Correlation of Telomere Length and Telomerase Activity with Occult Ovarian Insufficiency

Abstract

Background: Occult ovarian insufficiency is associated with infertility, impaired response to ovarian stimulation, and reduced live birth rates in women treated with assisted reproductive technologies. Although a decline in ovarian follicle number is expected with age, the proximate causes of occult ovarian insufficiency in young women remain poorly understood. Abnormalities in telomere length and telomerase activity in human granulosa cells may serve as molecular markers for this condition.

Methods: A cross-sectional study was performed. Subjects (37 yr old or less) undergoing in vitro fertilization were classified as cases of occult ovarian insufficiency or controls with mechanical infertility (male or tubal factor). Granulosa cells were acquired at the time of oocyte retrieval to quantify telomere length and telomerase activity.

Results: Fifty-four women were enrolled. Human granulosa cell telomerase activity was demonstrated, and lack of granulosa cell telomerase activity was associated with occult ovarian insufficiency (odds ratio, 11.0; 95% confidence interval, 1.3–495.6; P = 0.02). Telomeres were shorter in women with occult ovarian insufficiency than in controls (relative telomere/single copy gene ratio, 1.88 vs. 3.15; P = 0.039).

Conclusions: Aberrant telomere homeostasis is associated with occult ovarian insufficiency in young women. This finding is consistent with the presence of telomeric attenuation that has been shown in multiple age-related conditions.

Occult ovarian insufficiency is associated with shortened telomeres and diminished telomerase activity in human granulosa cells.

Primary ovarian insufficiency describes a spectrum of declining ovarian function and reduced fecundity due to a decrease in initial follicle number, an increase in follicle destruction, or poor follicular response to gonadotropin stimulation (1,2). Three categories of ovarian insufficiency have been described that span a continuum from least to most severe dysfunction. Reduced fecundity, regular cycles, and premenopausal FSH levels characterize occult ovarian insufficiency. The development of elevated FSH levels indicates the presence of biochemical ovarian insufficiency. Women with overt ovarian insufficiency have both elevated FSH levels and amenorrhea (1,3). Diminished ovarian reserve (DOR) is a term that has been used to characterize women at risk for poor performance with assisted reproductive technologies (4,5,6,7,8,9). The term is used to predict prognosis as a function growing follicle number and the reproductive potential of harvested oocytes (4,5,6,7,8,9). Compared with other infertility diagnoses, DOR is associated with the lowest odds of pregnancy with assisted reproductive technologies because affected women develop few oocytes with hyperstimulation and have low implantation rates after embryo transfer (3,4,7,10). Poor responsiveness to gonadotropin stimulation as seen in DOR patients may also have serious consequences beyond refractory infertility and has been shown to predict early menopause in several series (11,12,13). Primary ovarian insufficiency has been suggested as the preferred term to describe DOR because it allows for an accurate characterization of the severity of ovarian dysfunction regardless of the underlying etiology. Accordingly, we have chosen to use occult ovarian insufficiency throughout this report to capture the fundamental clinical features of DOR: regular menstrual cycles, reduced fecundity, and elevated FSH levels below the menopausal range.

Only a small number of genetic and molecular pathways have been described that cause derangements in human follicle number and function severe enough to result in primary ovarian insufficiency (1,2,14,15,16,17,18). Fewer still are the number of specific environmental and chemical exposures with evidence of toxicity to ovarian follicles sufficient to cause early menopause (1,19,20,21,22). The establishment of the initial follicular cohort—a critical element in the determination of ovarian function and reproductive life span—appears to be governed by two key factors: the efficiency of germ cell proliferation in utero and the trajectory of follicular depletion thereafter (1,2,14). It is universally accepted that the peak number of approximately 6 million germ cells is achieved at 20 wk gestation and that this number can never be replenished (1,23,24,25). Follicle number decreases continually from this point forward due to programmed cell death of oocytes and granulosa cells (14,26,27). Any individual who achieves less than the peak number of follicles (as in blepharophimosis, ptosis, epicanthus inversus syndrome) (5,14,15,16,17) and/or experiences extremely rapid follicular atresia (as in Monosomy X, chemotherapy exposure, or Fragile-X mental retardation-1 premutation carriers) (18,19,20,21,22,23,24,25,26,27,28,29) is at risk for premature ovarian insufficiency. Unfortunately, the majority of young women with ovarian insufficiency do not have a discrete cause that can be identified.

