Chronic Immune Activation in Systemic Lupus Erythematosus and the Autoimmune PTPN22 Trp620 Risk Allele Drive the Expansion of FOXP3+ Regulatory T Cells and PD-1 Expression

Abstract

In systemic lupus erythematosus (SLE), perturbed immunoregulation underpins a pathogenic imbalance between regulatory and effector CD4⁺ T-cell activity. However, to date, the characterization of the CD4⁺ regulatory T cell (Treg) compartment in SLE has yielded conflicting results. Here we show that patients have an increased frequency of CD4⁺FOXP3⁺ cells in circulation owing to a specific expansion of thymically-derived FOXP3⁺HELIOS⁺ Tregs with a demethylated FOXP3 Treg-specific demethylated region. We found that the Treg expansion was strongly associated with markers of recent immune activation, including PD-1, plasma concentrations of IL-2 and the type I interferon biomarker soluble SIGLEC-1. Since the expression of the negative T-cell signaling molecule PTPN22 is increased and a marker of poor prognosis in SLE, we tested the influence of its missense risk allele Trp⁶²⁰ (rs2476601C>T) on Treg frequency. Trp⁶²⁰ was reproducibly associated with increased frequencies of thymically-derived Tregs in blood, and increased PD-1 expression on both Tregs and effector T cells (Teffs). Our results support the hypothesis that FOXP3⁺ Tregs are increased in SLE patients as a consequence of a compensatory mechanism in an attempt to regulate pathogenic autoreactive Teff activity. We suggest that restoration of IL-2-mediated homeostatic regulation of FOXP3⁺ Tregs by IL-2 administration could prevent disease flares rather than treating at the height of a disease flare. Moreover, stimulation of PD-1 with specific agonists, perhaps in combination with low-dose IL-2, could be an effective therapeutic strategy in autoimmune disease and in other immune disorders.

Keywords: FOXP3; PD-1; PTPN22 Arg620Trp; autoimmunity; immunotherapy; regulatory T cells (Tregs); systemic lupus erythematosus (SLE); type I interferon.

Figure 1

Regulatory CD4⁺ T cell (Treg) frequency is increased in the circulation of SLE patients. (A) Gating strategy for the delineation of CD4⁺ regulatory T cells (Tregs). (B,C) Scatter plots depict the frequency (geometric mean ± 95% CI) of CD127^lowCD25^hi Tregs (B), and CD127^lowCD25^hi FOXP3⁺ Tregs (C) in PBMCs isolated from SLE patients (depicted in red) and matching healthy volunteers (depicted in blue). (D) Gating strategy for the delineation of the CD45RA⁺ naïve and CD45RA⁻ memory CD4⁺ T-cell compartments. (E,F) Scatter plots depict the frequency (geometric mean ± 95% CI) of memory CD45RA⁻ (E), and naïve CD45RA⁺ (F) CD127^lowCD25^hi FOXP3⁺ Tregs. Data were obtained from a discovery clinic-attending cohort (cohort 1) consisting of 34 SLE patients and 24 healthy volunteers, and from a population-based replication cohort (cohort 2) consisting of 41 SLE patients and 112 healthy volunteers. P-values were calculated using two-tailed Student's t-tests comparing the frequency of the assessed immune subsets in patients and controls. SLE, systemic lupus erythematosus; HC, healthy control; ns, not significant.

Figure 2

Frequency of FOXP3⁺ T cells in SLE patients is increased in both CD25^low and CD25^hi CD4⁺ T-cell subsets. (A) gating strategy for the delineation of the CD25^lowFOXP3⁺ Treg subset. (B,C) Scatter plots depict the frequency (geometric mean ± 95% CI) of CD25^lowFOXP3⁺ Tregs in SLE patients (red) and matching healthy volunteers (blue), defined as the frequency of FOXP3⁺ cells among CD127^lowCD25^low T cells (B) or as the frequency of CD127^lowCD25^lowFOXP3⁺ T cells among total CD4⁺ T cells (C). (D) Scatter plot depicts the frequency (geometric mean ± 95% CI) of total CD4⁺FOXP3⁺ Tregs in SLE patients (red) and matching healthy volunteers (blue), defined as the frequency of FOXP3⁺ cells among total CD4⁺ T cells. Data were obtained from a discovery clinic-attending cohort (cohort 1) consisting of 34 SLE patients and 24 healthy volunteers, and from a population-based replication cohort (cohort 2) consisting of 41 SLE patients and 112 healthy volunteers. P-values were calculated using two-tailed Student's t-tests comparing the frequency of the assessed immune subsets in patients and controls. SLE, systemic lupus erythematosus; HC, healthy control.

