi therpies re the identifiction of efficcious gene trgets nd the successful delivery of RNAi molecules into ffected cells to induce trget gene knockdown. In this study, we hve otined dt suggest tht could e n effective trget for treting PH1, devstting renl disese. The led HAO1 DsiRNA ws efficiently delivered in lipid nnoprticle to heptocytes nd durly reduced trget gene expression in mice nd monkeys. In Agxt -/- mice, suppression of protein levels, in turn, reduced the conversion of glycolte to glyoxylte, leding to decrese in urinry oxlte levels, n expected corresponding elevtion of glycolte, nd elimintion of COx deposition in Agxt -/- nimls who were treted with EG, model which recpitultes key fetures of humn PH1. No helth effects were noted in either Ho1-deficient mice 28 or mice tht received multiple doses of HAO1 DsiRNA over the course of 15 dys. Tken together, the results support the trgeted knockdown of HAO1 y DsiRNA s potentilly sfe nd effective strtegy for the tretment of PH1. Potentil difference in glyoxylte nd oxlte metolism in rodents, NHPs, nd humns We hve demonstrted tht ws le to reduce urinry oxlte nd prevent renl dmge in PH1 mouse model Moleculr Therpy vol. 24 no. 4 pr. 216 775
DsiRNA Trgeting HAO1 Prevents Hyperoxluri % HAO1 mrna remining versus predose 1 1 75 5.1.3.1.3 Dy 7 Dy 14 1 3 d Glyolte/cretinine rtio (mg/g) 2 15 1 5 5 3. mg/kg 1. mg/kg.3 mg/kg 5 Urine glycolte 1 15 2 Dose (mg/kg) Dys (mg/kg).1.3.1.3 1. 3. Dy 4 Dy 7 Dy 14 e f Clycolte conc. (μmol/l) 8 6 4 2 5 1 Plsm glycolte 3. mg/kg 1. mg/kg.3 mg/kg 5 Individul niml 1 Dys 15 2 c % HAO1 mrna remining 1 75 5 1 2 3 4 5 6 7 8 9 1 11 12 5 1 15 2 3 % HAO1 mrna knockdown reltive to predose 8 6 R 2 =.72 4.3 mg/kg Q3W 1; Q3W 4 2.3 mg/kg Q1W 3; Q3W 4 1 15 2 3 Dys % glycolte levels reltive to predose Figure 6 HAO1 DsiRNA displys specific nd potent ctivity in NHPs. A selected HAO1 DsiRNA formulted in LNP, designted DCR-PH1, ws i.v. injected into cynomolgus monkeys with vrying doses s indicted. Liver iopsies were collected for nlysis of mrna nd protein t vrying time points s indicted. Urine nd plsm smples were lso collected t vrying time points s indicted nd then nlyzed with LC/MS for glycolte nd cretinine levels. () Results of rel-time reverse trnscription PCR indicted tht HAO1 DsiRNA inhiits HAO1 mrna expression in dose-dependent mnner. Expression of ech individul niml ws plotted. Ech group contins four nimls. () The results of western lot nlysis indicted tht HAO1 DsiRNA inhiits protein expression on dys 7 nd 14. Expression in ech individul niml ws nlyzed. Note tht only three iopsies were recovered from the 3 mg/kg dose group t dy 7, nd lne 5 ws empty with no smple. (c) The results of rel-time reverse trnscription PCR indicte tht HAO1 DsiRNA, when injected t.3 mg/kg durly inhiits HAO1 mrna expression (men ± SD). (d) LC/MS nlysis shows tht urinry glycolte levels (normlized with cretinine) incresed in HAO1 DsiRNA-treted nimls, in dose-dependent mnner (men ± SD). (e) Similrly, plsm glycolte ws elevted with comprle kinetics nd dose dependency in treted nimls. (f) Animls receiving HAO1 DsiRNA t.3 mg/kg once every 3 weeks were nlyzed for their HAO1 mrna expression nd plsm glycolte levels t stedy stte. There ws notle correltion (R 2 =.72) etween the knockdown of HAO1 mrna nd elevtion of plsm glycolte in individul nimls. This indictes tht plsm glycolte could e explored s potentil phrmcodynmic mrker for trget knockdown. LNP, lipid