NR4A family members regulate T cell tolerance to preserve immune homeostasis and suppress autoimmunity
Ryosuke Hiwa, … , Ravi Mandla, Julie Zikherman
JCI Insight. 2021. https://doi.org/10.1172/jci.insight.151005.
Research In-Press Preview Immunology
Graphical abstract
Find the latest version: https://jci.me/151005/pdf 1 NR4A family members regulate T cell tolerance to preserve immune homeostasis and
2 suppress autoimmunity
3
4 Ryosuke Hiwa1, Hailyn V. Nielsen1, *, James L. Mueller1, *, Ravi Mandla2
5 and Julie Zikherman1
6
7 Affiliations: 1Division of Rheumatology, Rosalind Russell and Ephraim P. Engleman
8 Arthritis Research Center, Department of Medicine, University of California, San
9 Francisco, CA, 94143
10 2Cardiology Division, Department of Medicine, University of California, San Francisco,
11 CA, 94158
12 *These authors contributed equally: Hailyn V. Nielsen, James L. Mueller
13 Address correspondence to:
14 Dr. Julie Zikherman, 513 Parnassus Avenue, Room HSW1201E, Box 0795, San
15 Francisco, CA 94143-0795. Phone: (415)-476-1689; E-mail address:
17 Funding: The Uehara Memorial Foundation Research Fellowship (RH), NIGMS
18 Molecular and Cellular Immunology Program T32AI00733432 (HVN), NIAMS R01
19 AR069520 (JZ), NIAID R01 AI148487 (JZ), Rheumatology Research Foundation (JZ)
20 Conflict of interest: JZ is a scientific advisor for Walking Fish Therapeutics.
21 Key words: NUR77, NOR-1, Nr4a1, Nr4a3, Treg, T cell tolerance, anergy, negative
22 selection
23 Running title: NR4A family in T cell tolerance and immune homeostasis
1 24 ABSTRACT
25 The NR4A family of orphan nuclear receptors (Nr4a1-3) plays redundant roles to
26 establish and maintain Treg identity; deletion of multiple family members in the thymus
27 results in Treg deficiency and a severe inflammatory disease. Consequently, it has been
28 challenging to unmask redundant functions of the NR4A family in other immune cells.
29 Here we use a competitive bone marrow chimera strategy, coupled with conditional
30 genetic tools, to rescue Treg homeostasis and unmask such functions. Unexpectedly,
31 chimeras harboring Nr4a1–/– Nr4a3–/– (DKO) bone marrow develop autoantibodies and a
32 systemic inflammatory disease despite a replete Treg compartment of largely wild-type
33 origin. This disease differs qualitatively from that seen with Treg-deficiency and is B cell-
34 extrinsic. Negative selection of DKO thymocytes is profoundly impaired in a cell-intrinsic
35 manner. Consistent with escape of self-reactive T cells into the periphery, DKO T cells
36 with functional, phenotypic, and transcriptional features of anergy accumulate in
37 chimeric mice. Nevertheless, we observe upregulation of genes encoding inflammatory
38 mediators in anergic DKO T cells, and DKO T cells exhibit enhanced capacity for IL-2
39 production. These studies reveal cell-intrinsic roles for the NR4A family in both central
40 and peripheral T cell tolerance, and demonstrate that each is essential to preserve
41 immune homeostasis.
42
43 INTRODUCTION
44 Since the initial discovery of regulatory T cells (Treg) and their recognition as a
45 distinct T cell lineage dependent upon the transcription factor FOXP3, it has been
46 shown that they are absolutely essential for immune homeostasis and tolerance to self
2 47 (1). Indeed, Foxp3-deficient mice and mice with a loss-of-function mutation in Foxp3
48 (Scurfy) rapidly develop an autoimmune disease characterized by cytokine storm,
49 immune cell proliferation and infiltration, autoantibody production, and death typically by
50 4 weeks of age (1-4). Conversely, re-introducing Treg is sufficient to prevent this
51 disease (2). However, extensive cell-intrinsic mechanisms that operate in other immune
52 cell lineages are also essential to maintain tolerance to self, including processes such
53 as deletion and hypo-responsiveness of self-reactive lymphocytes (termed anergy) (5).
