Systemic lupus erythematosus (SLE) remains a poorly understood multifactorial autoimmune disease, with genetic, environmental, and hormonal factors believed to contribute to its etiology (1). The clinical manifestations of SLE are diverse. Due to the complexity of the clinical manifestations and the lack of specific diagnostic criteria, the diagnosis and treatment of SLE are extremely difficult (2). SLE patients often exhibit T lymphocytopenia, reduced inhibitory T cell function, and hyperplasia of B cells, resulting in the production of large numbers of autoantibodies against ds-DNA and other nuclear antigens, which subsequently trigger immune complex-induced inflammatory responses in various organs. In recent years, interference with B-cell activation targeting B-cell stimulators (BAFF/BLyS) or B-cell monoclonal antibodies targeting CD20 has been used to treat SLE, and CAR-T (Chimeric Antigen Receptor T) cell immunotherapy has also enabled SLE patients to achieve rapid remission with high safety (3). B cell depletion has emerged as an effective treatment strategy for SLE. However, clinical observations reveal that a subset of patients does not respond favorably to B-cell depletion therapy, with 52-week response rates for belimumab and rituximab reported as 53.8% and 29.6% respectively (4, 5), and for telitacicept the 52-week response rate was 82.6% (6). Although cellular therapy has had some success in patients with SLE, the presence of extremely complex immune pathways in patients with SLE has necessitated the search for new ways to treat SLE.

T cell exhaustion (TEX), commonly referred to as CD8+T cell exhaustion, represents a state of diminished functionality within T lymphocytes (7). T lymphocytes have a variety of biological functions, ranging from the direct elimination of target cells, assisting or inhibiting B cells to produce antibodies, responding to specific antigens and mitogen, such as producing cytokines, etc. Central to the body’s defense against disease infections and tumor formation, T cells orchestrate cellular immunity, which operates through two primary mechanisms: specific binding and destruction of target cell membranes leading to direct cell killing, and the release of lymphocytokines that amplify and potentiate the immune response. TEX arises from prolonged antigenic stimulation, often triggered by the persistent presence of invasive pathogens in chronic infection states or the exhaustion of T cells resulting from chronic antigenic exposure in tumorous environments (8). Blanco et al. (9) demonstrated that CD8+T cells exhibited a more prominent cytotoxic phenotype and functions during SLE onset than during remission. Furthermore, the frequency of such cells was correlated with SLE Disease Activity Index (SLEDAI) score. In the study conducted by Winchester et al. (10). tubulointerstitial nephritis associated with CD8+T cell infiltration was identified as a risk factor for the progression of lupus nephritis (LN). This is consistent with the previous research results of Couzi et al. (11)., which indicated that CD8+T cell can also generate CD4-CD8-double negative (DN) T cells in an inflammatory environment, subsequently secrete IL-17, infiltrate the kidney and induce tissue damage (12). Collectively, these studies suggest a pivotal role for CD8+T cells in the pathogenesis of SLE. However, despite studies indicating a relationship between TEX and systemic autoimmune diseases, the exact therapeutic mechanism of TEX in SLE remains relatively scarce.

To screen biomarkers for SLE, we performed differential expression profiling, machine learning techniques, enrichment analysis, immune infiltration assessment, regulatory network construction, chemical compounds prediction, molecular docking simulation, and expression validation on datasets and genes from the public database. These findings offer promising potential targets for clinical diagnosis of SLE and prognostic predictions, while also providing a theoretical foundation for further elucidating the intricate mechanisms underlying SLE.

TEX diminishes the immune response mediated by T cells, leading to decreased secretion of effector cytokines, poorer proliferative capacity, reduced persistence, and enhanced expression of cell surface receptors (e.g., PD-1 and CTLA-4 on CD 8+T cell surface) (32). And in autoimmune diseases, patients with TEX show a better prognosis (33, 34). Therefore, there is an urgent need to better understand the detailed mechanisms of TEX to provide new strategies for the treatment of SLE.

