Proinflammatory Features of Stem Cell–like Memory T Cells from Human Patients with Rheumatoid Arthritis

Stem cell–like memory T (Tscm) cells, the least differentiated type of memory T cells, have the same characteristics as stem cells and are capable of self-renewal and differentiation into subsets of effector T cells (1, 2). Tscm cells were first identified in mice and have also been found in nonhuman primates and humans (3–5). They are designated as CD3⁺ and CD4⁺ cells or as CD8⁺, CD45RO⁻, CCR7⁺, CD45RA⁺, CD62L⁺, CD27⁺, CD28⁺, CD127 (IL-7Rα)⁺, CD122⁺, and CD95^hi cells in humans, according to Gattinoni et al. (2). They play important roles as reservoirs of T cells in various diseases. For example, they provide long-term immunity to patients infected with yellow fever virus and help chimeric Ag receptor-T (CAR-T) cells self-renew in vivo (6, 7). Tscm cells also act as reservoirs of pathogenicity. The number of CD8⁺ Tscm cells is elevated in patients with immune thrombocytopenia (8) and in subjects with acquired aplastic anemia (9). We recently reported the pathogenic role of Tscm cells in systemic lupus erythematosus (SLE) (10).

Rheumatoid arthritis (RA) is a chronic systemic inflammatory polyarticular disease characterized by synovial hyperplasia and prominent infiltration by inflammatory cells (11). Although the pathophysiology of RA remains unclear (12), CD4⁺ T cells are thought to play a critical role. In addition, inflammatory cytokines, including TNF, IL-6, and IL-1, are important, as confirmed by the success of biological agents that target them (13–16). The recent success of JAK inhibitors as treatment for RA further supports the role of inflammatory cytokines in this disease (17, 18). However, RA remains almost impossible to cure, even with the best combinations of novel drugs (19), which indicates that an undetected eradicable pathway may be involved in its pathogenicity.

Previously, we reported a pathogenic role of Tscm cells in SLE, such that Tscm cells generate pathogenic follicular helper T (Tfh) cells (10). In the current study, to explore the pathogenetic role of Tscm cells in RA as reservoirs and eternal producers of pathogenic T cells, we compared the proinflammatory characteristics of Tscm cells from RA patients with those from healthy controls (HCs).

The present study included 109 RA patients and 85 HCs. We used previously published data describing the percentages of CD4⁺ and CD8⁺ Tscm cells within the CD4⁺ and CD8⁺ T cell populations of 53 HCs (10). For all other experiments, we gathered data from newly recruited RA patients and HCs. PBMCs and synovial fluid mononuclear cells (SFMCs) were isolated on a Ficoll-Hypaque (GE Healthcare, Princeton, NJ) gradient. All RA patients met the 2010 classification criteria of the American College of Rheumatology/European League Against Rheumatism (20). Laboratory investigations included erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), rheumatoid factor (RF), and anti-citrullinated protein Ab (ACPA). We evaluated disease activity by measuring the ESR and CRP level in each patient. The disease activity score in 28 joints (DAS28)–ESR and –CRP was available for RA patients. Remission status was defined according to American College of Rheumatology/European League Against Rheumatism Boolean criteria (21). Clinical characteristics and treatments for the enrolled patients are listed in Table I.

This study was performed in accordance with the guidelines of the Declaration of Helsinki, and written informed consent was obtained from all subjects. The protocol was approved by the institutional review board of Seoul National University Hospital.

The data sets analyzed in the current study are available from the corresponding author upon reasonable request.

For cell surface phenotyping, Fc receptors on PBMCs (1 × 10⁷ cells/ml) were blocked with purified mouse anti-human IgG (BD Biosciences, San Jose, CA), and cells were stained with adequate fluorescent Abs. The detailed flow cytometry gating strategy is described in (Fig. 1A. To measure the absolute number of CD4⁺ or CD8⁺ Tscm cells, we used Accucheck Counting Beads (Thermo Fisher Scientific, Waltham, MA). For Tscm cells, we stained for CD95, which is a distinctive Tscm cell marker, and for CXCR3 and LFA-1 to test the validity of the gating strategy against CD95; Tscm cells defined as CD95⁺ T cells also expressed CXCR3 or LFA-1 (Supplemental Fig. 1A).

