Anti-Müllerian Hormone Deficiency in Females With Fanconi Anemia

Abstract

Context:

In females with Fanconi anemia (FA), infertility is often accompanied by diminished ovarian reserve and hypergonadotropic amenorrhea before the age of 30 years, suggesting primary ovarian insufficiency (POI). POI is typically diagnosed only after perimenopausal symptoms are observed.

Objective:

The objective of the study was to assess whether serum anti-Müllerian hormone (AMH) levels can serve as a cycle-independent marker for the diagnosis of POI in patients with FA.

Design and Setting:

This observational study used the National Cancer Institute's inherited bone marrow failure syndrome cohort at the National Institutes of Health Clinical Center.

Participants:

The study included 22 females with FA, 20 unaffected female relatives of patients with FA, and 21 unrelated healthy females under 41 years of age.

Main Outcome Measure:

Serum AMH, a marker of ovarian reserve, was measured in all participants.

Results:

Females with FA had very low AMH levels (median 0.05 ng/mL; range 0–2.32 ng/mL; P < .001) when compared with unaffected relatives (median 2.10 ng/mL; range 0.04–4.73 ng/mL) and unrelated healthy females (median 1.92 ng/mL; range 0.31–6.64 ng/mL). All patients with FA older than 25 years of age were diagnosed with POI and had undetectable AMH levels.

Conclusions:

AMH deficiency appears to be a shared trait across this heterogeneous FA cohort. Substantially reduced AMH levels in females with FA suggest a primary ovarian defect associated with reduced fertility. Measurement of AMH at the time of FA diagnosis and subsequent monitoring of AMH levels at regular intervals may be useful for the timely management of complications related to POI such as subfertility/infertility, osteoporosis, and menopausal symptoms.

In Fanconi anemia (FA), biallelic autosomal recessive mutations in the FA/BRCA DNA repair pathway result in congenital anomalies, bone marrow failure, increased frequency of malignancy, and infertility (1). Common anomalies include short stature; microcephaly; thumb and radial malformations; renal, cardiac, and other skeletal malformations; hearing deficits; and developmental delay (1). FA is predicted to occur in 1 in 131 000 live births in the United States (2). Of the 16 known FA genes, mutations in FANCA, FANCC, and FANCG account for approximately 85% of all cases in the Western world. Median survival is 29 years of age with complications from bone marrow failure and cancer as leading causes of death (1). FA proteins are highly expressed in the ovaries, and inadequate repair of double-strand DNA breaks in ovarian follicles has been reported to prematurely diminish ovarian reserve in humans, highlighting the association between DNA repair mechanisms and ovarian aging (3, 4).

Anti-Müllerian hormone (AMH) is a measure of ovarian reserve, which is currently used as a menstrual cycle-independent marker for the clinical diagnosis of polycystic ovarian syndrome and for evaluating responsiveness to in vitro fertilization and assisted reproduction techniques (5, 6). In females, AMH is produced exclusively by the granulosa cells of small growing follicles within the ovaries and is correlated with antral follicle count (AFC) (7). As a peptide hormone from the TGF-β family of growth factors, AMH controls cell proliferation, cell cycle, cell differentiation, and apoptosis as well as being responsible for the apoptotic regression of Müllerian ducts (cervix, fallopian tubes, uterus, upper one third of the vagina) in male fetuses, thus conferring a male phenotype. Extremely low levels of AMH are detectable in healthy females, beginning at approximately the 36th week of gestation. AMH levels rise just prior to puberty, remain elevated throughout a female's reproductive years (normal range 1–8 ng/mL), and then decline to undetectable levels 5 to 10 years prior to menopause (8, 9).

Reduced fertility in women with FA has been associated with premature menopause, now more appropriately termed primary ovarian insufficiency (POI) (10, 11). In the general (unaffected) population, genetics play a crucial role in the establishment and maintenance of ovarian reserve because the follicular pool is derived in utero and is often similar among first-degree female relatives (5, 7). POI results when one or a combination of the following occurs prior to the age of 40 years: 1) establishment of a suboptimal follicular pool, 2) follicular dysfunction, or 3) accelerated depletion of the follicular pool (11, 12). The diagnosis of POI uses criteria similar to that of menopause: two measures of elevated FSH and amenorrhea for more than 4 months (11–13). Although a variety of genetic defects have been associated with POI, the mechanisms responsible for POI remain unknown for most affected women, both with and without FA (11, 12). AMH has been favorably evaluated for the use in a POI diagnosis because it is a better marker of diminished ovarian reserve when compared with FSH (14–16). We hypothesized that AMH could serve as a cycle-independent marker for the diagnosis of POI in females with FA.