nnoprticle. in which mice were treted with EG. While this result suggests tht potentil reduction of urine oxlte nd disese progression could e chieved in humns, there re differences in the glyoxylte metolism pthwys of mice nd humns tht mke it chllenging to fully predict the effect of HAO1 inhiition y DsiRNA in humns (Figure 1). For exmple, D-mino cid oxidse, which converts glycine ck to glyoxylte, is expressed in humn ut not in mouse liver (Biogps.org). Thus, it is possile tht oxlte production in mice is more sensitive thn tht in PH1 ptients to the effects of protein reduction. Additionlly, mice, ut not humns nd other species, re cple of de novo synthesis of scoric cid, which cn e nonenzymticlly converted into oxlte. 3 These species-specific differences in oxlte metolism indicte the limittions for trnslting mouse models to humn responses nd prompted us to verify the HAO1 DsiRNA pproch in NHPs. Our results confirm tht HAO1 DsiRNA efficiently reduces HAO1 mrna nd protein levels in NHP s well s induces n increse in glycolte, further iomrker of ctivity. However, urine oxlte levels in norml NHP re low to the point tht ny potentil further reduction cnnot e ccurtely detected with our current methodology. Other potentil therpeutics for PH1 Smll-molecule inhiitors of hve lso een explored for treting PH1. (L)-oxothiozlidine-4-croxylte (OTZ) ws effective in hyperoxluric rt model 24 ut showed very limited effect on urine oxlte in PH1 ptients.,26 Another inhiitor of, novel 4-sustituted-3-hydroxy-3-pyrroline-2,5-dione, hs een recently 776 www.moleculrtherpy.org vol. 24 no. 4 pr. 216
DsiRNA Trgeting HAO1 Prevents Hyperoxluri under investigtion for clinicl utility in ptients. 27 Other potentil trgets for PH1 include inhiition of hydroxyproline dehydrogense (HYPDH or PRODH2) (Figure 1). HYPDH controls the first step of hydroxyproline ctolism, which is hypothesized to e n importnt source of glyoxylte nd oxlte. 31,32 Potentil therpeutics for PH2 nd PH3 There re currently no effective tretments for either PH2 or PH3. In PH2 nd PH3 ptients, muttions in the genes encoding glyoxylte reductse (GRHPR) nd 4-hydroxy-2-oxoglutrte ldolse (HOGA), respectively, underlie the disese, cusing the overproduction of glyoxylte nd susequent oxlte elevtion (Figure 1). Blocking protein production reduces the contriution y glycolte to the pool of glyoxylte nd thus should reduce the glyoxylte pool tht is ville for oxlte production. Therefore, inhiition could lso hve potentil s therpeutic strtegy for PH2 nd PH3. However, the reltive contriutions of the vrious pthwys to glyoxylte production in helth nd disese re not yet fully chrcterized nd elucidting them further will e of gret interest. Exploring the effect of HAO1 inhiition on oxlte metolism in mouse models of PH2 nd PH3 should provide insights to help identify new therpeutic strtegies for these PH types. Oxlte concentrtions nd excretion rte to predict kidney disese The Agxt -/- mice, despite their high urinry oxlte concentrtions, show very low proility of developing kidney crystls. However, mle Agxt -/- mice provided with excess glycolte from EG to further increse urinry oxlte concentrtions developed kidney COx deposits with high frequency. 