54 Prior work has implicated a small family of orphan nuclear hormone receptors
55 (encoded by Nr4a1-3) in several of these processes. Most notably, NR4A family
56 members play redundant roles upstream of Foxp3 to maintain Treg identity and
57 function; deletion of multiple family members in the thymus results in profound Treg
58 deficiency and a severe “Scurfy-like” disease that phenocopies Foxp3-deficient mice (6,
59 7). As a result, it has been difficult to isolate redundant functions of this family in other
60 immune cell populations. Yet this remains an important area to explore since the NR4A
61 family are widely expressed and thought to be druggable targets that may facilitate
62 manipulation of immune cell function in the context of autoimmune disease, tumor
63 immunotherapy, and hematologic malignancies (8-11).
64 Nr4a1-3 (encoding NUR77, NURR1, and NOR-1, respectively) are rapidly
65 upregulated in response to mitogenic stimuli, including antigen receptor ligation, and are
66 thought to function as constitutively active transcription factors without a confirmed
67 endogenous ligand (12). As a result, not only are these family members upregulated in
68 T and B cells after acute antigen encounter, but also in Treg in the steady-state, in
69 thymocytes undergoing negative selection, and in self-reactive, anergic, or exhausted
3 70 lymphocytes in response to chronic antigen stimulation (6, 9, 13-19). Indeed, the NR4A
71 family has been argued to play a tolerogenic role in all these contexts. The NR4A family
72 selectively restrains the survival and expansion of B cells that encounter antigen (signal
73 1) in the absence of co-stimulation (signal 2) (15, 20). Similarly, overexpression of either
74 Nr4a1 or Nr4a3 mediates antigen-induced cell death in the thymus, while a dominant-
75 negative transgenic (Tg) construct had the opposite effect (13, 21, 22). However,
76 Nr4a1–/– mice exhibit extremely subtle defects in thymic negative selection (23, 24),
77 suggesting possible redundancy among the family members. Nr4a1 and Nr4a3 also
78 play non-redundant roles in peripheral conventional T cells (Tconv): most notable
79 among these are roles for Nr4a1 in CD4+ T cell anergy (17) and an additive role for all 3
80 family members in CD8+ T cell exhaustion (9). Finally, it has been argued that Nr4a1
81 and Nr4a3 redundantly maintain myeloid homeostasis since, in their absence, a
82 myeloproliferative disease is observed (25). However, unmasking redundancy between
83 NR4A family members in many of these contexts has been hampered by profound
84 immune dysregulation that develops in the absence of functional Treg.
85 We sought to bypass this obstacle by generating competitive bone marrow (BM)
86 chimeras harboring both wild-type (WT) cells (that could reconstitute a functional Treg
87 compartment) and DKO cells (lacking both Nr4a1 and Nr4a3) to isolate cell-intrinsic
88 immune functions for the NR4A family. Unexpectedly, mixed chimeras harboring both
89 WT and DKO BM rapidly developed anti-nuclear autoantibodies (ANA) and a systemic
90 inflammatory disease, despite a replete Treg compartment of largely WT origin. The
91 disease that developed in BM chimeras was B cell-extrinsic and qualitatively different
92 from that in germline DKO mice. We found that negative selection of DKO thymocytes in
4 93 competitive chimeras was profoundly impaired in a cell-autonomous manner. DKO
94 Tconv cells with phenotypic, functional, and transcriptional features of antigen
95 experience and anergy accumulate in these chimeras, suggesting escape of self-
96 reactive T cells into the periphery. However, self-reactive DKO CD4+ Tconv cells
97 nevertheless exhibit expression of inflammatory mediators and exaggerated capacity for
98 IL-2 production, suggesting that anergy is defective. Our findings unmask essential,
99 redundant roles for the NR4A family in both central and peripheral T cell tolerance to
100 maintain immune homeostasis.