The following Abs were obtained from BD Biosciences: CD3 (clone SK7), CD4 (SK3), CD8 (SK1), CD45RO (UCHL1), CD62L (DREG-56), CD45RA (HI100), CD27 (M-T271), CXCR5 (RF8B2), ICOS (DX29), CXCR3 (1C6/CXCR3), and PD-1 (EH12.1).

Anti-CD28 (CD28.2) Ab, IL-4 (MP4-25D2), IL-17A (eBio64DEC17), and IL-2 (MQ1-17H12) were obtained from eBioscience (San Diego, CA), and Abs specific for CD127 (A019D5), CD122 (Tu27), CD95 (DX2), Ki-67 (Ki-67), LFA-1 (m24), IFN-γ (B27), TGF-β1 (TW4-2F8), and CCR7 (G043H7) were obtained from BioLegend (San Diego, CA).

Stained cells were analyzed with an LSR Fortessa flow cytometer (BD Biosciences) and sorted with a FACSAria instrument (BD Biosciences). All data were analyzed with FlowJo (FlowJo, Ashland, OR).

Sorted Tscm cells were stimulated for 6 d with anti-CD3/CD28–coated beads (Dynabeads; Thermo Fisher Scientific) at a 1:1 ratio. Their capacity to differentiate into other daughter T cell subsets (central memory T [Tcm], effector memory T [Tem], terminal Tem [Temra], and Tfh cells) were examined with staining for CD3, CD4, CD45RO, CCR7, CD45RA, CXCR5, ICOS, and PD-1.

Transcription factors associated with Th subsets were stained. Fc receptors were blocked, and cells were stained for CD3, CD4, CCR7, and CD45RA. After surface staining, cells were permeabilized with Fixation/Permeabilization solution (eBioscience) and stained for transcription factors T-bet (4B10), GATA3 (16E10A23; BioLegend), RORγt (Q21-559; BD Biosciences), and FOXP3 (PCH101; eBioscience).

FACS-sorted T cells (naive T, Tscm, Tcm, or Tem cells) were cultured with 10% RPMI media, rIL-2 (PeproTech, Cranbury, NJ), and anti-CD3 Abs (Thermo Fisher Scientific). Th-polarizing differentiation was then performed. To induce Th0 differentiation, anti–IL-4 Abs (BioLegend) and anti–IFN-γ Abs (BioLegend) were added. To induce Th1 differentiation, recombinant protein IL-12 (BioLegend) and anti–IL-4 Abs were added. To induce Th2 differentiation, recombinant protein IL-4 (PeproTech) and anti–IFN-γ Abs were added. For Th17 differentiation induction, recombinant protein IL-1β, IL-6 (both from PeproTech), anti–IFN-γ Abs, and anti–IL-4 Abs were added. To induce regulatory T (Treg) cells, rTGF-β1 (PeproTech), anti–IFN-γ Abs, and anti–IL-4 Abs were added. Cells were cultured for 5 d without changing the media. After 5 d, intracellular transcription factors and cytokines were measured.

Plasma was collected from RA patients and HCs and stored at −80°C. Plasma IL-6 levels were measured with the Quantikine HS Human IL-6 ELISA kit (R&D Systems, Minneapolis, MN). All procedures were performed according to the manufacturer’s instructions. Standards were measured in triplicate, and samples were measured in duplicate.

Sorted Tscm cells (4 × 10⁴ cells) were stimulated overnight with 50 ng/ml recombinant human IL-6 protein (PeproTech) and also with anti-CD3/CD28–coated beads (Thermo Fisher Scientific) if applicable. On the next day, culture supernatants were stored for the cytokine assay, and cells were stained for activation markers, including CD69, CD25 (IL-2Rα), and CD154 (CD40L).

The cytokine assay was performed with a cytometric beads assay. IFN-γ, IL-4, IL-17A, and TNF cytokine levels were measured according to the manufacturer’s protocol (LEGENDplex; BioLegend) using an LSR II flow cytometer (BD Biosciences).

Frozen PBMCs were surface-stained with Tscm markers (CD4, CCR7, CD45RA, CD27, CD95) and cross-linked with bis[sulfosuccinimidyl]suberate (BS³; Thermo Fisher Scientific). Mouse spleen cells were added to the solution, and telomeres were stained with a peptide nucleic acid probe and incubated for 10 min at 80°C, 700 rpm in a thermomixer (Eppendorf, Hamburg, Germany). After incubation, the cells were incubated for 2 h at room temperature, washed, and analyzed by flow cytometry.