We have chosen to consider the association of granulosa cell telomere homeostasis with occult ovarian insufficiency in women 37 yr of age and younger. Telomeres consist of repetitive arrays of (5′-TTAGGG-3′)n that cap the ends of all eukaryotic chromosomes (30,31,32). Telomeres shorten with successive rounds of DNA replication because most somatic cells lack the proper machinery to replicate the ends of lagging strand DNA and replenish telomere repeats (30,31,32,33,34,35). Over successive cell divisions, this results in critically short, dysfunctional telomeres that trigger replicative senescence through p53/RB intracellular pathways (36,37). Telomere attrition has been postulated to play a role in the development of age-related disease by mediating the onset of cellular senescence and imposing a limit on the proliferative life span of certain cell types (31,33,34,38,39,40).

One of the critical components of telomere homeostasis is the activity of telomerase, a ribonucleoprotein enzyme capable of extending telomeres in cell populations such as embryonic, germline, and stem cells (34,35,41,42). Moreover, telomerase has been shown to enhance the proliferation of certain stem cell populations independent of its role in maintaining telomere length (42,43,44).

In the human, normal ovaries express telomerase (45), and human oocytes and embryos have been found to exhibit telomerase activity (46,47). Lavranos et al. (48) demonstrated telomerase activity in the granulosa cells of bovine follicles and suggested that active telomerase permits the proliferation of granulosa cells required for proper follicle formation.

We hypothesized that telomere length homeostasis in human granulosa cells is associated with the presence of occult ovarian insufficiency. To test this hypothesis, granulosa cell telomere length and telomerase activity were compared between cases of ovarian insufficiency and controls with mechanical infertility (male factor or tubal factor).

Subjects and Methods

Subjects and clinical procedures

Cases of occult ovarian insufficiency were defined as women age 37 yr or less with infertility and a history of FSH elevation (in the premenopausal range) before initiation of in vitro fertilization (IVF). An early follicular phase (d 2–4 of the menstrual cycle) FSH of 11.4 IU/liter or greater in conjunction with an estradiol level of less than 293.6 pmol/liter was used to diagnose occult ovarian insufficiency. This cutoff was based on data from our population of IVF patients demonstrating very few pregnancies in women with FSH values above this level (49). Controls were women undergoing IVF with normal FSH testing and either male factor infertility (abnormal semen analysis on more than one occasion) or tubal factor infertility (history of ectopic pregnancy, bilateral tubal ligation, bilateral salpingectomy, or tubal obstruction on hysterosalpingogram). Immunoassays to measure FSH and estradiol were performed on the Siemens Immulite 2500 (Siemens Healthcare Diagnostics, Deerfield, IL). The sensitivity for FSH using this assay is 0.5 IU/liter, and the total coefficient of variation is 4.8%. For the estradiol assay, the sensitivity is 110 pmol/liter, with a coefficient of variation of 6.2%.

This study was approved by the Institutional Review Board of the Hospital of the University of Pennsylvania. Women undergoing IVF treatment at Penn Fertility Care were approached at cycle initiation to provide informed consent to participate in the study. The stimulation regimen for subjects involved the administration of recombinant FSH (Gonal F; Serono Pharmaceuticals, Rockland, MA) in combination with pituitary down-regulation using a GnRH agonist (Luprolide Acetate; TAP Pharmaceuticals, Lake Forest, IL) starting either in the midluteal phase or early follicular phase (minidose flare). Follicular fluid was aspirated from follicles at the time of ultrasound-guided transvaginal oocyte retrieval. Granulosa cells were isolated from the follicular fluid.

