Predict Value of Relative Telomere Length, TERT, and Follicular Fluid Anti-mullerian Hormone Levels for IVF Outcome in Infertile Women

Background: Telomere attrition has been shown to play a critical role in the reproductive aging process in human beings. Telomere length (TL) is normally regulated by telomerase enzyme. Telomerase reverse transcriptase (TERT) is the main component of the telomerase. Anti-Mullerian hormone (AMH), a member of the transforming growth factor superfamily, is derived from the granulosa cells of early developing pre-antral and antral follicles. The aim of this study was to evaluate the associations between relative telomere length (RTL), TERT expression of granulosa cells (GC), follicular uid (FF) AMH levels and ovarian/embryonic performance in infertile women at different age. Moreover, whether they acting as predictors for probability of clinical pregnancy were also assessed. Method: A total of 160 women underwent their ﬁ rst fresh cycle of in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) were included in our study as follows: 100 women were enrolled for RTL measurement and 60 women were enrolled for TERT measurement. All these 160 women underwent FF AMH measurement. Correlations between RTL(cid:0)TERT expression, FF AMH levels and age, ovarian/embryonic performance and probability of clinical pregnancy were assessed. Results: There was a statistically signicant relationship between the expression levels of TERT, RTL, FF AMH levels and patient age(r = −0.20, P = 0.04; r=0.30, P=0.02; r=-0.191, P=0.003, respectively). Relationships between the expression levels of TERT, FF AMH levels and oocytes yield were signicant (P<0.001; P<0.05, respectively). However, no statistically correlation was observed between the RTL of GC samples and oocytes yield. All these three biomarkers had no correlation with blastocyst formation rate. There was signicant relationship between FF AMH levels and probability of clinical pregnancy in patients older than 35 years (OR=1.284, 95%CI=1.031-1.599, P=0.026).


Background
In the past decades, because of social and economic factors, more and more women postponed their pregnancy until an advanced age [1], and the increased maternal age caused infertility and diminished ovarian reserve (DOR) [2]. It is well known that ovarian aging has an important impact on reproductive capacity, altering the hormonal equilibrium, decreasing the quality of the oocytes and worsening the reproductive outcomes [1]. Hypotheses for age related decline in reproductive capacity and ovarian reserve involve oocyte aneuploidy [3], spindle defects [1], oxidative stress response [4], mitochondrial dysfunction [5], autophagy de ciency [6], altered in epigenetics pattern [1] and so on. Telomere attrition has also been shown to play a critical role in the reproductive aging process in human beings [7].
Telomere shortening is proposed as one of the possible mechanisms leading the decrease in quality of oocytes as it may affect chromosome segregation and genome stability [1].
Telomeres are highly preserved nucleoprotein structures of 5-15 kb-long tandem repeat hexanucleotide sequence (TTAGGG) repeats that form protective caps at the ends of eukaryotic chromosomes [8][9], playing a vital role in the sequential cycles of cell division and maintenance of chromosomal stability [10]. Progressive shortening of telomeres within granulosa cells because of inadequate DNA repair and long-term oxidative stress has been implied as a main driving factor underlying reproductive aging [11].
Telomere length (TL) is normally regulated by telomerase enzyme, a reverse transcriptase that adds hexameric sequences (TTAGGG) to telomeres, thus protecting them from the progressive erosion during cell division. One of the key roles of telomerase is to delay programmed telomere shortening for the purpose of ensure an accurate synthesis of the DNA replication [12]. Telomerase consists of two fundamental elements: telomerase reverse transcriptase (TERT) and telomerase RNA component (TERC).
The former is the main component of the telomerase and the latter serves as template for telomeric DNA synthesis. As the main component of the telomerase, TERT with reverse transcriptase activity plays a signi cant role in synthesizing hexameric sequence at the end of chromosomes. The regulation manner of TERT expression in cells is tissue-dependent [13]. The activity of TERT prevents telomere shortening with cell division, preventing cell aging and thus allowing extensive self-renewal capacities [8].
Some ndings suggested that decreased TERT and telomere-binding protein (TRF1, TRF2 and POT1) expression may cause the telomere shortening of ovaries with age, which may be related to female fertility decline [14]. In mice model, the study showed that telomeres shortening of ovaries lead to elevated embryo fragmentation, increased levels of cellular arrest, degeneration and chromosomal abnormalities [15]. In human studies, in vitro matured oocytes from the germinal vesicle (GV) stage were used as experimental materials, and telomere length negatively predicted fragmentation in day 3 preimplantation embryos [16]. Studies also demonstrated that telomeres in polar bodies from women who conceived after in vitro fertilization-embryo transfer (IVF-ET) was longer compared with those who did not conceived [17]. However, some studies found no signi cant associations between granulosa cells relative telomere length (RTL) and any IVF outcomes [11].The exact correlation between telomere length/telomerase and IVF outcomes remains unclear and deserve further investigation. The granulosa cells (GC) surrounding the oocytes are considered markers for oocyte quality and competence [18], and they assist at the beginning of embryo development [19].This study examines the roles of both GC telomere length and telomerase gene TERT as potential biomarkers of ovarian and embryo performance.
Hormone regulation of the maintenance of telomeres by telomerase in the ovary affects cell proliferation and ovarian aging. Anti-Mullerian hormone (AMH), a member of the transforming growth factor superfamily, is derived speci cally from the granulosa cells of early developing pre-antral and antral follicles [20]. The decrease in AMH secretion as the cycle progresses releases the follicles from its inhibitory effect on folliculogenesis and premature maturation [21][22][23].
The aim of the present study is to investigate the associations between relative RTL, TERT expression of granulosa cells, FF AMH levels and ovarian/embryonic performance, probability of clinical pregnancy in infertile women at different age undergoing IVF or ICSI treatment.