Materials and Methods

Design, setting, and participants

This is a cross-sectional study of prepubertal and reproductive-aged females with FA, their unaffected relatives, and healthy volunteers under 41 years of age. With the exception of two patients with FA whose blood samples were obtained in field studies, all patients with FA and their unaffected relatives were evaluated by medical history, physical examination, and clinical laboratory studies at the National Institutes of Health Clinical Center (Bethesda, Maryland) as part of the National Cancer Institute's (NCI) inherited bone marrow failure syndrome cohort (http://www.marrowfailure.cancer.gov; NCI 02-C-0052; clinicaltrials.gov; identifier NCT00027274). A detailed description of the NCI inherited bone marrow failure syndrome cohort was provided in a previous publication (Supplemental Figure 1, A and B) (17). This study was approved by the NCI Institutional Review Board, and all participants provided written informed consent.

FA was diagnosed by abnormal chromosome breakage in peripheral blood lymphocytes (or in skin fibroblasts for those suspected to have hematopoietic somatic mosaicism) after culture with DNA cross-linking agents (diepoxybutane and mitomycin C). Patients with somatic mosaicism had an event in hematopoietic stem cells in which one of the FA mutant alleles reverted to normal; the progeny of these stem cells have a selective advantage compared with FA cells in which both alleles are mutated (18). An FA diagnosis was confirmed by complementation studies and gene mutation analysis in all but one patient. FA-related anomalies including short stature; microcephaly; thumb and radial malformations; renal, cardiac, and other skeletal malformations; hearing deficits; and developmental delay were recorded. Patients were monitored prospectively to determine changes in the clinical status.

Fifteen of the 21 healthy females were self-reported to be healthy participants in the Occupational Health Service Normal Donor Program at the Frederick National Laboratory for Cancer Research (FNLCR) (Frederick, Maryland). An additional three healthy females received a full clinical evaluation and laboratory measures at the National Institutes of Health Clinical Center as part of a vaccine trial (clinicaltrials.gov; identifier NCT00798265). Serum from the remaining three healthy females was purchased from Equitech-Bio, Inc.

AMH measurement

Serum AMH was measured using the sensitive Gen II AMH ELISA from Beckman Coulter, Inc at the FNLCR according to the manufacturer's protocol. Excellent intra- and interplate reproducibility have been reported for this assay (19). Reproducibility of the AMH ELISA was confirmed in the FNLCR laboratory with intra- and interplate variability less than 10%.

Statistical analysis

Results are expressed as medians and ranges. GraphPad Prism was used for graphing and statistical analyses. Comparison of AMH levels and ages between groups was performed by Mann-Whitney U and Kruskal-Wallis tests, and P < .05 was considered significant.

Results

Serum AMH levels were measured in 22 females with FA (age range 7–37 y), 20 unaffected female FA relatives (age range 3–40 y), and 21 unrelated healthy females (age range 12–40 y), with a focus on pubertal and postpubertal females over the age of 10 years (47 of 63 subjects) (Table 1). FA-related anomalies were seen in 17 patients: 13 had three or more anomalies, four had one or two, and five had none. Fourteen patients with FA had mutations in FANCA, five FANCC, one FANCJ, one FANCF, and for one, the FA gene mutation was unknown. All postpubertal patients were FANCA or FANCC. Three postpubertal patients (ages 27, 27, and 33 y) with FANCA mutations had hematopoietic somatic mosaicism. Fifteen unaffected FA relatives were carriers of an FA mutation, three were noncarriers, and in two the mutation status was unknown.

Table 1.

Features of Participants

Parameters	FA	FA Relatives	Unrelated Controls	P Value
Subjects, n	22	20	21
Median age when serum drawn, y (range)a	15 (7–37)	33.5 (3–40)	27 (12–40)	.004
Subjects achieving menarche, n	12	16	20	.37
Median age at menarche, n (range)	13.5 (11–17)	12.5 (8–15)	NA	.09
Subjects over the age of 10 y, nb	15	18	21	.37
Median age of subjects over the age of 10 y, n (range)	20 (11–37)	34 (11–40)	27 (12–40)	.03

^aAll ages are reported in years.

^bTen years is the median age at which non-Hispanic white females enter puberty (20).

