LncRNA XIST alters the balance of peripheral blood immune cells in systemic lupus erythematosus by regulating the miR-17-92, OFLM4 and CEACAM8 axis

X-inactive-specic transcript (XIST) has been shown to silence linked genes on the X chromosome that may be related to the pathogenesis of systemic lupus erythematosus (SLE) in female patients. However, the function of XIST in SLE at other levels remains unclear. The present study aimed to clarify the correlations between XIST expression and SLE clinical features and the contribution of XIST to SLE pathogenesis at the transcriptome level. by quantitative-polymerase reaction. Bioinformatics methods used to explore of blood cell; Hb, hemoglobin; PLT, platelet; Treg, regulatory T cells; ANA: antinuclear antibodies; SSA, single-stranded DNA; anti-SSB, anti RNA-protein complex antibodies; RNP, ribonucleoprotein; RPP, Ribosomal P protein.

Abstract Background X-inactive-speci c transcript (XIST) has been shown to silence linked genes on the X chromosome that may be related to the pathogenesis of systemic lupus erythematosus (SLE) in female patients. However, the function of XIST in SLE at other levels remains unclear. The present study aimed to clarify the correlations between XIST expression and SLE clinical features and the contribution of XIST to SLE pathogenesis at the transcriptome level.

Methods
Expression of XIST in 79 SLE patients and 23 healthy controls was detected by quantitative-polymerase chain reaction. Bioinformatics methods were used to explore the function and regulatory mechanism of XIST.

Results
Expression of XIST was signi cantly upregulated in SLE patients compared with healthy controls, and had a high diagnostic value for SLE. Importantly, SLE patients with high expression of XIST tended to have elevated levels of total T cells and CD8 + T cells, but reduced levels of Treg cells and NK cells.
Bioinformatics analyses suggested that XIST may regulate the expression of OLFM4 and CEACAM8 by acting as a spongy body for miR-20a, miR-92a, miR-106a, and miR-449a. Furthermore, OLFM4 and CEACAM8 are signi cantly upregulated in SLE patients and had signi cant positive correlations with expression of XIST.

Conclusions
We propose that XIST may alter the balance of peripheral blood immune cells in SLE by acting as a spongy body for the miR-17-92 cluster and promoting the expression of OLFM4 and CEACAM8, resulting in immune dysregulation and tissue damage in SLE.

Background
Systemic lupus erythematosus (SLE) is one of the most heterogeneous autoimmune diseases, but mainly occurs in women and involves multiple systems [1]. The diversity of SLE brings great challenges to its clinical diagnosis and treatment. Various pathogenic factors, including environmental, genetic, epigenetic, hormonal, and immune, lead to a variety of clinical manifestations and organ damage in SLE [2].
However, the pathogenesis of SLE remains unclear. To better understand SLE and provide effective treatments for patients, it is important to clarify the mechanism of the disease.
Epigenetic and immune factors are susceptibility factors for SLE [2,3]. Regulation of gene expression by non-coding RNAs is one of the epigenetic processes. Long non-coding RNAs (lncRNAs) comprise a class of non-coding RNAs that exceed 200 nt and have a variety of biological functions, including growth and development, immune regulation, and tumour microenvironment regulation. Many studies have been shown that lncRNAs were involved in disease activity, immune regulation, and tissue injury in SLE [4][5][6].
However, the understanding of lncRNAs in SLE remains insu cient.
As a female-dominated disease, clari cation of the related genes on the X chromosome is important to achieve a deeper understanding of SLE. X-inactive-speci c transcript (XIST) is a lncRNA that can silence gene expression on the inactive X chromosome and lead to X chromosome inactivation [7]. Recently, several investigators have reported a role of XIST in skewed allelic expression on the X chromosome in lymphocytes and X-linked gene dose compensation in T cells of SLE patients [8,9]. These ndings strongly suggest that XIST is involved in the pathogenesis of SLE. However, the studies were mainly focused on gene silencing on the X chromosome and did not investigate the function of XIST at other levels such as the transcriptome level. In addition, the number of samples used for veri cation was small (≤ 10 samples).
In the present study, we used a large sample (79 SLE patients and 23 healthy controls) to detect the expression of lncRNA XIST in SLE and analyze its correlations with the clinical characteristics of SLE patients. In addition, through bioinformatics methods, the target miRNAs of XIST were predicted, and a competitive endogenous RNA (ceRNA) network was constructed to reveal the regulatory function of XIST at the transcriptome level. We found that XIST not only has the function of gene silencing on the X chromosome in SLE, but can also play a role in altering the balance of peripheral blood immune cells in SLE patients and promoting the onset and progression of the disease by regulating miRNA and gene expression levels. Our study expands the understanding of XIST in the pathogenesis of SLE and may provide a basis for the development of new diagnostic and therapeutic approaches.  Construction of a ceRNA network Cytoscape 3.8.0 software (https://cytoscape.org/) was used to construct and visualize the ceRNA network. The Cytohubba plug-in was used to identify hub genes [12].