101
102 RESULTS
103 Systemic immune dysregulation in mice with germline deficiency of Nr4a1 and
104 Nr4a3
105 Nr4a1, 2, and 3 are all expressed in thymocytes, Treg, and peripheral T cells, but
106 the expression of Nr4a2 is minimal under steady-state conditions (Supplementary
107 Figure 1A; immgen.org). To unmask redundant functions of the NR4A family, we
108 generated mice lacking germline expression of both Nr4a1 and Nr4a3. We used
109 Nr4a1fl/fl mice to generate Nr4a1-deficient mice with germline excision of the loxp-
110 flanked locus and bred this with a CRISPR-generated Nr4a3–/– line that we recently
111 described (20). An independently-generated line of Nr4a1–/– mice in widespread use
112 has been reported to express a truncated NUR77 protein encoded by exon 2 of Nr4a1
113 (23, 26). Our newly generated Nr4a1–/– Nr4a3–/– mice (germline DKO, denoted as gDKO
fl/fl 114 here-in) do not express exon 2 of Nr4a1 consistent with the prior analysis of Nr4a1
115 mice (26).
5 116 gDKO mice were born at Mendelian ratios but exhibited severe runting (Figure
117 1A) and invariably died before 4 weeks of age, consistent with observed mortality in an
118 independent gDKO line generated with distinct Nr4a1 and Nr4a3 null alleles (25). As
119 previously reported for CD4-cre Nr4afl/fl Nr4a3–/– mice (6), our gDKO mice exhibit near-
120 complete loss of FOXP3+ Treg in both thymus and periphery (Figures 1B-D). gDKO
121 mice also exhibit severe thymic atrophy, but loss of peripheral Treg is disproportionate
122 relative to a more modest reduction of total splenocytes (Supplementary Figures 1B,
123 C). Concurrently, we observed expansion of a unique population of CD4+CD25+
124 FOXP3– T cells in the thymus and also in the periphery that may represent cells that
125 have either lost or failed to upregulate expression of FOXP3, described elsewhere
126 (Figures 1B, Supplementary Figures 1D-G) (6, 27, 28). Importantly, and consistent
127 with prior reports, neither Nr4a1–/– nor Nr4a3–/– single knockout (SKO) mice exhibit Treg
128 loss, expansion of this unique cell population, or frank disease (Figures 1B-D,
129 Supplementary Figures 1B-G) (6).
130
131 Cell intrinsic Treg defect in the absence of Nr4a1 and Nr4a3
132 Systemic inflammatory disease and associated thymic atrophy preclude the
133 study of thymic development and mature Tconv cells in gDKO mice (Supplementary
134 Figure 1B). Similar mortality observed in both germline (25) and CD4-cre conditional
135 mouse lines (6) suggested to us that disease in gDKO animals might be due to Treg
136 deficiency in both settings. We reasoned that restoring functional Treg could unmask
137 cell-intrinsic functions of NR4A family in other immune cell populations. To do so, we
138 generated competitive chimeras to allow WT donor BM to reconstitute a functional Treg
6 139 compartment. Equal proportions of congenically marked donor BM from CD45.2 gDKO
140 and CD45.1/2 WT mice were transplanted into lethally irradiated CD45.1 BoyJ
141 recipients (Figure 1F). In parallel, we generated control chimeras in which CD45.1
142 hosts were reconstituted with a mixture of CD45.2 WT and CD45.1/2 WT BM (Figure
143 1E). In addition, we also generated mixed chimeras with a low proportion of gDKO
144 donor BM (1:5 ratio) to further ensure development of a WT Treg compartment (Figure
145 1G). We assessed reconstitution and immune phenotypes of chimeras at sequential
146 time points between 6-14 weeks post-transplant.