RNA was extracted from FACS-sorted Tscm cells with an RNeasy Mini Kit or RNeasy Plus Micro Kit (QIAGEN, Hilden, Germany). RNA-sequencing libraries were produced with NEXTflex Rapid Directional mRNA-Seq Kit Bundles (Bioo Scientific, Austin, TX) and sequenced on an Illumina HiSeq 2500 platform (Illumina, San Diego, CA). The sequence reads were analyzed by alignment to the Human Ensembl Archive Release 90 using STAR (22) with ENCODE options and quantification using RSEM (23).

Data are expressed as means ± SEMs. For continuous variables, a Student t test (unpaired, two-tailed) was used when the sample size was >30, and normality of the data was confirmed with Shapiro–Wilk test; the Mann–Whitney U test was used when the sample size was <30. All graphs were drawn with Prism (GraphPad Software, La Jolla, CA). For RNA transcriptome analyses, we identified differentially expressed genes (DEGs) using DESeq2 (24) with Wald test and the likelihood-ratio test with the Benjamini–Hochberg correction (false discovery rate [FDR] < 0.05) and one-way ANOVA (p < 0.01) corrected with FDR < 0.05. Enrichment analyses of Gene Ontology (GO) terms and pathways were performed with Metascape (25), and volcano plots were drawn with the R package EnhancedVolcano (version 1.01). All statistical analyses were performed with SPSS (IBM, Armonk, NY) unless otherwise stated.

We compared the percentage of Tscm cells in PBMCs isolated from 50 RA patients and 53 HCs [data from the HCs were obtained in a previous study (10)]. The clinical characteristics of the analyzed patients are summarized in Table I. The flow cytometry gating strategy used to measure the percentage of Tscm cells is presented in (Fig. 1A. The percentage of CD4⁺ or CD8⁺ Tscm cells in the total CD4⁺ or CD8⁺ T cell population was significantly higher in RA patients than in HCs (CD4⁺ T cells: 0.9 ± 0.1% [RA] versus 0.4 ± 0.0% [HC], p < 0.001; CD8⁺ T cells: 1.3 ± 0.2% [RA] versus 0.8 ± 0.1% [HC], p < 0.01; (Fig. 1B). However, the absolute Tscm cell count was similar in RA patients and HCs (Supplemental Fig. 1B).

The synovium is the target tissue in RA, and T lymphocytes play a pivotal role in the pathogenesis of the disease. To determine whether Tscm cells are present in synovium, we measured the proportion of Tscm cells in nine SFMC samples. Tscm cells were present in both the CD4⁺ and CD8⁺ T cell compartments of SFMCs (CD4⁺ T cells: 0.4 ± 0.3%; CD8⁺ T cells: 0.9 ± 0.6%; (Fig. 1C).

Stem cells can self-renew and differentiate into daughter cell subsets (2). To determine whether Tscm cells in RA patients also have these abilities, we stimulated FACS-sorted Tscm cells (purity =90.5%) with anti-CD3/CD28–coated beads for 6 d. After stimulation, Tscm cells from RA patients differentiated into Tcm (CCR7⁺ CD45RA⁻), Tem (CCR7⁻ CD45RA⁻), and Temra (CCR7⁻ CD45RA⁺) cells (Fig. 2A). The differentiation pattern was similar in RA patients and in HCs (Fig. 2A, graph). Although the frequency of self-renewed Tscm cells was not higher than that of self-renewed Tcm or Tem cells, Tscm cells produced naive-like T (CD45RA⁺) cells more frequently than Tcm or Tem cells throughout the division triggered by TCR activation, suggesting they can act as stem cells (Supplemental Fig. 2). The differentiated naive-like T (CCR7⁺ CD45RA⁺) cell population derived from the Tscm cells consisted mostly of Tscm cells (Fig. 2B); this is in accordance with a previous study in which Tscm cells self-renewed and differentiated into Tcm and Tem cells (2). An examination of transcription factors T-bet (for Th1 cells), GATA3 (for Th2 cells), RORγt (for Th17 cells), and FOXP3 (Treg cells) expressed by Tem cells differentiated from CD4⁺ Tscm cells under anti-CD3 stimulation revealed that expression of these transcription factors was similar in the two groups (Fig. 2C). To determine the differentiation pattern of Tscm cells more specifically, Tscm cells were stimulated under different Th-polarizing conditions. Under these conditions, Tscm cells differentiated into Th1, Th2, Th17, and Treg cells (Supplemental Fig. 3). Overall, the differentiation pattern of RA Tscm cells did not differ from that of HC Tscm cells; however, RA Tscm cells differentiated less frequently toward FOXP3⁺ Treg cells than HC Tscm cells under the Treg-differentiation condition (26).