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis
An online tool, WebGestalt (www.webgestalt.org/), was used for Gene Ontology enrichment analysis of the target genes. ClueGO + CluePedia, a plug-in of Cytoscape, was used for KEGG enrichment analysis and the results were shown in a visual network.
Statistical analysis IBM SPSS Statistics 25 software (IBM Inc., Chicago, IL, USA) was used to analyze the clinical data of all samples, draw receiver operator characteristic (ROC) curves, and calculate the areas under the ROC curves (AUCs). GraphPad Prism 8.0 software (GraphPad Inc., San Diego, CA, USA) was used to draw scatter diagrams and perform correlation analyses. Student's t-test was used to compare the differences between the two groups where applicable. P < 0.05 was considered statistically signi cant.

Expression of XIST in SLE patients and its correlations with clinical traits
Initially, we detected the expression of XIST in the SLE patients and healthy controls by q-RT-PCR. We found that XIST was signi cantly upregulated in SLE patients compared with healthy controls (Fig. 1A, p = 0.0043). Through a ROC curve analysis, we found that XIST had a high diagnostic value for SLE (Fig. 1B, AUC = 0.762, 95% CI: 0.658 to 0.867, p = 0.000136). We further analyzed the correlations between XIST expression and clinical characteristics of SLE patients. Compared with male patients and older patients, XIST was signi cantly upregulated in female patients and younger patients ( Fig. 1C and 1D). Moreover, XIST was highly expressed in patients with arthralgia, anti-Ro-52 antibody positivity, and elevated WBC (normal range, 4-10 g/L) (Fig. 1E-1G). Interestingly, XIST expression was increased in patients with more T cells or fewer NK cells (Fig. 1H and 1I).
XIST may regulate the balance of peripheral blood immune cells To better explore the relationships between XIST and peripheral blood immune cells, we obtained detailed immune cell test results for the patients and divided the patients into two groups based on their XIST expression. Consistent with our observations, total T cells ( Fig. 2A, p = 0.041) were elevated in the high expression group, while NK cells were reduced (Fig. 2G, p = 0.0082). In addition, we found that CD3 + CD8 + T cells (Fig. 2B, p = 0.0301) were elevated in the high expression group, while CD4/CD8 ratio (Fig. 2D, p = 0.045), Treg cells (Fig. 2E, p = 0.0308), and activated CD25 + lymphocytes (Fig. 2F, p = 0.0389) were reduced. However, CD3 + CD4 + T cells (Fig. 2C) and total B cells (Fig. 2H) did not differ signi cantly between the two groups. We further found that total T cells ( Construction of a ceRNA network and functional enrichment analysis of the target genes A lncRNA exerts its biological functions through its target miRNAs and target genes. In the present study, we obtained 16 target miRNAs through prediction by online tools (Fig. 4A). According to the ceRNA hypothesis, an lncRNA competitively binds to an miRNA to regulate gene expression [13]. Through a literature review, we found that 8 of the 16 miRNAs were signi cantly downregulated in PBMCs from SLE patients, namely miR-17, miR-19b, miR-20a, miR-92a, miR-106a, miR-125b, miR-196a, and miR449a [14][15][16][17][18]. Next, the target genes of these downregulated miRNAs were predicted by four online tools. The intersected results for the target genes were further intersected with our previous transcriptome sequencing results (GEO accession: GSE162828; we extracted 1508 upregulated differentially expressed genes in SLE). Finally, we obtained 115 upregulated target genes and constructed a ceRNA network that included 1 lncRNA, 8 miRNAs, 115 mRNAs, and 150 edges (Fig. 4B). Hub transcripts (XIST, 8 miRNAs, and 6 mRNAs) were identi ed by the MCC algorithm of Cytohubba (Fig. 4C) [12]. The 6 hub mRNAs were OLFM4, CEACAM8, SH3TC2, PRR11, NR4A2, and IRF1. We further performed GO and KEGG enrichment analyses for the 115 upregulated target genes, and the results are shown in Fig. 4D and 4E. Most enriched GO terms were immune effector process and leukocyte-mediated immunity and immune response. Signi cant pathway enrichment terms were Fibronectin matrix formation, MAP2K and MAPK activation, Bladder cancer, C-type lectin receptor signalling pathway, HIF-1 signalling pathway, and Nuclear receptor transcription pathway.

