Abstract

   Polycystic ovarian syndrome (PCOS) is one of the most common
   endocrinopathies among reproductive women worldwide, contributing
   greatly on the incidence of female infertility and gynecological
   cancers. It is a complex health condition combining of multiple
   symptoms like androgen excess, uncontrolled weight gain, alopecia,
   hirsutism, etc. Conventionally PCOS was associated with obesity while
   it is often found among lean women nowadays, making the disease more
   critical to diagnose as well treatment. The disorder has an impact on
   several signal transduction pathways, including steroidogenesis,
   steroid hormone activity, gonadotrophin regulation, insulin secretion,
   energy balance, and chronic inflammation. Understanding the aetiology
   and pathophysiology of PCOS is difficult due to its multiple causes,
   which include environmental factors, intricate genetic predisposition,
   and epigenetic modifications. Despite research supporting the role of
   familial aggregations in PCOS outcomes, the inheritance pattern remains
   unknown. Henceforth, to reduce the burden of PCOS, it is inevitably
   important to diagnose at early ages as well as intervene through
   personalized medicine. With this brief background, it was imperative to
   elucidate the genetic architecture of PCOS considering BMI as an
   controlling factor. This study aims to investigate the genetic basis
   behind obesity-mediated PCOS, focusing on both obese and lean
   individuals. It uses a comprehensive bioinformatics methodology to
   depict pathways and functionality enrichment, allowing for
   cost-effective risk prediction and management. In the present research,
   the representative study participants (N = 2) were chosen from a
   cross-sectional epidemiological survey, based on their anthropometric
   parameters and confirmation of PCOS. Upon voluntary participation and
   written consent, biological fluids (whole blood and buccal swab) were
   taken from where DNA was extracted. The clinical-exome sequencing was
   performed by the Next-generation Illumina platform using the Twist
   Human Comprehensive Exome Kit. A comprehensive bioinformatics
   methodology was employed to identify the most important, unique, and
   common genes. A total of 26,550 variants were identified in clinically
   important exomes from two samples, with 5170 common and 2232 and 2322
   unique among PCOS lean and obese phenotypes, respectively. Only 262 and
   94 variants were PCOS-specific in lean and obese PCOS. Three filters
   were applied to shortlist the most potent variants, with 4 unique
   variants in lean PCOS, 2 unique variants in obese PCOS, and 5 common
   variants in both. The study found that leptin signalling impairment and
   insulin resistance, as well as mutations in CYP1A1, CYP19A1, ESR1, AR,
   AMH, AdipoR1, NAMPT, NPY, PTEN, EGFR, and Akt, all play significant
   roles in PCOS in the studied group. Young women in West Bengal, India,
   are more likely to have co-occurring PCOS, which includes estrogen
   resistance, leptin receptor insufficiency, folate deficiency, T2DM, and
   acanthosis nigricans, with obesity being a common phenotypic
   expression.

   Keywords: Polycystic ovarian syndrome (PCOS), Next-generation
   sequencing (NGS), Clinical-exome sequencing (CES), Comorbidity
   analysis, Mutation-phenotype interactions, Molecular screening

   Subject terms: Biological techniques, Genetics, Endocrinology,
   Molecular medicine

Introduction

   Polycystic ovarian syndrome (PCOS), one of the most common
   heterogeneous, severe endocrinopathies that affect 15–20% of
   reproductive women globally, has been linked to infertility and even
   some gynaecological cancers^[26]1–[27]3. The condition is classified as
   a syndrome rather than a disease due to its diverse phenotypic
   symptoms, such as ultrasonographic polycystic ovarian morphology
   (PCOM), clinical or biochemical hyperandrogenism, hirsutism, acne,
   acanthosis nigricans, obesity, infertility, and hair loss^[28]4,[29]5.
   Understanding the aetiology and pathophysiology of PCOS has been a
   difficult task because of its plentiful causes, including environmental
   variables, genetic predisposition, and epigenetic changes^[30]2,[31]3.
   On the other hand, despite having sufficient evidence to confirm the
   importance of familial aggregations in PCOS outcomes^[32]6–[33]10, the
   inheritance pattern remains unclear. PCOS affects multiple signal
   transduction pathways, including steroidogenesis, steroid hormone
   activity, gonadotrophin control, insulin secretion, energy balance, and
   chronic inflammation^[34]2–[35]4. According to the
   database^[36]11,[37]12, 533 potential genes and 145 SNPs have been
   linked to PCOS, and countless studies are being conducted throughout
   the world to comprehend the molecular signals and their intricate
   underlying mechanisms.