23 Our results suggest tht urinry oxlte levels need to rech certin threshold level nd durtion for kidney deposits to occur. This is consistent with reports tht, in PH ptients, oxlte excretion rtes t dignosis cn predict the proility of progression to ESRD. 33 In study monitoring PH ptients over period of 3 yers, those who hd urinry oxlte excretion rtes lower thn 1.3 mmol/1.73 m 2 /24 hours hd much lower chnce of developing ESRD thn ptients with urinry oxlte excretion rtes higher thn 1.3 mmol/1.73 m 2 /24 hours. The dt lso suggested tht mintining the oxlte concentrtion under threshold level is sufficient to prevent ESRD. Consistent with this hypothesis, we demonstrted tht is le to prevent kidney dmge in hyperoxluric mice even though their oxlte levels hve not een reduced to the level of wild-type nimls (Figure 5 e). From this, we conclude tht HAO1 DsiRNA tretment my e le to chieve reduction of oxlte excretion rte tht flls elow the threshold point, thus preventing kidney crystl formtion. Reltionship of Ho1 mrna nd protein knockdown to oxlte reduction We oserved tht the durtion of oxlte reduction is shorter thn tht of Ho1 mrna nd protein reduction (Figures 3 nd 4). One potentil explntion is tht protein expression must e reduced to elow certin level in order to effectively prevent genertion of glyoxylte from glycolte nd to chieve oxlte reduction. Given the differences etween mice nd humns in the glyoxylte metolism pthwy, the threshold level of reduction in humns required to chieve chnges in urinry oxlte concentrtions my e different thn tht in mice. Glycolte s potentil noninvsive, direct, nd predictive iomrker Ho1 -/- mice show elevted levels of urinry glycolte without renl dmge or other phenotypic consequences. 28 In ddition, glycolte oxidse deficiency in humns hs een shown to cuse symptomtic glycolic ciduri. 34 These findings re consistent with the concept tht glycolte is highly solule nd it cn e excreted y the kidneys without pthologicl consequences. We oserved tht urinry glycolte levels correlte inversely with oxlte chnges in -treted Agxt -/- mice. This suggests tht glycolte could potentilly e used s phrmcodynmic mrker to evlute HAO1 mrna knockdown using esy, noninvsive procedures to monitor urinry nd plsm glycolte levels. In summry, our preclinicl studies demonstrte tht HAO1 DsiRNA is promising pproch s potentil PH1 therpeutic. MATERIALS AND METHODS Selection of led HAO1 DsiRNA. ws identified through process of lrge-scle screening of sirnas for mrna knockdown ctivity in vitro (Supplementry Figure S2). RNA strnds for sirna duplexes were synthesized purified t Integrted DNA Technologies (Corlville, IA). The sir- NAs in the primry screen hd /27-mer duplex RNA structure, with -nucleotide sense strnd composed of ll RNA except for two 3 -terminl DNA nucleotides, nneled to 27-nucleotide ll RNA ntisense strnd. The sirnas in the primry screen were prtilly 2 -OMe modified. RNA preprtion nd rel-time PCR. Smple tissues were homogenized in gunidinium thiocynte-phenol using TissueLyser II (Qigen, Vlenci, CA). RNA ws then purified ccording to mnufcturer instructions (Promeg, Mdison, WI). Trnscriptor first strnd cdna kit (Roche) ws used to prepre cdna nd mouse specific Ho1 nd Hprt1 (Integrted DNA Technologies) primers were used for PCR on CFX96 cycler-rel Time PCR Detection System (Bio-Rd Lortories, Hercules, CA). In situ hyridiztion using ViewRNA. Mnufcturer s protocol (Affymetrix, Snt Clr, CA) ws followed with the exceptions noted elow. Briefly, formlin-fixed/prffin-emedded tissue sections were cut to 5 μm, ked t 6 C for 6 minutes. Liver tissue were pretreted for 2 minutes t 95 C nd digested with protese for 15 2 minutes t 4 C. Signl intensity nd re