Treatment-naïve RA patients were obtained from the first affiliated hospital of Harbin Medical University, Department of Rheumatology. Recruited RA patients adhered to the Helsinki Declaration. All the three early treatment-naïve RA patients were seropositive for rheumatoid factor (RF) and or/anti-citrullinated peptide antibodies (ACPA), fulfilled 2010 ACR/EULAR criteria for RA and had active disease with a DAS28 > 3.2. The three early treatment-naïve RA patients were collected for RNA-seq and six treatment-naïve RA patients were collected for validation by qRT-PCR (Table 1). All the RA patients fulfilled the ACR criteria, according to clinical and radiological imaging (Aletaha et al., 2010). All participants gave written informed consent according to the Declaration of Helsinki. Ethics approval for the study was obtained from Harbin Medical University Research and Ethics Committee, Henan Provincial People’s Hospital Research and Ethics Committee, Zhengzhou University Research and Ethics Committee.

Naïve CD4⁺ T cells were isolated from early treatment-naïve RA patients and cultured in the presence or absence of As₂O₃ with anti-CD3/CD28 activation and IL-2 exists. The detailed procedures as described in our previous study (Li et al., 2019).

For intracellular cytokine detection, cells were stimulated with the corresponding Cell Activation Cocktail (with Brefeldin A) (Biolegend, San Diego, CA) for 6 h. The detailed procedures described in our previous study (Li et al., 2019).

The cytokines IL-17A and IL-10 (D1700 and D1000, respectively, purchased from R&D systems, Minneapolis, United States) and MMP13 (E-EL-H0134c, Elabscience Biotechnology, Wuhan, China) were determined using enzyme-linked immunosorbent assay (ELISA). The ELISA was performed according to the manufacturer’s instructions.

Cultured cells were fixed in 4% paraformaldehyde, followed by penetrating and blocking serum for 1 h, Rabbit anti-STAT3 (ab68153, Abcam, Cambridge, MA, United States) and Rabbit anti-Foxp3 (BA 2032–1, Boster, Wuhan, China) were used as the primary antibody. Samples were washed three times and incubated with FITC secondary antibodies (PV6001, ZSGB-BIO, Beijing, China), and DAPI (Sigma) was used for staining nuclei. Images were captured using a microscope (Leica, Mannheim, Germany).

Treg cells were purified from PBMCs by high-speed cell sorter system (Moflo XDP, Beckman coulter, United States). The detailed procedures are described in our previous literature (Li et al., 2019). The working schematic model is shown in Figure 1A.

RNA degradation and contamination were monitored on 1% agarose gels. RNA purity was checked using the NanoPhotometer spectrophotometer (IMPLEN, CA, United States). RNA concentration was measured using Qubit RNA Assay Kit in Qubit 2.0 Flurometer (Life Technologies, CA, United States). RNA concentration and integrity were assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, United States).

To assess the function of identified DEGs, the functional analysis and clustering tool from Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.7 was used to perform Gene Ontology (GO) function enriched analysis based on DEGs obtained from RA Treg cells. GO enrichment analysis of DEGs was implemented by the GOseq R package, in which gene length bias was corrected. DAVID provides a comprehensive set of functional annotation tools for investigators to understand the biological meaning based on a large list of genes (http://david.ncifcrf.gov/). For any given gene list, DAVID tools can identify enriched biological themes (GO terms), discover enriched functional-related genes cluster, and visualize genes on BioCarta and KEGG pathway maps (Huang et al., 2009). Therefore, DAVID was carried out to identify the enriched GO functions including the biological processes (BPs), molecular functions (MFs), and cell components (CCs). KEGG Orthology Based Annotation System (KOBAS 2.0) (http://kobas.cbi.edu.cn) was employed to identify biological pathways from the identified DEGs involved in the diseases. It performs statistical tests to identify statistically significantly enriched pathways and diseases using biological knowledge from five well-known pathway databases and GO (Xie et al., 2011). Thus, the KOBAS 2.0 was used to identify the enriched KEGG pathway based on adjusted p values. KEGG pathways including five or more DEGs genes were considered as the biologically meaningful analysis.