   It is terrible that women are still not completely aware of their
   reproductive health. In our previous survey, we discovered that the
   majority of the population is unaware of their menstrual health while
   experiencing a few symptoms such as irregular menstruation,
   dysmenorrhea, and so on^[38]13. The condition stays the same in both
   developing and developed nations, and this could be related to
   traditional and cultural views^[39]14–[40]16. Over the last 2 decades,
   the global burden of PCOS-related infertility has grown
   dramatically^[41]17. A study showed the same observation as that of
   Dhar et al.^[42]13, at 30 years of age and those with a prior diagnosis
   of PCOS were 4 times more likely to have reported difficulties becoming
   pregnant than those undiagnosed and regularly sought medical
   treatment^[43]18. Contradictory to this, an earlier study reported that
   POM was present among 14% of healthy women and it was more prevalent
   among women in the age group < 35 years^[44]19. In addition, irregular
   menstrual bleeding is less trustworthy until it is prolonged, but
   moderate hair growth and chronic anovulation are normal in late puberty
   and early adolescence, which can lead to a delayed diagnosis^[45]20 and
   failure in PCOS management. So, accumulating shreds of evidence
   reinforces the importance of detection of PCOS early in youth or before
   puberty to ensure timely diagnosis and adequate health care
   (Fig. [46]1).

Fig. 1.

   [47]Fig. 1
   [48]Open in a new tab

   Gaps & scope of the present research hypothesis.

   In the era of medical advent, the introduction of next-generation
   sequencing (NGS) has caused a paradigm shift in genomics research,
   providing unequalled capabilities for analysing DNA and RNA molecules
   in a high-throughput manner. It provides insights into genome
   structure, genetic variants, gene expression levels, and epigenetic
   alterations, broadening research into uncommon genetic illnesses,
   cancer genomics, and precision medicine^[49]21. Clinical exome
   sequencing (CES) is a cutting-edge molecular diagnostic technology that
   rapidly discovers disease-causing genetic abnormalities in any human
   gene and is gaining popularity in clinical settings. Its high
   diagnostic yield makes it cost-effective and suitable for routine
   diagnosis in patients with a variety of phenotypes^[50]22. Since the
   last few years, a plethora of studies have been conducted to elucidate
   the genetic architecture of PCOS, with the majority of the studies
   using whole-exome sequencing methodology^[51]8,[52]9,[53]23–[54]26.

   The current advances in genomics research of PCOS have primarily
   focused on family cohorts to better understand the disease’s
   inheritance pattern, whereas the genetic mechanism behind
   obesity-mediated PCOS remains unknown. With this background, the
   present study has attempted to explore the clinically relevant
   mutational landscape of PCOS emphasizing the obese and lean category
   subject. In addition to this, this study has tried to elaborately
   represent the pathways enrichment, and functionality enrichment using a
   holistic bioinformatics approach so that further risk can be predicted
   and managed cost-effectively.

Materials and methods

Selection of study subjects

   This study was conducted in urban and peri-urban areas of West Bengal,
   India, based on the epidemiological survey^[55]13. For subject
   selection, Rotterdam criteria were employed for confirmation of PCOS
   and two representative probands (N = 2) were included in this study.
   The reason behind the selection of these subjects due to their opposite
   anthropometric features.

   To start with, all demographic and lifestyle details have been
   collected. Pathological screening like blood parameters and pelvic
   ultrasonography have been done along with other phenotypic assessments
   such as hirsutism, acne, alopecia,acanthosis nigricans, skin
   tags/moles. Pedigree data was also taken to understand familial
   aggregation. Subjects provided written informed consent for the genetic
   research studies, which were performed following the study protocol
   approved by the Institutional ethical committee of the University of
   Calcutta (Ref. No. 07/ET/20-21/1777) and in concordance with the
   Helsinki Declaration of 1975, as revised in 2008.

Biological sample collection and preparation

   3 mL of whole blood was collected from each subject and genomic DNA was
   extracted from peripheral blood mononuclear cells (PBMCs) through
   density gradient centrifugation using HiSep™ (Himedia, India, LS001),
   followed by TriZol (Invitrogen™, United States) method. The
   concentration and purity of extracted DNA were assessed using a
   NanoDrop instrument (NanoDrop Technologies, Wilmington, DE, USA).