of signl were mesured using Nikon Elements softwre. Western nlysis of tissue smples. Tissue lystes were prepred using Qigen Tissue Lyser with lysis uffer (2 mmol/l Tris ph 8., 15 mmol/l NCl, 1% NP-4) nd protese inhiitor cocktil (Sigm-Aldrich, St Louis, MO). Tissue lystes were resolved y 4 15% Tris-glycine SDS PAGE (Bio- Rd Lortories). The lots were locked nd incuted with rit nti- primry followed y nti-rit-ir8 (Li-Cor Biosciences, Lincoln, NE) nd with mouse nti- or tuulin (Sigm-Aldrich) followed y nti-mouse-ir6 (Li-Cor Biosciences). The signl intensity ws determined y Odyssey Infrred Imging System (Li-Cor Biosciences). In vivo studies nd smple collection in mice. Mice were treted in ccordnce with institutionl ethicl guidelines of niml cre, hndling, nd termintion; ll studies were pproved y Dicern s Institutionl Review Bord. Mice were kept in pthogen-free fcility, with free ccess to stndrd chow nd wter. C57BL6 mice (Hrln Lortories, Indinpolis, IN) were used to screen the led. Agxt -/- nimls were used to evlute efficcy of on urine oxlte nd.7% EG chllenge experiments. All dosing with or ws injected i.v.. Urine smples were collected mnully unless stted otherwise. The urine smples were cidified efore storing. Moleculr Therpy vol. 24 no. 4 pr. 216 777
DsiRNA Trgeting HAO1 Prevents Hyperoxluri Histologicl nlysis. Tissue were fixed in 1% neutrl uffered formlin for 24 hour nd then trnsferred into 7% ethnol nd emedded in prffin for sectioning. COx stining ws performed on kidney sections s descried. 35 The imges were tken using Olympus BX51 microscope nd processed using Imge Pro Premier 9.1 imge nlysis softwre. Enzymtic oxlte nlysis nd cretinine ssy. Urine oxlte ws mesured using clinicl oxlte oxidse ssy (Trinity Biotech, Wicklow, Irelnd) modified for use with murine smples. Urine cretinine ws mesured using stndrd cretinine detection kit (Enzo Life Sciences AG, Lusen, Switzerlnd). Oxlte, glycolte, nd cretinine ssys y LC/MS. Oxlte, glycolte, nd cretinine were determined y methods sed on liquid chromtogrphy comined with mss spectrometry detection (LC/MS) on triple- qudrupole instrument (API 55, AB Sciex, Frminghm, MA) y electrospry ioniztion nd multiple rection monitoring. Mouse nd monkey urine smples were nlyzed for oxlte, glycolte, nd cretinine fter smple dilution y hydrophilic interction liquid chromtogrphy (HILIC) nd MS detection in the negtive electrospry ioniztion mode. Glycolte ws determined in monkey plsm smples y protein precipittion, followed y chemicl derivtiztion nd nlysis y reversed-phse LC nd MS detection in the positive electrospry ioniztion mode. Stle-lel internl stndrds ( 13 C 2 - oxlte, 13 C 2 -glycolte, nd 2 H 3 -cretinine) were utilized for quntifiction. Clirtion curves were mde in surrogte mtrix. Qulity control smples were nlyzed in triplictes t minimum of three concentrtion levels in ech nlyticl tch. The oserved inccurcy (%Bis) nd precision (%CV) for qulity control smples were less thn 15% for ll nlytes. NHP study. Tretment of the nimls ws conducted y certified contrct reserch orgniztion in ccordnce with the testing fcility s stndrd operting procedure, which dheres to the regultions outlined in the United Sttes Deprtment of Agriculture Animl Welfre Act (9 CFR, Prts 1 3) nd the conditions specified in the Guide for the Cre nd Use of Lortory Animls (ILAR puliction, 1996, Ntionl Acdemy Press). Animls were given single i.v. infusion over 2 minutes of lipid nnoprticle-formulted DsiRNAs. Tissue