Twenty of the 21 unrelated healthy females, 16 of 20 unaffected female relatives, and 12 of 22 females with FA achieved menarche prior to AMH measurement. Nine of the 12 postmenarchal females with FA were at least age 13 years at the time of menarche (median 13.5 y; range 11–17 y; n = 12), whereas half of the unaffected postpubertal relatives were younger than 13 years at the time of menarche (median 12.5 y; range 8–15 y; n = 16); however, the age at menarche did not differ significantly between the two groups (P = .09) (Table 1). Overall, 15 patients older than age 10 years entered puberty prior to AMH measurement and reached Tanner staging of 3 or greater; the progression of pubertal milestones was similar to that in the general population (20).

Five of the 12 postmenarchal females with FA were hypothyroid; two were diagnosed at the time of sampling, one was hypothyroid despite thyroid replacement, and two others were euthyroid on thyroid replacement. Three had possible hypothalamic hypogonadism, with one of the three also having multiple endocrinopathies including GH deficiency. Two patients were taking estrogen to aid with pubertal development, three were taking hormone replacement, and another two were on chronic androgen therapy for bone marrow failure (Table 2). None of the unaffected relatives had endocrine abnormalities.

Table 2.

Clinical Details for FA Patients Over the Age of 10 Years

UPN	Age at Study, y	Age at Menarche, y	Menses	Hormonal Therapy	Seeking Fertility	POI	AMH, ng/mL	Number of Anomalies	FA Gene	FA Mosaic	BMT Prior to Study	Age at BMT, y	Time Since BMT, y	Prevalent Cancer	Incident/Recurrent Cancer
NCI-213-1	11	NA	NA	Estradiol	No	Yes	0	3	FANCA	No	Yes	9a	1	No	No
NCI-246-1	11	NA	NA	Oxymetholone	No	No	0.07	2	FANCA	No	No			No	No
NCI-25-1	12	NA	NA	Premarin	No	No	0	8	FANCC	No	Yes	6a	6	No	No
NCI-98-1	14	11	Irregular		No	No	0.02	6	FANCA	No	Yes	8a	6	No	No
NCI-59-1	16	14	Regular		No	No	0	0	FANCA	No	No			No	No
NCI-111-1	19	15	Regular		No	No	1.18	0	FANCA	No	No			No	No
NCI-12-1	20	14	Irregular		No	No	0.02	0	FANCA	No	No			No	No
NCI-331-1	20	12	Regular		No	No	1.64	0	FANCA	No	No			No	No
NCI-19-1	22	17	Irregularb	Danazol	No	No	0.23	3	FANCC	No	No			No	Basal cell skin
NCI-169-1c	23	13	Irregular		No	Yes	0.26	3	FANCC	No	Yes	7d	16	No	Vulvar, anocervical
NCI-73-1	27	13	Irregular		Yes	Yes	0	4	FANCA	Yes	No			No	Breast
NCI-73-2	27	13	Irregular		Yes	Yes	0	7	FANCA	Yes	No			No	No
NCI-33-1c	30	12	Irregular	Provera, estrogen patch	No	Yes	0	8	FANCC	No	Yes	10	20	Vulvar, tongue	Scalp
NCI-61-1	33	14	Irregular	Combipatch	Yes	Yes	0	5	FANCA	Yes	No			Tongue, skin	Esophageal
NCI-144-1c	37	14	Irregular	Drospirenone ethinyl estradiol, nandronolone	No	Yes	0	1	FANCA	No	No			Vulvar with perianal spread	Recurrent perianal, finger

Abbreviations: NA, not applicable; UPN, unique patient number. Years are rounded down to the preceding integer.

^aTotal body irradiation.

^bRegular menses returned when androgen dose was decreased, subsequent to this study.

^cSuspected hypothalamic hypogonadism.

^dTotal abdominal irradiation.