Clinical exome sequencing

   For quality control, genomic DNA was further quantified using a Qubit
   assay (Invitrogen, USA). Exome regions were captured using the Twist
   Human Comprehensive Exome Kit (Twist Bioscience, San Francisco, CA,
   USA). Paired-end sequencing (2 × 150 bp) was performed using an
   Illumina HiSeqX / NovaSeq next-generation sequencing platform (San
   Diego, CA, USA), yielding a minimum read depth of 80-100X. The
   sequences were aligned with the human reference assembly (hg38). The
   standard pipeline using the GATK tools for exome sequencing data
   analysis which has been shown to yield results of higher quality was
   used for analysis.

Variants screening using bioinformatics approach

   Ultra-filtration was done in different approaches using VEP, Variation
   origin, OMIM report and Variation class. The functional enrichment
   analysis in terms of both pathways and ontology enrichment was done
   using the EnrichR bioinformatic tool^[56]27–[57]29 and visualization
   was done using Appyters
   ([58]https://appyters.maayanlab.cloud/Enrichment_Analysis_Visualizer)
   tool. For pathway enrichment analysis, 6 websites viz. Reactome 2022,
   KEGG 2021, Elsevier Pathway Collection, WikiPathway 2023, Biocarta 2016
   and Panther 2016 has been used. For ontology analysis, Biological
   process, Cellular components and Molecular functions has been taken
   into consideration. Protein–protein interaction (PPI) network was
   created using another web-based tool viz. NetworkAnalyst v3.0
   ([59]https://www.networkanalyst.ca/NetworkAnalyst/).

Results

Description of the study participants

   This study encompasses two Indian Hindu Bengali families with 2 PCOS
   probands, one PCOS lean phenotype and another PCOS obese phenotype.
   Demography and anthropometric features have been defined in Table
   [60]1.

Table 1.

   Demographic and anthropometric details of the Subjects participate in
   the present study.
   Features PCOS-Lean PCOS-Obese
   Age (years) 28 29
   Residential status Urban Peri-urban
   Economic status Higher-middle class
   Literacy Yes
   Food habit Non-vegetarian
   Any medications Yes (OCP + Metformin + Folate supplement) Yes (only
   OCP)
   Physical activity High Low to moderate
   Family history No Yes
   Blood pressure 136/89 118/78
   Pulse rate 94 101
   Body weight (kg) 45 79.4
   Total fat content (%) 17.7 37.6
   Visceral fat content (%) 0.5 9
   BMI (kg/m^2) 19.5 29.5
   Resting metabolism 919 1532
   Body age (years) 18 46
   Subcutaneous mass content (%)
    Whole body 14 34.5
    Trunk 9.7 30.4
    Arm 28.6 52.2
    Leg 24.4 50.2
   Skeletal muscle mass content (%)
    Whole body 29.4 23.4
    Trunk 25.8 17.3
    Arm 37.1 20.5
    Leg 39.6 37.1
   [61]Open in a new tab

   Proband-1 (PCOS lean phenotype): At the time of this study, the age of
   the subject was 28 years and the age of onset was 22 years. The pelvic
   ultrasonography clearly showed the bilateral polycystic ovaries; the
   right ovary was enlarged (16.8 cc vol.) and the left ovary was normal
   in size (11.6 cc vol.) with peripheral pea-sized multiple follicles.
   The level of Thyroid stimulating hormone (TSH), Prolactin, Leutinizing
   hormone (LH) Follicle stimulating hormone (FSH), Anti-mullerian hormone
   (AMH) and fasting insulin were 1.3 µIU/mL, 8.9 ng/mL, 1.9 mIU/mL, 2.7
   mIU/mL, 7.29 ng/mL and 9.3 U/mL respectively. Other phenotypic
   characteristics including acne, acanthosis nigricans, and hirsutisms
   were not seen, but a slight alopecia was observed. Therefore, the
   subject belongs to Phenotype-D, i.e., Non-hyperandrogenic PCOS
   category. The family pedigree is depicted in Fig. [62]2a.

Fig. 2.

   [63]Fig. 2
   [64]Open in a new tab

   Familial aggregation chart. (a) Lean PCOS subject, (b) Obese PCOS
   subject.

