NR4A family members regulate T cell tolerance to preserve immune homeostasis and suppress autoimmunity

Abstract

The NR4A family of orphan nuclear receptors (Nr4a1-3) plays redundant roles to establish and maintain Treg identity; deletion of multiple family members in the thymus results in Treg deficiency and a severe inflammatory disease. Consequently, it has been challenging to unmask redundant functions of the NR4A family in other immune cells. Here we use a competitive bone marrow chimera strategy, coupled with conditional genetic tools, to rescue Treg homeostasis and unmask such functions. Unexpectedly, chimeras harboring Nr4a1-/- Nr4a3-/- (double-knockout, DKO) bone marrow developed autoantibodies and a systemic inflammatory disease despite a replete Treg compartment of largely WT origin. This disease differs qualitatively from that seen with Treg deficiency and is B cell extrinsic. Negative selection of DKO thymocytes is profoundly impaired in a cell-intrinsic manner. Consistent with escape of self-reactive T cells into the periphery, DKO T cells with functional, phenotypic, and transcriptional features of anergy accumulated in chimeric mice. Nevertheless, we observed upregulation of genes encoding inflammatory mediators in anergic DKO T cells, and DKO T cells exhibited enhanced capacity for IL-2 production. These studies reveal cell-intrinsic roles for the NR4A family in both central and peripheral T cell tolerance and demonstrate that each is essential to preserve immune homeostasis.

Keywords: Autoimmune diseases; Autoimmunity; Immunology; T cell development; Tolerance.

Figure 1. Systemic immune dysregulation and Treg deficiency in mice with germline deficiency of Nr4a1 and Nr4a3.

(A) Nr4a1^–/–Nr4a3^–/– (gDKO) mouse (red arrow) compared with healthy littermate, 4 weeks; representative of n = 8. (B) Flow plots show splenic CD4⁺ T cells with FOXP3⁺ Treg gate in mice of each genotype. Representative of 5 mice/genotype. (C and D) Quantification of thymic (C) and splenic (D) Treg cell number (n = 5, 3- to 4-week-old gDKO and 5- to 6-week-old mice with other genotypes). (E–G) Competitive BM chimera design. (H and I) Flow plots show thymic CD4 single-positive (CD4SP) (H) or splenic CD4⁺ T cell (I) subpopulations in 1:1 DKO:WT chimeras. Representative of 6 (H) or 10 (I) chimeras. (J–O) Quantification of thymic (J) or splenic (M) Treg cell number in 1:1 chimeras. Ratio of CD45.2 to CD45.1/2 for thymic (K and L) or splenic (N and O) Treg, CD25⁺FOXP3^–, and CD25^–FOXP3^– cells in 1:1 chimeras, normalized to double-positive (DP) thymocytes. n = 3 (J–L) or 6 (M–O), pooled from 2 sets of independently generated chimeras 6–10 weeks posttransplant. (P and Q) Flow plots show thymic CD4SP (P) or splenic CD4⁺ T cell (Q) subpopulations in 1:5 DKO:WT chimera. Representative of ≥3 chimeras from 1 chimera setup. Graphs depict mean ± SEM. Statistical significance was assessed by 1-way ANOVA with Tukey’s test (C, D, K, L, N, and O) or 2-tailed unpaired Student’s t test (J and M). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. NS, not significant.

Figure 2. DKO thymocytes have a cell-intrinsic defect in negative selection.

(A) Flow plots show thymic subsets in WT, Nr4a3^–/–, Nr4a1^–/–, and gDKO mice. Representative of n ≥ 4 mice/genotype. (B) Quantification of thymic subset cell number as gated in A; (n ≥ 4, 3 to 4-week-old gDKO and 5- to 6-week-old mice of other genotypes). (C) Ratio of CD45.2 to CD45.1/2 thymocytes among thymic DN3a and DN3b subsets (as gated in Supplemental Figure 2B), normalized to DN2 subset (n = 3–4 chimeras). (D) Flow plots show thymic subsets in competitive chimeras. Representative of ≥3 mice/genotype. (E) Ratio of CD45.2 to CD45.1/2 thymic subsets as gated in D normalized to DP subset (n ≥ 3). Data in C–E were from 6 to 7 weeks posttransplant chimeras pooled from 3 sets of independently generated chimeras. (F–K) Thymocytes from 1:1 DKO:WT chimeras were cultured with varying doses of plate-bound anti-CD3 and 2 μg/mL of anti-CD28 for 24 hours. Cells were stained to detect CD4/CD8 surface markers, followed by permeabilization and detection of active Caspase3 (aCasp3). Representative plots show aCasp3 expression in WT CD45.1/2 and DKO CD45.2 DP (F) and CD4SP (I) thymocytes from 1:1 DKO chimeras cultured with 10 μg/mL anti-CD3. Quantification percentage aCasp3⁺ cells among DP (G and H) or CD4SP (J and K) in 1:1 DKO:WT (G and J) or 1:1 WT:WT (H and K) chimeras (n = 3 from 1 chimera setup). Graphs depict mean ± SEM. Statistical significance was assessed by 1-way (B) or 2-way (C and E) ANOVA with Tukey’s test or 2-tailed unpaired Student’s t test with the Holm-Šídák method (G, H, J, and K). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. NS, not significant.

