Abstract. Introduction. RBMOnline - Vol 18. No Reproductive BioMedicine Online; on web 20 March 2009

Transcription

1 RBMOnline - Vol 18. No Reproductive BioMedicine Online; on web 20 March 2009 Article Anti-Müllerian hormone in pregnant women in relation to other hormones, fetal sex and in circulation of second trimester fetuses Melissa Lutterodt studied medicine at three different Universities in Denmark and Sweden. During her medical training, she worked as scientific assistant at the Occupational Health department, University of Århus, Denmark. She graduated as MD from the University of Copenhagen in 2004 and is currently attending a PhD programme at the Laboratory of Reproductive Biology, Rigshospitalet, University of Copenhagen, Denmark. Her PhD thesis concerns the possible effects of cigarette smoke exposure in utero on first trimester fetal gonads. In 2007 she became responsible for the collaboration of 10 different scientific groups in relation to her PhD. Dr Melissa Lutterodt Melissa Lutterodt 1, Anne Grete Byskov 1, Sven Oluf Skouby 2, Ann Tabor 3 Claus Yding Andersen1,4, 1 Laboratory of Reproductive Biology, The Juliane Marie Centre for Women, Children and Reproduction, University Hospital of Copenhagen; 2 Department of Obstetrics and Gynaecology, Herlev Hospital and Frederiksberg Hospital, Faculty of Health Sciences, Copenhagen University; 3 Department of Fetal Medicine and Ultrasound, Copenhagen University Hospital, Copenhagen, Denmark 4 Correspondence: yding@rh.dk Abstract The aim of this study was to investigate the regulation of anti-müllerian hormone (AMH) blood concentrations in mother and fetus during pregnancy. Serum concentrations of AMH, gonadotrophins, oestradiol and progesterone were measured in pregnant women in the first trimester and AMH concentrations in second-trimester fetuses, and these were compared in relation to the sex of the fetus. A total of 153 women undergoing elective termination of a first-trimester pregnancy and seven second-trimester pregnant women undergoing cordocentesis were included. Concentrations of AMH in the serum of first-trimester pregnant women were similar to non-pregnant women and were unrelated to the very high concentrations of human chorionic gonadotrophin and the undetectable concentrations of FSH and LH. Serum concentrations of oestradiol and progesterone were unrelated to the concentrations of AMH and the sex of the fetus. Serum concentrations of AMH of four, second trimester, male fetuses ranged from 64 to 92 ng/ml, whereas it was undetectable in female fetuses. It appears that AMH serum concentrations in first-trimester pregnant women seem to be independent of gonadotrophin concentrations and fetal sex. The concentration of AMH in the circulation of male fetuses is higher than previously reported and is a highly sensitive marker for fetal sex. Keywords: anti-müllerian hormone, fetal circulation, pregnancy, steroids Introduction 694 Anti-Müllerian hormone (AMH) is known to be one of the two classic hormones determining sex differentiation in the developing fetus (Jost, 1947; Lee and Donahoe, 1993; Josso et al., 1993). Immediately after the testis becomes morphologically recognizable (at around day 40 after conception in humans) it starts to secrete AMH, which, in concert with testosterone, ensures male development of the secondary sex characteristics. In contrast, the female fetus does not secrete these two hormones and develops femaleness according to the classical paradigm. However, low circulating concentrations of AMH are present in girls during childhood until puberty, when the concentration rises and reaches concentrations similar to those of boys of comparable age (Lee et al., 1996). AMH expression is localized to granulosa cells of all follicle types with the exception of primordial and large antral/pre-ovulatory follicles, where AMH expression appears to be absent (Visser et al., 2006). Receptors for AMH have been demonstrated in the rat granulosa and theca cells by in situ hybridization (Baarends et al., 1995). The exact 2009 Published by Reproductive Healthcare Ltd, Duck End Farm, Dry Drayton, Cambridge CB23 8DB, UK