   Proband-2 (PCOS obese phenotype): The age of the subject (at the time
   of this study) was 29 years and the age of onset was 16 years. The
   hormonal profile and fasting glucose data were 4.76 µIU/mL,
   26.70 ng/mL, 10.46 mIU/mL, 4.74 mIU/mL, 2.85 ng/mL and 110.5 mg/dL
   respectively. The pelvic features showed bilateral polycystic ovaries
   with both enlarged ovaries; the right ovary was 13 cc vol. and the left
   ovary was 14 cc vol. with peripheral pea-sized multiple follicles.
   Subject exhibited extreme hirsutism and slight alopecia whereas acne
   and acanthosis nigricans were not observed. Therefore, the subject
   belongs to Phenotype-A, i.e., Full-blown PCOS category. The family
   history is depicted in Fig. [65]2b.

Variants summary

   A total of 26,550 variants were identified in the clinically important
   exomes from two samples, among which 93.56% were SNPs and the rest were
   InDels. After filtering and eliminating the synonymous variations, we
   identified 92.68% missense variants and the rest contained nonsense,
   startloss, stoploss, frameshift-InDels and inframe-InDels.

   After passing through variant filtering, 5170 variations were found to
   be common in all samples whereas 2232 and 2322 variations were unique
   among PCOS lean phenotype and PCOS obese phenotype respectively. Among
   5170 variants, only 262 variants from 131 genes were PCOS-specific,
   according to the PCOSKb database
   ([66]https://pcoskb.bicnirrh.res.in/index.php). Similarly, 120 variants
   from 70 genes were PCOS-specific in lean PCOS and 94 variants from 70
   genes in obese PCOS.

Classification of variants and functional enrichment analysis

   Figure [67]3 depicts the distribution of various types of mutations. To
   increase the chance of detecting biological processes related to the
   above genetic variants, all variants, regardless of the expected
   effect, were included in functional and pathway enrichment analysis.
   Functional enrichment analysis on 131 genes to which the above 262
   variants (Supplementary Fig. 1 and 2) were assigned showed predominant
   biological processes like positive regulation of protein
   phosphorylation, cytokine production, protein kinase B signalling,
   regulation of MAPK cascade, etc. which involved in molecular functions
   like steroid hydroxylase activity, heme binding, IGF-binding,
   transcription regulatory region nucleic acid binding, Wnt receptor
   activity, etc. However, in the lean-PCOS subject, analysis of 70 genes
   (Supplementary Fig. 3) showed the most important biological processes
   like cellular response to peptide hormone stimulus, response to
   insulin, regulation of protein phosphorylation, regulation of glucose
   import, etc. thereby participating in molecular functions such as
   transmembrane receptor protein tyrosine kinase activity,
   sequence-specific dsDNA binding, steroid hydroxylase activity, ATP
   binding, transcription cis-regulatory region binding, etc.
   (Fig. [68]4). Similarly, in obese-PCOS subject, analysis of 70 genes
   (Supplementary Fig. 4) showed important biological processes like
   regulation of glucose import, negative regulation of cell
   differentiation, cellular response to insulin stimulus, regulation of
   fat cell differentiation, etc. thereby participating in molecular
   functions such as LDL particle binding, phosphatidylinositol
   phospholipase activity, IGF binding, Ca^2+ binding, protein
   homodimerization activity, etc. (Fig. [69]5).

Fig. 3.

   [70]Fig. 3
   [71]Open in a new tab

   Distribution of different types of mutations found in the study
   participants.

Fig. 4.

   [72]Fig. 4
   [73]Open in a new tab

   Functional enrichment analysis in Lean PCOS. (a,c,e) Bar charts of top
   10 terms from the GO-2023 database gene set library, based on a
   combined score of p-value and q-value; (b,d,f) Volcano plots of terms
   from the GO-2023 database gene set; each point represents a single
   term, plotted by the corresponding odds ratio in X-axis and -log[10]
   p-value in Y-axis from the enrichment results of the input query gene
   set; larger and darker-coloured the point, the more significantly
   enriched the input gene set for the term.

Fig. 5.