Figure 3. Myeloproliferative disorder in DKO mice is a non–cell-autonomous effect of NR4A deficiency.

(A–D) Lymph node cells (A and B) and thymocytes (C and D) from WT and gDKO mice (A and C) or 1:1 DKO:WT chimeras (B and D) were stained to detect CD11b and Gr1 (Ly6G/Ly6C) expression. Shown are representative plots of ≥ 5 mice. (E–H) Quantification of CD11b⁺Gr1^– cells in lymph nodes (E and F) and thymocytes (G and H) from WT, Nr4a3^–/–, Nr4a1^–/–, and gDKO mice (E and G) (n ≥ 5, 3- to 4-week-old gDKO and 5- to 6-week-old mice of other genotypes) and from WT:WT = 1:1 and DKO:WT = 1:1 chimeras (F and H) (n ≥ 3 pooled from 2 sets of independently generated chimeras). (I) Ratio of CD45.2 to CD45.1/2 for CD11b⁺Gr1^– cells in lymph nodes, thymus, and spleen from WT:WT = 1:1 and DKO:WT = 1:1 chimera (n ≥ 3 pooled from 2 sets of independently generated chimeras). Graphs depict mean ± SEM. Statistical significance was assessed by 1-way ANOVA with Tukey’s test (E and G), 2-tailed unpaired Student’s t test with (I) or without (F and H) the Holm-Šídák method. **P < 0.01; ****P < 0.0001. NS, not significant.

Figure 4. Abnormal B cell homeostasis in DKO mice is a non–cell-autonomous effect of NR4A deficiency.

(A) Representative flow plots showing CD69 expression on splenic B cells from WT (shaded gray histogram) and overlaid Nr4a3^–/–, Nr4a1^–/–, or gDKO mice. (B) Quantification of CD69 MFI as in A (data in A and B represent n ≥ 5, 3- to 4-week-old gDKO and 5- to 6-week-old mice of other genotypes). (C) Quantification of CD69 MFI on splenic B cells of each donor genotype in competitive 1:1 chimeras (n = 3 from 1 chimera setup). (D and G) Representative flow plots show FAS^hiGL7⁺ GC B cells pregated on B220⁺IgD^lo splenocytes (D) and CD138⁺ splenocytes (G) from competitive chimeras. (E–I) Frequency of GC B cells among total B cells (E), ratio of CD45.2 to CD45.1/2 GC B cells normalized to B220⁺IgD^hi naive B cells (F), ratio of CD138⁺ to B220⁺ splenocytes (H), ratio of CD45.2 to CD45.1/2 CD138⁺ cells normalized to B220⁺CD138^– cells (I) from competitive chimeras as gated in D and G (data in D–I represent n ≥ 6 pooled from 3 sets of independently generated chimeras). (J–M) Representative flow plots show GC B cells (J) and CD138⁺ cells (L) in spleen from host chimeras transplanted with either mb1-cre or mb1-cre Nr4a1^fl/fl Nr4a3^–/– (cDKO) BM after 40 weeks. Frequency of GC B cells among total B cells (K) and CD138⁺ cells among splenocytes (M) (n ≥ 3). Graphs depict mean ± SEM. Statistical significance was assessed by 1-way ANOVA with Tukey’s test (B) or Dunnett’s test (E, F, H, and I) or 2-tailed unpaired Student’s t test with (C) or without (K and M) the Holm-Šídák method. *P < 0.05; **P < 0.01; ****P < 0.0001. NS, not significant.

Figure 5. Reconstitution of WT Treg compartment does not restore CD8⁺ T cell homeostasis in competitive chimeras.