Previous studies have shown that MX1 is highly expressed in the glomeruli and that this gene is associated with renal fibrosis (35). Peripheral blood MX1 gene expression has been used as a sensitive biomarker for LN therapy (36). The immune systems of mice and humans share some similarities, yet they also exhibit numerous differences in physiological structure and function. Despite these variations, observing the production of specific antibodies and local inflammatory responses in the kidney of mice may offer insights that are more readily discernible (37). In experiments with lupus mice, the expression level of Ly-6A/E on splenic lymphocytes was positively correlated with disease severity (38). The expression of Ly6E in proximal convoluted renal tubules is elevated due to proteinuria (39), suggesting a potential role for LY6E in nephropathophysiology. OASL expression in PBMCs and CD19+ B cells is upregulated in patients with active SLE with renal disease compared with patients without clinical manifestations (40). IFI44, as a LN specific biomarker, can distinguish between patients with active and inactive LN (41). In our study, four biomarkers MX1, LY6E, OASL and IFI44 were screened, and they are ISGs (42). These biomarkers, which are all upregulated in SLE patients, are closely linked to the activity of SLE disease. This is consistent with previous studies (43–45). A previous study comparing lymphocyte subtypes in healthy and SLE patients by two-color immunofluorescence flow cytometry analysis showed a reduction in circulating CD8+ T cells in SLE patients (491(146-1136) VS 331 (92-1401), p < 0.01) (46), but there are no relevant studies at the genetic level. This is also the innovation of this study.

Interferon-stimulated genes (ISGs) are produced by induction of interferon (IFN), and the activities of IFN-Is are potentiated by hundreds of ISGs. ISGs include antimicrobial proteins, chemokines, cytokines and other immune mediators that induce recruitment of immune cells and inflammation. ISGs play diverse roles in many cellular processes such as migration, antigen processing and presentation, cellular activation, differentiation, mitosis and apoptosis (47). Previous in vitro experiments have shown that anti-LY6E antibody treatment leads to CD3ζ chain tyrosine phosphorylation and blocks TCR-mediated T cell apoptosis (48). It indirectly reflects the role of LY6E in T cell apoptosis. In a cytological study by Perez et al., it was found that SLE patients experienced a decrease in Naive CD4+ T cells and an expansion of cytotoxic GZMH+ CD8+ T cells, as well as increased expression of ISG signals. It is speculated that IFN-1 causes monocytes to explicitly express ISGs and inhibits the basis of CD4+ T cell export to lymphoid tissue, leading to the observed decrease in circulating naive CD4+ T cells (49). This is consistent with the clinically observed lymphocytopenia in SLE patients and our findings. In studies of COVID-19, strong infiltration of CD8 T cells into lung tissue and manifestations of TEX have been shown in patients with low local expression of ISG (50). Previous scholarly endeavors have established that a substantial quantity of endogenous DNA, present in the cytoplasm beyond the confines of the nucleus and mitochondria, possesses the capacity to activate the cGAS-STING signaling pathway (51). Upon activation, STING meticulously orchestrates the recruitment of TANK-binding kinase 1 (TBK1), which, in turn, activates and phosphorylates interferon regulatory factor 3 (IRF-3), thereby stimulating the production of IFN-I. These IFN-1s subsequently engage with the IFN-I receptor, igniting the JAK/STAT signaling cascade and prompting the transcription of ISGs (52). Notably, the STING gene itself is classified as an ISG, paving the way for a potential self-sustaining positive feedback loop (53). Recent research has further demonstrated that elevated STING expression in T cells triggers an enhanced response associated with the induction of IRF-3-dependent and p53-dependent pro-apoptotic genes, ultimately culminating in the programmed cell death or apoptosis of T cells (54). During chronic infections and perhaps some cancers, IFN-α/β and IFN-γ can remain elevated and induce the expression of IL-10, IDO, PD-L1, TIM-3, LAG-3 and other negative regulators of T cell responses (32, 55, 56). CXCL9, CXCL10 are primarily secreted by monocytes, endothelial cells, fibroblasts, and cancer cells in response to IFN-γ, CXCL10 can also be strongly induced by IFN-α/β (57), and may also play a role in TEX (32). IFN-α/β can also foster attrition of activated T cells via Fas/FasL-mediated T cell death and perhaps other mechanisms (58, 59). There is also evidence that high IFN-α/β signaling can promote terminal exhaustion by antagonizing the TEX progenitor pool through effects on the transcription factor T cell factor-1 (Tcf-1) (60). Chen et al. showed that chronic IFN-I stimulation disrupted lipid metabolism and REDOX balance in TEX cells, resulting in abnormal lipid accumulation and elevated oxidative stress, and these defects promoted lipid peroxidation, thereby enhancing TEX cell metabolism and directly contributing to the terminal exhaustion of CD8+ T cells. At the same time, in vitro studies showed that IFN-I treated TEX cells exhibit greater gene enrichment associated with TEX, including MX1, OSAL (61). Studies have also found that long-term exposure to IFN-I increases NAD+ consumption of CD8+ T cells, triggering mitochondrial changes in CD8+T cells, resulting in impaired mitochondrial respiration and reduced cell viability, and promoting CD8+T cell death after TCR reactivation (62). At the same time, IFN can up-regulate the expression and secretion of galectin-9 (Gal-9), and promote TIM-3 mediated T cell death (63). At present, there are no relevant studies on how ISGs affects the specific molecular pathways of TEX in SLE patients. This study is the first to screen TEX-related genes from the perspective of TEX, providing a new approach and basis for the treatment of SLE.

