Maternal Age at Menarche Gene Polymorphisms Are Associated with Offspring Birth Weight

In this study, the association between maternal age at menarche (AAM)-related polymorphisms and offspring birth weight (BW) was studied. The work was performed on a sample of 716 pregnant women and their newborns. All pregnant women underwent genotyping of 50 SNPs of AAM candidate genes. Regression methods (linear and Model-Based Multifactor Dimensionality Reduction (MB-MDR)) with permutation procedures (the indicator pperm was calculated) were used to identify the correlation between SNPs and newborn weight (transformed BW values were analyzed) and in silico bioinformatic examination was applied to assess the intended functionality of BW-associated loci. Four AAM-related genetic variants were BW-associated including genes such as POMC (rs7589318) (βadditive = 0.202/pperm = 0.015), KDM3B (rs757647) (βrecessive = 0.323/pperm = 0.005), INHBA (rs1079866) (βadditive = 0.110/pperm = 0.014) and NKX2-1 (rs999460) (βrecessive = −0.176/pperm = 0.015). Ten BW-significant models of interSNPs interactions (pperm ≤ 0.001) were identified for 20 polymorphisms. SNPs rs7538038 KISS1, rs713586 RBJ, rs12324955 FTO and rs713586 RBJ–rs12324955 FTO two-locus interaction were included in the largest number of BW-associated models (30% models each). BW-associated AAM-linked 22 SNPs and 350 proxy loci were functionally related to 49 genes relevant to pathways such as the hormone biosynthesis/process and female/male gonad development. In conclusion, maternal AMM-related genes polymorphism is associated with the offspring BW.


Introduction
BW is one of the key indices of a newborn's conditions/pregnancy outcome and has meaningful prognostic value as a biological predictor of future health [1][2][3]. Newborns with sufficiently high/low BW values are characterized by a considerably raised risk of health disorders (including mortality) in their life-course in comparison with newborns with average BW values [1][2][3][4]. BW has strong phenotypic/genetic correlations of reverse orientation with the risk of later-life cardiometabolic phenotypes/disorders such as coronary artery disease, systolic blood pressure, type 2 diabetes mellitus/glycemic traits and

Study Subjects
The study design outline is presented in Figure 1. The participants of our study were 716 women recruited at the Belgorod Regional Clinical Hospital (Perinatal Center) in the period 2008-2017. All methods/procedures of this study were approved by the Ethics Council of Belgorod State University. Signed informed consent from the participants was a prerequisite for inclusion in the study. All subjects of the current study participated in previously conducted clinical and genetic studies of pregnancy outcomes/disorders (preeclampsia, fetal growth restriction) (detailed biomedical, anamnestic information, data on the course of a real pregnancy and its outcomes were collected about each subject; BW parameter was taken from the medical record of childbirth) [43][44][45][46][47][48][49][50]. The group of subjects included women who met the following criteria: singleton pregnancy (24-41 weeks of gestation) ending in a live birth, Russian origin [self-reported], born in the Central region of Russia [51,52]. Women with parameters such as age younger than 16, multiple pregnancies, anomalies of the uterus/cord/fetus, childbirth <24 weeks of pregnancy did not participate in the study (exclusion criteria). The detailed characteristics of the sample are presented in Table 1.

DNA Extraction, AMM-Involved SNPs Selection, Genotypes Testing
The maternal venous blood (taken from the cubital vein) was exploited to gain DNA (the phenol/chloroform/ethanol DNA extraction/purification method has been described in detail in [53]). The quality/quantity of extracted DNA was evaluated on a NanoDrop2000 micro volume spectrophotometer and then DNA samples were stored in a kelvinator at −80 • C.