2 physiological mechanism of action of AMH in the ovary is as yet undetermined. Several studies suggest that AMH interacts with gonadotrophins and it has been shown that AMH decreases granulosa cell sensitivity to FSH stimulation and thereby limits the number of follicles that develop to the pre-ovulatory stage (Visser et al., 2006; Pellatt et al., 2007). However, FSH does not seem to affect AMH secretion directly as shown in cultures of human granulosa cells from small antral follicles (Pellatt et al., 2007) and by the inability of FSH administration to acutely affect the AMH concentration in women (Wachs et al., 2007). Whether absence of FSH activity, affects AMH concentrations is still not clear, although a recent study found that follicular fluid from prepubertal girls, aged years, contained AMH concentrations similar to those of adult women (Yding Andersen and Byskov, 2006; Yding Andersen et al., 2008). Based on a strong inverse correlation between the concentration of AMH and oestradiol in fluid from small human antral follicles, these studies suggested that AMH decrease FSH-dependant expression of the aromatase enzyme and possibly the LH receptor (Yding Andersen and Byskov 2006; Yding Andersen et al., 2008) confirming studies in cultures of human granulosa lutein cells (Grossman et al., 2008) and in fetal ovaries from several animal species (Vigier et al., 1989). AMH concentrations have also been found to correlate positively with serum concentrations of both androstenedione and testosterone in normal women and women with polycystic ovaries (PCO; Cook et al., 2002; La Marca et al., 2004; Pigny et al., 2006) and it has recently been suggested that androgens cause the increased AMH production in women with PCO (Pellatt et al., 2007). This could potentially indicate a role for LH or an LH-induced signal in the regulation of AMH expression. The aim of the present study was to investigate the regulation of AMH secretion by combining the effect of AMH expression from the fetus and from the mother by studying first-trimester pregnant women in comparison with nonpregnant controls. The circulating hormone concentrations in pregnant women, especially during the first trimester of pregnancy, differ profoundly from those found in the natural menstrual cycle, providing an opportunity to study the interaction of AMH secretion with gonadotrophins, sex steroids and the sex of the fetus. Additionally, in the present study, the concentration of AMH in the circulation of secondtrimester fetuses was investigated. Materials and methods Patients Legislation in Denmark allows women to have free therapeutic termination of pregnancy during the first trimester (i.e. prior to fetal week 11/ gestational week 13). If a woman requests the procedure it is performed in a public hospital free of charge. In the present study, 153 women pregnant in the first trimester, who requested termination of their pregnancy [aged 28.0 ± 0.5 years (mean ± SEM), range years, median 27.5 years] were included. Women with a chronic disease (e.g. diabetes or asthma) or those taking permanent medication during their pregnancy were excluded. Oral and written information was given about collection of a blood sample prior to the procedure, and collection of fetal material to be used for determination of the sex of the fetus. Informed consent was obtained from all participating women. The project was approved by The Regional Committee on Biomedical Research Ethics Copenhagen and Frederiksberg Counties (J. no. KF (01) ). As a control group for the pregnant women, 33 non-pregnant normal menstruating patients (aged 27.8 ± 0.3 years, range years) had their serum analysed for concentrations of AMH. All had both ovaries present and were without malignancies related to the ovaries. The serum sample was collected at a random time during their menstrual cycle and stored at 20 C until analysis. Hormone measurements AMH was measured using a specific enzyme-linked immunosorbent assay (ELISA) kit according to the manufacturer s instructions (DSL ; Diagnostic System Laboratories, USA). Maternal serum samples were tested undiluted, whereas fetal serum samples were initially tested undiluted and subsequently at a dilution of 1:10 for those having an absorbance exceeding the standard curve. Inter-assay variation of a sample containing 7.6 ng/ml AMH was 4.4% (n = 12) and intra-assay variation was 3.3% (n = 5) for a sample containing 0.45 ng/ml. Serum -human chorionic gonadotrophin (ßHCG), FSH, LH, oestradiol and progesterone concentrations were measured, using the KRYPTOR analyser a random access immunofluorescence analyser (Brahms GmbH, Germany) using time-resolved amplified cryptate emission (TRACE) technology and Brahms