GSEA analysis showed that the expression of these biomarkers was significantly positively correlated with immune-related pathways such as NLR signaling pathway, TLR signaling pathway, and RLR signaling pathway. Numerous studies have consistently demonstrated that IFN-I, particularly IFN-α, serves as the principal pathogenic mediator in SLE (64–66). The majority of patients with SLE exhibit overexpression of genes associated with the IFN-I pathway in peripheral blood cells (67–69). Studies have shown that TLR signaling exists in SLE (70, 71). TLRs, especially TLR7 and TLR9, are involved in the activation of plasmacytoid dendritic cells (pDC) to produce IFN-I (72). The increased TLR7 expression is associated with the risk of SLE and the severity of lupus in mice (73, 74). Some studies have shown that RIG-I is significantly expressed in kidney tissue samples from patients with LN, in addition (75). Other studies have demonstrated that the expression level of RIG-I mRNA in urinary sediment is significantly elevated in patients with LN compared to those with IgA nephropathy (76), hinting at a potential role of RIG-I in the underlying pathogenesis of LN. At present, there are few studies on SLE and NLR signaling pathway. All the 4 biomarkers studied in our study were ISGs, which may play a role in SLE by interfering with IFN signaling pathway.

We showed by immune infiltration analysis that naive B cells, resting memory CD4 T cells and resting NK cells exhibited significant downregulation. This is consistent with most of the available findings. Research has demonstrated that immune imbalance is an important factor in the emergence and progression of SLE (77). Studies have shown that the proportion of naive B cells is significantly lower in patients with SLE (78). Most recent studies indicate a significant reduction in the number of NK cells present in the peripheral blood of SLE patients (79–81). However, the underlying cause remains elusive. Cells differentiated from a subpopulation of CD4+ memory T cells may rapidly secrete effector cytokines to promote memory responses to reinfection (31). In the study conducted by Kosalka et al., a notable decrease in the absolute counts of naive CD4+ T cells and T central memory cells within the lymph nodes was observed, with the decrement being particularly pronounced in patients experiencing active disease (82). Currently, the fundamental mechanisms of information transmission and functionality of these cells in SLE remain largely unexplored. Research has demonstrated that SLE patients exhibit abnormal distribution of CD4+ memory T cell subsets, along with a diminished proliferative response to in vitro stimulation and a notable elevation in cellular apoptosis (83, 84). However, the results of various studies on memory CD4+ T cells are now inconsistent (85).

Immunoinfiltration analysis showed upregulation of regulatory T (Treg) cells, monocytes, activated dendritic cells and resting mast cells. Increased accumulation of DCs in affected tissues is associated with tissue inflammation and damage (86). In response to microenvironmental stimuli, monocytes can polarize into M1-like or M2-like macrophages. The imbalance between the M1 and M2 phenotypes correlates with SLE pathogenesis, and disease activity (87, 88). Although most reports have concluded that circulating Treg numbers are reduced or functionally impaired (89), some research groups have found elevated Treg levels in patients with SLE that correlate with disease activity (90). These findings are consistent with our immune infiltration results.