When selecting polymorphisms for this genetic study, special criteria were taken into account [24,[28][29][30] such as; (1) association with AAM and/or phenotypes that share pathways with AAM (weight/height/BMI/obesity/vitamin D metabolism, etc.) (Supplementary Table S1); (2) functionality (epigenetic role, association with gene expression etc.); (3) the frequency of the polymorphic variant is not less than 5%. Using these criteria, the literature data on this issue (presented in detail in Supplementary Table S1) and the materials of our previous studies [24,[28][29][30], a total of 50 loci of AAM candidate genes were included in the study. All these loci had important functional significance (Supplementary  Table S2) (according to the information provided in the HaploReg database [54]). Among  included in the study, 13 SNPs were associated with AAM according to GWAS  data and 26 SNPs were AAM-link according to the results of other associative studies  (Supplementary Table S1); 15 SNPs were associated with various an anthropometric parameters (Supplementary Table S1); 10 SNPs out 50 loci (such as rs4633, COMT; rs1544410, VDR; rs2252673, INSR; rs222003, rs222020, GC; rs7766109 F13A1; rs1884051, rs3020394 ESR1, rs3756261 and EGF; rs12324955 FTO) were not directly related to AAM, but they correlated with AAM-significant phenotypes/disorders [e.g., metabolism of vitamin D; polycystic ovary syndrome, etc. (Supplementary Table S1)] and had significant opportunity functionality (Supplementary Table S2). Importantly, we previously used this SNP panel in conducting associative studies of AAM and anthropometric parameters (height/BMI) of women [24], proliferative uterine diseases (fibroids/endometriosis/endometrial hyperplasia) [28][29][30] in the studied population (Central Russia), which allows us to conduct a correct comparative analysis of the role of BW-associated polymorphisms in the formation of other phenotypes/diseases of the female reproductive system in this work.
Baseline characteristics The DNA SNPs were identified by Sequenom MassARRAY system ("Seqeunom", San Diego, CA, USA) in Tomsk National Research Medical Center of the Russian Academy of Sciences (Tomsk, Russia). Quality control of SNPs genotyping was accomplished based on the following indicators: call rate (>90%); success rate of duplicate check (>99%); success rate of the blank check (>90%) [30,55]. In total, 49 out of 50 analyzed loci passed the quality control procedure (for rs11724758 FABP2, the call rate was 84.25%, so it was excluded from the subsequent analysis).

SNPs Association Analysis
For all studied AAM-involved SNPs, the compliance of their distribution (alleles/genotypes) with the Hardy-Weinberg law was checked [56,57]. The link of BW with polymorphisms and their haplotypes was evaluated in the gPlink (Java-tied versia 2.050) program [58] by linear regression (three genetic models [additive; dominant; recessive] [59] were considered). Methods and programs of the same name MDR [60,61] and MB-MDR [24,62] were applied to modeling BW-significant interlocus interactions. For genetic computations, transformed BW values were used (real BW values with non-normal distribution were reduced to a normal distribution by QQ-plot function in the R program [24]). The calculations used covariates (age at menarche, pre-pregnancy BMI, presence in the anamnesis of hypertension and fetal growth restriction, presence of current pregnancy complications (fetal growth restriction, preeclampsia and a combination of fetal growth restriction with preeclampsia; date from Table 1) (covariates were included as independent variables in the regression analyses) and procedures that minimize the risk of false positive results (permutation test was applied [63,64]). Importantly, when selecting BW-significant interlocus interaction models for permutation procedures, we introduced additional Bonferroni corrections (the quantity of possible combinations of 49 SNPs was taken into account) and as a result, values of the significance level such as p < 4 × 10 −5 (<0.05/1176; two-locus models), p < 3 × 10 −6 (<0.05/18424; three-locus models) and p < 2 × 10 −7 (<0.05/211876; four-locus models) were set as the "threshold". Indicators with p perm < 0.017 (for individual SNPs; an additional Bonferroni correction was introduced, equal to 3 according to the number of calculated genetic models) and p perm ≤ 0.001 (for models of interlocus interactions) were taken as threshold values when assessing the statistical significance of the established associations. In our sample of subjects (n = 716), with a planned study capacity of at least 80%, it is possible to detect differences in BW at the level of 86-142 g (additive model), 138-156 g (dominant model), 146-623 g (recessive model). Statistical power for BW-associated SNPs was calculated by Quanto 1.2.4 [65].