Study subjects
A total of 160 women underwent their rst fresh cycle of IVF or intracytoplasmic sperm injection (ICSI) between March 2019 and December 2020 were recruited from Reproductive Center of the First A liated Hospital of Anhui Medical University. Among them, 100 women were enrolled for RTL measurement and 60 women were enrolled for TERT measurement. All the160 women underwent follicular uid AMH measurement.
The women met the following criteria: i) Age: =35 years old (young age group) or 35 years old (old age group); ii) the patients underwent their rst fresh cycle of IVF or ICSI. The exclusion criteria included: i)Patients with chromosomal abnormalities; ii) Patients with gynaecological disease, endocrine diseases and autoimmune diseases such as polycystic ovary syndrome, hyperprolactinaemia, endometriosis, adenomyosis, hyperthyroidism, systemic lupus erythematosus and so on; iii) Patients with history of radio-or chemotherapy, ovarian surgery; iv) Patients diagnosed with premature ovarian failure. Granulosa cells and follicular uid were collected from all participants. Clinical outcomes referred only to the rst fresh embryo transfer cycles or the rst frozen embryo transfer cycles in cases with no fresh transfer within 3 months after oocyte retrieval.
This study was approved and reviewed by the Ethics Committee of Anhui Medical University (Ethics Number: 20190228) and written informed consents were obtained from all participants.
Oocyte retrieval and sample collected Controlled ovarian stimulation was achieved using a standard long pituitary downregulation protocol with gonadotropin-releasing hormone agonists and recombinant follicle stimulating hormone (FSH, Gonal-F, Merck Serono, Germany). Oocyte retrieval was carried out with transvaginal ultrasound-guided needle aspiration. When at least three leading follicles reached 18 mm in diameter, 250µg recombinant human chorionic gonadotropin(r-hCG, Merck Serono, Germany) was administered. A pooled collection of follicular uid was obtained from each patient on the day of oocyte retrieval. The rst and last tubes of follicular uid were not selected because of containing follicle ushing uid (ASP, Vitrolife, Sweden). A tube of 10-15 ml volume of follicular uid was obtained from each enrolled patient then centrifuged for 15 min at 1500 rpm. The supernatants were collected and stored at -80℃ for future AMH measurement.
Identi ed cumulus oocyte complexes (COCs) were isolated from the pooled follicular uid and placed in a separate falcon 3001 dish. Excess granulosa cells were mechanically separated from the COCs with 1ml syringe needle (BD, USA). The separated cells were washed and resuspended in phosphate buffer saline (PBS). Surplus COCs were then moved to a falcon 3037 dish containing fertilization medium for continued culture. The granulosa cells were stored at − 80°C for measurement of telomere length and TERT expression.