147 Consistent with studies of CD4-cre chimeras, we observed a profound cell-
148 intrinsic disadvantage for DKO Treg in the thymus and spleen when compared to
149 CD4SP thymocytes (Figures 1H-O) (6). Similar results were reproduced with DKO:WT
150 1:5 chimera (Figure 1P, Q, Supplementary Figures 1H-K). FOXP3 and CD25
151 expression in DKO Treg was reduced (Supplementary Figures 1L-O), consistent with
152 a role for the NR4A family in ‘maintenance’ of Treg identity (18). Most importantly, the
153 Treg compartment was restored and Treg number was comparable between DKO:WT
154 chimera and WT:WT chimera (Figures 1J, M, Supplementary Figures 1J, K). We also
155 confirmed Treg were largely reconstituted from WT donor in DKO:WT 1:5 chimera
156 (Figures 1P, Q). This allowed us to explore the cell-intrinsic roles of the NR4A family in
157 other immune cell types.
158
159 Thymic atrophy is partially rescued in competitive chimeras
160 gDKO mice exhibit severe thymic atrophy with marked reduction of all thymocyte
161 subsets and disproportionate loss of DP thymocytes (Figures 2A, B). We postulated
7 162 that this might be an indirect consequence of Treg deficiency and systemic inflammation
163 in gDKO mice, since DP thymocytes are sensitive to glucocorticoid-induced apoptosis
164 (29). Indeed, profound thymic atrophy is partially rescued in DKO:WT 1:1 chimeras and
165 fully rescued in DKO:WT 1:5 chimeras within the first 6 weeks of reconstitution
166 (Supplementary Figure 2A). However, progressive thymic atrophy was observed over
167 time in DKO:WT chimeras (relative to WT:WT control chimera). This led us to focus on
168 early time points to isolate cell-intrinsic roles for the NR4A family during thymic
169 development (6 weeks post-transplant).
170
171 NR4A expression is dispensable for thymic β-selection
172 We previously showed, using a fluorescent reporter of Nr4a1 transcription
173 (NUR77-eGFP), that GFP is upregulated at the β-selection checkpoint during thymic
174 development, raising the possibility that Nr4a1 and family members might play a
175 functional role here (14). Immature double negative (DN) thymocytes (lacking both CD4
176 and CD8 expression) recombine the TCRβ chain, which pairs with pre-TCRα to signal in
177 an antigen-independent manner at the ‘β-selection’ checkpoint (30). This occurs during
178 the DN3 stage of development; pre-selection DN3a thymocytes are CD25hiCD44lo and
179 forward scatter (FSC)-low, while DN3b thymocytes that have traversed this checkpoint
180 successfully express the same surface markers but are larger (FSC-high) (31)
181 (Supplementary Figure 2B). We probed β-selection in both 1:1 and 1:5 DKO:WT
182 competitive chimeras, but identified no advantage for either DKO or WT CD45.2 cells
183 relative to competitor CD45.1/2 WT cells (Figure 2C).
184
8 185 DKO thymocytes have a profound cell-intrinsic defect in negative selection
186 Studies of two independent NUR77-eGFP reporter lines as well as transcriptional
187 analysis have shown that Nr4a genes are upregulated at the positive selection
188 checkpoint and are especially enriched among thymocytes undergoing negative
189 selection (14, 18, 32). Overexpression of full-length and truncated dominant-negative
190 constructs suggested that NUR77 and NOR-1 redundantly mediate thymic negative
191 selection (13, 22, 33, 34), yet Nr4a1–/– mice exhibit only subtle defects (23, 24). We
192 reasoned that DKO:WT competitive chimeras could unmask cell-intrinsic, redundant
193 functions of the NR4A family during thymic selection. Indeed, we observe a striking
194 advantage for DKO cells in CD4SP and CD8SP subsets relative to DP in both 1:1 and
195 1:5 chimeras but not at an earlier stage (Figures 2D, E), suggesting either enhanced
196 positive selection or impaired negative selection. However, we do not see an advantage
197 for DKO cells in post-selection DP thymocytes relative to pre-selection DP thymocytes,
198 arguing against a role during positive selection (Supplementary Figures 2C, D).
