We performed a search in the Gene Expression Omnibus (GEO) database using several keywords, including “Healthy Control,” “Wild Type,” “Mice,” “Mus musculus,” “Th9,” “Th2,” and “Expression profiling by array” from January 1, 2012, to December 17, 2020, and selected 2 gene series expressions (GSEs) data for further study: GSE99166 and GSE123501. GSE99166 contained 4 samples of Th2 wild-type cells (GSM2634701, GSM2634702, GSM2634711, and GSM2634712) and 5 samples of Th9 wild-type cells (GSM2634695, GSM2634703, GSM2634704, GSM2634713, and GSM2634714) from the spleen. GSE123501 contained 2 samples of Th2 wild-type cells (GSM3505597 and GSM3505602) and another 2 samples of Th9 wild-type cells (GSM350598 and GSM3505603) from the spleen (Figure 1).

The differentially expressed genes (DEGs) were obtained from the 13 samples of 2 different data sets (GSE99166 and GSE123501) using the GREIN (GEO RNA-seq Experiments Interactive Navigator) platform (BD2K-LINCS Data Coordination and Integration Center). This interactive online web tool analyses GEO RNA-seq data [18]. The DEGs extracted from the data sets comprised genes from Th2 and Th9 cells. As we wanted to assess the alteration of genes during the conversion of Th2 to Th9 cells, the analysis was performed with DEGs of Th2 cells as the standard to which DEGs of Th9 cells were compared. The workflow for the data processing and analysis is portrayed in Figure 2.

The DEGs were considered upregulated when the expression of genes in Th9 cells was higher than that in Th2 cells. The cutoff for the selection was kept at P<.05, and overlapping DEGs between 2 data sets (GSE99166 and GSE123501) on comparison of Th2 and Th9 cells were identified by the Venn diagram tool [19,20]. In addition, the common upregulated, downregulated, and oppositely regulated DEGs of these 2 data sets (GSE99166 and GSE123501) were identified. The fold change expression distribution was visualized by a heat map and violin plot using the Linear Models for the Microarray Data (limma) Package of R (R Foundation for Statistical Computing) and Orange Data Mining (University of Ljubljana) [21,22].

The codifferential genes were divided into 3 parts, namely, (1) common upregulated, (2) common downregulated, and (3) common, oppositely regulated. The top ranked ontological features of all DEGs were analyzed with STRING. The Gene Ontology (GO) terms included the following 3 categories: biological processes, cellular components, and molecular functions. The significant GO terms regulating genes are presented in a radar graph with a negative log10 (false discovery rate). We defined P<.05 as a significant value.

We searched the functionally significant Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways for top ranked significantly altered DEGs using the STRING and WikiPathways databases. We identified important genes participating in each pathway, and selected the top 7 pathways based on negative log10 (false discovery rate) and P values (<.05) that were important for further study.

We downloaded the complete gene list of the top ranked 7 individual pathways with an interaction network from the KEGG database. We revisualized and constructed the pathway with the help of Cytoscape (Cytoscape Team/Institute for Systems Biology; an open-source software platform for visualizing complex networks and integrating these with any type of attribute data) [23] and marked the DEGs that play a significant role in the differentiation of Th2 to Th9 cells.

The top 10 microRNAs (miRs) that targeted the hub genes were predicted by the well-established miR target prediction database miRNet version 22.0 [24], with special emphasis on the selected organism. Default values for the degree of interaction and betweenness were selected. Common miRs and their targeted messenger RNAs (mRNAs) of all groups were sorted by the Venn diagram.

The DEGs that were identified to play a significant role in Th2 to Th9 differentiation were further analyzed for their involvement in various pathways pertaining to specific diseases using DisGeNET (IBI Group) [25], a discovery platform that describes genes, transcription factors (TFs), chemokines, and IL in association with various specific diseases.

The study was approved by the Institutional Ethics Committee of All India Institute of Medical Sciences (AIIMS) Jodhpur (certificate reference number AIIMS/IEC/2019-20/792).

