Identification of immune cell infiltration and effective biomarkers of polycystic ovary syndrome by bioinformatics analysis

Polycystic ovary syndrome (PCOS) is a common reproductive endocrine metabolic disease in women of childbearing age and is often characterised by chronic anovulation and hyperandrogenaemia. Its clinical manifestations include menstrual disorders, hirsutism, acne, and polycystic ovarian changes, and these symptoms can be accompanied by metabolic diseases, such as obesity, insulin resistance (IR), and dyslipidaemia [1]. However, the pathogenic mechanisms of PCOS are still unclear. In addition to the ovarian-pituitary-hypothalamic-gonadal axis, studies of PCOS pathogenesis must also consider ovarian local cytokines, immunology, and genetics.

Recent studies have found that the inflammatory immune mechanism is closely related to the occurrence and development of PCOS. Under physiological conditions, appropriate inflammatory stress is conducive to the growth and development of oocytes [2, 3]. However, under pathological conditions, the inflammatory response is enhanced, the development of oocytes may be limited, and the further development of chronic inflammation leads to decreased ovum quality, thereby affecting ovulation [4]. Systemic and ovarian cytokines (e.g., tumour necrosis factor [TNF]-α, interleukin [IL]-6, and IL-18) can change the local microenvironment in the ovary, regulate ovarian function, induce excessive androgen production, and promote IR through various mechanisms [5]. Inflammation in the follicular microenvironment may be involved in the dysfunction of the hypothalamic-pituitary-gonad axis and the occurrence and development of follicular dysplasia.

In studies of the distribution of white blood cells in the ovary and cytokine mRNA expression in the follicular fluid (FF) of patients with PCOS and non-PCOS women undergoing fertilisation embryo transfer in vitro, Wu et al. [6] showed that T lymphocytes play important roles in the local pathological mechanisms of PCOS. T lymphocytes secrete various inflammatory and immunomodulatory molecules that participate in the regulation of ovarian function. T-cell subsets have also been shown to be dysregulated in the peripheral blood and ovaries of patients with PCOS owing to disruption of sex hormone levels in these patients [7].

The pathogenesis of PCOS is multifactorial and complex. In addition to reproductive abnormalities, the pathogenic mechanisms also include interactions between the immune system and reproduction, resulting in a variety of changes to cytokines and immune cells. Therefore, evaluation of immune cell infiltration in patients with PCOS based on changes in the expression levels of genes may be essential for elucidating the immunological mechanisms of PCOS and identification of novel biomarkers.

In this study, we aimed to identify the roles of immune cell subsets and related gene expression changes in the pathogenesis of PCOS. We downloaded four PCOS datasets from the Gene Expression Omnibus (GEO) database and analysed differentially expressed genes. Two different machine learning algorithms were then used to further identify PCOS biomarkers. We also studied PCOS from the perspective of immunology using single-sample gene set enrichment analysis (ssGSEA) to evaluate the differences in the compositions of 28 immune cell subsets between patients with PCOS and healthy women of reproductive age. In addition, the relationships between PCOS biomarkers and immune cell infiltration were studied to improve our understanding of the immunological mechanisms of PCOS occurrence and development.

PCOS is one of the most common ovo-derived endocrine diseases in women of reproductive age. Its basic pathophysiological features include hyperandrogenaemia and IR caused by disruption of the local internal environment, cytokine expression, and ovary function.

In this study, we identified DEGs between PCOS and normal controls and showed that the combination of TMEM54 and PLCG2 was a biomarker of PCOS. We also found that central memory CD4⁺ T cells, central memory CD8⁺ T cells, effector memory CD4⁺ T cells, γδT cells, and Th17 cells may influence the occurrence of PCOS. PLCG2 was highly correlated with γδT cells and central memory CD4⁺ T cells.

The follicular microenvironment is composed of FF and granulosa cells (GCs). GCs regulate the local microenvironment of FF through various mechanisms, participate in the metabolism of oocytes, and protect oocytes from the invasion of components in the external environment [20]. The composition of the oocyte GC regulatory loop and follicular microenvironment is critical for the coordination of reproductive activities, and any changes in the composition of FF/GC molecules may affect the quality of oocytes [21]. Therefore, analysis of the immune infiltrating microenvironment of GCs can provide insights into the microenvironment of FF.

As the most important adaptive immune response cells in the immune system, T lymphocytes secrete various inflammatory and immunomodulatory molecules and are involved in regulating multiple ovarian functions, such as follicle formation, ovulation, and luteal degeneration. Activated lymphocytes secrete inflammatory cytokines, including IL-1, IL-6, IL-12, TNF-α, and insulin-like growth factor-1 [22]. In recent years, many domestic and foreign studies have reported that the levels of inflammatory factors, such as IL-6, TNF-α, and C-reactive protein, are higher in patients with PCOS than in normal women [23]. Moreover, inflammatory factors can mediate IR directly or indirectly through adipose tissue, suggesting that the occurrence and development of PCOS are closely related to inflammatory factors [24, 25]. Mature T cells can be categorised as CD4⁺ T cells or CD8⁺ T cells according to their CD molecular phenotype. Studies have shown that the ratio of CD4+/CD8 + is related to IL-2 [26, 27], IL-7 [28] and IL-16 [29]. CD4⁺ T cells and CD8⁺ T cells have different functions, which play synergistic or restrictive roles, and the ratio of CD4⁺/CD8⁺ T cells reflects changes in immune regulation.