Zygote, embryo assessment and Embryo transfer
Quality assessment of oocytes, zygotes and embryos was carried out according to routine guidelines described in our previous study [24]. Serum β-HCG levels were measured to con rm biochemical pregnancy 14 days after embryo transfer. Clinical pregnancy was de ned as a visible gestational sac by ultrasonography with a fetal heartbeat 35 days after embryo transfer. Luteal support was initiated with daily injections of 60 mg progesterone.
Relative Telomere length measurement of GC Genomic DNA was extracted from granulosa cells using DNeasy Tissue Kit (QIAGEN, Inc). The relative telomere length was determined by a modi ed quantitative polymerase chain reaction (Qpcr) method Telomere length was determined by calculating T/S ratio using ΔCt (Ct telomere/Ct single copy gene). The T/S ratio of each sample[x] was normalized relative to the mean T/S ratio of the reference sample [25].

TERT expression measurement of GC
For TERT gene measurement, RNA was extracted from granulosa cells using the RNeasy Mini kit (QIAGEN, Inc). RNA was reverse transcribed into cDNA, which was quanti ed by real-time PCR. QPCR SYBR Green Master Mix was used to detect the expression levels of TERT. Gene expression was quanti ed relative to the housekeeping gene GAPDH using the 2 −ΔΔCT method [27]. The following primer sequences were used for the TERT and GAPDH: TERT-F, 5'-CTCCCATTTCATCAGCAAGTTT − 3' and TERT-R, 5'-CTTGGCTTTCAGGATGGAGTAG − 3', and for the housekeeping gene GAPDH-F, 5'-GGAAGCTTGTCATCAATGGAAATC-3' and GAPDH-R, 5'-TGATGACCCTTTTGGCTCCC − 3'.
Follicular uid AMH levels measurement FF AMH concentrations were measured by enzyme-linked immunosorbent assay (ELISA), using a sensitive diagnostic kit for the quantitative detection of AMH, according to the manufacturer's instructions [28]. Standards covered a range from 0.06 to 18.0 ng/ml, and test accuracy was less than 10%. The intra-and inter-assay coe cients of variation were less than 10% and less than 15%, respectively.

Statistical Analysis
The normallities of data were distribution normally and were assessed by the Kolmogorov -Smirnov test. The relative telomere length was LG-transformed because of a skewed distribution. Descriptive statistics were expressed using mean ± SD for normally distributed variables, and Student's t-test was used for measurement data. A binary logistic regression model was used to determine the pregnancy probability.
Pearson correlations for normal distribution data or spearman correlations for skewed data were used respectively. A two-sided P < 0.05 was considered statistically signi cant. A linear regression model was