Study Participants' Characteristics
The study group included women aged 16-45 years, with a single pregnancy ending in a live birth, Russian origin, born in the Central region of Russia. The main characteristics of the women/newborn subjects included in our study (n = 716) are presented in Table 1. As follows from Table 1, a lower BW has been associated with maternal parameters such as late menarche (p = 0.02), low pre-pregnancy BMI (p < 0.0001), a history of hypertension (p = 0.0002) and fetal growth restriction (p < 0.0001), as well as with complications of the current pregnancy (fetal growth restriction and a combination of fetal growth restriction with preeclampsia (p = 0.0001)). The above phenotypic characteristics were covariates when evaluating the association of AAM-involved SNPs with BW.

SNPs/Haplotypes Association Analysis
The allele/genotypic distribution of the considered AAM-involved 49 SNPs was in full respect of the Hardy-Weinberg law (Supplementary Table S3 Table 2). The link of haplotypes with BW was not confirmed by the results of permutation testing (p perm > 0.05) (Supplementary Table S4).

BW-Involved SNP × SNP Interactions
For 20 SNPs out 50 examined loci, we modeled ten BW-significant SNP-SNP interplay models (p perm ≤ 0.001) ( Table 3). Three loci such as rs7538038 KISS1, rs713586 RBJ, rs12324955 FTO and two-locus interaction rs713586 RBJxrs12324955 FTO have been included in 30% of BW-significant models. The most appreciable exposure on BW (Wald value was maximum) was registered for a model consisting of four SNPs such as rs713586 RBJ, rs3020394 ESR1, rs1782507 FSHB and rs12324955 FTO (WH BW = 56.83 [p perm < 0.001]). The combinations of genotypes determining a higher BW have the greatest statistical . The most perceptible phenotypic (BW) effect was also noted for the BW-enhancing combination of genetic variants (rs10441737 CC ZNF483xrs1073768 Table S5). Three-order interaction models (threshold level p < 3 × 10 −6 , real level p < 5 × 10 −8 ) Four-order interaction models (threshold level p < 2 × 10 −7 , real level p < 3 × 10 −13 ) Note: The results were obtained using the MB-MDR method with adjustment for covariates; NH, number of significant high risk genotypes in the interaction; beta H, regression coefficient for high risk exposition in the step2 analysis; WH, Wald statistic for high risk category; NL, number of significant low risk genotypes in the interaction; beta L, regression coefficient for low risk exposition in the step2 analysis; WL, Wald statistic for low risk category; P perm , permutation p-value for the interaction model (1000 permutations).
Amid the BW-significant epistatic interplays (Figure 1), antagonistic liaisons (indicated in Figure 1 in blue) play a noteworthy role, in which the SNPs rs7538038 KISS1, rs7589318

Functional of BW-Candidate SNPs
Using the in silico methodology, the potential functionality of 22 BW-associated loci and 350 proxy SNPs (their connections with missense mutations; epigenetic modifications; regulation of expression/splicing; common pathways) was evaluated. The in silico approach allows using modern bioinformatic tools (as a rule, they include previously obtained experimental results on the functional value of SNP/genes) to obtain data on the putative functional role of the studied SNP/genes in the body. SNPs associated with missense mutations lead to amino acid substitutions in the encoded polypeptide, which can determine the change/disruption of the function of this protein. Loci associated with epigenetic modifications can regulate the transcriptional activity of genes (by influencing the promoter/enhancer regions of the genome, interaction with transcription factors and regulatory proteins, etc.), which ultimately determines the amount of protein product of these genes. Polymorphisms affecting the regulation of gene expression and splicing determine the level of transcription of mRNA from a given gene and the level of its alternative splicing, which as a result will affect the quantitative representativeness of proteins in the body controlled by these genes.
