Modifiers of Ovarian Function in Girls and Women With Classic Galactosemia

Abstract

Context:

Classic galactosemia is a potentially lethal genetic disorder resulting from profound impairment of galactose-1P uridylyltransferase (GALT). More than 80% of girls and women with classic galactosemia experience primary or premature ovarian insufficiency despite neonatal diagnosis and rigorous lifelong dietary galactose restriction.

Objective:

The goal of this study was to test the relationship between markers of ovarian reserve, cryptic residual GALT activity, and spontaneous pubertal development in girls with classic galactosemia.

Design and Setting:

This was a cross-sectional study with some longitudinal follow-up in a university research environment.

Patients:

Patients included girls and women with classic galactosemia and unaffected controls, <1 month to 30 years old.

Main Outcome Measures:

We evaluated plasma anti-Müllerian hormone (AMH) and FSH levels, antral follicle counts ascertained by ultrasound, and ovarian function as indicated by spontaneous vs assisted menarche.

Results:

More than 73% of the pre- and postpubertal girls and women with classic galactosemia in this study, ages >3 months to 30 years, demonstrated AMH levels below the 95% confidence interval for AMH among controls of the same age, and both pre- and postpubertal girls and women with classic galactosemia also demonstrated abnormally low antral follicle counts relative to age-matched controls. Predicted residual GALT activity ≥0.4% significantly increased the likelihood that a girl with classic galactosemia would demonstrate an AMH level ≥0.1 ng/mL.

Conclusions:

A majority of girls with classic galactosemia demonstrate evidence of diminished ovarian reserve by 3 months of age, and predicted cryptic residual GALT activity is a modifier of ovarian function in galactosemic girls and women.

Classic galactosemia is a potentially lethal metabolic disorder that affects more than 1 in 60 000 live births in many populations (reviewed in Ref. 1). Affected infants may appear normal at delivery but develop life-threatening complications within days to weeks of exposure to breast milk or a milk-based formula due to their inability to metabolize galactose. With the benefit of early diagnosis, often by newborn screening, and rigorous dietary galactose restriction, most infants with classic galactosemia survive and grow. However, most will experience a broad constellation of long-term complications (2) including premature ovarian insufficiency (POI) (reviewed in Ref. 3), which affects >80% of galactosemic young women and may present as primary amenorrhea and lack of secondary sex characteristics or, in milder cases, as irregular menses, diminished ovarian reserve, infertility, menopausal symptoms, and ultimately secondary amenorrhea in the second or third decade of life (3, 4).

One accessible biomarker of ovarian function in both pre- and postpubertal girls and women is anti-Müllerian hormone (AMH), also called Müllerian-inhibiting substance (MIS) (reviewed in Refs. 5 and 6). AMH is a glycoprotein secreted by granulosa cells in the developing ovarian follicle. Population studies demonstrate that circulating AMH levels in healthy newborn girls are low to undetectable but rise to ≥1 ng/mL by 1 year and ≥2 ng/mL by 8 to 25 years (7, 8). AMH then declines slowly over time, falling to undetectable levels near menopause (7, 8). Previous studies of healthy women (9, 10) and also patients with Turner syndrome (TS) (7) have demonstrated that even single AMH measurements can have robust positive predictive value regarding future menses or POI (11).

Like AMH, FSH also serves as an indicator of both normal and abnormal ovarian function, but the relationship between FSH and ovarian status is complex. In typically developing girls, the hypothalamic-pituitary-ovarian (HPO) axis inactivates after about age 18 months, effectively uncoupling FSH from ovarian function until the HPO axis reactivates shortly before puberty. In young girls with primary ovarian insufficiency, FSH levels may therefore remain deceptively within the normal range despite a complete lack of ovarian function.

Girls with profound POI may require hormone replacement therapy (HRT); however, deciding whether and when to initiate HRT in a prepubertal girl can be challenging. Furthermore, given recent advances in cryopreservation of oocytes or ovarian tissue (12–14), young girls and their families may want to know whether POI is imminent so they can make informed decisions about potential intervention before the window of opportunity has closed.