automated immunofluorescence assays (Brahms GmbH, Germany). For ßHCG there was no high dose hook effect up to 1,000,000 IU/l; inter- and intra-assay variations for a sample containing 846 IU/l were 1.9% and <0.9% (n = 30), respectively. FSH results showed a detection limit of 0.4 miu/ ml, inter-assay variation of 6.5% with a sample containing 3.3 miu/l and intra-assay variation of 5.4% of sample containing 3.4 miu/l (n = 35). For LH the detection limit was 0.3 miu/ ml, inter- and intra-assay variations of a sample containing 3.0 miu/l (n = 22) were 7.0% and 5.1%, respectively. For oestradiol the inter-assay variation was 6.9% with a sample containing 194 pmol/l (n = 51) and intra-assay variation was 8.0% with a sample containing 94 pmol/l (n = 30). For progesterone the inter-assay variation was 7.1 with a sample containing 91.0 nmol/l (n = 12) and intra-assay variation was 1.4% with a sample containing 40 nmol/l (n = 30). Gender determination A small piece of fetal tissue was snap frozen in liquid nitrogen for later determination of the chromosomal sex. Sex determination was performed using polymerase chain reaction with X-Y homologous primers according to Nakahori et al. (1991). Fetal blood samples from secondtrimester pregnancies In order to confirm the result of a prenatal genetic diagnosis from amniotic fluid cells or in conjunction with an intrauterine transfusion, a total of seven individual ultrasound-guided 695

3 696 umbilical cord blood samples (i.e. cordocentesis) were performed in fetal weeks (gestational weeks 17 30), independent of the present study. The diagnoses included spina bifida, Turner mosaicism and rhesus immunization. The cordocentesis sample comprised a total volume of a few hundred microlitres, which after isolation of the serum was stored at 80 C until analysis. Immunohistochemistry for detection of AMH expression From each recovered fetus in which one or two gonads were localized, a small piece of gonadal tissue was fixed in Bouin s fixative for 2 h, dehydrated and processed for histological sectioning (5 μm thick sections). Immunohistochemistry for detection of AMH was performed accordingly to the method of Durlinger et al. (2002) with some modifications. Sections were deparaffinized, treated with 0.3% hydrogen peroxide for 20 min to block endogenous peroxidase activity, washed and boiled in 0.01 M citrate buffer for 3 5 min for damasking AMH. After blocking with 5% (v/v) rabbit serum (DAKO A/S, Glostrup, Denmark) in 5% (w/v) bovine serum albumin (BSA; Sigma- Aldrich, Denmark) in phosphate-buffered saline (PBS) for 15 min, sections were incubated at 4 C for 20 h with goat serum polyclonal antibody raised against AMH (C-20, sc-6886, Santa Cruz, AH Diagnostics, Aarhus, Denmark) diluted 1:2000 in 5% (w/v) BSA in PBS for 20 h. According to the manufacturer, the primary antibody was affinity purified and raised against a peptide mapping at the carboxy terminus of AMH of human origin. After the wash in PBS, sections were incubated with secondary antibody (i.e. biotinylated rabbit-anti-goat antibody (DAKO A/S) diluted 1:400 in 5% (w/v) BSA in PBS at room temperature for 30 min. After washing with PBS three times, sections were incubated with streptavidin-biotin-peroxidase complex ABComplex (DAKO A/S) at room temperature for 30 min, washed and developed using 0.07% 3,3 -diaminobenzidine tetrahydrochloride (DAB, DAKO) for 5 min, dehydrated and mounted with Eukitt mounting medium (code no G; WvR-Bie & Berntsen, Copenhagen, Denmark). The negative control used the blocking peptide (sc-6886p, Santa Cruz, AH Diagnostics, Denmark). Statistics Comparison of hormone concentrations between subgroups (n > 2) was analysed with the analysis of variance (ANOVA) test. Comparison of two independent groups was performed using Student s t-test. Results From the 153 women who had a blood sample taken prior to termination of pregnancy, a total of 101 had fetal tissue collected in which it was possible to determine the sex of the fetus. The overall mean concentrations of the hormones measured in the serum sample taken before the procedure are given in Table 1. A characteristic picture of first trimester serum concentrations was found with extremely high concentrations of HCG while concentrations of FSH and LH were below the detection limit (respectively <0.4 IU/l and 0.3 IU/l) except for a few cases (FSH, two samples of values: 3.2 and 0.42 IU/l, and LH, six samples of values: 