Total RNA was extracted from Treg cells according to the instructions of RNA extraction kit (Trizol Reagent, Invitrogen, Carlsbad, CA, United States). cDNA obtained from the reverse transcriptase reaction and subjected to quantitative RT-qPCR using SYBR Green PCR Master Mix (Bio-Rad, California, United States) and using the ABI Prism 7500 Sequence detection system (Applied Biosystems). Primers for human glyceraldehyde-3-phosphate dehydrogenase (GAPDH), cyclin-dependent kinase inhibitor 3(CDKN3), sushi domain containing 4 (SUSD4), histone cluster 4 H4 (HIST4H4), ubiquitin-specific peptidase 7 (USP7), histidine ammonia lyase (HAL), protein tyrosine phosphatase, non-receptor type 13 (PTPN13), DNA fragmentation factor subunit beta (DFFB), receptor interacting serine/threonine kinase 1 (RIPK1), and methyltransferase like 3 (METTL3) were purchased from Takara. The primer and concentrations were optimized according to the manufacturer’s instructions in SYBR Green PCR Master Mix Protocol. Relative expression levels of the nine selected genes were calculated by using the 2^−ΔΔCt method. The detailed following methods were performed according to our previous literature (Zhang et al., 2017).

Statistical significance was determined using GraphPad Prism Software (Version 6 for Windows; GraphPad Prism, San Diego, CA, United States). Simple comparisons were made using unpaired, two-tailed Student’s t-test for parametric data or Mann–Whitney test for nonparametric data, as indicated. Multigroup comparisons of the means were carried out by one-way analysis of variance test with post hoc contrasts by Tukey test. p values of 0.05 or less was considered statistically significant. All data are presented as mean ± S.E.M. Transcripts with significantly differential expressions of Treg cells by As₂O₃ were identified using hypergeometric test. The resulting p-values were adjusted using the Benjamini and Hochberg’s method for controlling the false discovery rate (FDR). DEGs were identified by applying the Benjamini and Hochberg method with adjusted p values of <0.05. DAVID v6.7 was used to carry out GO function enriched analysis based on the DEGs. KOBAS 2.0 was used to identify the enriched KEGG pathway based on the adjusted p values using Benjamini and Hochberg method.

The immunomodulatory activity of As₂O₃ on PBMCs cells responses was also determined by the reduction of the growth factor IL-2 compared with IL-2-producing anti-CD3/CD28 activated PBMCs (Figure 2A). This effect was not due to an induction of the cell death, as assessed by annexin V staining that was used as a marker for apoptosis in combination with propidium iodide (PI), to distinguish between apoptotic and necrotic cells. After treatment 24 h, As₂O₃ treatment of PBMCs did not induce cell apoptosis, ruling out the potential cytotoxic role of As₂O₃ (Figures 2B,C). The decrease of IL-2 did not result in a significant decrease in PBMCs proliferation.

Th17 and Treg subsets are both derived from naïve CD4⁺ T cells in peripheral blood upon antigen stimulation and specific polarizing cytokines. Because As₂O₃ dampened the inflammatory response of IL-17–producing cell from PBMCs, we next investigated whether As₂O₃ could directly affect their differentiation from naïve CD4⁺ T cells into Th17 cells lineages. To obtain this aim, a standard naïve CD4⁺ T cells differentiation assay was performed by polyclonal stimulation with anti-CD3/CD28 and specific polarizing cytokines in the presence of As₂O₃ (Figure 3A).