Previously, we reported that AMH levels are markedly diminished or undetectable in >80% of girls and women with classic galactosemia (4) but not in girls with Duarte galactosemia (15). However, our initial data were insufficient to address the timing of onset of ovarian dysfunction or the short-term longitudinal stability of circulating AMH levels. We also did not have clinical information to assess the relationship between AMH level and ovarian outcome in classic galactosemia. Finally, we were not able to explore candidate modifiers of ovarian outcome in girls with classic galactosemia. Here we have addressed each of these questions using substantially expanded cross-sectional datasets from cases and controls together with outcome information from a subset of volunteers. This work has implications for understanding the mechanism of POI in classic galactosemia and for the clinical care of affected girls and women.

Subjects and Methods

Subjects

Blood samples were collected from girls and women with classic galactosemia and from age-matched controls aged <1 month to 30 years. Data from a subset of these girls and women were included in 1 or both of 2 previous publications (4, 15). This study was approved by the Institutional Review Board at Emory University (eIRB protocol 00024933, principal investigator J.L.F.-K.).

Sample collection and handling

Details of our methods for sample handling and AMH and FSH assays were previously published (4).

AMH and FSH measurements

AMH and FSH assays were performed at the Biomarkers Core Laboratory of the Yerkes Primate Research Center at Emory University (Atlanta, Georgia) or at the Reproductive Endocrine Unit Reference Laboratory, Massachusetts General Hospital (Boston, Massachusetts).

Predicted residual galactose-1P uridylyltransferase activity

Yeast models of 28 known galactose-1P uridylyltransferase (GALT) missense or nonsense mutations present in the study volunteers were generated using site-directed mutagenesis of the wild-type human GALT open reading frame within the centromeric yeast expression vector pMM22.hGALT, essentially as described previously (16).

Pelvic ultrasounds

Ultrasounds were performed as described in Supplemental Materials and Methods (published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.orgj). Controls >18 years of age were not recruited because sufficient data have been reported previously to establish normal ranges for adults (17).

Statistical analyses

Before analysis of the biochemical outcomes, we dichotomized AMH as ≥0.1 ng/mL (clearly detectable) or <0.1 ng/mL (borderline or undetectable) and transformed FSH to approximate normality using an inverse-normal transformation. We examined the relationship between each outcome and disease status using appropriate generalized estimating equation models that can account for more than 1 clinical value per study volunteer. For AMH, we stratified analyses by different age categories (0–0.25, 0.25–1.5, 1.5–4, 4–8, and 8–30 years). For FSH, we limited analyses to those volunteers in the age range 1.5 to 8 years. Additionally, we also applied generalized estimating equation models to examine the relationship between AMH and FSH among study volunteers >1.5 years old.

For the analysis of physiological outcomes, we examined the relationship between disease status and antral follicle count (AFC) values using logistic regression models adjusted for age. To study the relationship between spontaneous menses (dichotomized as presence or absence) and AMH (dichotomized based on being above or below 0.1 ng), we performed both logistic regression and Fisher's exact test (to ensure valid analysis in the presence of small sample size). We also performed similar analyses to examine the relationship between our dichotomized spontaneous menses variable and predicted GALT activity (dichotomized based on being above or below 1% activity). All analyses were performed using the R programming language.

Results

Study volunteers

Study volunteers (Table 1) who provided information about ovarian outcome or ultrasound data all participated with informed consent under Emory eIRB protocol 00024933 (principal investigator J.L.F.-K.). Some plasma samples were ascertained as clinical lab discards collected under a Health Insurance Portability and Accountability Act (HIPAA) waiver associated with eIRB00024933.

Table 1.

Study Volunteers

Study Population	Classic Galactosemia, n	Control, n
Total	158	132
Age at initial test for this study
0–3 mo	10	27
3 mo to 1.5 y	17	21
1.5–4 y	18	20
4–8 y	28	15
8–12 y	16	13
12–30 y	69	36
Testing for this study
Pelvic ultrasoundsa	14	36
Volunteers contributing hormone test results	158	96
Total no. of AMH values	300	96
Total no. of FSH values	298	80

^aTotal number of readable ultrasounds from which an AFC of at least 1 ovary could be determined.