Granulosa cell isolation

Total fluid from all follicular puncture sites at the time of oocyte retrieval was pooled for each subject. The follicular fluid was centrifuged for 15 min at 300 × g and a temperature of 4 C to recover cellular material. The cells were layered over of a 50% Ficoll gradient (Ficoll-Paque; Amersham Biosciences, Buckinghamshire, UK) and centrifuged at 300 × g for 15 min at 4 C. Granulosa cells were isolated from the interface layer, washed once with PBS, and then resuspended in PBS. This process resulted in a population of cells with minimal lymphocyte contamination (less than 5%). The cells were stored at −80 C until use.

Assessment of telomere length

DNA was extracted from granulosa cells using the DNeasy Tissue Kit (QIAGEN Inc., Mississauga, Ontario, Canada). Thereafter, telomere length was assessed using a modification of the quantitative PCR method described by Cawthon (50,51,52,53,54).This technique measures the factor by which the ratio of telomere repeat copy number to single-gene copy number differs between an unknown sample and a reference DNA sample. As a result, this method generates a relative Telomere/Single Copy Gene (T/S) ratio that is proportional to average telomere length. For each specimen, two PCRs were carried out—one to amplify telomere repeats (Tel PCR), and the other to amplify the β-globin gene (S PCR). The SYBR Green Jumpstart Taq Ready Mix (Sigma Aldrich, St. Louis, MO) was used for all quantitative PCRs. Each reaction was carried out in triplicate using the Opticon 2 detector system on the MJ Research PTC200 DNA Engine thermal cycler (Waltham, MA). Before starting PCR, a mixture of sample DNA (5 ng), Escherichia coli DNA (40 ng), 10X Taq polymerase buffer (magnesium free), and diethylpyrocarbonate water was made and heated at 94 C for 30 min to fully denature the DNA.

For telomere PCR, the following conditions and primers were used; 100 nm of primer Tel 1b (5′-CGGTTTGTTTGGGTTTGGGTTTGGGTTTGGGTTTGGGTT-3′) and 900 nm of primer Tel 2b (5′-GGCTTGCCTTACCCTTACCCTTACCCTTACCCTTACCCT-3′) were used to amplify sample DNA under the following cycling conditions: an initial polymerase activation step at 94 C for 1 min followed by 30 cycles of 95 C for 15 sec and 56 C for 1 min. For β-globin gene PCR, 300 nm of primer hgb 1 (5′-CTTCTGACACAACTGTGTTCACTAGC-3′) and 700 nm of primer hgb 2 (5′-CACCAACTTCATCCACGTTCACC-3′) were used to amplify sample DNA with the following cycling conditions: an initial polymerase activation step at 94 C for 1 min, followed by 33 cycles of 95 C for 15 sec, 58 C for 20 sec, and 72 C for 20 sec. Negative controls using water in lieu of sample DNA were processed during each PCR run. Serial dilutions (seven concentrations from 25 ng to 1 ng in water) of reference DNA were used to construct standard curves in both Tel PCR and β-globin gene PCRs, allowing for relative quantification of samples.

Method of telomerase activity assessment

Telomerase activity in samples was measured using a PCR-based assay (TRAPeze Telomerase Detection Kit; Millipore, Billerica, MA). Granulosa cells were first exposed to 1X CHAPS lysis buffer to extract protein. Protein concentration of each extract was determined using the Bradford method after establishing a standard curve of protein concentration. Extracts with a concentration range of 10–750 μg were suitable for assay. Each extract was then combined with TRAPeze reagents to perform the two-step PCR-based assay. Because telomerase is a heat-sensitive enzyme, each sample extract was heat treated before PCR to generate a sample-specific negative control. Polyacrylamide gel electrophoresis was used to distinguish telomerase extension products amplified by PCR. Telomerase activity (positive or negative) was determined based on the presence or absence of a characteristic laddering pattern of telomerase products compared with positive and negative controls (Fig. 1). The sensitivity of this assay under optimal conditions is 50 telomerase-positive cells.

Figure 1.

TRAP assay results in three test samples using polyacrylamide gel electrophoresis. Telomerase activity was demonstrated in two control samples (and absent in heat-inactivated negative controls) and absent in a sample from a case with occult ovarian insufficiency. Lane 1, Telomerase-positive control cells; lane 2, buffer, negative control; lanes 3 and 4, case subject: heat-inactivated extract and untreated extract; lanes 5 and 6, control subject: heat-inactivated extract and untreated extract; lanes 7 and 8, control subject: heat-inactivated extract and untreated extract.