To better understand whether Tscm cells from RA patients differentiate into Tfh cells, we stained stimulated Tscm cells for Tfh-associated markers. After stimulation with anti-CD3/CD28, Tscm cells from RA patients differentiated into Tfh cells (designated CD3⁺ CD4⁺ CXCR5⁺ PD-1⁺ ICOS⁺; Supplemental Fig. 4A). However, the percentage of differentiated Tfh cells in the CD4⁺ T cell population was not higher in RA patients than in HCs. This was confirmed by results showing that the percentage of Bcl-6⁺ cells in CD4⁺ T cells was similar in RA patients and HCs (Supplemental Fig. 4B).

In addition to having the capacity for differentiation, Tscm cells self-renew in response to anti-CD3/CD28 stimulation. Tscm cells (CD4⁺ CFSE^low CCR7⁺ CD45RO⁻ CD62L⁺ CD95⁺) were consistently detected in the proliferated Tscm cell population (denoted by low CFSE level) in both RA patients and HCs (Supplemental Fig. 2). The ratio of Tscm cells increased gradually as the number of divisions increased (Fig. 2D), which indicates that Tscm cells have a greater proliferative capability than other T cells. Additionally, we analyzed telomere length and the expression of Ki-67. The telomere lengths of Tscm cells were maintained to a similar extent to those of naive T cells, but the telomere lengths of Tcm and Tem cells were significantly shorter (Fig. 2E). Ki-67 was highly expressed in Tscm cells but barely expressed in naive T cells (Fig. 2F). These data show that the Tscm cell is a distinct subset of T cells from naive T, Tcm, or Tem cells with respect to its replicative history and proliferative capacities.

We compared the number of Tscm cells in RA patients with active disease and those in remission. The percentage of Tscm cells was significantly lower in RA patients in remission (n = 6) than in RA patients with active disease (CD4⁺ T cells: 0.26 ± 0.05% [RA remission] versus 0.9 ± 0.1% [active RA], p = 0.034; CD8⁺ T cells: 0.04 ± 0.1% [RA remission] versus 1.3 ± 0.2% [active RA], p = 0.023; (Fig. 3A). When the patients were divided into those with more than 2% Tscm cells and less than 2% Tscm cells, there was no significant difference between the clinical features (Table II). Moreover, when we analyzed the association between Tscm cells and expression of disease activity markers in RA patients, we found that the percentage of CD4⁺ Tscm cells in the CD4⁺ T population correlated with DAS28-ESR and DAS28-CRP scores (Fig. 3B). However, the percentage of CD8⁺ Tscm in the CD8⁺ T cell population was not associated with disease activity markers (Fig. 3C).

IL-6, an important cytokine in the pathogenesis of RA, activates various leukocytes and osteoclasts and mediates B cell differentiation to generate autoantibodies (11). In the current study, plasma IL-6 levels were higher in RA patients than in HCs (Fig. 4A). To evaluate the effects of IL-6 on Tscm cells, we compared the activation status of Tscm cells after stimulation with IL-6 and/or anti-CD3/CD28 beads. The expression of surface activation markers CD69, CD25 (IL-2Rα), and CD154 (CD40L) increased after stimulation with IL-6 plus anti-CD3/CD28 (Fig. 4B). Tscm cells were more activated in RA patients than in HCs (Fig. 4C). Thus, IL-6 (the key inflammatory cytokine in the pathogenesis of RA) acts cumulatively with the TCR to activate Tscm cells in RA patients.