Most galactosemic girls demonstrate significantly low AMH by age 3 months

As a marker of ovarian reserve, we measured AMH in 396 samples representing 158 girls and women with classic galactosemia and 96 controls, all <1 month to 30 years old (Figure 1 and Supplemental Figure 1). Data points connected by lines indicate longitudinal measurements from the same individual. All control values were single measurements derived from different individuals.

Figure 1.

AMH and FSH levels in girls and women with classic galactosemia vs controls. A, Symbols represent 396 AMH measurements from 158 girls and women with classic galactosemia (●) and 96 controls (○). B, Symbols represent 298 FSH measurements from 106 girls and women with classic galactosemia (●) and 78 controls (○). In both panels, all study volunteers were between the ages of <1 month to 30 years, and symbols representing longitudinal measurements from the same individual are connected by solid lines.

Consistent with previous reports (7), AMH was low to undetectable in most newborns (Figure 1 and Supplemental Figure 1) but rose quickly in controls. By 18 months, almost all control girls had detectable AMH, and by 3 years, most had AMH ≥1.5 ng/mL (10.7 pmol/L). In contrast, for the vast majority of girls with classic galactosemia, AMH remained very low to undetectable, although for a small number, AMH rose into the detectable or even the normal range. Comparisons of AMH levels between cases and controls across 5 age ranges (birth to <3 months, 3 months to <1.5 years, 1.5 to <4 years, 4 to <8 years, and 8 to <30 years) demonstrated significant differences in all age categories; P values ranged from .01 for the youngest to 1.9 × 10⁻⁷ for the oldest. Of all AMH values measured for girls with classic galactosemia in the study ≥3 months old, fewer than 27% were within the age-specific range reported for 95% of the normal population (7).

Inactivation of the HPO axis is delayed for many young girls with classic galactosemia

In typically developing girls, the HPO axis is active from just after birth until about 18 months of age, at which point it inactivates until the girl approaches puberty (18, 19). In girls with normal ovaries, this means that serum FSH levels may be either low or elevated in infants and toddlers (see Figure 2A), remain low through childhood, and then rise slightly once pulsatility alters just before puberty. In older women approaching menopause, FSH begins to rise as ovarian function decreases, and in postmenopausal women, FSH can exceed 100 mIU/mL (20).

Figure 2.

Relationship between AMH and FSH in young girls with classic galactosemia and controls. A, In girls less than 18 months old, elevated FSH values are seen in both cases (●) and controls (○). B, In girls between 18 months and 6 years old, many of the cases but none of the controls demonstrate elevated FSH.

For young girls whose ovaries are absent or hypofunctional, the timing of HPO axis inactivation and reactivation may vary. For example, girls with TS, a chromosomal aneuploidy associated with POI, can demonstrate elevated FSH that persists until age 4 to 5 years (7). Similarly, girls with Nijmegen chromosome breakage syndrome (NBS), a severe chromosomal instability disorder associated with POI, also demonstrate markedly elevated FSH past age 4 years (21).

We observed a persistence of elevated FSH in 18-month to 8-year-old girls with classic galactosemia (Figure 1B), consistent with previous reports (reviewed in Ref. 22), indicating a delay in HPO inactivation (P = 4.4 × 10⁻⁶). The fact that FSH levels in so many girls with classic galactosemia, TS, and NBS remain elevated for so long is striking. Although there are probably central mechanisms that inactivate the HPO axis 1 to 2 years after birth, the lack of estradiol and inhibin feedback from the hypofunctioning ovary may contribute to the prolonged elevation of FSH (19).

Longitudinal measurements of AMH in girls and young women with classic galactosemia

To explore the long-term implications of AMH measurements in girls and young women with classic galactosemia, we tested whether plasma AMH values were stable in individual volunteers over time. We analyzed 177 longitudinal plasma samples representing 53 girls and young women with classic galactosemia between the ages of 3 months and 25 years. These values are plotted in Figure 3 with lines connecting data points derived from individual girls or women. Forty-four of the 53 volunteers demonstrated AMH values that never climbed to within the established normal range for age-matched controls; these data points are presented in Figure 3 as gray symbols connected by gray lines. Nine of the 53 volunteers demonstrated 1 or more AMH values within the control range, and of these, 3 demonstrated AMH values that were consistently within the control range. The remaining 6 had some AMH values within the control range and other AMH values below the control range (Figure 3).