Results
Patient general characteristics of the study subjects 160 women range from 23 to 45 years old met our inclusion and exclusion criteria. 100 women were enrolled for RTL measurement and 60 women were enrolled for TERT measurement. All these 160 women underwent follicular uid AMH measurement. No clinical parameters, including age, BMI, D3 serum FSH, total gonadotropin (Gn) dose and IVF outcome measures distinguished among the RTL group, the TERT group and FF AMH group. Patient characteristics were described in Table 1.
Relative telomere length in relation to age, and ovarian/embryonic outcome Linear regression analysis revealed that a negative relationship between the RTL of GC samples and patient age, and the difference was statistically signi cant (P 0.05). This result was represented in  Table 1. The detail results of the statistical analyses evaluating the relationship between relative GC telomere length and outcome measures were displayed in Table 3.
Relative TERT expression in relation to age and ovarian/embryonic outcome  Fig. 3A). Linear regression analysis revealed that signi cant negative correlation between the FF AMH levels and patient age (P 0.01). This result was represented in Fig. 3B (FF AMH levels =-0.175×year + 11.286; R2 = 0.054; r=-0.191; P = 0.003). When evaluating the correlation between FF AMH levels and ovarian and embryonic outcome, a statistically signi cant positive relationship was observed between FF AMH levels and number of oocytes retrieved (r = 0.193; P < 0.05) and number of MII oocytes retrieved (r = 0.163, P < 0.05). A statistically signi cant negative relationship was observed between FF AMH levels and serum FSH level (r=-0.162, P < 0.05). No statistically correlation was noted between the FF AMH levels and blastocyst formation rate (P = 0.508). The relationship between FF AMH levels and outcome measures was illustrated in Table 3.
Correlations between relative telomere length, relative TERT expression, follicular uid AMH levels and probability of clinical pregnancy Based on the clinical pregnancy outcome, all patients (> 35 years old and =35 years old) were divided into 2 groups: pregnant group and non-pregnant group. We analysed relative telomere length, relative TERT expression and follicular uid AMH levels in pregnant group and non-pregnant group. Pregnant group had higher FF AMH levels in older patients (Fig. 4,P 0.05). The correlations between relative telomere length, relative TERT expression, follicular uid AMH levels and the probability of clinical pregnancy were analysed by binary logistic regression and the results were assessed by the odds ratios (ORs) and 95% CIs. As a result of that age may have confounded the associations with the pregnancy, logistic regression was performed in elderly and young age group respectively. There was signi cant relationship between FF AMH levels and probability of clinical pregnancy in old age group. But no signi cant relationship between FF AMH levels and probability of clinical pregnancy in young age group was observed and there were no signi cant relationships between relative telomere length, relative TERT expression and probability of clinical pregnancy whatever in old age and young age group (shown in Table 4). Predictive value of follicular uid AMH levels for clinical pregnancy in old age group was depicted by ROC curve shown in Fig. 5.