   [74]Fig. 5
   [75]Open in a new tab

   Functional enrichment analysis in Obese PCOS. (a,c,e) Bar charts of top
   10 terms from the GO-2023 database gene set library, based on a
   combined score of p-value and q-value; (b,d,f) Volcano plots of terms
   from the GO-2023 database gene set; each point represents a single
   term, plotted by the corresponding odds ratio in X-axis and -log[10]
   p-value in Y-axis from the enrichment results of the input query gene
   set; larger and darker-coloured the point, the more significantly
   enriched the input gene set for the term.

   The pathway enrichment analysis of 131 common genes was depicted in
   Supplementary Fig. 5 and 6. In lean-PCOS (Supplementary Fig. 7), HIF-1
   signalling, PI3K-Akt signalling, AMPK signalling, complement &
   coagulation cascading, etc. play the major role in PCOS manifestation
   (Fig. [76]6) whereas in obese-PCOS (Supplementary Fig. 8), longevity
   regulating pathway, pathways in cancer, FoxO signalling, GPCR
   downstream signalling plays a crucial role (Fig. [77]7).

Fig. 6.

   [78]Fig. 6
   [79]Open in a new tab

   Pathway enrichment analysis in Lean PCOS. (a,c,e,g,i,k) Bar charts of
   top 10 terms from pathway analysis database gene set library, based on
   a combined score of p-value and q-value; (b,d,f,h,j,l) Volcano plots of
   terms from the pathway analysis database gene set; each point
   represents a single term, plotted by the corresponding odds ratio in
   X-axis and -log[10] p-value in Y-axis from the enrichment results of
   the input query gene set; larger and darker-coloured the point, the
   more significantly enriched the input gene set for the term.

Fig. 7.

   [80]Fig. 7
   [81]Open in a new tab

   Pathway enrichment analysis in Obese PCOS. (a,c,e,g,i,k) Bar charts of
   top 10 terms from pathway analysis database gene set library, based on
   a combined score of p-value and q-value; (b,d,f,h,j,l) Volcano plots of
   terms from the pathway analysis database gene set; each point
   represents a single term, plotted by the corresponding odds ratio in
   X-axis and -log[10] p-value in Y-axis from the enrichment results of
   the input query gene set; larger and darker-coloured the point, the
   more significantly enriched the input gene set for the term.

   The PPI network for both subjects has been demonstrated in Fig. [82]8.
   Genes with higher scores in terms of node degree and betweenness were
   considered ‘hub’ genes. The hub genes include EGFR, NFKB1, AR, NCOR1,
   ERBB2, PIK3R1, INSR, VEGFA, RUNX2 and AKT3 in the case of PCOS subject
   with lean phenotype while AKT1, YAP1, PTEN, PIK3CA, BRCA1, TLR2, INSR
   and TLR2 in the case of PCOS subject with obese phenotype.

Fig. 8.

   [83]Fig. 8
   [84]Open in a new tab

   Protein–protein interaction network. (a) Lean PCOS, (b) Obese PCOS.

Variants prioritization

   To shortlist the most potent clinically relevant variants, 3 filters
   have been applied to the clinical-exome sequencing data, i.e. only
   germline mutations with ‘High’ VEP and OMIM ID have been selected.
   Therefore, 4 unique variants were found in the lean-PCOS subject, 2
   unique variants were found in the obese-PCOS subject and 5 common
   variants were found in both (Table [85]2). Among them, VDR (rs2228570),
   CYP21A2 (rs7755898), KISS1 (rs71745629) and XDH (rs773456900) were
   already found as PCOS-associated genes.

Table 2.