Figure 3.

Longitudinal AMH values from individual girls and women with classic galactosemia. Symbols representing repeated AMH measures from the same individual are connected by lines. Symbols and lines representing girls and women whose AMH values were consistently below the control range are presented in gray. Symbols outlined in black and connected by black lines represent girls and women for whom at least 1 AMH value was within the age-appropriate control range; these 9 volunteers are identified by letters a to i. For these 9 volunteers, open symbols indicate AMH values within the age-appropriate control range; gray filled symbols indicate AMH values below the age-appropriate control range. Volunteers a and c are represented by squares and triangles, respectively, only to distinguish them from each other and from volunteer b, whose symbols were in close proximity on the graph.

Two patterns were clear. First, as expected from previous studies (4) and from the larger data set (Figure 1), more than 83% of the 53 volunteers in the longitudinal cohort had AMH levels consistently below the control range. Second, both the volunteers (d, g, and i), who demonstrated the most consistently normal AMH values and also the volunteers (a, b, c, e, f, and h) whose AMH values were sometimes within and sometimes below the control range were not clustered but spanned almost the full age range (3–20 years).

Girls with classic galactosemia show low AFCs

As a quantitative visual measure of ovarian function in both pre- and postpubertal girls and young women with classic galactosemia, we performed pelvic ultrasound scans to assess AFCs (Figure 4). We obtained high-quality images for a total of 38 controls (of whom 35 showed clearly visible ovaries) and 26 classic galactosemics (of whom 14 showed clearly visible ovaries).

Figure 4.

AFCs in girls and young women with classic galactosemia vs controls. Upper left panel, Average AFC per ovary visualized in girls and young women by ultrasound, plotted relative to age of the volunteer. Filled circles represent cases, and open circles represent controls. Left lower panel, Cases and controls, plotted by age, whose ovaries could not be identified by ultrasound. Right panels, Representative images of ovaries visualized by ultrasound in cases (upper image) and controls (lower image).

Two patterns were apparent (Figure 4). First, although 2 young girls with classic galactosemia had AFCs within the control range, the other 12 all had AFCs below the control range; this difference was significant (P = .0041). Second, although not identifying ovaries by transabdominal ultrasound in any girl or young woman is not itself meaningful, that this occurred in <10% of controls but in almost half of cases was striking. Presumably, this is because the ovaries in the galactosemic girls were small and more difficult to visualize, as noted recently by Gubbels and colleagues (23). Of note, uterine volume was not significantly different between cases and controls (data not shown).

Relationship between AMH level and spontaneous vs HRT-assisted menarche in girls with classic galactosemia

To address whether girls with classic galactosemia who exhibit higher AMH levels are more likely to achieve spontaneous puberty than their counterparts with lower AMH levels, we assessed AMH values (taken within 5 years of menarche) for 28 volunteers for whom we had the relevant data (Figure 5A). Of the 11 volunteers for whom hormone replacement was required to achieve menarche, only 2 had detectable AMH, and both of those levels were very low. Of the 17 volunteers who achieved spontaneous menarche, 3 had normal or near-normal AMH, and another 5 had AMH levels that were detectable above baseline. Although the pattern was interesting, the AMH distribution between the 2 cohorts was not statistically different (P = .098 using a Fisher's exact test), likely reflecting the limited sample size.

Figure 5.

Box and whisker plots showing the relationships between AMH, menarche, and predicted residual GALT activity. A, Elevated AMH appears to cosegregate with spontaneous menarche in girls and young women with classic galactosemia, but some girls with undetectable AMH also achieve spontaneous menarche. B, Girls and young women with predicted residual GALT activity ≥0.4% were significantly more likely to demonstrate AMH values ≥0.1 ng/mL than were girls and young women with predicted residual GALT activity <0.4% (P = .0031), and adjustment for age did not compromise the significance of this relationship (P = .0050). All data included in this figure were derived from girls and women between the ages of 3 months and 30 years. WT, wild type.