Statistical analysis

Continuous variables were described as means (± sd or ± sem) if they were normally distributed or medians if they were not. Graphical methods were employed to evaluate the normality of all continuous variables. Because the relative T/S ratios denoting telomere length were not normally distributed in this population, these values were log-transformed to use parametric statistical tests to interpret these data.

Associations between categorical variables were tested using the Fisher’s exact test. Comparisons of continuous variables between groups were made using the Student’s t test or the Wilcoxon rank sum test where appropriate. Spearman’s coefficient was used to determine correlations between continuous variables. A P value of less than 0.05 was considered statistically significant for nominal comparisons of variables, and a Bonferroni correction was applied to minimize the chance of a type I error in the multiple comparisons made in IVF cycle outcomes. Odds ratios (ORs) were calculated to determine the association between telomerase activity (present or absent) and occult ovarian insufficiency. Logistic regression was used to create a model for occult ovarian insufficiency and control for multiple confounders. All statistical tests were performed using STATA 9 software (College Station, TX).

The following definitions were used to categorize pregnancy outcomes. Positive pregnancy was defined as a serum B-human chorionic gonadotropin level greater than 5 mIU/ml measured 2 wk after oocyte retrieval, and clinical pregnancy was defined as the presence of a gestational sac on ultrasound after a positive pregnancy test. A live birth was defined as delivery of a viable fetus at or after 24 wk gestation.

Our sample size of 54 patients was based on the following assumptions: a difference of 45% in the proportion of cases lacking telomerase compared with controls, 80% power, α error of 0.05, and a 3:1 ratio of controls to cases.

Results

A total of 54 patients consisting of 12 cases of occult ovarian insufficiency and 42 controls undergoing IVF for the first time at Penn Fertility Care were enrolled for study participation. Baseline characteristics of the study population are listed in Table 1.

Table 1.

Baseline characteristics of study participants

	Occult ovarian insufficiency	Controls
n	12	42
Median age (range)	34.5 (30–37)	33 (23–37)a
White	10 (83%)	28 (67%)a
Non-white	2 (17%)	14 (33%)a
Median gravity (range)	0 (0–2)	0 (0–5)a
Median parity (range)	0 (0–1)	0 (0–3)a

^aP = Not significant. 

IVF cycle outcomes between cases and controls were compared. Numbers of follicles, numbers of oocytes, maximum estradiol concentration, and total amount of gonadotropin administered were significantly different between the groups (Table 2). When the adjusted P value for the 13 comparisons made in IVF cycle outcomes was applied (new P value = 0.004 or 0.05/13), all comparisons remained statistically significant except for the number of mature oocytes, number of atretic oocytes, and number of eight-cell embryos.

Table 2.

IVF cycle outcomes in cases compared to controls

	Occult ovarian insufficiency	Controls
n	12	42
Median maximum pre-IVF FSH (range)	14.3 IU/liter (11.4–21.6)	6.05 IU/liter (3.9–9.9) (P = 0.0001)
Median total units of total gonadotropin administered (range)	6,487.5 (2,587.5–6,900)	2,887.5 (1,537.5–5,775) (P < 0.0001)
Mean maximum estradiol ± sd	8,006.8 pmol/liter ± 3,786	13,853.1 pmol/liter ± 5,627.2 (P = 0.002)
Mean no. of follicles ± sd	11.4 ± 5.2	24 ± 9.7 (P = 0.0001)
Median no. of oocytes retrieved (range)	8.5 (2–17)	16.5 (3–29) (P = 0.0004)
Mean no. of mature oocytes ± sd	6.8 ± 4.2	11.4 ± 5.8 (P = 0.01)
Mean percentage fertilization	70%	69% (P = 0.37)
Median atretic eggs (range)	0 (0–2)	1 (0–9) (P = 0.03)
Median no. of eight-cell embryos (range)	0 (0–4)	1 (0–6) (P = 0.04)
Pregnancy rate per retrieval	33.3%	45.2% (P = 0.53)
Clinical pregnancy rate per retrieval	25%	40.5% (P = 0.5)
Miscarriage rate	33.3%	17.6% (P = 0.51)
Live birth rate per retrieval	16.7%	33.3% (P = 0.47)

The IVF response data for the 12 cases of occult ovarian insufficiency are elaborated in detail in Table 3. These figures highlight two key features of ovarian insufficiency: 1) the challenge of consistently achieving significant numbers of mature oocytes even when considerable gonadotropin stimulation is used; and 2) the poor pregnancy outcomes in this population.