Correlations between XIST expression and hub genes
We randomly selected two of the six hub genes, OLFM4 and CEACAM8, for further veri cation in samples from 40 SLE patients and 23 healthy controls. Consistent with the sequencing results, OLFM4 and CEACAM8 were signi cantly upregulated in SLE patients ( Fig. 5A and 5B). Next, we performed correlation analyses between the expression of these two mRNAs and XIST in SLE patients. Interestingly, we found that expression of XIST was signi cantly positively correlated with expression of OLFM4 (Fig. 5C, r = 0.3767, p = 0.0166) and CEACAM8 (Fig. 5D, r = 0.5441, p = 0.0003).
These ndings suggest that XIST may play a biological role in SLE through regulation of OLFM4 and CEACAM8.

Discussion
The pathogenesis of SLE is complex and heterogeneous, rendering diagnosis and treatment of the disease very di cult. SLE also has characteristics of female dominance, immune dysregulation, multisystem involvement, and organ damage [2]. Therefore, understanding the mechanisms of the immune dysregulation and tissue damage in female SLE patients is of particular importance. LncRNAs have many functions, including involvement in the progression of various diseases such as SLE [4,19,20]. Previous studies demonstrated that changes in expression of the lncRNA XIST can silence linked genes on the X chromosome that may be related to the pathogenesis of SLE in female patients [8,9]. However, the relationships between XIST expression and SLE clinical features and the contribution of XIST to the pathogenesis of SLE at other levels remain unclear.
In the present study, we found the lncRNA XIST was signi cantly upregulated in SLE patients and had a good diagnostic value for SLE. Furthermore, XIST expression was elevated in young or female patients and patients with arthralgia or anti-Ro-52 antibody positivity. It was reported that SLE mainly affects young women of reproductive age, with initial signs of physical symptoms, rash, and arthritis [1]. This suggests that XIST may be a potential biomarker for the diagnosis of early SLE. However, the sample size in the present study was not su ciently large, and the results require veri cation in a large prospective cohort. Nevertheless, the present study lays a partial foundation for future cohort research.
To explore the effects of XIST on SLE at different levels, we analyzed the correlations between XIST expression and clinical features. Interestingly, we found that patients with elevated XIST expression tended to have high levels of total T cells and CD8 + T cells, but reduced levels of Treg cells, activated CD25 + lymphocytes, and NK cells. T cells contribute to the initiation and persistence of immunity in SLE and are involved in organ damage in the disease [21]. Blanco et al. [22] found that activated CD8 + T cells were signi cantly increased, had a cytotoxic effector T cell phenotype, and generated high levels of soluble nucleosomes and granzyme B in patients with active SLE. Furthermore, CD8 + T cells not only had a role in the blood, but also accumulated and in ltrated in the glomerular guard region of the kidney, leading to tissue injury and organ pathology [23]. On the contrary, in patients with active SLE, the number of CD4 + CD25 + Treg cells was reduced, and their inhibitory function was insu cient, allowing overactivation of other T cells to cause tissue in ammation and damage [24,25]. Like Treg cells, the number of NK cells was signi cantly decreased in patients with active SLE [26]. However, Suárez-Fueyo et al. [27] reported that NK cells in SLE patients exhibited increased cytotoxicity and