Relationship between predicted residual GALT activity and AMH level in girls and women with classic galactosemia

As a first step toward identifying modifiers of ovarian outcome in classic galactosemia, we asked whether girls and women with GALT genotypes predicted to encode residual GALT activity in a yeast expression system (16, 24) were more likely to have AMH values ≥0.1 ng/mL. Recently, we reported that school-aged boys and girls with classic galactosemia predicted to express ≥0.4% wild-type GALT activity demonstrated significantly higher scholastic achievement in math than did their counterparts predicted to express <0.4% wild-type GALT activity (25).

We therefore compared the AMH levels detected in 172 plasma samples from 91 girls and women with classic galactosemia predicted to express <0.4% wild-type GALT activity with AMH levels detected in 77 plasma samples from 39 girls and women with classic galactosemia predicted to express ≥0.4% wild-type GALT activity (Figure 5B). All volunteers included in this comparison were between the ages of 3 months and 30 years. A majority of volunteers in both groups demonstrated low to undetectable AMH values; however, the distributions were significantly different (P = .0031, Figure 5B). Of note, when the analysis was adjusted to include a variable for age (7), the difference remained highly significant (P = .0050).

It is also interesting to note the predicted residual GALT activities of the 9 study volunteers whose longitudinal AMH levels each showed at least 1 value in the control range (Figure 3). Of the 3 with consistently high AMH (d, g, and i), 2 had predicted GALT activity ≥0.4% and 1 was unknown. Of the 2 who each presented with 3 AMH values in the control range and 1 low (e and h), both had predicted GALT activity levels ≥0.4%. Of the remaining 4 girls (a, b, c, and f) who each had 1 AMH reading within the control range and 4 or more low readings, 3 had no predicted residual GALT activity and 1 was of unknown. Although these data are anecdotal, a pattern is clear.

Discussion

Despite the benefit of early diagnosis and rigorous dietary restriction of galactose, POI continues to affect >80% of girls and women with classic galactosemia (3); the underlying mechanism remains unclear. Previously, we reported the results of a small study looking at AMH and FSH in girls and women with classic galactosemia (4); here we have both confirmed and extended from that work. We now demonstrate that a reduced AMH level, consistent with diminished or arrested ovarian reserve, is present for many affected girls before the age of 3 months, that inactivation of the HPO axis is delayed for many affected girls, and that once AMH falls to undetectable levels it rarely rises again. Finally, we report the first significant modifier of ovarian outcome severity in classic galactosemia-predicted residual GALT activity. Combined, these results shed light on mechanism and offer potential tools for clinical practice.

Implications for mechanism

Our results demonstrate 3 separate indicators of ovarian dysfunction in prepubertal girls with classic galactosemia: strikingly low AMH in most, elevated FSH in some, and low AFCs, or ovaries that cannot be identified by ultrasound, in many. Although these data cannot conclusively define a mechanism, they do offer some intriguing clues.

One possible explanation for POI could be precocious maturation of ovarian follicles leading to increased atresia and diminished reserve. Were this occurring, we might have expected to see increased rather than decreased AFCs in the youngest cases; we did not. Of course, it is possible a wave of early maturation and atresia occurred before the time frame of our scans, perhaps even prenatally. Of note, our data demonstrate that even infants younger than 3 months have plasma AMH levels lower than those seen in age-matched controls (P = .01). We also cannot rule out early follicular maturation arrest because primordial follicles cannot be seen by ultrasound and do not secrete AMH. There is currently no biomarker available to quantify primordial follicles. Previous (4, 23, 26, 27) and current data do argue against a defect in FSH or FSH receptor function; women with abnormal FSH receptor function generally demonstrate elevated FSH with normal AMH (28), whereas in galactosemia, we see elevated FSH with low AMH.