0.39, 0.55, 0.43, 0.52, 0.37 and 0.52 IU/l). The mean AMH concentration was 2.3 ng/ml in an environment in which the concentration of progesterone and oestradiol was increased compared with non-pregnant women (oestradiol normal range pmol/l; progesterone normal range 3 60 nmol/l). In the group of non-pregnant women, concentrations of AMH averaged 2.4 ± 0.33 ng/ml(mean ± SEM), the range was ng/ml. Immunohistochemical staining of AMH and morphology of the fetal gonadal tissues confirmed that gonads of male fetuses exhibited strong AMH staining by the Sertoli cells whereas the ovaries of the female fetuses were unstained (Figure 1). A possible relationship between AMH and the other hormones measured was evaluated (Table 2). The measured concentrations of AMH were unrelated to the corresponding HCG, progesterone and oestradiol concentrations. The presence of an association between fetal age and measured hormones was evaluated by comparing hormone concentrations in four fetal age groups (Table 3). Whereas the concentration of AMH was not related to fetal age, concentrations of HCG, progesterone and oestradiol each showed a highly significant association with fetal age (P < , P < and P < respectively). The possible production of AMH in male fetuses was not reflected in the maternal circulation since concentrations of AMH were similar in mothers carrying either a male or a female fetus (Table 4). Concentrations of HCG, progesterone and oestradiol remained similar in these two groups. Concentrations of AMH in serum from seven second-trimester fetuses, are given in Table 5. Whereas AMH was undetectable in the three samples from female fetuses, in the four samples from male fetuses an increase exceeding one order of magnitude was found in the fetal circulation compared with the average of that seen in the mothers. Discussion The present study demonstrated that concentrations of AMH in serum from first-trimester pregnant women were similar to those of non-pregnant women and were independent of low concentrations of FSH/LH and high concentrations of HCG. FSH was undetectable in nearly all serum samples, showing that a lack of FSH stimulation does not affect AMH secretion. Thus, the present results corroborate other studies indicating that FSH does not have an effect on AMH expression (La Marca et al., 2005; Wachs et al., 2007; Yding Andersen et al, 2008). Further, the very high concentrations of HCG, which may be expected to create an androgenic milieu in the antral follicles present in the ovaries of the pregnant women, do not have any observable effect on AMH secretion. The high expression of AMH in the fetal testis and its absence in fetal ovaries was reflected in the fetal circulation in which AMH was below the detection limit in females and several orders of magnitude higher in males, confirming earlier results (Josso et al., 1993). The AMH found in the maternal circulation is likely to result from the pool of small ovarian antral follicles, irrespective of whether the woman is pregnant or not (La Marca et al.,

4 Table 1. Serum concentrations of anti-müllerian hormone (AMH), gonadotrophins, progesterone and oestradiol from first trimester pregnant women and serum concentrations of AMH from adult non-pregnant women. Pregnant women (n = 153) Non-pregnant women (n = 33) AMH (ng/ml) ( ) ( ) HCG (IU/l) ( ) FSH (IU/l) <0.4 LH (IU/l) <0.3 Progesterone (nmol/l) 93 3 (19 194) Oestradiol (pmol/l) ( ) Values are mean SEM (range). HCG = human chorionic gonadotrophin. Figure. 1 (A) Five-μM section of an 8-week-old human embryonic ovary stained for anti-müllerian hormone. A faint background staining of oogonia is seen, otherwise no staining is present in the ovary. Scale bar = 50 μm. (B) Five-μM section of an 8-week-old human embryonic testis stained for anti- Müllerian hormone. The Sertoli cells of the newly differentiated testis are heavily stained but otherwise staining is absent in the tissue. Scale bar = 50 μm. Table 2. Serum concentrations of anti-müllerian hormone (AMH) from first trimester pregnant women in relation to human chorionic gonadotrophin (HCG), progesterone and oestradiol. AMH concentration <1 (n = 28) 1 2 (n = 48) >2 3 (n = 41) >3 (n = 36) (ng/ml) HCG (IU/l) Progesterone (nmol/l) Oestradiol (pmol/l) Values are mean SEM. Analysis by ANOVA showed no statistically significant differences. 697