Discussion
In this study, we rst compared relative telomere length and TERT expression in GC, AMH levels in follicular uid between elderly patients and young patients in IVF/ICSI treatment. We found that there were signi cant differences between these three biomarkers with regard to age. Correlations between relative telomere length, relative TERT expression, follicular uid AMH levels and the ovarian/embryonic outcome, probability of clinical pregnancy were also assessed.
The current study demonstrates that TERT in GC and AMH levels in follicular uid predicts ovarian response better than RTL during IVF treatment. However, all of them have no potential predictors of embryonic development for infertile women undergoing fertility treatment. In addition, the present study also indicates that FF AMH is associated with basal serum FSH. Finally, The ROC analyses indicated that higher FF AMH levels may predict a higher probability of clinical pregnancy in advanced women.
It is well established that TL and telomerase activity are known biomarkers for biological aging [29]. Low telomerase activity is associated with unexplained ovarian dysfunction in women [30], and a positive correlation is found between the success of IVF treatment and telomerase activity in granulosa cells [10,31]. TERT encodes the catalytic subunit of telomerase, thereby counters telomere shortening during cell division [32].The TERT mRNA codi es for the catalytic component of telomerase, with the other enzyme component TERC which is a RNA template [33]. Both components constitute active telomerase, which elongates the gradual shortening of telomere with each round of DNA replication by maintaining the telomere sequences and conferring sustained proliferation capacity to growing cells [34]. GC cells supporting the oocytes during follicular growth and maturation, exploring knowledge about aging of the GC gives us important insights into oocyte aging [35]. Telomerase activity in GC is important for their proliferation and differentiation capacity [36], as it plays versatile roles in various reproduction pathways. Therefore, we select GC cells which are typically discarded during IVF treatment as research samples.
Here, we analysed the correlation between the expression level of major telomerase gene TERT and telomere length in ovarian and embryonic outcome. The results reveals that the relative telomere length of GC samples in the present study is not associated with any measure of ovarian or embryonic performance but age, which was not in accordance with the previous report that GC relative telomere length does not change signi cantly with aging [11]. Rather, TERT is associated with age and ovarian response but not embryonic performance. This nding stands in contrast to a number of previous studies that have demonstrated variations in embryonic outcomes in the setting of shorter or longer GC telomere length [37,38]. All these ndings indicate that, TERT appears to be a more signi cant biomarker for predicting the ovarian response among infertile women, both TERT and RTL have no role in predicting embryonic development potential as no statistically correlation is noted between TERT/RTL and blastocyst formation rate. Findings of some researchers support the decline in TERT expression and activity in mouse oocytes with reproductive aging [34] and its association with telomere shortening and decline in oocyte quality [39]. Interestingly, our data indicated an increasing in TERT expression in GC with aging and our unpublished RNA-sequencing data showed no signi cant changes of TERT expression between old and young human oocytes. These seemingly discordance results demonstrate that the regulation manner of TERT expression is tissue-dependent [13].
In addition to its established role of compensating for cell replication-dependent telomere shortening, TERT exhibits multiple physiological activities beyond its canonical action. For instance, TERT is shown to modulate mitochondrial and ubiquitin proteasomal function or act as a transcription co-factor [40], to regulate gene expression, to promote cell survival and growth, to protect cells from apoptosis [32], and to display the RNA-dependent RNA polymerase activity [40].A pleiotropic role for TERT in regulating the epigenetic clock [32] and intrinsic DNA methylation age during cell proliferation on advancing cells are also indicated [29].Together, all these TERT effects are actively involved in biological activities, ultimately affecting aging process. In the present study, relative telomere length became shorter while TERT mRNA expression increased with aging in GC. TERT possessing activities dependent or independent on telomere maintenance may contribute to this aging process.
GC cells produce the female hormones. Telomerase activity in GC is under the control of growth factors and steroid hormones [36]. The microenvironment of granulosa cell contains oestrogens reported to in uence telomerase activity and telomere length [35,41]. Enhancement of TERT expression may increase the steroidogenic genes expression which is important for various physiological functions such as reproduction. These confer the importance of endocrine in ovarian microenvironment. Although the telomere theory of reproductive aging has mostly focused on the effect of shorter telomeres on meiotic errors and higher risk of aneuploidies [29,41], it is also possible that the length of telomeres or telomerase interaction with endocrine in ovarian microenvironment may play a role in ovarian reserve.
AMH is produced by ovarian granulosa cells, primarily in pre-antral and small antral follicles less than 8 mm in diameter [24]. It plays a pivotal role in folliculogenesis [35]. The expression of AMH [both mRNA and protein] in human granulosa cells wanes with increasing size of the human follicles [42][43], reaching a stable low level in GC isolated from pre-ovulatory follicles before ovulation induction and at oocyte retrieval [44]. Currently, serum AMH has been widely used as a golden maker for evaluating ovarian reserve of females, particularly in the eld of assisted reproduction [45]. and it was associated with increased cumulative ongoing pregnancy in a study [46]. However, FF AMH can also not be ignored. A prospective study investigated AMH level in the follicular uid and mRNA expression pattern in granulosa cells in relation to age, and an increase in FF AMH level in the older than in the younger age group was found [47]. This result was in contrast with our ndings that FF AMH levels was lower from the older age group than from younger group. Whatever, these suggest that FF AMH may be involved in follicular aging process. Sacha et al. found higher FF AMH appears to be associated with a signi cantly higher probability of clinical pregnancy,but in their study, the patients were not grouped by age [24]. In our study, patients were grouped by age and ROC curve showed that predictive value of follicular uid AMH levels for probability of clinical pregnancy was signi cant just in old age group.
Some researchers investigated whether follicular uid AMH levels were predictive of IVF outcomes but got inconsistent results [48][49][50][51]. In present study, we found FF AMH was related to ovarian response but not embryonic development.
TERT interaction with FF AMH in ovarian microenvironment may play a role in ovarian aging process. A distinctive epigenetic aging pro le in human granulosa cells refers to AMH gene [35]. Some researchers found age-related changes in the AMH gene in human granulosa cells [52]. The cross-talking between TERT and FF AMH may lead to the interaction of TERT with key signaling pathways, which is deserve to investigate and discuss in future [40].
There are still some de ciencies must be referred in the research. First of all, due to the limitation of the sample size, there might be certain bias. Secondly, GC samples are pooled from all of cumulus-oocyte complex of each patient and rather than being analyzed individually. Moreover, we do not measure all tubes of follicular uid AMH levels in each patient on the day of oocyte retrieval. Although follicular uid AMH concentrations show little variability between pre-ovulatory follicles in a retrospective cohort study [24], analyzing relationship between individual cumulus-oocyte complex and corresponding follicular uid AMH levels might lead to more reliable results. Last but not least, TERT measurement with qPCR in the study is at the levels of transcription, other methods of TERT assessment such as western blot can be performed at translation level so as to con rm the results.