Under specific polarizing conditions, highly purified naïve CD4⁺ T cells displayed significantly higher amounts of intracellularly produced and extracellularly released IL-17 and MMP13, as compared to nonpolarized (Th0) cells (Figure 3C), in particular, in non-skewed Th0 cells, which produce low level IL-17 and MMP13. However, As₂O₃ significantly reduced Th17 generation, acting both on intracellular production and extracellular release from Th17 cells, suggesting that As₂O₃ affects not only Th17 cells induction but also specific functional properties. To address whether Th17 polarization was associated with the acquisition of their typical features, we also measured the mRNA encoding for the transcription factor RORγt known to be critical for their differentiation. As we expected, Th17 condition induced the highest expression of their specific transcription factors. The presence of As₂O₃ during Th17 polarization led to decreased RORγt in Th17 cells (Figure 3E). These findings support a pivotal role of As₂O₃ in hindering de novo Th17 differentiation.

In light of the role of As₂O₃ in resolving inflammation and because Treg cells is an important cell subset involved in modulating and maintaining self-regulation of the immune system, we also investigated whether As₂O₃ could affect the generation of induced Treg cells. This cell subset develops from naïve CD4⁺ T cells upon antigen stimulation and transforming growth factor-β (TGF-β) exposure. To obtain this aim, highly purified naïve CD4⁺ T cells were cultured under Treg-inducing conditions in the presence of As₂O₃ (Figure 3B). We found that As₂O₃ potentiated Treg differentiation, with As₂O₃ enhancing STAT5 expression compared to control Treg cells (Figure 3F). As₂O₃-induced de novo generation of Treg cells was also paralleled by their capacity to increase their suppressive cytokines IL-10 and TGF-β1 (Figure 3D), suggesting that As₂O₃ affect not only Treg induction but also specific functional properties.

To investigate the effect of As₂O₃ on CD4⁺CD25⁺ Treg cells proliferation, CD4⁺CD25⁻ T and CD4⁺CD25⁺ Treg cells were isolated from PBMCs in treatment-naïve RA patients (Figure 4A). CD4⁺CD25⁺ Treg cells generally represent anti-inflammatory subtype. To examine whether As₂O₃ directly affect CD4⁺CD25⁺ Treg-cell proliferation, CD4⁺CD25⁺ Treg cells were cultured with physical condition anti-CD3/CD28 stimulation, as well as As₂O₃ was added in the well for 3 days. Interestingly, we noticed that As₂O₃ dramatically increased the proportion of CD4⁺CD25⁺ Treg cells compared with without As₂O₃ treatment of CD4⁺CD25⁺ Treg cells (Figure 4B). To address whether As₂O₃ affects CD4⁺CD25⁺ Treg-cell polarization was associated with the acquisition of their typical features, we also measured the mRNA encoding for the transcription factor known to be critical for their differentiation Foxp3. As expected, As₂O₃ induced higher expression of Foxp3 compared with CD4⁺CD25⁺ Treg cells alone (Figure 4C).

To examine the effect of As₂O₃ in the phenotypic characteristics of Treg cells, we used flow cytometry to assess the expression of Foxp3 and surface markers known to be associated with either Th17 or Treg cells. CCR6 is human Th17 cell marker. The expression of CCR6 was slightly but significantly lower in CD4⁺CD25⁺ Treg cells than in CD4⁺CD25⁻ Treg cells (mean fluorescence intensity (MFI): 2.18 ± 0.24 vs 3.84 ± 0.27, p < 0.05), whereas the expression levels of the Treg-associated regulatory molecule CD127 and Foxp3 were significantly higher in CD4⁺CD25⁺ Treg cells than in CD4⁺CD25⁻ Treg cells (MFI: 58.4 ± 2.65 vs 20.2 ± 1.7 and 22.2 ± 1.02 17.6 ± 2.5, respectively). Interestingly, we noticed that As₂O₃ downregulated CCR6 expression, while upregulated CD127 and Foxp3 expression in CD4⁺CD25⁺ Treg cells (Figures 5A,B). These results show that As₂O₃ enhances Treg-cell immunosuppression by regulating phenotype of Treg and Th17 cells.