Importantly, all four BW-causal variants have appreciable purported regulatory actions, e.g., rs757647 KDM3B has been located at the DNAase-hypersensitive province in cultured trophoblast cells, in the enhancer-impact DNA positions in mesenchymal stem and adipocyte cells, fatty nuclei, placenta, amnion, ovaries, a genome binding site to MAFK (regulatory protein) and CTCF (TF) (Supplementary Table S2 Table S8)  Interestingly, the BW-raising allele A rs7589318 POMC has been correlated with lower RBJ mRNA production in blood (Z = −4.94) (Supplementary Table S7) and higher splicing level of ADCY3 in subcutaneous adipose (β = 0.35) and thyroid (β = 0.26) (Supplementary  Table S11). BW-increasing allele T rs757647 KDM3B was linked with low expression REEP2 in subcutaneous adipose (β = −0.25) (Supplementary Table S9) and higher CDC25C splicing value (β = 0.77) in EBV-transformed lymphocytes (Supplementary Table S11). The network of BW-significant protein-protein interactions (P-Pi) obtained in the STRING program is shown in Figure 3 and indicates a significant level of these interactions (p < 10 −16 ). The basis of these P-Pi has been the pathways listed above according to the Gene Ontology resource. In addition, interestingly, there were founded common RNA expression patterns for two pair BW-causal genes such as CDC25C and KIF20A (RNA coexpression score 0.796; these genes were functionally (sQTL/eQTL) related to rs757647 KDM3B), NKX2-1 and SFTA3 (RNA coexpression score 0.473; these genes were functionally (sQTL/epigenetic modifications) related to rs999460 NKX2-1). P-Pis have been clustered into three groups ( Figure 4D). The first group ( Figure 4A

Discussion
This work has been the first study to report a link between maternal AAM-associated genes SNPs with offspring BW. Four loci independently (rs7589318 POMC; rs757647 KDM3B; rs1079866 INHBA; rs999460 NKX2-1) and 20 SNPs in interlocus interplays have been BW-correlated. BW-associated AAM-linked 22 SNPs and 350 proxy loci were functionally related to 49 genes relevant with pathways such as the hormone biosynthesis/process and female/male gonad development. The study results outline is presented in Figure 5.

Discussion
This work has been the first study to report a link between maternal AAM-associated genes SNPs with offspring BW. Four loci independently (rs7589318 POMC; rs757647 KDM3B; rs1079866 INHBA; rs999460 NKX2-1) and 20 SNPs in interlocus interplays have been BW-correlated. BW-associated AAM-linked 22 SNPs and 350 proxy loci were functionally related to 49 genes relevant with pathways such as the hormone biosynthesis/process and female/male gonad development. The study results outline is presented in Figure 5.
In this work, it was found that the maternal allele T rs757647 KDM3B has the largest associations with increased offspring BW (β = 0.323). According to earlier genetic studies, the allele T rs757647 KDM3B has been linked with early AAM [22,80]; in addition, rs17171818 KDM3B, which linked to rs757647 KDM3B (r 2 = 0.96), also was correlated with AAM [23]. Thus, the genetic factor of early maternal AAM was associated with a high offspring BW. Importantly, AAM is a key parameter of the woman (it reflects the state of the hypothalamic-pituitary-ovarian system) and is strongly correlated with woman's fertility [17]. Early AAM is risk factor of various woman's health-related disorders in adulthood such as impaired glucose metabolism/insulin resistance, obesity, metabolic syndrome, diabetes, cardiovascular diseases, etc. [16,21]. The limited literature on this topic also indicates the connection of early maternal AAM with an increased BMI of offspring [33][34][35]. Correlations of this orientation were found in the studies of Wang et al. among the Chinese population (a sample of 17,571 children showed a correlation between early maternal menarche and the offspring BMI of both boys and girls) [33]. In the works of Basso et al. in the USA population (with a sample of 31,474 black and white children was demonstrated that children of mothers with menarche <12 years were taller from age 1 and had higher BMI at ages 7-8 years) [34] and Ong et al. among the UK population (using a sample of 6009 children demonstrated an between association early AMM and characteristics of offspring such as higher height, weight, BMI, fat mass index at the age of 9 years) [35]. In this work, it was found that the maternal allele T rs757647 KDM3B has the largest associations with increased offspring BW (β = 0.323). According to earlier genetic studies, the allele T rs757647 KDM3B has been linked with early AAM [22,80]; in addition, rs17171818 KDM3B, which linked to rs757647 KDM3B (r 2 = 0.96), also was correlated with According to our data, SNP rs757647 KDM3B affects the interaction of DNA with the regulatory protein MAFK and TF CTCF, modulates gene expression due to localization in the enhancer regions of the KDM3B gene in amnion, placenta, ovary, adipose (mesenchymal stem cells, adipose nuclei, etc.); BW-raising allele T rs757647 KDM3B was correlated with low REEP2 transcription in adipose (subcutaneous) and high CDC25C splicing in EBV-transformed lymphocytes. Additionally, rs757647 KDM3B has been high-linked with 18 sQTL SNPs in EBV-transformed lymphocytes (the splicing of two genes such as FAM53C,CDC25C was regulated by these polymorphisms).