Patients with FA had significantly decreased and mostly undetectable AMH levels (median 0.05 ng/mL; range 0–2.32 ng/mL; P < .001) compared with unaffected relatives (median 2.10 ng/mL; range 0.04–4.73 ng/mL) and unrelated healthy controls (median 1.92 ng/mL; range 0.31–6.64 ng/mL) (Figure 1, A–C). There were no significant differences between AMH levels in the unaffected relatives and unrelated healthy volunteers (P = .87). When subjects were restricted to pubertal females over the age of 10 years, patients with FA (n = 15; median age 20 y) had significantly decreased (P < .001) and mostly undetectable levels of AMH (median 0 ng/mL) compared with unaffected relatives (n = 18; median 2.25 ng/mL; median age 34 y) and unrelated healthy controls (n = 21; median 1.92 ng/mL; median age 27 y) (Figure 1D). All postpubertal females with FA had AMH concentrations less than 0.26 ng/mL with the exception of two patients (aged 19 and 20 y) whose AMH levels were 4- and 6-fold higher (1.18 and 1.64 ng/mL) than those of any other postpubertal patient. Interestingly, these two women did not have any FA-related congenital anomalies, whereas those with multiple anomalies had negligible or undetectable AMH levels (Table 2).

Figure 1.

AMH levels in all subjects stratified by age. A, All FA patients (n = 22). B, All FA relatives (n = 20). C, All unrelated healthy controls (n = 21). D, All pubertal and postpubertal subjects older than 10 years of age. Numbers include the following: FA patients, n = 15; FA relatives, n = 19; unrelated healthy controls, n = 21. Horizontal line within each boxplot represents the median AMH level. Upper and lower lines of the box represent the 25th and 75th percentiles. Significant differences (*, P < .001) were found between the median AMH level in FA patients compared with each of the control groups.

Five of 22 patients with FA underwent bone marrow transplantation (BMT) prior to AMH measurement (1, 6, 6, 16, and 20 y prior to sampling) at age 10 years or younger (Table 2). Of these, three received total body and one received total abdominal irradiation as part of their BMT conditioning regimen. Four patients achieved puberty after BMT, with three having menses. One patient was diagnosed with POI at age 12 years and the remaining patient is still prepubertal. When comparing patients who underwent BMT (n = 5; age at study 11, 12, 14, 23, 30 y) with those of comparable ages who did not undergo BMT (n = 4; age at study 11, 16, 22, 27 y), AMH levels were not significantly different (P = .73) (Table 2).

All 10 women with FA over the age of 18 years had irregular periods or had become amenorrheic except for the two who lacked FA-related anomalies and had AMH levels greater than 1 ng/mL. Seven of the 10 were sexually active. Three patients consulted a doctor after failing to conceive for more than 1 year. All three women seeking fertility were experiencing irregular menses and had undetectable AMH levels (0 ng/mL). Transvaginal ultrasound in two of these patients showed diminished numbers of antral follicles (less than eight total) than expected for their age group (6). The third woman wishing to conceive had active cancer, which precluded childbearing; thus, AFC was not obtained. Two other patients developed squamous cell cancers (SCCs) of the vulva. The three oldest patients had cancer at the time of study followed by incident or recurrent disease: one at age 30 years (vulvar and tongue, followed by skin SCC), one at age 33 years (tongue and skin, followed by esophageal SCC), and one at age 37 years (vulvar with perianal spread, which recurred, as well as a finger SCC). Three other patients developed cancer after AMH measurement: ages 24 years (vulvar), 25 years (skin), and 30 years (breast) (Table 2). Unaffected female relatives of a similar age did not have cancer.

None of the women with FA in this series ever became pregnant, whereas 12 of 14 postpubertal unaffected relatives reported two to six pregnancies each (median three pregnancies; median age at birth of first child 23 y; range 17–26 y). Four of the five oldest women with FA reported hot flashes and vaginal dryness, indicative of perimenopausal hypoestrogenemia. All five of these women had undetectable AMH levels and had POI as confirmed by elevated serial FSH measures, irregular menses, and absent fertility (11).

Discussion

This cohort study demonstrated that AMH levels in patients with FA were low in prepubertal girls, nearly absent in pubertal girls prior to menarche, and nearly absent or undetectable in most postmenarchal adolescent girls and women. In contrast, similarly aged unaffected female relatives and unrelated healthy controls had normal AMH levels. Ovarian reserve, as measured by AMH, was compromised in FA, regardless of the number of FA-related abnormalities, history of bone marrow failure, history of BMT, hormonal therapy, hematopoietic somatic mosaicism, or FA genotype. Our data suggest that most females with FA fail to produce appropriate levels of AMH at any time in their lives. These findings suggest that ovarian defects are a common factor in this otherwise heterogeneous clinical disease.