Conclusion
In conclusion, our ndings demonstrate that RTL, relative TERT expression in GC and AMH levels in follicular uid are age-related, but all of them fail to predict embryonic outcomes. Relative TERT expression and FF AMH levels appear to be as potential biomarkers for predicting ovarian response among infertile women during IVF/ICSI treatment. Moreover, FF AMH level is a predictor of probability of clinical pregnancy in patients older than 35 years old. Further studies are needed to better understand the mechanisms behind telomere, TERT and FF AMH biology in aging and larger sample researches are needed to con rm these results.

Declarations
Yan Hao, Mingrong Lv and Jing Peng were responsible for the sample collection, relative telomere length and relative TERT expression measurement and manuscript drafting.   Figure 1 Relative Telomere length in relation to age. A the comparison of relative telomere length between old and young age group. Signi cantly shorter RTL was observed in old age group (**P 0.01); B a negative relationship between T/S ratio of granulosa cells and patient age. T/S ratio represents relative telomere length which is shown as blue circles.

Figure 2
Relative TERT expression in relation to age. A the comparison of relative telomere length between old and young age group. Signi cantly higher level of relative TERT expression was observed in old age group (*P 0.05); B a positive relationship between relative TERT expression of granulosa cells shown as red circles and patient age.

Figure 3
Follicular uid AMH levels in relation to age. A the comparison of follicular uid AMH levels between old and young age group. Follicular uid AMH level was signi cantly lower in old age group (**P 0.01); B a negative relationship between follicular uid AMH levels shown as black circles and patient age. Relative telomere length, relative TERT expression and follicular uid AMH levels in pregnant and nonpregnant group. A no signi cant difference of RTL between pregnant and non-pregnant group in young age patients (P 0.05); B no signi cant difference of RTL between pregnant and non-pregnant group in old age patients (P 0.05); C no signi cant difference of relative TERT expression between pregnant and nonpregnant group in young age patients (P 0.05); D no signi cant difference of relative TERT expression between pregnant and non-pregnant group in old age patients (P 0.05); E no signi cant difference of FF AMH levels between pregnant and non-pregnant group in old age patients (P 0.05); F signi cant difference of FF AMH levels between pregnant and non-pregnant group in old age patients (*P 0.05)