The Mus musculus (C57BL/6) mRNA expression profiles of GSE99167 and GSE123501, which were selected for this study, included the expression profiles of Th2 and Th9 cells obtained from the spleen. We extracted and compared mice spleen samples from 2 different studies to identify genes that are involved in the differentiation of Th2 to Th9 cells. In both groups, 254 common mRNAs were identified, and 634 common DEGs were identified, of which 236 were downregulated and 260 were upregulated. We performed a quality assessment of the selected samples for our expression profiles (Figures 3A-3I; see Tables S1 and S2 in Multimedia Appendix 1, and Multimedia Appendix 2 for larger version of figures).

Our analysis identified genes encoding various ILs and receptors whose differential expression may determine the differentiation of Th2 to Th9 cells. Some important ILs identified were CCL1 (chemokine [C-C motif] ligand 1), CCL20 (chemokine [C-C motif] ligand 20), IL-13, IL-4, IL-12A, and IL-9. The important receptors identified in our analysis were IL-12RB1, IL-4RA (interleukin 4 receptor alpha), CD53 (cluster of differentiation 53), CD6 (cluster of differentiation 6), CD5 (cluster of differentiation 5), CD83 (cluster of differentiation 83), CD197 (cluster of differentiation 197), IL-1RL1 (interleukin 1 receptor-like 1), CD101 (cluster of differentiation 101), CD96 (cluster of differentiation 96), CD72 (cluster of differentiation 72), CD7 (cluster of differentiation 7), CD152 (cytotoxic T lymphocyte–associated protein 4), CD38 (cluster of differentiation 38), CX3CR1 (chemokine [C-X3-C motif] receptor 1), CTLA2A (cytotoxic T lymphocyte–associated protein 2 alpha), CTLA28, and CD196 (cluster of differentiation 196). In addition, the differential expression of various TFs such as FOXP3 (forkhead box P3), IRF8 (interferon regulatory factor 8), FOXP2 (forkhead box P2), RORA (RAR-related orphan receptor alpha), AHR (aryl-hydrocarbon receptor), MAF (avian musculoaponeurotic fibrosarcoma oncogene homolog), SMAD6 (SMAD family member 6), JUN (Jun proto-oncogene), JAK2 (Janus kinase 2), EP300 (E1A binding protein p300), ATF6 (activating transcription factor 6), BTAF1 (B-TFIID TATA-box binding protein associated factor 1), BAFT (basic leucine zipper transcription factor), NOTCH1 (neurogenic locus notch homolog protein 1), GATA3, SATB1 (special AT-rich sequence binding protein 1), BMP7 (bone morphogenetic protein 7), and PPARG (peroxisome proliferator–activated receptor gamma) may influence the differentiation of Th2 to Th9 cells. The expression of the aforementioned immune regulators is represented by a heat map and Venn diagram in Figures 4A-4F (also see Tables S3-S5 in Multimedia Appendix 1, and Multimedia Appendix 2 for larger version of figures).

A GO analysis of DEGs classified them into 3 functional classes (Figures 4G-4I; see Multimedia Appendix 2 for larger versions of figures): cellular component, biological process, and molecular function.

The enrichments for the 3 DEG classes with significantly altered expression are shown in Tables S6-S8 in Multimedia Appendix 1. In the KEGG pathway enrichment analysis, the identified genes were enriched in various KEGG pathways such as cytokines-cytokines interaction, Th1 and Th2 cell differentiation, inflammatory bowel disease (IBD), Th17 cell differentiation, the Fc epsilon RI signaling pathway, the T-cell receptor signaling pathway, and pathways in cancer (Figure 4J and Tables S9 and S10 in Multimedia Appendix 1; see Multimedia Appendix 2 for larger version of figures).

We downloaded the complete protein-protein interaction (PPI) network of the identified KEGG pathways from the KEGG database. The Cytoscape software was used for the construction of the network. The significantly altered DEGs of cytokines, chemokines, receptors, and TFs were highlighted in the respective networks. Our analysis of the KEGG pathway enrichment and PPI network demonstrated that the genes that had a significantly altered expression in Th9 cells when compared with Th2 cells also played a significant role in other immune regulating pathways. These affected pathways were mainly involved in cytokines-cytokines interaction, Th1 and Th2 differentiation, CTLA4 (cytotoxic T lymphocyte–associated protein 4) regulation, T-cell receptor signaling, Fc epsilon signaling, Th17 cell differentiation, IBD, and cancer. The concurrent presence of these genes in the aforementioned pathways highlights the significance of the differentiation of Th2 to Th9 in diseases where these pathways are affected. The role of the identified DEGs in these pathways and their interaction with other genes has been depicted in Figures 5-7. See Multimedia Appendix 2 for larger images.