(A) Splenocytes from WT, Nr4a3^–/–, Nr4a1^–/–, and gDKO mice were stained to detect CD8⁺ T cell subsets on the basis of CD44 and CD62L expression. Plots are representative of ≥5 mice/genotype. (B) Quantification of splenic CD44^hiCD8⁺ T cells as gated in A (n ≥ 5, 3- to 4-week-old gDKO and 5- to 6-week-old mice of other genotypes). (C) Flow plots showing the peripheral CD8⁺ T cell subsets in competitive chimeras, as described for A above. Representative of ≥7 chimeras of each type. (D) Quantification of splenic CD44^hiCD8⁺ T cells from chimeras as gated in C at varied time points posttransplant (n ≥ 3). (E) Ratio of CD45.2 to CD45.1/2 for CD8⁺CD44^hi population as gated in C, normalized to naive CD8⁺CD44^loCD62L^hi gate (n ≥ 3). Data in C–E pooled from 2 sets of independently generated chimeras. (F and G) Thymocytes and splenocytes from CD8-cre and CD8-cre Nr4a1^fl/fl Nr4a3^–/– (cDKO) mice were stimulated with PMA and ionomycin (PMA/Io) for 2 hours. Flow plots show intracellular NUR77 expression following fixation and permeabilization within thymic and splenic T cell subsets (F). Quantification of NUR77 MFI in T cell subsets (G) (n = 3 mice/genotype). (H) Quantification of splenic CD8⁺CD44^hi T cells from CD8-cre, Nr4a3^–/–, CD8-cre Nr4a1^fl/fl, and CD8-cre cDKO mice (n = 3 mice/genotype). Graphs depict mean ± SEM. Statistical significance was assessed by 1-way ANOVA with Tukey’s test (B and H) or 2-way ANOVA with Dunnett’s test (D and E). ***P < 0.001; ****P < 0.0001. NS, not significant.

Figure 6. Accumulation of anergic DKO CD4⁺ T cells in competitive chimeras.

(A) Splenocytes from WT, Nr4a3^–/–, Nr4a1^–/–, and gDKO mice were stained to detect CD4⁺ T cell subsets on the basis of CD44 and CD62L expression. Plots are representative of ≥5 mice/genotype. (B) Quantification of splenic CD4⁺CD44^hiCD62L^lo T cells as gated in A (n ≥ 5, 3- to 4-week-old gDKO and 5- to 6-week-old mice of other genotypes). (C) Quantification of splenic FOXP3^–CD4⁺CD44^hiCD62L^lo T cells as gated in Supplemental Figure 6A from competitive chimeras at indicated time points posttransplant (n ≥ 3, pooled from 2 sets of independently generated chimeras). (D) Splenocytes from 12 weeks posttransplant DKO:WT = 1:5 chimera were stained to detect anergic T cell subsets. Flow plots depict CD73^hiFR4^hi (anergic) T cells within CD44^loCD62L^hi (naive), CD44^hiCD62L^hi, and CD44^hiCD62L^lo (memory) compartments of CD4⁺FOXP3^– cells of each donor genotype. Representative of 7 chimeras, generated in 1 set. (E–G) Ratio of CD45.2 to CD45.1/2 within CD73^hiFR4^hi gate among naive (E), CD44^hiCD62L^hi (F), or memory (G) CD4⁺ T cell compartments, as gated in D. Shown are WT:WT = 1:1 and DKO:WT = 1:5 chimeras at indicated time points posttransplant (n ≥ 3 pooled from 2 sets of independently generated chimeras). Ratios were normalized to naive CD4⁺ T cells. Graphs depict mean ± SEM. Statistical significance was assessed by 1-way (B) ANOVA with Tukey’s test, 2-way ANOVA with Dunnett’s test (C), or 2-tailed unpaired Student’s t test with the Holm-Šídák method (E–G). *P < 0.05; **P < 0.01; ****P < 0.0001. NS, not significant.

Figure 7. Functional and transcriptional characteristics of anergic CD4⁺ T cells in competitive chimeras.