Figure 3

Expanded FOXP3⁺ Tregs in SLE patients are predominantly HELIOS⁺. (A,B) Gating strategy for the delineation of the CD45RA⁻ CD25^low (A) and CD25^hi (B) Treg subsets stratified by the expression of the canonical Treg transcription factors FOXP3 and HELIOS. (C,D) Scatter plots depict the frequency (geometric mean ± 95% CI) of CD25^low FOXP3⁺HELIOS⁺ (C), and CD25^hi FOXP3⁺HELIOS⁺ (D) Tregs in SLE patients (red) and matched healthy volunteers (blue). (E,F) Scatter plots depict the frequency (geometric mean ± 95% CI) of CD25^low FOXP3⁺HELIOS⁻ (E), and CD25^hi FOXP3⁺HELIOS⁻ (F) Tregs in SLE patients and matched healthy volunteers. Data were obtained from a discovery clinic-attending cohort (cohort 1) consisting of 34 SLE patients and 24 healthy volunteers, and from a population-based replication cohort (cohort 2) consisting of 41 SLE patients and 112 healthy volunteers. P-values were calculated using two-tailed Student's t-tests comparing the frequency of the assessed immune subsets in patients and controls. SLE, systemic lupus erythematosus; HC, healthy control; ns, not significant.

Figure 4

Single-cell RNA-sequencing reveals phenotypic similarities between CD25^lowFOXP3⁺ T cells and conventional CD25^hiFOXP3⁺ Tregs. (A) Uniform manifold approximation projection (UMAP) plot depicting clustering of all captured CD4⁺ single cells using the combined proteomics and transcriptomics data. Data were extracted from a targeted single-cell RNA-sequencing dataset (ref. 29), and depicts the analysis of pre-sorted CD127^lowCD25^hi (N = 7,711) and CD127^lowCD25^low (N = 7,115) T cells from one SLE patient and two control donors, including one type 1 diabetes (T1D) patient and one healthy donor. (B) Expression levels of the CD45RA (black to green) and CD45RO (black to red) isoforms using oligo-conjugated antibodies. (C) Expression of the canonical Treg transcription factors FOXP3 and HELIOS in the identified resting CD4⁺ T-cell clusters. (D) Detection of FOXP3⁺ cells (as assessed by the expression of >=1 copy of FOXP3 in each single cell) in either the pre-sorted CD127^lowCD25^hi (blue) or CD127^lowCD25^low (red) populations. Relative frequencies of FOXP3⁺ cells in SLE and control donors is indicated in the figure. (E) Frequency of FOXP3⁺ cells from either the CD127^lowCD25^hi or CD127^lowCD25^low populations in each identified regulatory T cell (Treg) or effector T cell (Teff) cluster. Clusters were ordered according to their relative positioning along the naïve to memory (Mem) differentiation axis. (F,G) Volcano plots depict the differential expression of the assessed genes at the mRNA (N = 397 genes) and protein (N = 24) level between: (i) CD127^lowCD25^low FOXP3⁺ vs. CD127^lowCD25^low FOXP3⁻ T cells (F); and (ii) CD127^lowCD25^hi FOXP3⁺ Tregs vs. CD127^lowCD25^low FOXP3⁻ T cells (G).

Figure 5

Memory FOXP3⁺HELIOS⁺ Tregs from SLE patients display hallmarks of recent immune activation. (A–D) Representative histograms and summary scatter plots depict the frequency (geometric mean ± 95% CI) of: (i) PD-1⁺ (A); (ii) CD25 mean fluorescence intensity (B); (iii) TIGIT⁻ (C); and (iv) Ki-67⁺ (D) cells within CD45RA⁻ CD25^hi FOXP3⁺HELIOS⁺ Tregs. Data were obtained from a discovery clinic-attending cohort (cohort 1) consisting of 34 SLE patients (depicted in red) and 24 healthy volunteers (depicted in blue) and from a population-based replication cohort (cohort 2) consisting of 41 SLE patients and 112 healthy volunteers. Samples from cohorts 1 and 2 were processed within a period of 11 months of each other. For each parameter an illustrative histogram is provided as a reference depicting the expression of the marker in the CD45RA⁺ FOXP3⁺HELIOS⁺ naïve Treg population (depicted in gray). P-values were calculated using two-tailed Student's t-tests comparing the frequency of the assessed immune subsets in patients and controls. SLE, systemic lupus erythematosus; HC, healthy control; ns, not significant.