Next, we measured the concentration of cytokines in culture supernatants of activated Tscm cells using a cytometric beads assay. The following cytokines were measured: IFN-γ (for Th1 cells), IL-4 (for Th2 cells), and IL-17A (for Th17 cells). Tscm cells from RA patients produced more cytokines when stimulated with IL-6 and anti-CD3/CD28 beads than when stimulated only with anti-CD3/CD28 beads (TNF, p = 0.013; IL-4, p = 0.002; IFN-γ, p = 0.002; IL-17A, p = 0.019; (Fig. 4D). Tscm cells from RA patients secreted more TNF than cells from HCs when stimulated with anti-CD3/CD28 beads (p = 0.014). In addition, Tscm cells from RA patients released more IL-17A and IL-4 than Tscm cells from HCs when stimulated with anti-CD3/CD28 beads, regardless of the presence of IL-6 (IL-17A: p = 0.029 when stimulated with anti-CD3/CD28, p = 0.049 when stimulated with both anti-CD3/CD28 beads and IL-6; IL-4: p = 0.037 when stimulated with anti-CD3/CD28 beads, p = 0.027 when stimulated with both anti-CD3/CD28 beads and IL-6).

RNA transcriptome patterns were examined to detect inherent differences between Tscm cells from RA patients and HCs. RNA transcription patterns were compared in CD4⁺ Tscm cells from patients with active RA (n = 3), RA patients in remission (n = 3), and HCs (n = 2). First, we compared the transcriptome patterns of Tscm cells from RA patients (those with active RA and those in remission) to those of cells from HCs to identify RA-specific transcripts. We identified 332 DEGs (FDR < 0.05, fold change > 2), of which 120 DEGs were upregulated in RA patients and 212 were upregulated in HCs (Fig. 5A). Enrichment analyses revealed that the GO terms for upregulated DEGs in RA were classified as protein glycosylation, transcription, positive regulation of MAPK cascade, positive regulation of mitotic nuclear division, and defense response to virus (Fig. 5B). When a gene expression heatmap of active RA and HC signatures was constructed, 153 genes showed higher expression in patients with active RA. The 153 genes were classified into the GO terms cellular response to cytokine stimulus, leukocyte activation, and regulation of fibroblast proliferation (Fig. 5C).

Next, the transcriptome of Tscm cells from patients with active RA was compared with that of cells from RA patients in remission. Only 15 DEGs were identified (FDR < 0.05, fold change > 2). Many genes (n = 38) were highly expressed in patients with active RA, moderately expressed in RA patients in remission, and minimally expressed in HCs (Fig. 5D). Principal component analysis revealed that Tscm cells in all three groups showed distinct gene expression profiles; the profile of Tscm cells from RA patients in remission was between those of patients with active RA and HCs (Fig. 5E). These results indicate that Tscm cells in patients with active RA are distinct from those in HCs and RA patients in remission, whereas Tscm cells from RA patients in remission are slightly different from those in HCs.

The results of the current study demonstrate the proinflammatory nature of Tscm cells in RA patients. Tscm cells in RA patients have the capacity for self-renewal and differentiation into memory and effector T subsets, which mainly contribute to synovial inflammation. They are more easily activated to produce inflammatory cytokines when stimulated by TCRs in the presence of IL-6. The proinflammatory nature of Tscm cells is embedded in RNA. The RNA transcriptomes of Tscm cells from RA patients are clearly different from those of HCs. Note that, although the percentage of Tscm cells was low, these cells showed the RNA transcriptome signature of active RA, even in a state of clinical remission, which indicates that Tscm cells play a role in the persistence of RA.

CD4⁺ T cells are involved in the pathogenesis of RA. Inflammatory cytokines secreted by T cells recruit and activate various inflammatory cells and induce the proliferation of synoviocytes; this results in the formation of pannus, which can ultimately destroy joints (27). Tscm cells, because of their capacity for self-renewal, can be a source of CD4⁺ T cells. In the current study, the percentage of Tscm cells was increased in RA patients compared with HCs. Consistent with our results, CD4⁺ Tscm cells were expanded in CD4⁺ T lymphocytes of RA patients, and citrullinated vimentin-specific CD4⁺ Tscm cells were increased in patients with active RA (28). The higher percentage of Tscm cells can be explained by the greater proliferative capacity of RA Tscm cells. Similar to another study (2), self-renewal was more pronounced as the proliferation of Tscm cells increased.