Drawing upon our literature review, we postulate that SLE can be triggered by the augmentation of the IFN pathway following the IFN-induced expression of ISGs. Additionally, prolonged exposure to IFN can lead to the expression of STING, IL-10, IDO, PD-L1, TIM-3, LAG-3 and other negative regulatory factors, which induced the apoptosis of T cells. Meanwhile, IFN also triggered alterations in the mitochondrial metabolic pathway of CD8+ T cells, leading to the death and exhaustion of chronically activated CD8 T cells, thereby contributing to the pathogenesis of SLE.

In order to fully understand the potential mechanism of action between the 4 biomarkers and TEX, the TF-mRNA network and ceRNA network were constructed in this study, and the results showed that the 4 biomarkers were regulated by different TFs. FOXC1 and USF2 are co-regulated TFs. At present, the main research on the 8 micrornas suggested in the study focuses on the secretion of cellular inflammatory factors, cell metabolism, apoptosis, proliferation, invasion and migration, especially in tumor (91). However, there are few studies on SLE. LncRNAs are critical in regulating lymphocyte activity, T-helper cell polarization, and adaptive immune cells. BPA, BaP, Tetrachlorodibenzodioxin (TCDD) are not used as drugs, but they are also present in our lives. BPA is an industrial chemical widely used in the polymerization of plastics and in non-food-related materials, such as thermal paper receipts, eyeglasses and children’s toys (92). >90% of human BPA exposure occurs through the oral route, with absorption occurring in the gastrointestinal tract (93). BPA closely resembles the physiological impacts of estrogen (94). BPA induces structural alterations in DNA in vitro and is strongly recognized by SLE IgGs and induces high titer antibodies in rabbits (95). Lee et al. demonstrated that the exposure of mouse embryonic stem cells to BPA and progesterone led to a reduction in LY6E mRNA levels, indicating the suppressive effect of these compounds on LY6E expression (96). BPA may act on SLE by interfering with different signaling pathways. BaP is found in smoky environments, processed meat products such as smoked and barbecued. It has been extensively studied and is thought to be associated with an increased risk of many types of cancer and is listed as toxic or dangerous by several countries (97). TCDD is generally considered to be the most toxic synthetic molecule known. TCDD used to be a contaminant of some herbicides and is a harmful contaminant produced by industrial processes (98). Based on the findings of this study, we recommend that patients adopt appropriate preventive measures. These measures include selecting environmentally friendly products free from BPA, avoiding the use of dioxin-containing pesticides and other harmful compounds, and minimizing exposure to environmental pollutants such as industrial emissions, traffic exhaust, waste incineration, barbecuing, and tobacco smoke. Utilizing molecular docking, we predicted chemical compounds that specifically target biomarkers, and in light of the current limited research, we postulate that these chemical compounds have the potential to bind to and disrupt mRNA expression via the interaction with these biomarkers.

In conclusion, based on bioinformatics analysis, we identified 4 biomarkers that were highly correlated with T cell depletion in SLE patients, including LY6E, OASL, MX1, and IFI44. Our findings provide potential therapeutic targets and shed light on the pathogenesis of SLE. However, there are some limitations to our study. First, the patients in the validation cohort were made up of Chinese. Therefore, our findings may not generalize across racial and ethnic groups of SLE patients. Second, there are limited samples available for RT-qPCR validation and insufficient validation has been performed. Therefore, more mRNA expression that would validate these biomarkers is needed. In the forthcoming research undertakings, our principal emphasis will be placed on conducting comprehensive studies encompassing SLE patients from various racial and ethnic backgrounds. These studies will strive to meticulously assess the universal applicability of biomarkers MX1, LY6E, IFI44, and OASL across various global populations. Third, addressing the challenge of insufficient validation resulting from limited sample sizes, we intend to collect samples from SLE patients at various stages of the disease. This will enable a more comprehensive validation of the role these biomarkers play in the initiation and progression of SLE. Simultaneously, we will integrate other methodologies, including gene silencing and overexpression assays, flow cytometry, and immunoimprinting, to explore the molecular regulatory mechanisms underlying the interplay between these biomarkers and T cell exhaustion. Furthermore, we aspire to embark on multi-center, interdisciplinary collaborative research initiatives, ultimately enhancing the quality of life for SLE patients and fostering advancements in this research domain.