We performed a gene similarity analysis to find similar genes in all the 7 KEGG pathways identified with the help of the Venn diagram and calculate the percentage of similarity among the genes that were altered. We observed that 7/13 (54%) genes were similar between the “Th1 and Th2 cell differentiation” and “IBD” pathways, whereas 7/14 (50%) genes were similar between the “Th1 and Th2 cell differentiation” and “Th17 cell differentiation” pathways (Figure 7A; see Multimedia Appendix 2 for larger images).

To explore the posttranscriptional regulation of the identified DEGs, we predicted the miRs that could target the identified DEGs. We identified the following 53 common miRs that could target the DEGs listed in our analysis: let-7b-5p, let-7f-5p, let-7g-5p, let-7i-5p, miR-1-3p, miR-101-3p, miR-103a-3p, miR-107, miR-10a-5p, miR-10b-5p, miR-122-5p, miR-124-3p, miR-129-3p, miR-130a-3p, miR-133a-3p, miR-139-5p, miR-147a, miR-155-5p, miR-16-5p, miR-17-5p, miR-186-5p, miR-195-5p, miR-200b-3p, miR-20a-3p, miR-20a-5p, miR-20b-5p, miR-21-3p, miR-21-5p, miR-22-3p, miR-224-5p, miR-24-3p, miR-26a-5p, miR-26b-5p, miR-27-5p, miR-27a-3p, miR-302a, miR-30a-5p, miR-30c-5p, miR-30d-5p, miR-320a, miR-34-5p, miR-374-5p, miR-426, miR-429, miR-618, miR-6499-3p, miR-93-5p, miR-98-5p, miR-103a-3p, miR-139-5p, miR-147a, miR-195-5p, and miR-27a-5p.

Subsequent to the identification of pathways affected as a result of the alteration of DEGs found in our analysis, we further searched for possible diseases whose pathogenesis is affected by alterations in these pathways. We listed the diseases where alterations in the aforementioned 7 pathways have already been documented, and these were as follows: metabolic diseases (eg, diabetes mellitus, diabetic nephropathy, hyperactive behavior, hypertensive disease, ischemic stroke, steatohepatitis, liver fibrosis, obesity), allergic and pulmonary diseases (eg, asthma, chronic obstructive airway disease, chronic rhinosinusitis, nasal polyps, pulmonary hypoplasia, hay fever), carcinomas (eg, acute leukemia and myelocytic leukemia, B-cell lymphomas, lymphoma, adenocarcinoma, breast carcinoma, carcinoma of the lung, cervical cancer, colorectal carcinoma, glioblastoma and glioma, liver carcinoma, malignant neoplasm of the stomach, melanoma, neuroblastoma, osteosarcoma, pancreatic carcinoma, prostate carcinoma, stomach carcinoma), neuropsychiatric disorders (eg, mental depression, schizophrenia, Alzheimer disease), autoimmune diseases (eg, Graves disease, Crohn disease, colitis, psoriasis, systemic lupus erythematosus [SLE], systemic scleroderma, rheumatoid arthritis, multiple sclerosis [MS], IBD, atopic dermatitis, eczema), and infectious diseases (eg, sepsis, septicemia, tuberculosis, hepatitis, herpes simplex infections, malaria; Figure 8).