5 Table 3. Fetal age in relation to the serum concentrations of anti-müllerian hormone (AMH), human chorionic gonadotrophin (HCG), progesterone and oestradiol from first-trimester pregnant women. Fetal age a <40 (n = 23) >40 50 (n = 74) >50 60 (n = 34) >60 (n = 22) P-value b AMH (ng/ml) NS HCG (IU/l) < Progesterone (nmol/l) < Oestradiol (pmol/l) < Values are mean SEM. a Days post conception, b Analysed by ANOVA, NS = not statistically significant. Table 4. Serum concentrations of anti-müllerian hormone (AMH) human chorionic gonadotrophin (HCG), progesterone and oestradiol in first trimester pregnant women in relation to fetal sex. Parameters Male fetus Female fetus (XY) (n = 56) (XX) (n = 45) AMH (ng/ml) HCG (IU/l) Progesterone (nmol/l) Oestradiol (pmol/l) Values are means SEM. Analysis by Student s t-test found no statistically significant differences. Table 5. Serum concentrations of anti-müllerian hormone (AMH) from second-trimester human fetuses. Fetal week (weeks Sex AMH after conception) (ng/ml) 15 XX < XX < XX < XY XY XY XY ). Fluid from such small antral follicles obtained in nonpregnant women has recently been shown to contain AMH concentrations of several hundred ng/ml, two to three orders of magnitude higher than in the circulation (Yding Andersen and Byskov, 2006; Yding Andersen et al., 2008). Follicular fluid from these small human antral follicles shows a highly significant inverse correlation with the content of oestradiol, progesterone and inhibin-b. It has been suggested that FSH and AMH, at least in non-pregnant conditions, may interact via an AMH-induced inhibition of granulosa cell expression of the aromatase enzyme and possibly also the LH receptor (Yding Andersen and Byskov, 2006; Yding Andersen et al., 2008). Despite the higher concentration of both oestradiol and progesterone in the circulation of pregnant women compared with non-pregnant women, the present study was unable to find a similar correlation in maternal circulation. This indicates that much higher concentrations of steroids, such as those found in follicular fluid, irrespective of the women being pregnant or not (Westergaard et al., 1990), are required to affect AMH secretion. Taken together, the present study suggests that regulation of AMH expression occurs locally within the follicle. This points to a highly specific role for AMH within the intrafollicular regulation of development and corroborates the hypothesis that AMH reduces aromatase and possibly LH-receptor expression (Grossman et al., 2008; Vigier et al., 1989). The present study found similar AMH concentrations between first-trimester pregnant and non-pregnant women, supporting an earlier study that evaluated the entire length of pregnancy on a relative small number of patients (La Marca et al., 2005) and found no alternation of AMH concentrations in maternal serum. Conversely, results of this study did not support a previous report of AMH concentration more than two times higher in pregnancy serum than that seen during the menstrual cycle or in pre-ovulatory follicular fluid (Al-Qahtani et al., 2005). The pronounced differences in AMH concentrations seen in the circulation of second-trimester fetuses according to the sex of the fetus were not reflected in the maternal circulation confirming previous results (Al-Qahtani et al., 2005). This shows that AMH expression occurs independently of the maternal system and may be regulated through different mechanisms in mother and fetus. It confirms that the placenta does not allow passage of this protein either from maternal to fetal circulation or vice versa, and that fetal AMH is unlikely to be produced by the placenta. Thus, the fetal concentration is a direct reflection of conditions within the fetus and confirms that the fetal AMH concentration may be used to detect pathological conditions that potentially involve AMH (Rey et al., 1999). This study extends the early observation of AMH concentrations in serum from second-trimester fetuses confirming that AMH expression is absent in the female fetus (Josso et al., 1993) and is the first to provide data with a validated highly sensitive AMH ELISA. AMH concentrations in the four male fetuses spanned a range of ng/ml and were all higher than each individual value reported earlier (Josso et al., 1993). Furthermore, these concentrations seem to be considerably higher than previously reported for newborn boys (Lee et al., 1996) although both early studies used a different ELISA procedure compared with the present work.