   Summary of the important candidate variants from clinical-exome
   sequencing of PCOS subjects.
   Subject Gene (rsID) Variation class Inheritance Clinical significance
   Zygosity Associated phenotypes 1000G gnomAD SAS
   Both TMEM216 (rs10897158) Intronic splice acceptor Autosomal recessive
   Benign Homo.^1 Joubert syndrome, Meckel syndrome 0.719649 0.819658
   Both VDR (rs2228570) Start-loss Autosomal recessive Likely pathogenic
   Hetero.^2 Vitamin-D deficiency with or without alopecia 0.671526
   0.738688
   Both CTU2 (rs11278302) Intronic splice donor Autosomal recessive Benign
   Homo Microcephaly, facial dysmorphism, renal agenesis, and ambiguous
   genitalia syndrome 0.773762 0.844626
   Both SON (rs34373121) Frameshift insertion Autosomal dominant Benign
   Homo ZTTK syndrome 1 0.999793
   Both SON (rs34377180) Frameshift deletion
   L-PCOS CHIT1 (rs150192398) Nonsense Autosomal recessive Benign Hetero
   Chitotriosidase deficiency 0.289137 0.427436
   L-PCOS FUT6 (rs145035679) Nonsense NA Uncertain Hetero
   Fucosyltransferase-6 deficiency 0.0289537 0.0532974
   L-PCOS Cyp21A2 (rs7755898) Nonsense Autosomal recessive Pathogenic
   Hetero Hyperandrogenism, congenital adrenal hyperplasia due to
   21-hydroxylase deficiency NA^3 0.0128205
   L-PCOS Fam20C (rs774848096) Nonsense Autosomal recessive Benign Hetero
   Raine syndrome NA 0.317136
   Ob-PCOS KISS1 (rs71745629) Stop-loss Autosomal recessive Benign Hetero
   Congenital idiopathic hypogonadotropic hypogonadism 0.221845 0.194549
   Ob-PCOS XDH (rs773456900) Frameshift insertion Autosomal recessive
   Pathogenic Hetero Xanthinuria NA 0.0008285
   [86]Open in a new tab

   ^1Homozygous

   ^2Heterozygous

   ^3Not found

   Bold represents the PCOS-specific genes according to PCOSkb database.

   To find out candidate mutations in mitochondrial DNA, 7 unique variants
   were found in the lean-PCOS subject, 3 unique variants were found in
   the obese-PCOS subject and 5 common variants were found in both (Table
   [87]3).

Table 3.

   Summary of shortlisted variations in mitochondrial DNA of PCOS women.
   Subject Gene (rsID) Variation class Variation impact Gene type Zygosity
   Allele frequency Ontology
   Both CYB (rs527236041) Missense (Thr7Ile) Moderate Protein Homo NA
   Oxidoreductase and ubiquinol-cytochrome-c reductase activity
   CYB (rs2853508) Missense (Thr194Ala) Moderate Protein Homo NA
   RNR1 (rs2001030) Exonic-NC Modifier rRNA Homo NA NA
   RNR1 (rs2853518) Exonic-NC Modifier rRNA Homo NA
   ATP6 (rs2001031) Missense (Thr112Ala) Moderate Protein Homo NA ATP
   hydrolysis activity and proton transmembrane transporter activity
   L-PCOS ATP6 (rs879150284) Missense (Thr45Ala) Moderate Protein Homo NA
   ATP6 (novel) Missense (Thr13Ala) Moderate Protein Homo NA
   TL2 (rs2853498) Exonic-NC Modifier tRNA Homo NA NA
   ND1 (rs200180511) Missense (Thr263Ala) Moderate Protein Homo NA NADH
   dehydrogenase (ubiquinone) activity
   ND2 (rs878939965) Missense (Trp239Cys) Moderate Protein Homo NA
   ND5 (rs28359178) Missense (Ala458Thr) Moderate Protein Homo NA
   RNR2 (novel) Exonic-NC Modifier rRNA Homo NA NA
   O-PCOS RNR2 (rs200040509) Exonic-NC Modifier rRNA Hetero NA
   TC (rs9659239) Exonic-NC Modifier tRNA Homo NA NA
   CO1 (novel) Missense (Ile416Thr) Moderate Protein Homo NA Fe-ion
   binding and electron transfer activity
   [88]Open in a new tab

Comment

Principle findings

   PCOS has become a global concern due to its significant role in female
   infertility, as well as ovarian and/or endometrial cancer. It is one of
   the most complex transgenerationally inherited endocrine-related
   metabolic disorders and is also difficult to detect at an early stage
   of life due to its intricate phenotypic nature. Hitherto, a plethora of
   studies have already revealed the genetic association of PCOS and the
   familial aggregation of PCOS, but the genetic architecture of the
   disease is still not clear. The present study highlights the important
   genetic variations in PCOS in the Bengali population from West Bengal,
   India and attempts to understand the molecular crosstalk of PCOS while
   stratified by BMI. The results may imply the genetic difference between
   a lean PCOS group and an obese PCOS group, which in turn, helps in
   precision medicine invention.