In this study, we compared 2 different data sets (GSE99166 and GSE123501) that have compared the mRNA expression in Th2 and Th9 cells. We identified common DEGs that have significantly altered expression between Th2 and Th9 cells from these 2 data sets. Sequential assessment of the DEGs and miRs that had significantly altered expression between Th2 and Th9 cells allows to identify disease states that affect the differentiation process. Although this analysis does not answer whether differentiation of Th2 to Th9 is the cause or the effect of the disease state, it does unravel the possibility of hitherto unknown associations between various diseases and the process of differentiation of Th2 to Th9 cells. Our analysis indicates that differentiation of Th2 to Th9 may play a crucial role via the alteration of DEGs (Table 1) and miRs (Table 2) in various metabolic diseases, allergic and pulmonary diseases, carcinomas, neuropsychiatric disorders, autoimmune diseases, and infectious diseases. In concordance with the existing literature, it was revealed that Th9 cells might play a major role in erythematosus, MS, IBDs, and psoriasis. The role of Th9 cells in autoimmune disease has already been explored in multiple studies [26], including in Graves disease [27], Crohn disease [28-30], psoriasis [31], SLE [32-35], systemic scleroderma [36], rheumatoid arthritis [37-40], MS [26,36,41,42], IBD [26,29,30,43], and atopic dermatitis/eczema [44], which have demonstrated an increased level of differentiation of Th2 to Th9 cells. Th9 cells and IL-9 have been observed in peripheral blood mononuclear cells and synovial fluid from patients with rheumatoid arthritis. Toll-like receptor 2 (TLR2) stimulates naïve CD4⁺ T cells for IL-9 secretion and Th9 differentiation by increasing the expression of TFs BATF and PU.1. TLR2 activation results in increased expression of IL-33 and its receptor ST2, augmenting IL-9 gene expression and Th9 cell development [45]. Similarly, in patients with SLE, Th9 cell differentiation is suppressed by repression of IRF4 expression [46]. Although the role of Th9 has been explored in experimental models of MS and IBD, there is insufficient evidence regarding its role in humans. Th9 cells have been shown to play a pathogenic role in experimental autoimmune encephalomyelitis, an animal model of MS [47]. However, only limited studies have assessed Th9 cells in human patients with MS. The skin toxicity of Th9 cells makes them a crucial link in the pathophysiology of multiple skin diseases [48]. Our study highlights the possibility of Th9 playing a crucial role in the pathophysiology of various autoimmune skin diseases such as eczema, atopic dermatitis, psoriasis, and dermatitis. A predominant expression of IL-9 from Th9 cells was observed to be a characteristic immunologic signature in psoriatic arthritis [49]. Similarly, IL-9 and PU.1 gene expressions in atopic dermatitis were higher and associated with disease severity [50]. In addition, the Th9 cell percentage in patients with atopic dermatitis correlated with serum IgE levels, highlighting the link between allergy and the development of Th9 cells [51]. Our in silico analysis further reiterated the involvement of Th9 in various autoimmune pathways. The involvement of IL-9 and Th9 cells in allergic response can also be seen in other diseases. One such allergic disease in which Th9 cells have been recently explored is asthma. Patients with allergic asthma have increased peripheral blood Th9 cells and elevated levels of serum IL-9 [51]. SGK1 (serum/glucocorticoid regulated kinase 1) has been shown to enhance the differentiation of Th9 by modulating the nuclear factor kappa B (NF-κB) signaling pathway in patients with asthma [52]. The activation of MAPK (mitogen-activated protein kinase) has also been attributed to the activation of Th9 cells in mice models of asthma [53]. Interestingly, IL-9 and IL-13 have been elevated in patients with chronic obstructive airway disease compared with asthma [54]. However, so far, the Th9 cells have not been explored for their significance in the pathophysiology of chronic obstructive pulmonary disease. Interestingly, apart from asthma, our in silico analysis highlighted chronic obstructive airway disease, tuberculosis, and chronic rhinosinusitis with nasal polyps as major airway diseases in which Th9 cells may play a crucial role. Our findings are in sync with the study of Ye et al [55], which demonstrated tuberculous pleural effusion to be chemotactic for Th9 cells, while pleural mesothelial cells in tuberculosis stimulated the Th9 cell differentiation. This in silico analysis also highlights the possible role of Th9 in neuropsychiatric diseases. Very few studies have explored the role of Th9 in neuropsychiatric disorders. Saresella et al [56] have demonstrated an