iopsies, lood nd urine smples were collected for phrmcodynmics nd iomrker nlysis. SUPPLEMENTARY MATERIAL Figure S1. Results of HAO1 DsiRNA screen to identify the led HAO1 DsiRNA. Figure S2. reduces urinry oxlte nd elevtes urinry glycolte levels t similr rtes in mle Agxt-/- mice. Figure S3. reduces urine oxlte levels in the Agxt-/- PH1 mouse model. Figure S4. Body weight chnge nd wter consumption in the mle Agxt-/- mice supplied with drinking wter contining EG. Figure S5. chieves cumultive trget knockdown in mice. Figure S6. long-term tretment exhiits no ovious dverse effect in mice. ACKNOWLEDGMENTS We thnk Doug Fmrough, Pnkj Bhrgv, Ted Ashurn, Dvid Miller, Jim Weissmn, nd Jessic Gierut for their review of this mnuscript. We thnk Heidi Denmn nd Cristin Pz for technicl support nd help with niml cre. C.D., N.A.-C., N.P., M.L.C., B.H., R.D., W.Z., L.A., M.K., B.Y., D.C., X.S., U.S., W.A.C., A.S., N.N., W.W., M.A., H.D., B.D.B., C.L. re employees of Dicern Phrmceuticls, which is developing DsiRNAs s therpeutics. REFERENCES 1. Hoppe, B, Beck, BB nd Milliner, DS (29). The primry hyperoxluris. Kidney Int 75: 1264 1271. 2. Dnpure, CJ (24). Moleculr etiology of primry hyperoxluri type 1. Nephron Exp Nephrol 98: e39 e44. 3. Hrmt, J, Frgue, S, Bcchett, J, Acquviv, C nd Cocht, P (211). Primry hyperoxluri. Int J Nephrol 211: 86458. 4. Cocht, P, Frgue, S, Bcchett, J, Bertholet-Thoms, A, Sot, JF nd Hrmt, J (211). [Primry hyperoxluri]. Nephrol Ther 7: 249 9. 5. Leumnn, E nd Hoppe, B (21). The primry hyperoxluris. J Am Soc Nephrol 12: 1986 1993. 6. Cocht, P, Liutkus, A, Frgue, S, Bsmison, O, Rnchin, B nd Rollnd, MO (26). Primry hyperoxluri type 1: still chllenging! Peditr Nephrol 21: 175 181. 7. Cocht, P nd Rumsy, G (213). Primry hyperoxluri. N Engl J Med 369: 649 658. 8. Dnpure, CJ nd Jennings, PR (1986). Peroxisoml lnine:glyoxylte minotrnsferse deficiency in primry hyperoxluri type I. FEBS Lett 21: 2 24. 9. Dnpure, CJ nd Rumsy, G (24). Moleculr etiology of primry hyperoxluri nd its implictions for clinicl mngement. Expert Rev Mol Med 6: 1 16. 1. Cocht, P, Delorine, A, Rotily, M, Olive, F, Liponski, I nd Deries, N (1995). Epidemiology of primry hyperoxluri type 1. Société de Néphrologie nd the Société de Néphrologie Péditrique. Nephrol Dil Trnsplnt 1 (suppl. 8): 3 7. 11. Hoppe, B (212). An updte on primry hyperoxluri. Nt Rev Nephrol 8: 467 475. 12. Lorenzo, V, Alvrez, A, Torres, A, Torregros, V, Hernández, D nd Slido, E (26). Presenttion nd role of trnsplnttion in dult ptients with type 1 primry hyperoxluri nd the I244T AGXT muttion: single-center experience. Kidney Int 7: 1115 1119. 13. Hoyer-Kuhn, H, Kohrok, S, Vollnd, R, Frnklin, J, Hero, B, Beck, BB et l. (214). Vitmin B6 in primry hyperoxluri I: first prospective tril fter 4 yers of prctice. Clin J Am Soc Nephrol 9: 468 477. 14. Ymuchi, T, Quillrd, M, Tkhshi, S nd Nguyen-Kho, M (21). Oxlte removl y dily dilysis in ptient with primry hyperoxluri type 1. Nephrol Dil Trnsplnt 16: 247 2411. 15. Milliner, D (26). Tretment of the primry hyperoxluris: new chpter. Kidney Int 7: 1198 12. 16. Miyt, N, Steffen, J, Johnson, ME, Frgue, S, Dnpure, CJ nd Koehler, CM (214). Phrmcologic rescue of n enzyme-trfficking defect in primry hyperoxluri 1. Proc Ntl Acd Sci USA 111: 1446 14411. 17. Htch, M nd Freel, RW (28). The roles nd mechnisms of intestinl oxlte trnsport in oxlte homeostsis. Semin Nephrol 28: 143 151. 