The Relationship Between Genetic Variants Associated with Premature Menopause and Lipid Prole in Women Recruited from MASHAD Cohort Study

Background and aim: Premature menopause (PM) is dened by the occurrence of the menopause before the age of 40 years. It is often associated with cardiovascular disease (CVD). The purpose of this study was to explore the relationship between PM-associated genotypes cardio metabolic disorder risk factors. Methods: One hundred seventeen women with PM and one hundred eighty-three healthy women without PM were recruited in this study. DNA was extracted and analyzed using ASO-PCR or Tetra ARMS-PCR. Lipid proles were also assessed. Results: Multivariate logistic regression analysis showed that individuals with GG vs. TT genotype of the rs1046089 SNP were more likely to have high serum LDL risk (p = 0.03) compared to the control group. There was also a signicant association between low serum HDL risk and rs2303369 and rs4806660 SNP genotypes in the PM group. In the PM group, the percentage of those with a high total cholesterol was lower in those with a CC genotype compared to those with a TT genotype (p = 0.03). Conclusion: Some SNPs reported to be associated with PM appear to be independently associated with dyslipidemia. These results may be helpful to identify subjects with PM who may be susceptible to CVD.


Introduction
Menopause is de ned as occurring when menstruation stops for 12 consecutive months due to loss of ovarian follicular activity; this usually occurs around the age of 45-55 years 1 . Menopause that happens before the age of 40 years is called premature menopause (PM) or premature ovarian insu ciency (POI), which may be natural or due to or reproductive surgeries [2][3][4] . This condition is accompanied by amenorrhea, estrogen de ciency and an increase of gonadotrophin levels 5 . It has been reported that about 3.6% of women develop PM 6 . Various factors such as smoking 7 , certain medications 8 , infections and genetic and autoimmune disorders have been associated with PM 9 . Iatrogenic causes such as radiotherapy, chemotherapy, pelvic surgery are also associated with PM 5,[9][10][11][12][13][14] . About 60-90% of PM cases are idiopathic 14 . Several studies have also found that PM increases the risk of hypertension 15 , cardiovascular disease 3 , osteoporosis 16 cerebral infarction 17 , all-cause mortality 18 type 2 diabetes mellitus 15 , and other negative health consequences 16 .
Genetic factors also play a signi cant role in PM 19 . The heritability of menopausal age is estimated about 30-85% 20 . About 20-25% of PM cases are due to genetic causes 14 . Genome-wide association studies (GWAS) have identi ed a polymorphism (rs16991615) of minichromosome maintenance 8 homologous recombination repair factor (MCM8) gene involved in the age of natural menopause 21 . Also, rs1046089 and rs4806660, located on Proline Rich Coiled-Coil 2A (PRRC2A) and transmembrane (TMEM) gene, respectively are associated with the age at menopause 22,23 . GWAS identi ed several other variants that are associated with PM 24 .
Deleterious changes in risk factors for cardiometabolic disorders often occur around the age of menopausal [25][26][27][28] . Estrogen is involved in dilating blood vessels and helping blood ow 25 . Various studies have also shown that estrogen therapy in postmenopausal women reduces serum total cholesterol and low density lipoprotein (LDL) cholesterol concentrations, and increases serum highdensity lipoprotein (HDL) cholesterol and triglyceride concentrations 29,30 . Moreover, lack of ovarian function in the menopause is involved in the activation of the renin-angiotensin system, leads to immunode ciency, in ammation and endothelial dysfunction 25,31 . These are associated with obesity, diabetes and high blood pressure 25,31 . Several studies have shown that age at menopause is associated with cardiovascular disease 18, 32-34 . A Japanese study found women with early menopause had a higher risk for hypercholesterolemia 35 , and another study showed that early menopause is associated with hypertension 36 . Sarnowski et al. founded that genetic variants associated with early menopause are also associated with increased cardiovascular disease risk 37 . There appeared to be a need to evaluate the relationship between PM-related variants with lipid pro le and susceptibility to cardiometabolic disease risk factors. Few studies have been done on this subject. We aimed to explore the associations between PM-related variants with lipid pro le and susceptibility to cardiometabolic disease risk factors in Mashhad stroke and heart atherosclerotic disorders (MASHAD) cohort study population. Methods 117 women who had PM were included in the case group. Healthy women (n = 183) were recruited into the control group. All of participants were recruited as part of the MASHAD study. The MASHAD study is a cohort study from 2010-2020 that were included 9704 participants aged 35-65 years who will be followup exams every three years until 2020 38 . The inclusion criteria were as follows: the diagnostic criterion was based on the de nition of PM: 1) Women who go through menopause before the age of 40 years; 2) 12 continuous months have passed since the last bleeding; and 3) serum FSH > 40 IU/L. Exclusion criteria were: women over 40 years old, with a history of diseases and surgeries affecting menstruation (oophorectomy, hysterectomy), history of genetic abnormalities and syndromes associated with an early menopause is a part of their manifestation, history of using drugs affecting menstruation. Blood samples were collected into vacutainer® plain tubes, and were taken after 14 hours fast. Blood samples were centrifuged at 4° C in 5000 rpm for 15 minutes and the serum part was used for lipid pro le measurement. Body Mass Index (BMI) was measured using standard method 38 . Kidney, liver, and thyroid activity were normal in all participants.
DNA extraction and quality controls Participants' DNA was extracted from 200 µl blood or buffy coat samples using a DNA extraction kit (Pars Tous, Mashhad, Iran). Qualitative and quantitative quality control was performed by agarose gel electrophoresis (Pars Tous, Mashhad, Iran) and Nano drop 2000 (Thermo Fisher Scienti c, USA) in 280 and 260 nanometer wavelengths, respectively. Allele-speci c oligonucleotide polymerase chain reaction (ASO-PCR) The ASO-PCR reaction volume was 15 µl which included: 1.5 µl water, 2 µl genomic DNA, 1 µl of each primer, and 7.5 µl master mix (Pars Tous, Mashhad, Iran). First, to carry out PCR, we performed one cycle of denaturation for 7 minutes at 95°C. After that, 35 cycles include the following: 95°C for 30 sec, annealing for 30 s at 60°C, 72°C for 30 sec, and eventually one cycle of 7 min was done for nal extension.
Tetra ampli cation refractory mutation system PCR (ARMS-PCR) Tetra ARMS was carried out by the same method and the same composition of 15 µl reaction volume that performed in ASO-PCR. Primers were designed with Primer1 software 39 .
Lipid pro le measurements and dyslipidemia diagnosis Total cholesterol (TC), triglycerides (TG), and high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) levels were measured from serum which was taken from participants 12 hours fasting using standard method 40 Physical activity level assessment The equations of James and Scho eld for energy requirements, were used to assess physical activity of all participants. Questions regarding the physical activity level were based on the mentioned equations which were selected from World Health Organization MONICA project questionnaires 43 . The level of physical activity was calculated by total energy expenditure (TTE) and basal metabolism rate (BMR) during a whole day and night 44 .