(A) Splenocytes from DKO:WT = 1:5 chimera were stimulated with anti-CD3 for 30 seconds followed by secondary cross-linking antibody for 2 minutes, or alternatively with PMA for 2 minutes. Cells were fixed, permeabilized, and then stained to detect surface markers, FOXP3, and phosphorylated Erk (p-Erk). Representative histograms showing intracellular p-Erk expression in nonanergic (CD73^loFR4^lo; NA), intermediate anergic (CD73^intFR4^int; IA), or anergic (CD73^hiFR4^hi; A) among naive (CD44^loCD62L^hi) or memory (CD44^hiCD62L^lo) CD4⁺ T cells gated as in Supplemental Figure 7A. Dashed line shows the threshold of positive gate. Plots are representative of n = 6 mice. (B) Quantification of %pErk⁺ as in A above (n = 3 biological replicates, representative of n = 2 independent experiments from 1 chimera setup). Graphs depict mean ± SEM. Statistical significance was assessed by 2-tailed unpaired Student’s t test with the Holm-Šídák method. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (C and D) Naive or anergic CD4⁺ T cells from CD45.1 (WT) or CD45.2 (DKO) cells gated as in Supplemental Figure 7E were sorted directly into buffer RLT for RNA sequencing. ClustVis heatmaps depict expression of selected genes associated with anergy (C) or Tregs (D). (E and F) GSEA plots for the genes downregulated (E) or upregulated (F) by Nr4a1 (17) against differentially expressed genes (DEGs) in DKO and WT anergic cells. DEGs were defined as genes upregulated in DKO compared with WT anergic cells with P < 0.05. NES, normalized enrichment score; FDR, false discovery rate. (G) Heatmap shows expression of selected inflammatory mediators.

Figure 8. Transcriptional targets of NR4A family in acutely stimulated naive CD4⁺ T cells.

(A) ClustVis heatmap shows overlap PRGs and DEGs in TCR-stimulated naive CD4⁺ T cells. PRGs defined as genes upregulated in stimulated WT naive CD4⁺ T cells relative to ex vivo with FDR < 0.05, log CPM > 1, and log₂ fold change > 1.5. DEGs were defined as in Figure 7. (B and C) GSEA plots for the genes downregulated (B) or upregulated (C) by Nr4a1 (17) against DEGs in DKO and WT stimulated naive CD4⁺ T cells.

Figure 9. NR4A family negatively regulates IL-2 production in CD4⁺ T cells.

(A) CD4⁺ T cells were isolated by negative selection from lymph nodes and cultured in plates coated with indicated dose of anti-CD3 + anti-CD28 for 24 hours (left) or 48 hours (right). IL-2 concentration in supernatant was measured with ELISA (n = 3 biological replicates). (B) Lymph node cells from 10 weeks posttransplant DKO:WT = 1:1 chimera were cultured in plates coated with indicated doses of anti-CD3 for 20 hours. Then cells were restimulated with PMA, ionomycin, and brefeldin for an additional 4 hours. Representative histograms of 3 mice showing intracellular IL-2 in CD4⁺ cells of each donor genotype. (C) Quantification of %IL-2⁺ as described for B above (n = 3 biological replicates from 1 chimera setup). (D) Transcripts per million (TPM) of Il2 detected with RNA sequencing in WT and DKO cells sorted as described. Graphs depict mean ± SEM. Statistical significance was assessed by 2-way ANOVA with Tukey’s test (A), 2-tailed unpaired Student’s t test with the Holm-Šídák method (C), or a paired differential expression analysis with EdgeR comparing samples from the same chimeras (D). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. NS, not significant.

Figure 10. Restoring the Treg compartment in competitive chimeras alters autoantibody repertoire but does not restore tolerance.

(A–D) ANA immunofluorescence images — 1:40 diluted serum of indicated mice was applied to Hep-2 substrate slides, washed, and stained with FITC–anti–mouse IgG. Original magnification, 20×. Images are representative of biological replicates as quantified below (E and G). (E) Graphs depict frequency of negative, nuclear, or nuclear+cytoplasmic Hep-2 cell staining patterns in WT:WT = 1:1 chimera, DKO:WT = 1:1 chimera, and gDKO (1:40 dilution). Data include analysis of serum from 2 sets of independently generated chimeras 6 to 12 weeks posttransplant. Statistical significance was assessed by Fisher’s exact test. ****P < 0.0001. (F) Quantification of anti-dsDNA antibody from n = 9 WT:WT 1:1 chimeras and n = 17 DKO:WT 1:1 chimeras determined by ELISA, pooled from 2 sets of individually generated chimeras. Statistical significance was assessed by 2-tailed unpaired Student’s t test. **P < 0.01. (G) ANA titer determined with serial 2-fold dilution of serum from chimeras at indicated time points posttransplant stained as in A–D.