199 To test the hypothesis that DKO thymocytes escape negative selection, we
200 assessed antigen-induced apoptosis by detection of activated Caspase 3 (aCasp3) in
201 thymocytes from chimeras. We observed reduced aCasp3+ DKO relative to WT
202 thymocytes after in vitro TCR-stimulation (Figures 2F, G, I, J). By contrast, we saw no
203 difference between donors in control WT:WT chimeras (Figures 2H, K). Notably, we
204 also saw no significant difference in aCasp3 expression in SKO thymocytes relative to
205 co-cultured WT (Supplementary Figures 2E, F). Moreover, mixed chimeras generated
206 with either Nr4a1–/– or Nr4a3–/– SKO mice revealed only a small competitive advantage
207 for CD8SP cells, suggesting a largely redundant role for these family members during
9 208 negative selection that is only unmasked when both family members are lost (35)
209 (Supplementary Figure 2G). We conclude that DKO thymocytes escape negative
210 selection, and show for the first time that this is a profound effect in a physiological
211 context, independent of either a TCR Tg or NR4A misexpression.
212
213 Myeloproliferative disorder in DKO mice is a non-cell-autonomous effect of NR4A-
214 deficiency
215 Previous studies report that a severe myeloproliferative disorder develops in the
216 first weeks of life in independently generated gDKO mice (25). This was not seen in
217 SKO animals lacking only one Nr4a family member, although mice lacking three out of
218 four Nr4a alleles (i.e. Nr4a1+/–Nr4a3–/– or Nr4a1–/–Nr4a3+/–) did eventually succumb to a
219 similar disease at much later time points (36). Consistent with this, we observed
220 profound expansion of side-scatter (SSC)-high cells infiltrating all hematopoietic tissues
221 and lymphoid organs in gDKO mice; this included not only BM and spleen
222 (Supplementary Figure 3), but was especially pronounced in lymph nodes and thymus
223 (Figures 3A-D). These SSChi cells are CD11b+ but largely Gr1–. Since Treg-deficient
224 animal models like Scurfy and Foxp3-deficient mice similarly exhibit myeloid expansion
225 (2, 3, 37), we hypothesized that the myeloproliferative disorder observed in gDKO
226 animals was due to Treg deficiency. Consistent with this possibility, myeloid expansion
227 is observed in CD4-cre Nr4a1fl/fl Nr4a3–/– mice, but not in mixed chimeras generated
228 with WT donor BM (6). Resolving this question with gDKO cells has important
229 implications since it has been argued that the NR4A family may represent important
230 drug targets in myeloid leukemic diseases (8, 25, 36). Indeed, in our DKO:WT chimeras,
10 231 myeloid expansion is suppressed (even after 12 weeks of reconstitution) and DKO cells
232 exhibit no competitive advantage in these compartments (Figures 3A-I,
233 Supplementary Figure 3A, B, D, E). We did observe a minor infiltration of SSChi
234 CD11b+ cells into the thymus of 1:1 DKO:WT but not WT:WT chimeras, and here as
235 well the effect of Nr4a-deficiency was cell-extrinsic (Figures 3D, H, I). Taken together,
236 these data support our hypothesis that the myeloproliferative disorder observed in
237 gDKO animals is due to a non-cell-autonomous impact of Nr4a deletion.
238
239 Abnormal B cell homeostasis in DKO mice is a non-cell-autonomous effect of
240 NR4A-deficiency
241 Like other Treg deficient models, gDKO mice exhibit spontaneous polyclonal B
242 cell activation and differentiation under steady-state conditions (Figures 4A, B,
243 Supplementary Figures 4A-F). We recently identified a cell-intrinsic role for the NR4A
244 family in restraining Ag-induced B cell expansion in the absence of co-stimulation,
245 including in the context of B cell tolerance (15, 20). We therefore sought to determine to
246 what extent spontaneous B cell activation and differentiation in gDKO mice (under
247 homeostatic conditions) were attributable to a B cell-intrinsic role for the NR4A family.