18. Htch, M, Gjymishk, A, Slido, EC, Allison, MJ nd Freel, RW (211). Enteric oxlte elimintion is induced nd oxlte is normlized in mouse model of primry hyperoxluri following intestinl coloniztion with oxlocter. Am J Physiol Gstrointest Liver Physiol 3: G461 G469. 19. Grujic, D, Slido, EC, Shenoy, BC, Lngmn, CB, McGrth, ME, Ptel, RJ et l. (29). Hyperoxluri is reduced nd nephroclcinosis prevented with n oxlte-degrding enzyme in mice with hyperoxluri. Am J Nephrol 29: 86 93. 2. Hoppe, B, Beck, B, Gtter, N, von Unruh, G, Tischer, A, Hesse, A et l. (26). Oxlocter formigenes: potentil tool for the tretment of primry hyperoxluri type 1. Kidney Int 7: 135 1311. 21. Hoppe, B, Groothoff, JW, Hulton, SA, Cocht, P, Niudet, P, Kemper, MJ et l. (211). Efficcy nd sfety of Oxlocter formigenes to reduce urinry oxlte in primry hyperoxluri. Nephrol Dil Trnsplnt 26: 369 3615. 22. Ylä-Herttul, S (211). Gene therpy moves forwrd in 21. Mol Ther 19: 219 22. 23. Slido, EC, Li, XM, Lu, Y, Wng, X, Sntn, A, Roy-Chowdhury, N et l. (26). Alnine-glyoxylte minotrnsferse-deficient mice, model for primry hyperoxluri tht responds to denovirl gene trnsfer. Proc Ntl Acd Sci USA 13: 18249 184. 24. Bker, PW, Rofe, AM nd Bis, R (1998). The effect of (L)-cysteine nd (L)-2- oxothizolidine-4-croxylic cid (OTZ) on urinry oxlte excretion: studies using hyperoxluric rt model. J Urol 159: 2177 2181.. Holmes, RP (1998). Phrmcologicl pproches in the tretment of primry hyperoxluri. J Nephrol 11 (suppl. 1): 32 35. 26. Holmes, RP, Assimos, DG, Wilson, DM nd Milliner, DS (21). (L)-2-oxothizolidine- 4-croxylte in the tretment of primry hyperoxluri type 1. BJU Int 88: 858 862. 27. Crgoe, EJ, Rooney, CS, nd Willims, HWR (1985). Tretment of kidney nd ldder stones. Google Ptents. 28. Mrtin-Higuers, C, Luis-Lim, S nd Slido, E (215). Glycolte oxidse is sfe nd efficient trget for sustrte reduction therpy in mouse model of Primry Hyperoxluri Type I. Mol Ther (epu hed of print). 29. Dudek, H, Wong, DH, Arvn, R, Shh, A, Worthm, K, Ying, B et l. (214). Knockdown of β-ctenin with dicer-sustrte sirnas reduces liver tumor urden in vivo. Mol Ther 22: 92 11. 3. Prk, CH nd Boo, YC (26). Depletion method of lood plsm scorte. Google Ptents. 31. Knight, J, Jing, J, Assimos, DG nd Holmes, RP (26). Hydroxyproline ingestion nd urinry oxlte nd glycolte excretion. Kidney Int 7: 1929 1934. 32. Knight, J, Holmes, RP, Crmer, SD, Tkym, T nd Slido, E (212). Hydroxyproline metolism in mouse models of primry hyperoxluri. Am J Physiol Renl Physiol 32: F688 F693. 33. Zho, F, Bergstrlh, E, Meht, R, Seide, BM, Cogl, AG, Lieske, JC, et l. (214). Anlysis of urinry risk fctors for ESRD mong ptients with primry hyperoxluri. Astrct # TH-P31. : 18. 34. Frisherg, Y, Zehri, A, Lykhovetsky, R, Brgl, R nd Belostotsky, R (214). Muttions in HAO1 encoding glycolte oxidse cuse isolted glycolic ciduri. J Med Genet 51: 526 529. 35. Ysue, T (1969) Histochemicl identifiction of clcium oxlte. Act Histochem Cytochem 2: 83 95. This work is licensed under Cretive Commons Attriution-NonCommercil-NoDerivs 4. Interntionl License. The imges or other third prty mteril in this rticle re included in the rticle s Cretive Commons license, unless indicted otherwise in the credit line; if the mteril is not included under the Cretive Commons license, users will need to otin permission from the license holder to reproduce the mteril. To view copy of this license, visit http://cretivecommons.org/ licenses/y-nc-nd/4./ 778 www.moleculrtherpy.org vol. 24 no. 4 pr. 216