Table 3.

Demographics and IVF cycle characteristics of subjects with occult ovarian insufficiency

Subject no.	Age	Race	Peak FSH (IU/liter)	Pretreatment FSH (IU/liter)	Total units of gonadotropin administered	Peak estradiol (pmol/liter)	Oocytes	Mature oocytes	Atretic oocytes	Embryos	Live birth
1	37	White	15	5.9	6900	7,735	5	3	0	0	No
2	36	White	15	8.1	6300	9,306	10	10	0	9	No
3	30	White	14.9	5.8	6900	13,531	17	16	0	13	No
4	37	White	21.6	13.3	6675	14,269	12	10	1	9	No
5	35	White	11.6	6.2	7500	6,178	11	11	2	7	No
6	36	White	14.3	8.9	5550	6,578	3	3	0	2	No
7	33	White	14.3	5.1	6900	3,616	8	6	2	5	No
8	34	Asian	26.1	7.7	6300	4,064	2	2	0	1	No
9	30	White	11.4	6.1	2587.5	10,712	7	7	0	6	Yes
10	33	White	12.7	12.7	6900	4,460	9	6	0	5	No
11	36	Asian	12.4	10.3	5775	8,946	11	9	1	4	Yes
12	34	White	12.6	10.5	5700	3,106	2	1	0	1	No

There was a significant relationship between telomerase activity and the odds of occult ovarian insufficiency (Table 4). Women lacking granulosa cell telomerase activity were 11 times more likely to have occult ovarian insufficiency than those who exhibited telomerase activity [OR, 11; 95% confidence interval (CI), 1.3–495.6]. Granulosa cell telomeric shortening was also associated with occult ovarian insufficiency. Average telomere length (expressed as the log-transformed relative T/S ratio) was 1.88 in cases and 3.15 in controls (P = 0.03). Logistic regression was used to further characterize the influence of telomerase inactivity on the odds of occult ovarian insufficiency while controlling for telomere length, race (white vs. non-white), and categorical age (<35, 35–37). After adjusting for these factors, women lacking telomerase activity were 14.5 times more likely to have occult ovarian insufficiency than with telomerase activity (OR, 14.5; 95% CI, 1.4–149.5). There were no significant interactions between variables in the explanatory model.

Table 4.

Telomere length, telomerase activity, and OR for occult ovarian insufficiency based on absent telomerase activity

	Telomere length ± sem	No telomerase activity detected (%)	OR (95% CI)	Adjusted OR (95% CI)
Occult ovarian insufficiency (n = 12)	1.88 ± 0.69a	11 of 12 (92%)	11 (1.3–495.6)b	14.5 (1.4–149.5)c
Controls (n = 42)	3.15 ± 0.25	21 of 42 (50%)	Reference	Reference

^aP = 0.039, for difference in mean telomere length compared to controls. 

^bP = 0.02, OR for occult ovarian insufficiency based on absent telomerase activity. 

^cP = 0.024, OR for occult ovarian insufficiency based on absent telomerase activity adjusted for age, race, and telomere length. 

IVF cycle outcomes based on the presence or absence of telomerase are described in Table 5. There was a statistically significant association between lack of telomerase and mean number of follicles produced as well as median pretreatment FSH concentration. However, when the adjusted P value for the 13 comparisons made in IVF cycle outcomes was applied (new P value = 0.004), these differences were no longer statistically significant. Several other relevant outcomes, including total dose of gonadotropin administered, number of oocytes retrieved, number of atretic oocytes, percentage fertilization, and pregnancy rate per retrieval demonstrated a trend toward a significant relationship with absent telomerase activity.