increase in the activity of Th9 lymphocytes, while postthymic maturation pathways showed an accumulation of differentiated effector T lymphocytes (CD4⁺). In Alzheimer disease, schizophrenia, and multiple-episode schizophrenia, although IL-9 has been elevated, limited studies have been performed to assess the role of Th9 cells in the pathophysiology of the diseases [56,57]. In addition to the aforementioned diseases, this study identified malignancies as one of the disease states that could be affected by the development of Th9 cells. The role of Th9 cells in modulating immunity in cancer has been widely explored. Th9 cells contribute to antitumor immunity by enhancing the recruitment and activation of mast cells, natural killer cells, CD8 T cells, and dendritic cells in the tumor microenvironment. The antitumor effect of Th9 cells has been documented in various animal studies. Lu et al [58] have demonstrated the protective effects of IL-9 and Th9 on tumor development. The tumor-specific Th9 cells promoted the activation of CD8⁺ cytotoxic T lymphocytes by recruiting dendritic cells into tumor tissues and subsequently presenting tumor antigens in tumor-draining LNs. Th9 cells in tumor tissues mount an inflammatory response via CTL in a CCL20/CCR6 (chemokine [C-C motif] receptor 6)-dependent manner [59,60]. Wang et al [61] also demonstrated that Th9-enriched CD4⁺ T cells significantly increased the expansion of activated CD8⁺ T cells in a manner that was dependent on the expression of IL-9R (interleukin 9 receptor). Th9 thus seems to enhance antitumor immune response through T-cell cytotoxicity and play a crucial role in controlling the progression of cancer [62]. Apart from Th9 cells, the cytokine IL-9 has also been widely explored in cancers. Expression of IL-9 in the serum and circulating CD4⁺ T cells was significantly upregulated in patients with breast cancer compared with healthy controls [63]. Purwar et al [10] demonstrated that IL-9 depletion in RORγt-deficient mice promoted melanoma growth. Zheng et al [64] demonstrated that Th9 cells produce IL-9 to induce glioma cell apoptosis and inhibit tumor growth. Interestingly, tumor-specific Th9 cells displayed a unique PU.1-TRAF6-NF-kB activation–driven hyperproliferative feature, suggesting a persistence mechanism rather than an antiapoptotic strategy. This equips tumor-specific Th9 cells to become a more effective CD4⁺ T-cell subset for adoptive cancer therapy [65]. Although Th9 cells play an important role in tumor suppression, they have not been studied in various cancer subtypes. Our analysis suggests a possible role for Th9 in different cancer types such as malignant neoplasm of the stomach, melanoma, neuroblastoma, osteosarcoma, pancreatic carcinoma, and prostate carcinoma. Finally, our study also highlights the possible role of Th9 in different metabolic diseases. Interestingly, to our knowledge, no study has yet explored the role of Th9 in metabolic diseases such as diabetes and obesity. We want to highlight these lacunae to open up newer research attempts that would explore the role of Th9 in metabolic diseases. The insights into the role of Th9 in metabolic diseases would better help delineate the role of immunological dysregulation in developing metabolic diseases.

The main limitation of the study is that the analysis is based on an in silico method where only a few specific wild-type samples from data sets of previous studies were included; therefore, further validation of the identified genes and miRNAs is required in various animal models and human diseases. The data sets were compiled using different arrays on the Affymetrix platform, which may account for some of the variability in the results. However, the functional enrichment for the mRNAs highlighted some significant pathways related to immune regulation and its derangements.

This study identified common DEGs of ILs, receptors, and TFs that have significantly altered expression between Th2 and Th9 cells. The KEGG pathway enrichment analysis identified cytokines-cytokines interaction, Th1 and Th2 differentiation, T-cell receptor signaling regulation via CTLA4, Fc epsilon signaling, and Th17 cell differentiation as the significant pathways affected by the identified DEGs. Our study identified hitherto unexplored possible associations between Th9 and disease states. The interactome analysis also identified pathways that are involved in various metabolic diseases, allergic and pulmonary diseases, carcinomas, neuropsychiatric disorders, autoimmune diseases, and infectious diseases, where differentiation of Th2 to Th9 may play a crucial role. The scarcity of studies on the role of Th9 in metabolic diseases highlights the lacunae in this field. Thus, our study provides the rationale for exploring the role of Th9 in various metabolic disorders.