Another potential clue derives from the observation that many girls with classic galactosemia (22), TS, (7) or NBS (21) experience significantly elevated FSH well beyond 18 months of age, supporting the hypothesis that ovaries contribute to inactivation of the HPO axis in early childhood. Of course, other mechanisms are not excluded by these observations.

Implications for clinical practice

Three principle challenges face galactosemic girls and their healthcare providers with regard to ovarian function: 1) if and when to initiate HRT if spontaneous puberty is absent or delayed, 2) if and when to initiate HRT after spontaneous puberty if menstrual cycles become very irregular or symptoms of premature menopause occur, and 3) if, when, and how to deal with the likelihood of infertility.

With regard to spontaneous vs assisted puberty, the data presented here suggest that measuring plasma AMH and/or using ultrasound to estimate AFC in a prepubertal girl who has yet to present physical signs of puberty may assist in the decision of if and when to initiate HRT. AFC is a documented marker of ovarian reserve and should correlate with AMH (29–32); if ovaries are visible by ultrasound, conducting both tests might offer added confidence in the result. If AMH is detectable, it may pay to wait and see whether spontaneous puberty initiates. If AMH is undetectable, the risk-benefit ratio may favor intervention. For the postpubertal young woman facing irregular periods, either AMH or FSH should provide accurate insight into ovarian status, enabling a more informed decision about if and when to initiate HRT for cycle regulation, for the relief of menopausal symptoms, and/or to prevent accelerated loss of bone mineral density.

The question of fertility and potential intervention in girls and young women with classic galactosemia is more complex. First, it is important to note that spontaneous pregnancies do occur in women with classic galactosemia and with POI from other causes, even in the context of amenorrhea and a high FSH and/or undetectable AMH (33, 34). These pregnancies could reflect either residual follicles that were developing normally but simply few in number or follicles that had arrested and then resumed development. The probability of a woman with POI in general becoming pregnant is approximately 5% to 10%; however, although there are many case reports of interventions that have led to successful pregnancies in women with POI, no medical treatment has been rigorously proven to be of benefit (35, 36).

Oocyte and/or ovarian tissue cryopreservation can now be offered to young women with classic galactosemia if sufficient follicles are still present to harvest (37). However, by the time a young woman is old enough to make an informed decision about the risks and benefits of the procedure, her follicular depletion may be so profound as to make controlled ovarian hyperstimulation and mature oocyte cryopreservation impractical if not impossible. Clearly, this dilemma raises the question of whether parents should consider arranging for the procedure when their daughter is still very young and therefore more likely to have viable follicles or ovarian tissue to cryopreserve. Although potentially controversial, this is the recommended practice in some centers for prepubertal girls scheduled to undergo cancer treatment anticipated to result in POI (38). As in vitro maturation and tissue cryopreservation techniques improve, this may become an option for families with classic galactosemic daughters who continue to have ovarian reserve after infancy.

Predicted residual GALT activity is a modifier of ovarian function in girls and women with classic galactosemia

Finally, our data demonstrate a statistically significant relationship between girls and young women with classic galactosemia having a predicted residual GALT activity of ≥0.4% and a plasma AMH level of ≥0.1 ng/mL. Although this is a novel finding, it is not unexpected given our recent observation linking predicted residual GALT activity with scholastic achievement in math for school-aged children with classic galactosemia (25). By uncovering a similar relationship between predicted cryptic residual GALT activity and ovarian function, we suggest that the effect is not brain-specific and may be general. Nonetheless, it is also important to recognize that residual GALT activity cannot be the only modifier of ovarian outcome severity in classic galactosemia; some girls with <0.4% predicted residual GALT activity still demonstrate detectable AMH, and many girls with ≥0.4% predicted residual GALT activity demonstrate no detectable AMH. This point is further illustrated in Supplemental Table 2, which presents the different GALT genotypes observed among galactosemic study volunteers with AMH levels close to or far from the age-adjusted median for controls. As presented (Supplemental Table 2), hypomorphic alleles are common among girls and women with higher AMH levels, but the relationship is not absolute. Additional studies will be required to identify other factors that contribute to variability of long-term outcome among patients with classic galactosemia.