The KDM3B gene encodes lysine-specific demethylase 3B, which demethylates mono-and dimethylated histone lysine 9 H3 (H3K9me1/H3K9me2) [81]. This modification of histones is quite important for the processes of stem cells proliferation/differentiation/migration [82]. Liu et al. demonstrated in a mouse model with KDM3B knockout that this gene is expressed in ovarian, fallopian tubes, uterus (endometrium/myometrium) [83]. In females with KDM3B knockout, decreased expression of IGFBP-3 in the kidneys (by 53%) and a reduction in its content in the blood were observed, which led to accelerated degradation and a decrease in the IGF-1 blood concentration (by 36%). In addition, in the experiment, a decrease in the level of circulating 17beta-estradiol (>50%), irregular/prolonged estrous cycles, reduced ability to ovulate (by 45%) and fertilization (by 47%), decreased decidual response of the uterus (by 44%) were recorded [83]. The aforementioned data indicate the importance of KDM3B as a key H3K9 demethylase necessary to maintain circulating IGF-1 and 17beta-estradiol, the embryonic development processes, postnatal somatic growth and the functioning of the female reproductive system. It is assumed that KDM3B-mediated decrease in IGFBP-3 expression, leading to rapid degradation of IGF-1, may also cause small body sizes [83].
Interestingly, the CDC25C gene (the level of splicing of this gene depends on polymorphism rs757647 KDM3B and 18 proxy loci) encodes a cell division cycle 25C protein involved in the cell division regulation by participating in cyclin B-related processes and overwhelming the p53 suppressive effects [84]. The FAM53 (family with sequence similarity 53 member C) protein encoded by the gene of the same name (FAM53 splicing is controlled by SNP rs757647 KDM3B and 18 proxy variants), in interaction with transcription regulators, has a modifying effect on cell proliferation. There are GWAS data on the correlation of CDC25C/FAM53C genes polymorphisms with IGF-1 level, BMI, height [85][86][87].
The results of our work indicate the link between the maternal rs1079866 INHBA and offspring BW: the CC genotype was BW-raising (β = 0.110). Several previous genetic studies (GWAS, etc.) have showed association of rs1079866 INHBA with AAM [22,23,88], Tanner stage in girls [23], height adult [22,23,86], IGF-1 level [85], breast cancer risk [89]. Strongly coupled with rs1079866 INHBA SNP rs4141153 (r 2 = 0.63) was also AAM-(GWAS)-linked [90,91]. Additionally, the BW-raising C allele has been associated with early menarche and low height adult [22,23,88]. SNP rs1079866 INHBA and LD SNP rs1079866 INHBA determine the DNA interaction manner with six TFs (e.g., HNF4,YY1, etc.) and were gene enhancerrelated in adipose (mesenchymal stem cells; adipose nuclei; etc.). The INHBA gene encodes the βA subunit of the inhibin protein [84]. The inhibin βA subunit can also dimerize with the inhibin α and inhibin βB subunits to form inhibin A and activin AB, respectively. These hormones are involved in hypothalamic-pituitary-ovarian folliculogenesis and implantation processes. Inhibin A is synthesized by granulosa cells in the ovary and contributes to the suppression of gonadoliberin secretion in the hypothalamus and folliclestimulating hormone in the adenohypophysis by the mechanism of negative feedback. INHBA is a modulator BW-significant TGF-β signaling pathway: overexpression of INHBA upregulated TGF-β-related genes (TGF-β1/p-Smad2/p-Smad3/VEGF-A), and respectively INHBA low expression downregulated