248 Chimeras did not reveal a competitive advantage or disadvantage for DKO cells
249 during splenic B cell development apart from a subtle disadvantage in the MZ
250 compartment (Supplementary Figure 4G) (20). B cells in 1:1 DKO:WT chimeras
251 expressed higher levels of activation markers than B cells in WT:WT chimera (Figure
252 4C, Supplementary Figures 4H, I). However, this did not differ between donor
253 genotypes within individual chimeras, suggesting a B cell-extrinsic effect of NR4A-
11 254 deficiency. We observed expansion of GC B cells and CD138+ cells in DKO:WT
255 chimeras relative to WT:WT control chimeras, but this was similarly B cell-extrinsic
256 (Figures 4D-I). Indeed, no expansion of the GC or CD138+ compartment was evident
257 under steady-state conditions in mice lacking Nr4a1/3 exclusively in the B cell
258 compartment (mb1-cre Nr4a1fl/fl Nr4a3–/–), even when aged to 40 weeks (Figures 4J-
259 M). Nor could we detect an advantage for mb1-cre DKO B cells in a competitive setting
260 (Supplementary Figures 4J-L). We conclude that there is evidence of a spontaneous
261 polyclonal B cell activation and differentiation in gDKO chimeras, but it is a B cell-non-
262 autonomous effect of NR4A-deficiency.
263
264 Reconstitution of WT Treg in competitive chimeras does not rescue DKO CD8+ T
265 cell homeostasis
266 Progressive thymic atrophy (Supplementary Figure 2A) and spontaneous B
267 activation and differentiation in DKO:WT chimeras (Figure 4) suggested the
268 development of a systemic autoimmune and inflammatory state despite replete and
269 largely WT Treg compartment. We next probed the mature T cell compartment to
270 understand the source of this immune dysregulation. gDKO mice exhibit expanded
271 effector-memory compartment and nearly complete loss of naive CD8+ T cells (Figures
272 5A, B, Supplementary Figures 5A, B). However, despite reconstitution of a replete
273 Treg compartment of WT origin (Figure 1M), DKO:WT chimeras nevertheless exhibit
274 marked accumulation of CD44hiCD8+ T cells relative to WT:WT control chimeras
275 (Figures 5C, D), and moreover, DKO T cells accumulate in this compartment (Figure
276 5E). In addition, these CD44hiCD8+ DKO T cells upregulate PD-1 expression,
12 277 suggesting an exhausted state (Supplementary Figures 5C, D). These observations
278 reveal a T cell-intrinsic role for the NR4A family in CD8+ T cell homeostasis.
279
280 Abnormal DKO CD8+ T cell homeostasis in competitive chimeras is due to a cell-
281 intrinsic role for Nr4a1 and Nr4a3 during thymic development
282 To test whether abnormal DKO CD8+ T cell homeostasis reflects a requirement
283 for the NR4A family during thymic selection or exclusively in the periphery, we took
284 advantage of a CD8-cre construct driven by the E8I enhancer that expresses
285 specifically in mature CD8SP and peripheral CD8+ T cells in order to generate CD8-cre
286 Nr4a1fl/fl Nr4a3–/– mice (CD8-cre cDKO) (38) (Figures 5F, G). We can confirm that this
287 cre is not active until after thymic DP stage and positive selection checkpoint are
288 traversed, because NUR77 expression in the mature CD4 lineage of CD8-cre cDKO
289 mice is intact (Figures 5F, G). Since accumulation of CD44hiCD8+ T cells was not
290 observed in CD8-cre cDKO mice (Figure 5H, Supplementary Figure 5E), we conclude
291 that this phenotype must be attributable to a role for the NR4A family earlier in
292 development and likely reflects escape of self-reactive CD8+ T cells into the periphery
293 due to impaired negative selection.