Ethics
All steps of the study were approved by the Mashhad University of Medical Sciences (MUMS) Ethics Committee. Informed consents were obtained from all subjects and the procedure and possible risks were Polymorphisms' (SNPs') genotypes between PM cases and healthy controls were compared by Chisquare test. Assessment of normal distribution in quantitative data was performed by Kolmogorov-Smirnoff test. Man-Whitney test was used for comparing quantity of normally distributed values between the subgroups. Besides, we used multivariate logistic regression to prevent confounder's factors from affecting our results. P-value < 0.05 was statistically signi cant.

Results
The clinical characteristics of the population have been summarized in Table 1. In our cross-sectional analysis, participants had a mean age of 55 years averagely. Differences in lipid pro le factors serum level between different genotypes of PM-related polymorphisms were examined (Table 2). Furthermore, Table 3 and 4 show the results of multivariate logistic regression analysis before and after adjustment for age and physical activity level.
In the PM cases, serum total cholesterol was signi cantly different between various rs16991615, rs244715, rs4806660, and rs10183486 SNP genotypes; however, this association was not observed in the healthy controls. Interestingly, three of four of the investigated factors including serum total cholesterol, triglyceride, and HDL were substantially associated with different genotypes of rs4806660 SNP in participants with PM and this association was not detected in controls group.
These results for rs1046089 showed individuals with GG genotype were more likely to have low serum LDL (OR= 5.48, CI= 1.14-26.34, p = 0.02) risk than individuals with the AA genotype in control group using a multivariate logistic regression test. Also, the results demonstrate that there was a signi cant association with high TG risk in CC variant vs. TT in rs10183486 in PM group (OR= 4.63, CI= 1.17-18.29, p = 0.02). Furthermore, these results suggest that individuals carried the recessive homozygous genotype (CC) of rs2303369 SNP compared to individuals with dominant homozygous genotype (TT) had an increased risk of a high serum LDL and low serum HDL and the risk of high TC was decreased in control group The risk of low HDL was increased in individuals carrying rs23303369 variant (CT) compared to noncarriers (TT) in both studied groups, (OR, 6.6; 95% CI, 1.88-19.53, P=0.003 in PM group) and (OR= 4.21, CI= 1.17-15.18, p = 0.02 in control group).