activity of these genes [92]. Differential expression of the INHBA gene in the placenta has been shown [93], which may directly correlate with BW. Interesting experimental data on the potential mechanisms of the INHBA genes influence on the course of pregnancy (normal and complicated preeclampsia) were obtained in the work of Wang et al. [94]. The authors found that the INHBA gene together with the OPRK1 and TPBG genes are key inflammation-related genes that affect the course of pregnancy (the expression of INHBA and TPBG genes in complicated preeclampsia pregnancy is higher, and the OPRK1 gene is lower, compared with normal pregnancy). Wang et al. have shown that TFs such as SMARCC1, REL, ZNF274, RELB and RELA are involved in the regulation of the expression of INHBA (and TPBG), and the potential targets of the signaling pathways of these genes are microRNAs such as has-miR-1206 and has-miR-548k [94]. Increased expression of INHBA has been shown to contribute to human trophoblast invasion (via ALK3-BMPR2/ACVR2A-SMAD1/5/8-SMAD4 signaling pathways) whereby it has meaningful significance for early stages of embryonic implantation [95] and may have consequences for BW.
The ADCY3 gene (this gene-splicing level in adipose [subcutaneous] and thyroid depends on BW-involved polymorphism rs7589318 POMC and 11 proxy loci) encodes membranelinked enzyme, adenylyl cyclase 3, catalysing synthesis of cAMP (cyclic adenosine monophosphate) from ATP (adenosine triphosphate) in the process of signal transmission by the Pprotein [107]. cAMP is believed to be involved in the intracellular signaling of molecules such as leptin, glucagon-like peptide 1, orexins, ghrelin and α-melanocyte-stimulating hormone [107,108]. It is a generally recognized scientific fact that the ADCY3 gene/protein is involved in metabolic processes and obesity-significant pathways (data on this were obtained both on experimental models of ADCY3-knockout mice, Goto-Kakizaki rats and in numerous genetic associative studies in various human populations) [107,[109][110][111]. Obese subjects are characterized by reduced ADCY3 expression in adipose tissue [112]. According to our experimental/in silico data, BW-raising maternal allele A rs7589318 POMC was correlated with high ADCY3 splicing level in adipose [subcutaneous] and thyroid. Importantly, ADCY3 gene SNPs were GWAS-impact for BMI in children [113] and BMI across children aged 1 to 17 years [114].
In our study, maternal rs999460 NKX2-1's link with the offspring BW (β = −0.176) was revealed. In previous studies, rs999460 NKX2-1 polymorphism was associated with AAM [98], woman BMI [24] and endometrial hyperplasia [28]. The bioinformatic data obtained by us indicate the localization of this SNP in the promoter region of the NKX2-1 gene in cultured NK cells (CD56) of the ectoderm, the effect on the affinity of DNA to five TFs (AIRE; Arid5a; Foxa; STAT; Pax-45) and the regulation of the SFTA3 gene splicing in thyroid (BW-reducing allele A rs999460 NKX2-1 was linked with high SFTA3 splicing level). The NKX2-1 gene encodes a transcription factor (thyroid transcription factor 1 (TTF1)), which binds and activates promoters of the thyroglobulin, thyroperoxidase and thyroid-stimulating hormone receptor genes [84]. In addition, this protein suppresses the expression of the transcription factor NR1D1, which, in turn, inhibits the transcription of genes involved in the metabolism of lipids and bile acids, adipogenesis, gluconeogenesis and inflammatory reactions of macrophages [115].