Figure 6

FOXP3⁺ Treg expansion is a marker of recent autoimmune reaction. (A,B) Correlation between the frequency of CD25^lowFOXP3⁺ Tregs in blood and the circulating concentration of the plasma type I interferon (IFN) marker soluble SIGLEC-1 (sSIGLEC-1). Data shown depict the correlation in SLE patients (red) and matched healthy volunteers (black) in either the discovery clinic-attending cohort—cohort 1 (A) or the replication population-based cohort—cohort 2 (B). (C,D) Data shown depict the correlation between the total CD4⁺FOXP3⁺ CD45RA⁻ memory Tregs and the plasma concentration of sSIGLEC-1 in cohort 1 (C) and cohort 2 (D). (E,F) Data shown depict the correlation between the frequency of CD25^lowFOXP3⁺ (E) or total FOXP3⁺ (F) CD45RA⁻ memory Tregs and the plasma concentration of IL-2 measured in the 41 SLE patients from cohort 2. The correlation coefficients (r) and the respective P-values from the linear regression in SLE patients and healthy controls are shown in the plots.

Figure 7

The autoimmunity missense PTPN22 Trp⁶²⁰ risk allele is associated with increased frequency of CD127^lowCD25⁺ Tregs in circulation. (A) Gating strategy depicting the delineation of total CD127^lowCD25^hi Tregs in blood. Data were obtained by surface immunostaining of freshly isolated whole blood from 486 healthy subjects selected by genotype from the Cambridge BioResource. (B–D) Scatter plots depict the frequency (geometric mean ± 95% CI) of: (i) CD127^lowCD25^hi Tregs (B); (ii) CD45RA⁻ CD127^lowCD25^hi memory Tregs (C); and (iii) CD45RA⁺ CD127^lowCD25^hi naive Tregs (D) in PTPN22 Arg⁶²⁰/Arg⁶²⁰ (N = 362; red), Arg⁶²⁰/Trp⁶²⁰ (N = 85; blue), and Trp⁶²⁰/Trp⁶²⁰ (N = 39; green) donors. P-values were calculated by linear regression analysis. ns, not significant.

Figure 8

T cells from PTPN22 Trp⁶²⁰/Trp⁶²⁰ donors display an activated phenotype and increased proliferative capacity. (A,B) Scatter plots depict the frequency (geometric mean ± 95% CI) of PD-1⁺ cells within: (i) CD45RA⁻ CD25^hi FOXP3⁺HELIOS⁺ Tregs (A); and (ii) CD45RA⁻ CD25^int/low Teffs (B) from 152 healthy volunteers recruited from the Cambridge BioResource. Data were stratified according to the genotype of the autoimmune-associated PTPN22 Arg⁶²⁰Trp variant. (C) Treg suppressive capacity was assessed in a subset of 27 common Arg⁶²⁰ homozygous and 24 age- and sex-matched rare Trp⁶²⁰ homozygous donors by measuring the proliferation of autologous CD45RA⁻ CD25^int/low T effector cells (mTeffs) in co-culture after in vitro activation for 6 days with anti-CD2/3/28 beads, at a ratio of either 1:1 or 1:0.5 mTeff:Treg (D) Increased proliferation of mTeffs, in the absence of Tregs, from a given donor correlates with reduced suppression of mTeff proliferation when autologous Tregs were present at a ratio of 1 mTeff: 0.5 Treg. (E) Proliferative capacity of CD45RA⁻ memory Teffs and CD45RA⁺ naïve Teffs was assessed after 6 days in vitro stimulation with anti-CD2/3/28 beads in the absence of Tregs in culture. P-values were calculated using two-tailed student's t-tests comparing the frequency of the assessed immune subsets between the PTPN22 Arg⁶²⁰/Arg⁶²⁰ and Trp⁶²⁰/Trp⁶²⁰ genotype groups. ns, not significant; mTeffs, CD45RA⁻ CD25^int/low T effector cells; CPM, proliferation readings—[³H]thymidine incorporation rate (fmoles/g).