Discussion
Our ndings suggest that serum levels of several parameters in the fasted serum lipid pro le consisting of total cholesterol, LDL, and HDL, but not serum triglycerides, were associated with PM. Further analyses indicated that genotypes of polymorphisms, which were previously reported to be related with the incidence of PM, are substantially associated with the level of lipid pro le factors in PM cases. Moreover, we found that some genotypes in speci c polymorphisms including rs4806660, rs10183486, and rs2303369 SNPs were signi cantly related to abnormalities regarding total cholesterol, LDL and HDL level in the cases' serum.
Initially our results found signi cant difference between PM cases and control participants for serum HDL, LDL, and total cholesterol levels, and these factors were signi cantly higher in the PM group. Gulhan and his coworkers have also found that among the 4 lipid factors, only total cholesterol and LDL were substantially different between cases diagnosed with premature ovarian failure (POF) and control subjects. POF group in this study had higher levels of total cholesterol and LDL 46 . Gulhan et al. work had included only women with previous history of successful childbirth and without any hormone therapy within the last 6 months, this inconsistency in HDL serum level had happened. However, this difference might have occurred due to their use of a small sample size as it was as one third as our study or due to the role of age as a confounding factor and also the role of some genetic variants related to PM. Interestingly, a recent study on 3 Dutch university medical centers did not report any signi cant difference in lipid pro le between previously POI-diagnosed participants and population-based controls 47 . In their study, secondary amenorrhea (cessation of menstruation for at least 3 consecutive months) was one of the criteria for including POI cases; While, this period was too short compared to our criteria (12 consecutive months) and this important thing might have affected their results. Moreover, they have not indicated whether their participants had any previous history of surgeries, diseases, or taking medications related to female reproductive tract or not; Thus, this factor might have not been considered in their study which cause this disagreement.
Knauff et al. have reported changes in lipid pro le of cases with POF compared to the controls is potentially related to the rate of ovarian function 48 . Another study which included cases who enter menopause by surgical ovariectomy, clari ed that this intervention on female's reproductive tract, has caused impaired lipid metabolism 6 months post-surgery and substantial increase in all four lipid indicators (Total Cholesterol, Triglyceride, LDL, and HDL) 49 . Moreover it has been demonstrated that POI cases had signi cant lower level of Free Androgen Index (FAI) than people with regular menstrual cycles and there, it has been suggested that higher FAI was associated with high serum triglyceride and LDL in POI cases 50 . Overall, these results suggest that impairment in sexual hormones level, as a result of decreased females' reproductive system activity, is related to weaken lipid metabolism.
Our study has for the rst time investigated the association between PM-related polymorphisms' genotypes and lipid pro le status. In our analysis, we observed that genotypes of 2 different SNPs including rs4806660 (TMEM150B) and rs10183486 (TLK1) are substantially associated with abnormalities of lipid pro le. Regarding the rst one, low level of HDL was observed by 5.26 times higher in PM cases with TC genotype rather than CC participants. This association was not found in the controls. It is possible that these results are due to hormonal disorders that occur in postmenopausal individuals and consequently the lipid pro le in individuals is impaired. Furthermore, for the second mentioned SNP, hypercholesterolemia was found over 450 percent higher in cases who carried CC than those with TT genotype. Previous studies have reported that both rs4806660 51 and rs10183486 52 polymorphisms are associated with the incidence of PM. While it has been proved that TLK1 gene encodes nuclear serine-threonine kinases 52 in which actively transfers signals from receptors of estrogen within the cell membrane to the nucleus 53 , the exact function of TMEM150B gene product, called transmembrane protein 150B, has not yet been discussed.
Several SNPs, have been found to be associated with PM, a condition in which ovaries stop releasing sexual hormones, especially estrogen, making females infertile as early as before the age of 40. This estrogen hormone de ciency affects metabolism of lipids and thus, PM cases might face some lipid pro le abnormalities that could increase the risk of cardiovascular disorders in participants of our study.
Our study was mainly limited by its design as a cross-sectional study, and changes in the level of lipid pro le factors were not prospectively assessed. Data regarding pattern of the target population diet for the consumption of oils and meats, as the most common lipids, were not recorded in this study. We suggest future studies consider these limitations to achieve more reliable results. Moreover, several PMrelated SNPs were not investigated in our study that might be associated with the cases' metabolic status and it is highly recommended to include these polymorphisms in the analysis.
In conclusion, SNPs which are previously found to be attributed to PM, could cause impairment in cases' lipid pro le status through hormonal abnormalities. Present study clari es that TC, CC and, CT genotypes of rs4806660, rs10183486, and rs2303369 SNPs, respectively increase the rate of dyslipidemia by approximately 5 times compared to their reference genotype (rs4806660: CC; rs10183486: TT; rs2303369: TT) in PM cases. But, GG and CC genotypes of rs1046089 and rs2303369 SNPs, respectively increase the rate of dyslipidemia about 6 times compared to their reference genotype in healthy population.

Declarations
Ethics approval and consent to participate Informed consent was obtained from all subjects using protocols approved by the Ethics Committee of the Mashhad University of Medical Sciences. All participants were able to read and understand and were willing to provide written, informed consent.

Consent to publish
Not applicable

Availability of data and materials
The data that support the ndings of this study are available from the corresponding author upon reasonable request.

Competing interests
The authors declare no con ict of interests.