Interestingly, NKX2-1 gene product-TTF1 is actively expressed during embryonic development in various organs (lungs, ventral forebrain, diencephalon, hypothalamus, etc.), including in the thyroid gland, which is crucial for both the early stages of thyroid development and embryonic development in general [116]. During the development of the thyroid gland, TTF1 together with other thyroid-significant TF (e.g., HHEX, FOXE1, PAX8) are jointly expressed both in thyroid progenitor cells and in thyroid mature follicular cells, ensuring its development, differentiation, homeostasis and function at the proper level [116]. It is important that in the absence of TTF1, the progenitor cells of the thyroid endure apoptosis and vanish at the embryonic development early stages, which leads to a decrease in thyroid follicular cells' mass, its anomalous development and degradation [117]. Various abnormalities in the thyroid structure/functioning can lead to versatile endocrinemetabolic disorders in the neonatal period (e.g., congenital hypothyroidism) [Guan 2021], which directly correlate with BW (e.g., low-BW infants oftentimes demonstrate transient congenital hypothyroidism) [118]. Importantly, >50% of infants with low BW are diagnosed hypothyroxinemia at the thirtieth week of pregnancy [118].
In earlier genetic studies using the list of AAM-significant polymorphisms considered in this paper in the same population (Russian women of Central Russia), associations of a number of these polymorphic loci with AAM, BMI [24] and proliferative uterine diseases (endometriosis [29], endometrial hyperplasia (EH) [28], uterine fibroids (UL) [30]) were demonstrated. A comparative analysis of the results of this results and previously obtained data shows the link of BW-associated AAM-involved SNPs with AAM and BMI of women in this population (Supplementary Table S14): out of 22 BW-associated SNPs, eight loci were also BMI-associated (36.36%) and five loci were AAM-associated (22.73%); moreover, four loci such as rs4374421 LHCGR, rs4946651 LIN28B, rs1073768 GHRH and rs4633 COMT (18.18%) were associated with all three phenotypes under consideration (BW, AAM, BMI). Additionally, BW-associated AAM-involved SNPs were also involved in susceptibility to proliferative uterine diseases: out of 22 BW-associated SNPs, nine loci were also UL-associated (40.90%), 10 SNPs were EH-associated (45.45%) and eight loci were endometriosis-associated (36.36%); moreover, three loci such as rs4374421 LHCGR, rs12324955 FTO and rs1782507 FSHB (13.63%) were associated simultaneously with BW and the three proliferative diseases of the uterus indicated above (Supplementary Table S14). The aforementioned data indicate a significant similarity of the AAM-involved genetic determinants (polymorphisms) of the woman that determine the AAM, BMI and affect both the offspring BW and the development of the uterus proliferative diseases.
Among offspring BW-associated genes, 33.33% were maternal AAM-associated, 42.22% were maternal BMI-associated; while 31.11% (GHRH, ADCY3, EFR3B, ARVCF, COMT, HACE1, LHCGR, MROH8, LIN28B, LIN28B-AS1, MANBAL, POMC, RBJ, TXNRD2) were involved in the formation of all three phenotypes considered (offspring BW, maternal AAM and BMI) ( Figure 6, genes functionally related to BW-,AAM-,BMI-associated loci (22 SNPs [the data of this work], 13 and 14 SNPs [24] respectively) and proxy SNPs) were considered). At the same time, 65.22% of maternal AAM-associated genes and 61.29% of maternal BMIassociated genes have been offspring BW-associated genes ( Figure 4). Our data on the presence of common genetic factors offspring BW, maternal AAM and BMI correspond to modern literature materials [1,15]. Strong positive genetic correlations between BW and BMI/adult obesity [1] and causal genetic effect of BW on AAM (lower BW determined increased risk of earlier AAM) have been shown in several works [41,42]. The study by Tyrrell et al. convincingly showed that genetically raised maternal BMI was causally connected with increased offspring BW [15]. So, at the moment there is convincing genetic data on the association of higher pre-pubertal BMI (age 8 year) with earlier menarche [26,42], early menarche with high pubertal growth, pre-pubertal/adolescent/adult BMI [25][26][27]42], high maternal BMI with increased BW of the offspring [15]. At the same time, not high BW (as expected according to the above multi-stage correlation, such as high maternal BMI in childhood -> early maternal menarche -> high maternal BMI (adult) -> high offspring BW will directly correlate with maternal AAM, but on the contrary, low BW is genetically associated with early menarche [41,42]. The above literature materials indicate the presence of common unidirectional genetic factors underlying the high maternal BMI and the high offspring BW and the ambiguous (including multidirectional) genetic correlations of maternal AAM and offspring BW.