Central memory T cells have the ability to expand, differentiate, and self-renew like stem cells and can differentiate unidirectionally into effector memory T cells and short-term effector T cells to prevent chronic infection and cancer [30]. In our study, central memory CD8⁺ T cells and central memory CD4⁺ T cells were increased in patients with PCOS, whereas effector memory CD4⁺ T cells were decreased, albeit without statistical significance. In the local ovarian microenvironment of patients with PCOS, abnormal CD4⁺/CD8⁺ T cell ratios may disrupt immune regulation. However, the differentiation of central memory T cells into effector memory T cells may also be blocked, resulting in increased central memory T cells and decreased effector memory T cells.

The functions of γδT cells were first discovered and studied in the field of autoimmune rheumatism. Although γδT cells are a highly conserved T-cell subpopulation, they have important implications in various aspects of immunobiology [31]. Animal models have shown that γδT cells regulate classical autoantigen reactive αβT cells and B cells [32, 33] and play independent pro-inflammatory roles via direct secretion of IL-17 [34], TNF-α, and interferon (IFN)-γ [35] in a non-antigen-driven pattern. Many studies have confirmed that the levels of TNF-α, IFN-γ, and IL-18 are significantly increased in patients with PCOS and are positively correlated with IR [24, 36, 37]. In this study, the numbers of γδT cells were higher in patients with PCOS than in normal controls, this observation may be related to the observed increases in TNF-α and IFN-γ secretion by high numbers of γδT cells. Notably, CD3⁺CD4⁻CD8⁻ γδT cells are increased in women with recurrent abortion, contributing to foetal loss by regulating transforming growth factor-β and IL-17 secretion and promoting inflammation [38]. The risk of abortion after pregnancy in patients with PCOS is more than three times higher than that in normal women, which may be related to the presence of a chronic inflammatory state and autoimmune disorders, such as active autoimmunity.

The high correlation between the CXC motif chemokine receptor 1 (CXCR1) gene and various immune cells was notable, and the correlation between CXCR1 and neutrophils reached R = 0.72. The protein encoded by the CXCR1 gene is a member of the G-protein-coupled receptor family and acts as a receptor for IL-8. IL-8 is a powerful neutrophil chemokine that binds to receptors and promotes neutrophil activation. CXCR1 binds to IL-8 with high affinity and transduces signals through a second messenger system activated by G proteins [39, 40]. This explains the results of our analysis from a mechanistic level and supports the accuracy of our results.

One of the selected biomarkers, transmembrane protein 54 (TMEM54), is a member of the transmembrane protein family, which contains many proteins with unknown functions. Studies have shown that TMEM expression is different in tumour tissues than in adjacent healthy tissues, and some TMEM family members have been identified as potential prognostic biomarkers in different types of tumours [41]. In addition, TMEM proteins are tumour suppressors or oncogenes and have been shown to be associated with tumour progression and invasion [42, 43] or chemotherapy resistance [44, 45]. Although there are few studies on TMEM54, our current findings suggest that the TMEM54 gene may have important roles in the development of PCOS; further studies are warranted.

Phospholipase C gamma 2 (PLCG2) is a transmembrane signalling enzyme that is an important driver of many immunological aetiological diseases, such as inflammation, autoimmune diseases, immunodeficiencies, and allergies, as well as haematological malignancies. Some studies have demonstrated that point mutations in the PLCG2 gene may be an important cause of severe spontaneous inflammation and autoimmunity [46]. In a meta-analysis of gene expression in patients with rheumatoid arthritis (RA), PLCG2 was found to be upregulated in several datasets, including many pathways associated with RA inflammatory responses, e.g., inflammasome activation, platelet aggregation, and activation, indicating that PLCG2 is a potential target for the control of RA inflammation [47]. Moreover, PLCG2 is important in bone marrow cells, including monocytes, macrophages, NK cells, DCs, and mast cells, possibly because it promotes downstream signalling involving Fc receptors [48]. In a bioinformatics analysis of the tumour microenvironment in soft tissue sarcoma, PLCG2 was found to be an indicator of the tumour microenvironment and patient prognosis [49]. Additionally, CD8⁺ T cells, γδT cells, monocytes, and M1 macrophages were shown to be positively correlated with PLCG2 expression, consistent with our current results.

There were some limitations to this study. First, this study was only carried out from the perspective of gene transcriptome, and multi-omics and mechanistic studies were not performed. Further validation of our bioinformatics results through in vitro and in vivo experiments and clinical practice is needed.

In summary, in this study, ssGSEA was used for the first time to analyse immune infiltration into the follicle microenvironment in patients with PCOS. Central memory CD4⁺ T cells, central memory CD8⁺ T cells, effector memory CD4⁺ T cells, γδT cells, and Th17 cells may be involved in the occurrence of PCOS. In addition, differences in gene expression in ovarian tissues between patients with PCOS and healthy women of reproductive age were determined, and our findings showed that the combination of TMEM54 and PLCG2 was a biomarker of PCOS. PLCG2 was shown to be highly correlated with γδT cells and central memory CD4⁺ T cells. These findings provide a basis for further research on the immunological pathogenesis of PCOS.