pro-in ammatory phenotypes that were associated with down-regulation of CD3ζ. These observations are consistent with the present ndings. On the one hand, XIST may promote the increase and overactivation of total T cells and CD8 + T cells by affecting the number and function of Treg cells, leading to immune dysregulation and tissue damage in SLE. On the other hand, XIST may lead to stronger cytotoxicity and in ammatory phenotypes of NK cells by reducing the number of NK cells, and ultimately reducing their protective function. Therefore, we believe that XIST has multiple regulatory effects on the immune system in SLE patients, regulating not only adaptive immunity, but also innate immunity. This is achieved by altering the balance of immune cells in the peripheral blood of SLE patients.
To explore the molecular mechanism of lncRNA XIST involvement in the pathogenesis of SLE, we constructed a ceRNA network that can re ect the regulatory mechanism of an lncRNA at the transcriptome level [13]. A lncRNA can act as a sponge body for miRNAs to regulate gene expression [20].
Here, we found that XIST may affect the expression of 115 genes by regulating 8 miRNAs. Enrichment analyses for these 115 genes indicated that XIST may be involved in leukocyte-mediated immunity and immune response through MAP2K and MAPK activation, C-type lectin receptor signalling pathway, or HIF-1 signalling pathway. These ndings are consistent with our conclusion that XIST can change the balance of peripheral blood immune cells in SLE. Moreover, we selected two hub genes, OLFM4 and CEACAM8, to verify the accuracy and reliability of the predicted results. We found that both of these genes were signi cantly upregulated in SLE patients and had signi cant positive correlations with expression of XIST. These ndings not only indicate that our prediction results are relatively reliable, but also suggest that XIST may regulate the expression of OLFM4 and CEACAM8 by acting as a spongy body for miR-20a, miR-92a, miR-106a, and miR-449a, thereby exerting its functions. Interestingly, miR-20a, miR-92a, miR-106a, and two other miRNAs, miR-17 and miR-19b, belong to the miR-17-92 cluster or its paralog [28]. This family has been shown to play important roles in the immune system, cardiovascular system, and tumours, among others. In the adaptive immune system, miR-17-92 has a key role in the antigen response of T lymphocytes [29]. In Treg cells, miR-17-92 is critical for the function and accumulation of Treg cells during the autoimmune-mediated stress response. Once miR-17-92 is lost, Treg cells lose their regulatory function [30]. Similarly, both the number and function of Treg cells were partially de cient in SLE, which may be related to the low expression of miR-17-92. In CD8 + T cells, miR-17-92 decreased gradually during the differentiation of CD8 + T cells, which was associated with increased proliferation potential [31,32]. These lines of evidence also support our conclusion. However, the association between miR-17-92 and NK cells remains unclear. Further research is needed to explore this issue in the future. It will be an interesting and meaningful topic to pursue.

Conclusions
In conclusion, we propose that the lncRNA XIST may alter the balance of peripheral blood immune cells in SLE by acting as a spongy body for the miR-17-92 cluster and promoting the expression of OLFM4 and CEACAM8, resulting in immune dysregulation and tissue damage in SLE. However, the study still lacks some experimental data to con rm our conclusion, and thus further studies are warranted to supplement our ndings in the future.