We studied the effects of As₂O₃ on STAT3 in Th17 and Foxp3 in Treg cells by immunofluorescent staining. We determined the effects of As₂O₃ on STAT3 and Foxp3 function by observing nuclear translocation. As shown, STAT3 in nucleus was reduced by As₂O₃ treatment (Figure 6A), while the Foxp3 was increased (Figure 6B).

The pathway enrichment analysis identified 22 differentially expressed pathways. Considering that the biological functions identified and previous study results for Treg cells, eight potential critical pathways participate in regulation of Treg-cell function by As₂O₃ from early treatment-naïve RA patients (Table 2). One upregulated enriched pathway found to be involved in FoxO acetylation which affects Foxp3 transcription and Treg-cell development. Seven downregulated enriched pathways were also determined, including apoptosis, cytokine–cytokine receptor interaction, cell cycle, PI3K-Akt signaling pathway, nuclear factor kappa-B signaling pathway, calcium signaling pathway, and p53 signaling pathway (Table 2).

After normalization and correction for multiple testing, 100 genes were upregulated more than 2.0-fold in As₂O₃-treated cells compared to control cells, and 136 downregulated by the same factor (FC<-2.0). In total, 472 genes were found to be differentially expressed in As₂O₃-treated vs. control Treg cells in a statistically manner (FDR-corrected p value <0.05). We have identified 472 DEGs including 200 upregulated genes and 272 downregulated genes in treatment Treg cells compared to nontreatment Treg cells with As₂O₃ (Figure 1B). Twelve most strongly up- and downregulated genes are listed in Table 3.

The list of the most strongly upregulated genes include genes involved in the regulation of cell proliferation, immunoregulation, apoptosis, and amino acids and glycolysis metabolism. Among the most strongly downregulated genes are those linked to cell proliferation and differentiation, extracellular signal activation, and inflammation.

We selected nine significantly overrepresented BPs, including regulation of mitotic cell cycle, regulation of DNA recombination, and protein ubiquitination. There were nine significantly overrepresented MFs, including peptide transporter activity, RNA binding, glycogen synthase activity, kinase activator activity, and phosphatase activity. We also detected five significantly overrepresented CCs, including mitochondrial chromosome, cyclin E1–CDK2 complex, and DNA helicase complex (Figure 7). In summary, the significantly overrepresented BPs, MFs, and CCs were dramatically different between nontreatment Treg and treatment Treg cell with As₂O₃.

As chronic inflammatory autoimmune disease and changes in Treg cell frequency and dysfunction are central features in the pathogenesis, we set out to separately study genes linked to these processes (Table 4). Several proinflammatory factors such as hypoxia inducible factor 1 alpha subunit (HIF1α, fold change −8.4), matrix metallopeptidase 3 (MMP3, fold change −3.2), C-C motif chemokine receptor 6 (CCR6, fold change −7.6), and SMAD family member 3 (SMAD3, fold change −4.8) were significantly downregulated by As₂O₃, while the anti-inflammatory peroxisome proliferator activated receptor gamma (PPARγ, fold change 6.4) and interleukin 10 (IL-10, fold change 2.6) were upregulated. Genes affecting immune response, such as forkhead box O1 (Foxo1, fold change 11.8) and cytotoxic T-lymphocyte associated protein 4 (CTLA-4, fold change 4.6), were similarly upregulated.

RA is also known to be associated with metabolic syndrome. We separately studied genes for proteins participating in the main pathways of amino acid and carbohydrate metabolism (amino acid biosynthesis, glycolysis) (Weyand and Goronzy, 2017; Falconer et al., 2018). As₂O₃ did not have a marked (fold change>2.0) effect on any of these genes, with the single exception upregulation of acyl-CoA thioesterase 4 (ACOT4) (fold change 2.9).

Five upregulated genes (CDKN3, SUSD4, USP7, HAL, and HIST4H4) and four downregulated gene (DFFB, PTPN13, RIPK1, and METTL3) in early treatment-naïve RA patients were selected for RT-qPCR. The nine selected genes expression were consistent between RNA-seq and RT-qPCR analysis, confirming the accuracy of our data (Figure 8).