294
295 Cell-intrinsic accumulation of CD4+ DKO T cells with anergic phenotype in
296 competitive chimeras
297 gDKO mice exhibit an expanded CD4+ T cell effector-memory compartment that
298 is not evident in DKO:WT chimeras (Figures 6A-C). However, CD4+ T cell homeostasis
299 is not restored in these chimeras; rather DKO CD4+ T cells accumulate in the CD44hi
13 300 (memory) compartment and upregulate well-established markers of anergy (CD73 and
301 FR4) in a cell-intrinsic manner (Figures 6D-G) (39). Similarly, expansion of anergic
302 CD4+ T cells was exaggerated in 1:1 DKO:WT chimeras relative to both control
303 chimeras and SKO mice (Supplementary Figures 6A-H). Moreover, even
304 phenotypically ‘naive’ DKO CD44loCD62Lhi CD4+ T cells in mixed chimeras upregulated
305 CD73 and FR4, suggestive of antigen encounter (Figure 6D, E, Supplementary
306 Figures 6A, C, F) (19). Taken together, these data are consistent with escape of self-
307 reactive DKO CD4+ T cells from negative selection in the thymus (Figure 2) and
308 acquisition of an anergic phenotype in the periphery.
309
310 Impaired TCR signaling in anergic DKO T cells from competitive chimeras
311 Canonical functional features of anergic T cells include defective proximal TCR
312 signal transduction and impaired IL-2 production (40, 41). Therefore, we next assessed
313 TCR-induced Erk phosphorylation in T cells from DKO chimeras via a well-established
314 flow-based assay. Since anergic surface markers were largely preserved after TCR
315 stimulation and methanol permeabilization, we were able to gate cells according to
316 CD73 and FR4 expression, and on this basis defined cells as non-anergic, intermediate
317 anergic, or anergic (Supplementary Figure 7A). Consistent with this surface
318 phenotype, we observed progressively impaired TCR-induced Erk phosphorylation of
319 WT memory CD4+ T cells across these populations (Figures 7A, B). Within each gate,
320 DKO CD4+ T cells were even more refractory to TCR stimulation than WT cells from the
321 same chimera. Strikingly, unlike cells of WT origin, naive anergic DKO T cells were as
322 refractory as memory anergic T cells. These data suggest that DKO CD4+ T cells not
14 323 only upregulate markers of anergy (Figure 6), but acquired functional features of anergy
324 and did so to an even greater extent than WT.
325 Since we observed the accumulation of DKO CD44hiCD8+ T cells with increased
326 PD-1 expression in DKO:WT chimeras (Figure 5, Supplementary Figures 5C, D), we
327 utilized the same approach to assess functional characteristics of DKO CD8+ T cells.
328 We found that DKO CD44hiCD8+ T cells exhibit impaired Erk phosphorylation relative to
329 WT cells within the same chimera (Supplementary Figures 7B-D).
330 Of note, Erk phosphorylation downstream of PMA stimulation was intact in both
331 genotypes across all gated populations, suggesting a proximal rather than distal defect
332 in TCR signaling among ‘tolerant’ T cells (Figure 7A, Supplementary Figure 7C).
333 Importantly, TCR-induced Erk phosphorylation was robust in naive / non-anergic DKO T
334 cells (Figures 7A, B, Supplementary Figures 7C, D), implying that defective signal
335 transduction was an acquired feature of tolerant T cells. Collectively, these data suggest
336 that self-reactive DKO T cells escape negative selection and acquire both phenotypic
337 and functional features of antigen experience.
338
339 Upregulation of anergy-associated genes in DKO CD4+ T cells
340 To define the transcriptome of anergic DKO CD4+ T cells using an unbiased
341 approach, we undertook RNA-sequencing of DKO and WT cells sorted from 1:5
342 chimeras. We gated on CD4+CD25– cells to exclude a large fraction of FOXP3+ Treg
343 and sorted CD44hiCD62LloCD73hiFR4hi “anergic” cells (Supplementary Figure 7E). In
344 parallel, we also sorted naive CD4+ T cells from each genotype for analysis ex vivo and
345 following 3h in vitro TCR stimulation. Since we observed the expansion of anergic cells
15 346 in the naive CD4+ T cell compartment, we gated on CD73loFR4lo within the
347 CD44loCD62Lhi gate to collect the least self-reactive naive cells. Indeed, sorted “naive”
348 DKO cells do not express anergy-associated genes ex vivo (Supplementary Figure
349 7F, Figure 7C). By contrast, “anergic” DKO cells exhibited pronounced upregulation of
350 a subset of anergy-related genes relative to “anergic” WT cells (Figure 7C). This may
351 reflect a high degree of self-reactivity in this compartment due to escape from negative
352 selection. Importantly, although gene expression diverges between DKO and WT
353 anergic cells, much of their transcriptome is shared (Supplementary Figures 7F, G).