Initially, we anticipated that BMT would contribute to low AMH levels in our transplanted patients because chemotherapy and radiation regimens administered prior to BMT are known to decrease AMH levels, ovarian function, and fertility (21). However, similarly aged, nontransplanted patients also had low or undetectable AMH levels, suggesting that AMH deficiency occurred, regardless of BMT status. It is likely that BMT did not further affect AMH levels in our patients with FA because their ovarian reserve was compromised prior to the secondary insult of BMT conditioning. A low pregnancy rate has been reported among women with FA who underwent BMT (10%) and for those who did not undergo BMT (15%) (21, 22). The patients with FA who reported pregnancy had less severe disease as suggested by increased life span and less severe bone marrow failure (22). None of the patients in our cohort had been pregnant; however, all unaffected relatives seeking fertility achieved pregnancy. The two patients in our study who had AMH levels greater than 1 ng/mL also had regular menses and no FA-related congenital anomalies and did not require BMT. Even though these two patients were not seeking fertility, they appear to be less severely affected with FA, similar to the females with FA in the literature who achieved pregnancy.

Our study has several strengths as a result of the well-characterized NCI cohort of FA patients and unaffected relatives used in this study. All but two FA patients and relatives underwent a medical evaluation at the National Institutes of Health Clinical Center with longitudinal clinical follow-up. Additionally, our study included two control groups comprised of genetically related (unaffected relatives) and genetically unrelated (healthy volunteers) females. Conversely, our study was limited by the lack of transvaginal ultrasound AFC assessment in all postpubertal patients with FA, although others have reported a strong correlation between AMH levels and AFC (7). About half of our cohort's postpubertal women with FA were not sexually active, making vaginal ultrasound impractical. Transvaginal ultrasound was performed in only two patients desiring pregnancy. Importantly, women with FA may be less likely to attempt pregnancy because of the severity and continued treatment of hematopoietic disease, complications of BMT, use of androgens, development of gynecological or other malignancies at a young age, chronic illness, and social factors associated with FA, which may hamper their desire or ability to form intimate relationships. Our sample size is small due to the rarity of this syndrome; however, considering the low frequency of disease, our study reports on a very reasonable sample size using the NCI cohort. Additionally, our cohort of patients with FA is enriched for cases diagnosed in childhood as a result of obvious FA-related anomalies or bone marrow failure. The age distribution among unaffected relatives was skewed toward women over 30 years, most of whom were mothers of patients with FA. Although age matching would be preferential, the skewed design likely strengthens our findings by biasing controls to have lower AMH levels. Another limitation is that the AMH measurement was cross-sectional. Longitudinal assessments of AMH levels beginning at FA diagnosis are needed to corroborate our findings and to better understand AMH and the associated ovarian defect(s) observed in females with FA.

Women with FA are typically diagnosed with POI and cancer in their twenties and thirties, when the normal population is at peak fertility and health (1, 7). Genetic mutations and inactivation within the FA/BRCA DNA repair pathway in the general population also result in increased risks of cancers of the cervix, head and neck, ovary, and breast and influence the human follicular pool (3, 23, 24). Perhaps different mutations (null or hypomorphic) within FANC genes may be associated with the severity of infertility and AMH deficiency in patients with FA. Recently AMH has been shown to inhibit cell growth and metastasis in human breast, ovarian, endometrial, and cervical cancer cell lines and tumor development in mouse models of ovarian cancer (25). Low AMH may be associated with increased cancer risk because women with primary infertility have a 2- to 3-fold increased risk of developing cancers of Müllerian origin (26). Because low AMH levels are indicative of reduced fertility, those with low AFC and AMH levels associated with difficulty conceiving naturally during their reproductive years may also harbor an increased cancer risk. The recently elucidated anticancer properties of AMH warrant further research to determine whether AMH deficiency contributes to increased cancer risk in FA and, if so, whether AMH replacement could be a potential nontoxic adjuvant for cancer treatment in this population (27).

This study is the first to demonstrate AMH deficiency in females with FA. Most FA females who were later diagnosed with POI reached puberty naturally as adolescents, suggesting that patients with FA have a functional hypothalamic-pituitary-ovarian axis and produce sex hormones, despite negligible AMH levels (5, 28). Furthermore, because AMH levels do not significantly fluctuate across the menstrual cycle, AMH can be assessed at any time (29, 30), which is especially useful for females with FA, who largely experience irregular or absent menses. Thus, AMH measurement at the time of FA diagnosis and monitoring AMH levels thereafter may be useful for the management of complications associated with current or impending POI (infertility, osteoporosis, menopausal symptoms, etc), even in young prepubertal girls (10, 11, 31).