6 In conclusion, in light of the extreme concentrations of gonadotrophins in the circulation of pregnant women, AMH expression is unaffected and similar to that of nonpregnant women. Thus, the present study enforces that AMH expression occurs independently of gonadotrophins. The serum concentrations of AMH of first-trimester pregnant women are independent of fetal gender and similar to non-pregnant women, indicating that the protein is unlikely to cross, or to be produced by, the placenta. Acknowledgements This work was supported by The Danish Medical Research Council ( , AG Byskov). The excellent technical assistance by Tiny Roed and Anne Sørensen is gratefully acknowledged. References Al-Qahtani A, Muttukrishna S, Appasamy M et al Development of a sensitive enzyme immunoassay for anti-mullerian hormone and the evaluation of potential clinical applications in males and females. Clinical Endocrinology 3, Baarends WM, Uilenbroek JT, Kramer P et al Anti-Mullerian hormone and anti-mullerian hormone type II receptor messenger ribonucleic acid expression in rat ovaries during postnatal development, the estrous cycle, and gonadotropin-induced follicle growth. Endocrinology 11, Cook CL, Siow Y, Brenner AG et al Relationship between serum mullerian-inhibiting substance and other reproductive hormones in untreated women with polycystic ovary syndrome and normal women. Fertility and Sterility 1, Durlinger AL, Gruijters MJ, Kramer P et al Anti-Mullerian hormone attenuates the effects of FSH on follicle development in the mouse ovary. Endocrinology 11, Durlinger AL, Gruijters MJ, Kramer P et al Anti-Mullerian hormone inhibits initiation of primordial follicle growth in the mouse ovary. Endocrinology 3, Grossman MP, Nakajima ST, Fallat ME et al Müllerianinhibiting substance inhibits cytochrome P450 aromatase activity in human granulosa lutein cell culture. Fertility and Sterility 5 (Suppl. 1), Josso N, Lamarre I, Picard JY et al Anti-Mullerian hormone in early human development. Early Human Development 2, Jost A 1947 Recherches sur la differentiation sexuelle de l embryon de lapin. Archives d Anatomie Microscopique et de Morphologie Expérimentale 36, La Marca MA, Orvieto R, Giulini S et al Mullerian-inhibiting substance in women with polycystic ovary syndrome: relationship with hormonal and metabolic characteristics. Fertility and Sterility 4, La Marca MA, Giulini S, Orvieto R et al Anti-Mullerian hormone concentrations in maternal serum during pregnancy. Human Reproduction 6, Lee MM, Donahoe PK 1993 Mullerian inhibiting substance: a gonadal hormone with multiple functions. Endocrine Reviews 2, Lee MM, Donahoe PK, Hasegawa T et al Mullerian inhibiting substance in humans: normal levels from infancy to adulthood. Journal of Clinical Endocrinology and Metabolism 2, Nakahori Y, Hamano K, Iwaya M et al Sex identification by polymerase chain reaction using X-Y homologous primer. American Journal of Medical Genetics 4, Pellatt L, Hanna L, Brincat M et al Granulosa cell production of anti-mullerian hormone is increased in polycystic ovaries. Journal of Clinical Endocrinology and Metabolism 1, Pigny P, Jonard S, Robert Y et al Serum anti-mullerian hormone as a surrogate for antral follicle count for definition of the polycystic ovary syndrome. Journal of Clinical Endocrinology and Metabolism 3, Rey RA, Belville C, Nihoul-Fekete C et al Evaluation of gonadal function in 107 intersex patients by means of serum anti-mullerian hormone measurement. Journal of Clinical Endocrinology and Metabolism 2, Vigier B, Tran D, du Mesnil du BF et al Use of monoclonal antibody techniques to study the ontogeny of bovine anti- Mullerian hormone. Journal of Reproduction and Fertility 1, Vigier B, Forest MG, Eychenne B et al Anti-Mullerian hormone produces endocrine sex reversal of fetal ovaries. Proceedings of the National Academy of Science of the United States of America 10, Visser JA, de Jong FH, Laven JS et al Anti-Mullerian hormone: a new marker for ovarian function. Reproduction 1, 1 9. Wachs DS, Coffler MS, Malcom PJ et al Serum anti-mullerian hormone concentrations are not altered by acute administration of follicle stimulating hormone in polycystic ovary syndrome and normal women. Journal of Clinical Endocrinology and Metabolism 5, Westergaard LG, Yding Andersen C, Byskov AG et al Epidermal growth factor in small antral ovarian follicles of pregnant women. Journal of Endocrinology 2, Yding Andersen C, Rosendahl M, Byskov AG et al Concentration of anti-mullerian hormone and inhibin-b in relation to steroids and age in follicular fluid from small antral human follicles. Journal of Clinical Endocrinology and Metabolism 6, Yding Andersen C, Byskov AG 2006 Estradiol and regulation of anti- Mullerian hormone, inhibin-a, and inhibin-b secretion: analysis of small antral and preovulatory human follicles fluid. Journal of Clinical Endocrinology and Metabolism 10, Declaration: The authors report no financial or commercial conflicts of interest. Received 19 June 2008; revised and resubmitted 1 September 2008; refereed 23 September 2008; accepted 9 January