Results in the context of what is known

   Substantial reports have been found the increasing trends of PCOS
   prevalence worldwide. In 2016, Bozdag et al. revealed that the PCOS
   prevalence according to NIH, Rotterdam and AE-PCOS Society was 6%, 10%
   and 10% respectively^[89]27 while these proportions have become 5.5%,
   11% and 7.1% respectively in 2024^[90]28. Overall, global prevalence of
   hirsutism, hyper-androgenism, PCOM and oligo-anovulation were 13%, 11%,
   28% and 15% respectively^[91]27. Nonetheless, studies have also shown
   that the proportion of PCOS-A, PCOS-B, PCOS-C and PCOS-D phenotype were
   44.8%, 14.9%, 16.2% and 19.5% respectively in USA and Europe^[92]29
   whereas that of 23.9%, 46.3%, 21.6% and 8.2% respectively in
   Middle-east countries^[93]30 and PCOS-D was more prevalent in
   East-Asian countries^[94]31. In the last few years, studies have shown
   that the prevalence of PCOS phenotype-A has been decreased from 67.7%
   to 22.72% while that of PCOS phenotype-D has been sharply increased
   from 3.6% to 49.53% in India^[95]32,[96]33. However, Dadachanji and his
   group reported that prevalence of PCOS phenotype-D was highest (63.1%)
   followed by phenotype-A (30%) among lean subjects whereas the
   prevalence of phenotype-A (46.8%) and D (42.2%) were relatively closure
   among obese group^[97]34. In general, PCOS-A possesses insulin
   resistance, excess LH, T, 17αOHPG and ovarian estradiol by involving
   ovarian core pathways whereas PCOS-D possesses excess progesterone by
   involving adrenal regulatory pathways^[98]35. In the present study, the
   Proband-1 exhibit LH : FSH ratio ≤ 1 and increased AMH level suggesting
   the possibility of functional hypothalamic amenorrhea (FHA) which cause
   ovulation dysfunction and ovarian cyst formation^[99]36. Contrasting to
   this, high fasting glucose level (insulin resistance) might disrupt the
   ovarian steroidogenesis mechanism and thereby cause PCOS by initiating
   HPO feedback loop^[100]35.

   In previous studies, VDR polymorphism (rs2228570) aka Fok1 was found to
   be associated with systemic lupus erythematosus^[101]37, chronic
   spontaneous urticaria^[102]38, T2DM^[103]39, Parkinson’s
   disease^[104]40, autoimmune thyroiditis^[105]41, Vit-D
   deficiency^[106]42, and many other disorders^[107]43–[108]45. In PCOS
   individuals, rs2228570 showed significant association with phenotype C,
   i.e., hyperandrogenemia and PCOM^[109]46 and homozygous mutant genotype
   showed significant association with traits like hirsutism, infertility,
   alopecia and acne^[110]47. Lurie et al.^[111]45 reported that each copy
   of the mutant allele was associated with a modest 9% increased risk of
   invasive ovarian carcinoma. In our study, a heterozygous genotype of
   Fok1 has been found in both categories of PCOS women and thus can be
   correlated with the phenotypic features. Polymorphism in Cyp21A2
   (rs7755898) was found to be associated with Congenital Adrenal
   Hyperplasia due to 21-hydroxylase deficiency^[112]48. However, this
   rare variant (~ 1% of the population) has been found in Lean PCOS
   individuals for the first time in our study. The KISS1 gene or its
   polymorphisms have not been reported to be associated directly with
   PCOS yet. However, follicular development is controlled by both
   estrogen production and feedback, with low estrogen levels fine-tuning
   GnRH/LH pulses via negative feedback. Estrogen production rises with
   follicular development, resulting in a surge in GnRH/LH and ovulation
   via positive feedback. Kisspeptin neurons, which are connected to GnRH
   neurons, play an important role in LH preovulatory surge and ovulation
   defoliation. Kisspeptin’s positive feedback and raised LH levels play
   important roles in PCOS pathogenesis^[113]49. In this study, KISS1
   (rs71745629) has been found in Obese PCOS subject as a unique variant.
   It is noteworthy to mention that this stop-loss mutation was earlier
   found to be associated with metastatic cancer patients via
   VEGF-dependent angiogenesis signalling^[114]50. Besides this, another
   polymorphism in the XDH gene (rs773456900) has been found in our study
   which is a very rare variant and has not been reported to be associated
   with any disease. Apart from these PCOS-associated genes, TMEM216
   (rs10897158), CTU2 (rs11278302), SON (rs34373121 & rs34377180) were
   found in both PCOS subjects whereas CHIT1 (rs150192398), FUT6
   (rs145035679), Fam20C (rs774848096) were found in lean PCOS.