Tscm cells in RA patients show characteristics inherent to Tscm cells and have the capacity to differentiate into subsets of effector T cells. Despite similar profiles in differentiated naive-like T, Tcm, Tem, and Temra cells, the distribution of Tem cell subsets was different. When we stained Tem cells differentiated from Tscm cells for transcription factors characteristic of Th cells, we observed T-bet and RORγt expression and minimal expression of GATA3 (Fig. 2C). Th1 and Th17 subsets are thought to promote the pathogenesis of RA (11). In addition, Tscm cells differentiate into Tfh cells under appropriate conditions. The number of Tfh cells increases in patients with new-onset RA (29), and Tfh cells from RA patients contribute significantly to the chronic inflammation in the joints and are associated with disease activity (30). Notably, fewer FOXP3⁺ Treg cells were produced from RA Tscm cells than those in HCs under Treg-differentiation conditions. This result is consistent with previous reports showing fewer Treg cells in RA (31), which indicates that Tscm cells in RA patients differentiate more often to Th1 cells than Treg cells.

Similar to SLE (10), the percentage of Tscm cells showed almost no correlation with RA disease activity. Although correlations between the percentage of CD4⁺ Tscm cells in total CD4⁺ T cells and several features of RA disease activity (DAS28-ESR and DAS28-CRP) were statistically significant, the correlation coefficients were very low. Together with the results for SLE, this finding indicates that the role of Tscm cells in autoimmune diseases is to perpetuate the disease, and active effector T cells determine the disease activity.

Tscm cells of RA patients are easily activated and produce inflammatory cytokines when stimulated. When stimulated with TCR or IL-6, Tscm cells from RA patients expressed activation markers CD69, CD25, and CD154 (Fig. 4). Based on CD69 and CD154 expression, Tscm cells from RA patients were more easily stimulated than those from HCs. Regarding cytokines, the activated Tscm cells from RA patients secreted more TNF, IFN-γ, and IL-17A when stimulated with TCR or IL-6 (Fig. 4D). Specifically, IL-6 functioned synergistically with TCR to activate Tscm cells or to secrete cytokines. Thus, Tscm cells can contribute to the pathogenesis of RA in an RA microenvironment, which is abundant with proinflammatory cytokines such as IL-6. In the current study, IL-6 levels were significantly higher in the plasma of RA patients than HCs (Fig. 4A).

Transcriptome patterns of RA Tscm cells were clearly different from those of HCs (Fig. 5A). Transcriptome analyses showed upregulated patterns of protein glycosylation, transcription, positive regulation of the MAPK cascade, and positive regulation of mitotic nuclear division in RA patients (Fig. 5B). MAPKs are critical regulators of proinflammatory cytokines (IL-1, IL-6, IL-12, IL-23, and TNF) and contribute to worsening RA, such as joint destruction and inflammation, via signaling (32, 33). Genes governing cellular response to cytokine stimulus, leukocyte activation, and regulation of fibroblast proliferation were more prominent in RA patients (Fig. 5C). All of these genes and their associated pathways are related to the pathogenesis of RA, which indicates that Tscm cells already harbor the pathogenetic RA signature.

The percentage of Tscm cells in total T cells was significantly lower in patients in clinical remission than in patients with active disease (Fig. 3A). Although the percentage of Tscm cells was not greater in patients in remission than in HCs, the transcriptome pattern of RA patients in remission was different from that of HCs (Fig. 5E), which was more similar to that of patients with active RA (Fig. 5D). These results are consistent with the clinical finding that clinical remission in RA is different from cure (34). In addition, these findings indicate that a cure for RA may be associated with the normalization of transcriptome patterns in RA patients and that the transcriptome features of Tscm cells can be a biomarker of an RA cure in the future.

In the current study, the pathogenic nature of Tscm cells in RA was demonstrated. Because Tscm cells, like stem cells, have the capacity for self-renewal and differentiation, they are easily activated and secrete proinflammatory cytokines when stimulated in an RA-prone microenvironment. Their proinflammatory nature is embedded in the RNA transcriptome, which is not completely normalized in a state of clinical remission.

We thank all patients and HCs for providing blood for this study.

E.B.L. has acted as a consultant for Pfizer and received a research grant from GC Pharma (Gyeonggi-do, South Korea) and HANDOK (Seoul, South Korea).