AAM-associated genes and 61.29% of maternal BMI-associated genes have been offspring BW-associated genes ( Figure 4). Our data on the presence of common genetic factors offspring BW, maternal AAM and BMI correspond to modern literature materials [1,15]. Strong positive genetic correlations between BW and BMI/adult obesity [1] and causal genetic effect of BW on AAM (lower BW determined increased risk of earlier AAM) have been shown in several works [41,42]. The study by Tyrrell et al. convincingly showed that genetically raised maternal BMI was causally connected with increased offspring BW [15]. So, at the moment there is convincing genetic data on the association of higher pre-pubertal BMI (age 8 year) with earlier menarche [26,42], early menarche with high pubertal growth, pre-pubertal/adolescent/adult BMI [25][26][27]42], high maternal BMI with increased BW of the offspring [15]. At the same time, not high BW (as expected according to the above multi-stage correlation, such as high maternal BMI in childhood ─> early maternal menarche ─> high maternal BMI (adult) ─> high offspring BW will directly correlate with maternal AAM, but on the contrary, low BW is genetically associated with early menarche [41,42]. The above literature materials indicate the presence of common unidirectional genetic factors underlying the high maternal BMI and the high offspring BW and the ambiguous (including multidirectional) genetic correlations of maternal AAM and offspring BW. Maternal AAM-related polymorphisms (rs7589318 POMC, rs757647 KDM3B, rs1079866 INHBA, rs999460 NKX2-1) associated with offspring BW according to this work can be recommended for use in practical medicine for the purpose of selection among women at the pre-pregnancy/early pregnancy stages at risk of birthing newborns with sufficiently high/low BW values and implementation among these women complex preventive/therapeutic measures aimed at preventing offspring with high/low BW. It should be noted that it is necessary to continue a more active study of the role of maternal AAM genes in the formation of offspring anthropometric characteristics in various ethnically different populations in order to establish "universal" genetic determinants of these phenotypic traits for the population as a whole, as well as studies on the influence of this group of maternal candidate genes on the course of pregnancy and its complications (placental insufficiency, fetal growth restriction, preeclampsia, etc.). The data established in this work can be the basis for subsequent research in this area as comparison indicators, which are obtained for the Europeans of Central Russia. Maternal AAM-related polymorphisms (rs7589318 POMC, rs757647 KDM3B, rs1079866 INHBA, rs999460 NKX2-1) associated with offspring BW according to this work can be recommended for use in practical medicine for the purpose of selection among women at the pre-pregnancy/early pregnancy stages at risk of birthing newborns with sufficiently high/low BW values and implementation among these women complex preventive/therapeutic measures aimed at preventing offspring with high/low BW. It should be noted that it is necessary to continue a more active study of the role of maternal AAM genes in the formation of offspring anthropometric characteristics in various ethnically different populations in order to establish "universal" genetic determinants of these phenotypic traits for the population as a whole, as well as studies on the influence of this group of maternal candidate genes on the course of pregnancy and its complications (placental insufficiency, fetal growth restriction, preeclampsia, etc.). The data established in this work can be the basis for subsequent research in this area as comparison indicators, which are obtained for the Europeans of Central Russia.
The limitation of this study is that these scientific results were obtained only for one population (Europeans of the Central region of Russia) and further studies (replicative) are needed to confirm the revealed facts of the connection of maternal AAM-related polymor-

Conclusions
This study showed the dependence of the offspring BW on maternal AAM-associated gene polymorphisms that are functionally significant for hormone/gonad-related pathways. The data established in this work can be the basis for further research in this area and can also be used in the future in practical medicine.

Data Availability Statement:
The data generated in the present study are available from the corresponding author upon reasonable request.