   In this study, the family pedigree tree showed that T2DM is the most
   predominant metabolic disorder among both probands’ families indicating
   a genetic pattern of shared architecture with T2DM. In addition,
   hypertension has been frequently found in the case of the Lean-PCOS
   family, which denotes the cardiometabolic risk in the subject. However,
   PCOS was not detected in any of the family members of lean-PCOS
   although hypomenorrhea on the paternal side, but it was detected in the
   mother of obese PCOS, thereby confirming the non-mendelian inheritance
   pattern of PCOS. Dhar and her colleagues^[115]51 predicted mutations in
   11 loci, i.e. LepR, CAPN10, INSR, IRS1, CYP1A1, CYP11A1, CYP17A1,
   CYP19A1, ESR1, FSHR, and HSD11B1 which are functionally co-expressed
   with 5 loci (Lep, INS, FDX1, PTPN1 and PTPN11). In another study,
   mutations in 6 loci, i.e. ERBB4, GATA4, INSR, LHCGR, SUOX and YAP1 were
   predicted to have a potential effect on the manifestation of
   PCOS^[116]52. In this study, mutations in LepR, CAPN10, INSR, GATA4,
   SUOX, CYP1A1, CYP19A1 and ESR1 were observed in both category subjects
   while mutations in Lep, ERBB4 were found in lean PCOS and mutations in
   YAP1 was found in obese PCOS. Combining results suggest that leptin
   signalling impairment and insulin resistance thus play a major role in
   PCOS of the study population. Besides this, mutations in CYP1A1,
   CYP19A1 and ESR1 genes disrupt ovarian steroidogenesis and initiate a
   negative feedback loop, thereby leading to estrogen sensitivity and
   aromatase deficiency^[117]53,[118]54. Additionally, mutations in AR
   (affecting androgen signalling), AMH (affecting the secretion of AMH),
   AdipoR1 (affecting adiponectin signalling), NAMPT, NPY (affecting
   energy homeostasis), PTEN, EGFR, Akt (cancer signalling) were observed.

Clinical implications

   Comorbidity analysis using the data from the present study showed in
   Fig. [119]9. Based on genetic network of the study participants, it was
   found that the risk for estrogen resistance, leptin receptor
   deficiency, MTHFR/folate deficiency, T2DM and acanthosis to co-occur
   with PCOS was much higher while shared-gene based analysis showed very
   high risk for T2DM to co-occur with PCOS. Overall, this study may give
   a comprehensive overview of PCOS among young women from West Bengal,
   India emphasizing obesity as the major phenotypic expression.
   Henceforth, folate supplementation, insulin sensitizers can effectively
   act as a therapeutic drug upon the associated genetic risk factors.

Fig. 9.

   [120]Fig. 9
   [121]Open in a new tab

   Comorbidity analysis of PCOS individuals based on (a) network-based
   separation of shared genes, (b) shared genes.

Research implications

   Previously, a whole-genome association study on women from Asian and
   European populations revealed that DENND1A, THADA, FSHR, LHCGR, AMH,
   AMHR2, ADIPOQ, FTO, HNF1A, CYP19, YAP1, HMGA2, RAB5B, SUOX, INSR, and
   TOX3 were the common genes associated with PCOS^[122]55. In Table
   [123]4, the comprehensive summary of genomics study on PCOS subjects.
   With this background, our study, for the first time, attempted to
   identify and differentiate the genetic outlook of PCOS between Obese
   and lean subjects. Our findings support future exploration with a
   statistically significant sample size to validate the high-throughput
   data. On the other hand, the identified genetic variations in this
   study must be extrapolated using a sanger-sequencing technique in a
   large population size in order to establish the variant as an early
   indicator of PCOS. Nonetheless, protein–protein interaction network and
   comorbidity analysis, when combined, show the primary signalling
   pathways and their ultimate fate, implying the extent of risk detection
   at the pre-pubertal stage and protection from future or long-term
   issues.

Table 4.

   Comprehensive summary of identified genetic variations among different
   study populations.
   Study population Study design Methodology Sample size Observation
   References