Table 5.

IVF cycle outcomes in subjects with and without telomerase activity

	Telomerase activity present	Telomerase activity absent
n	22	32
Median maximum pre-IVF FSH (range)	5.7 IU/liter (3–14.3)	7.7 IU/liter (3.5–26.1) (P = 0.04)
Median total amount of total gonadotropin administered (units) (range)	2,887.5 (1,387.5–6,900)	4031.3 (1,537.5–7,500) (P = 0.08)
Mean maximum estradiol	13,157 pmol/liter ± 6,102.1	12,140 pmol/liter ± 5,614.4 (P = 0.55)
Mean follicles± sd	24.8 ± 8.9	19.1 ± 10.6 (P = 0.05)
Median oocytes (range)	19.5 (6–33)	12 (2–41) (P = 0.06)
Mean no. of mature oocytes ± sd	11.9 ± 5.3	9.3 ± 5.9 (P = 0.12)
Median atretic oocytes (range)	2.4 (0–9)	1 (0–7) (P = 0.07)
Mean percentage fertilization	78%	66% (P = 0.058)
Median no. of eight-cell embryos (range)	1.5 (0–5)	0.5 (0–6) (P = 0.1536)
Pregnancy rate per retrieval	59.1%	31.3% (P = 0.054)
Clinical pregnancy rate per retrieval	50%	28.1% (P = 0.15)
Miscarriage rate	20%	15.4% (P = 1.0)
Live birth rate per retrieval	40.9%	21.2% (P = 0.23)

Conversely, there were no statistically significant associations between mean granulosa cell telomere length and any of the IVF cycle outcomes described above (data not shown).

The correlation between total amount of gonadotropins administered during IVF and telomere length was assessed using the Spearman correlation coefficient. The Spearman’s rho was −0.0747 (P = 0.59).

Discussion

Although multiple investigations have examined the role of telomere-induced cellular dysfunction in age-related disease, this is the first in the English literature to evaluate the relationship between telomere length, telomerase activity, and occult ovarian insufficiency using human granulosa cells (55). Our finding of shortened telomeres in women with occult ovarian insufficiency is consistent with several studies that have demonstrated an association between telomere attrition and reproductive senescence (56,57). Compared with these investigations in which telomeres were measured in peripheral blood leukocytes, we investigated telomere length in the human follicle.

The association between occult ovarian insufficiency and granulosa cell telomeric shortening is likely multifactorial. It is possible that intrinsic telomere length differences exist in women with ovarian insufficiency, making their granulosa cells more likely to begin proliferation with shorter telomeres than women with normal ovarian function. Alternatively, telomeric attrition in young women with ovarian insufficiency could be a function of the kinetics of granulosa cell division in human follicles. Limited granulosa cell proliferative capacity has been demonstrated as a feature distinguishing women with ovarian insufficiency from those without it (58,59). In the follicles of women with ovarian insufficiency, more divisions may be required per competent granulosa cell to achieve a total population doubling. A greater number of replications per cell would naturally result in accelerated telomere erosion in the cells capable of dividing. The significantly higher amount of gonadotropin administered to women with occult ovarian insufficiency does not appear to explain the telomeric attrition in this group. The Spearman coefficient for correlation between total amount of gonadotropin and telomere length was very weak and not statistically significant.

The association between lack of granulosa cell telomerase activity and occult ovarian insufficiency could represent a mechanism by which telomere erosion is accelerated and replicative capacity is impaired. Telomerase activity is present in regenerative tissues with significant proliferative requirements where telomere loss would be poorly tolerated (34,41,48). Ovarian follicular development in humans requires an explosive increase in granulosa cell number from as few as three progenitors per follicle to tens of thousands of cells before ovulation (60). In the bovine model, telomerase activity is absent in pregranulosa cells of primordial follicles, activated when follicles begin to grow, and highest in preantral and small antral follicles with the most robust cell proliferation; telomerase activity declines in later stages of follicular development (46). Insufficient activation of telomerase in the early stages of follicular development could hasten telomere attrition and severely compromise cellular functions required for proper follicular maturation. Furthermore, because telomerase expression appears to have a role in supporting stem cell proliferation independent of its role in replenishing telomeres (42,43,44), its absence in association with occult ovarian insufficiency further supports its strength as a marker of granulosa cell dysfunction.

IVF outcomes as a function case status were also investigated. As expected, subjects with occult ovarian insufficiency achieved significantly lower estradiol levels, fewer numbers of follicles, and fewer oocytes than controls, despite receiving more gonadotropin stimulation during IVF (Table 2). Cases also produced fewer eight-cell embryos than controls (median of 0 and 1, respectively). Conversely, women with ovarian insufficiency had slightly fewer atretic oocytes than controls (median of 0 and 1, respectively). Given the relatively modest size of our sample, it is possible that chance played a role in the comparisons of atretic oocytes and eight-cell embryos between groups.

Clinical pregnancy rates, miscarriage rates, and live birth rates per retrieval were all lower in women with ovarian insufficiency than controls. Although these differences did not reach statistical significance, each outcome was more favorable in controls and was consistent with the published literature showing poor pregnancy rates in women with ovarian insufficiency (3,4,5,8,9). We cannot definitively conclude from the analysis of our data that women with ovarian insufficiency have diminished oocyte quality compared with unaffected women of comparable age. Whether or not young women with occult ovarian insufficiency have poor oocyte quality and greater odds of generating aneuploid embryos is an area of debate with some investigators suggesting a relationship (61,62,63,64) and others refuting this (65,66,67).

The association between telomerase activity and IVF outcomes was also tested. All of the outcomes measured (Table 5) were more favorable in subjects with telomerase activity than in those without it. Moreover, FSH values in women with telomerase activity were lower than in those lacking enzyme activity (5.7 and 7.7 IU/liter, respectively; P = 0.04). This finding suggests that subtle differences in biomarkers can have significant molecular correlates.

A strength of this investigation is the restriction of subjects to women age 37 or younger. The choice of this age criterion was based on the desire to investigate a population of cases who we believed were affected with occult ovarian insufficiency prematurely rather than those with physiological ovarian aging. We felt that using 37 as an upper age limit generates a population of cases in which the definition of occult ovarian insufficiency is unambiguous and the interaction of age is minimized.

The cross-sectional nature of this study imparts a limit on our ability to draw conclusions about the timing of changes in telomere homeostasis relative to the onset of ovarian insufficiency. Furthermore, we obtained our molecular exposures of interest from luteinized, terminally differentiated granulosa cells which are morphologically distinct from proliferating granulosa cells in early stage follicles. However, justification for our approach resides in the fact that occult ovarian insufficiency is a rare condition and that using granulosa cells obtained at the time of IVF is an established approach to investigating molecular pathways in infertile women (58,59,60,68,69). In our center, the prevalence of occult ovarian insufficiency in women pursuing IVF has ranged between 8 and 9% in women 37 yr old or less in recent years. Data from the Society for Assisted Reproductive Technology also provide reliable national estimates for the prevalence of occult ovarian insufficiency in infertile women pursuing IVF (categorized as DOR in the statistics compiled by the Centers for Disease Control and Prevention). The prevalence of occult ovarian insufficiency across all age categories has ranged from 3–9% over the past 10 yr (10,70,71,72,73,74,75,76).

In summary, granulosa cell telomeric shortening and diminished telomerase activity are associated with occult ovarian insufficiency in young women. The results of this investigation may have implications for women’s health beyond the realm of fertility outcomes. The follicular atresia that drives ovarian dysfunction is irreversible; only the rate is variable. If young women affected with occult ovarian insufficiency develop and maintain an accelerated rate of follicular loss, they may be at risk for frank ovarian failure and its clinical sequelae at earlier than expected ages. Ultimately, it is hoped that in studying granulosa cell function further, a better understanding of both oocyte quality and molecular reproductive dysfunction can be achieved. Further research aimed at the discovery of exposures that target important molecular pathways in human granulosa cells and that negatively influence granulosa cell competence could be an important step in predicting women at risk for poor reproductive outcomes, refractory infertility, and overt primary ovarian insufficiency.