354 These include negative regulators such as Lag3, Pdcd1/PD-1, Rnf128/GRAIL and
355 Spry1 that serve to suppress proximal TCR signaling and may account, at least in part,
356 for the defect in ERK phosphorylation we observed in DKO cells (Figure 7C). Indeed,
357 we confirmed upregulation of PD-1 and LAG3 expression in DKO anergic cells by flow
358 staining (Supplementary Figure 7H, I). By contrast, Treg-associated genes such as
359 Foxp3, Ikzf4/Eos and Lrrc32/GARP are not expressed in anergic DKO cells (Figure
360 7D), consistent with previous studies revealing an essential role for the NR4A family in
361 induction and maintenance of Treg fate (6, 7, 28, 42). Further validating our data set,
362 gene set enrichment analysis (GSEA) revealed that genes repressed by overexpression
363 of Nr4a1 (17) are enriched in DKO anergic cells (Figure 7E), and conversely, genes
364 upregulated by Nr4a1 overexpression (17) are enriched in WT anergic cells (Figure 7F)
365 (9, 17). Strikingly, a subset of genes encoding pro-inflammatory cytokines and
366 mediators is upregulated in anergic DKO cells (Figure 7G), including Th1-related genes
367 (Tbx21 and Ifng) and Th2-related genes (Gata3 and Il4). This is consistent with prior
368 reports identifying a role for the NR4A family in the repression of Th1 and Th2 cell
16 369 differentiation (7, 42) and may contribute to immune dysregulation observed in DKO:WT
370 chimeras.
371
372 Cell-intrinsic defect in peripheral CD4+ DKO T cell tolerance
373 In order to define how such dysregulated gene expression may arise and also to
374 control for self-reactivity of the DKO T cell repertoire, we next sought to identify the
375 immediate targets of the NR4A family in sorted CD73–FR4– “naive” CD4+ T cells
376 (Supplementary Figures 7E). We compared the expression of primary response genes
377 (PRGs) in sorted naive WT and DKO cells following TCR stimulation (Supplementary
378 Figure 8A). We selected an early 3 hour time point in order to capture peak NR4A
379 protein induction and to enrich for direct transcriptional targets (18, 43). Principle
380 component analysis (PCA) segregated DKO from WT cells following acute TCR
381 stimulation, though less robustly than for DKO and WT anergic cells (Supplementary
382 Figures 8B, 7G). We focused our attention on PRGs that were differentially expressed
383 between WT and DKO cells (Figure 8A). Among those under-induced in DKO cells, we
384 identified Bcl2l11/BIM, and negative regulators of TCR signaling including Cblb, Dusp4,
385 and Tnfaip3/A20. Conversely, we observed over-induction of inflammatory mediators
386 such as Ccl4, Il2 and Tnf in DKO cells. We found that genes downregulated by Nr4a1
387 over-expression are highly enriched in acutely TCR-stimulated naive DKO cells, while
388 the opposite is true for genes upregulated by Nr4a1 over-expression (Figures 8B, C).
389 Impaired IL-2 production is among the most characteristic features of anergic T
390 cells, while exogenous IL-2 can override anergy in some settings (40, 41), suggesting
391 its dysregulation in DKO T cells may contribute to disruption of T cell tolerance in our
17 392 chimeras. Indeed, Il2 has been previously implicated as a target of the NR4A family in
393 the context of anergy and exhaustion (9, 17). We first assessed secreted IL-2 in culture
394 supernatants of Nr4a1–/– or Nr4a3–/– SKO CD4+ T cells across a broad titration of TCR
395 stimulation (Figure 9A). Both SKO genotypes secrete higher amounts of IL-2 compared
396 to WT, suggesting an additive role for the NR4A family in IL-2 regulation. We next
397 sought to assess IL-2 responses by DKO T cells. To do so, we cultured T cells from
398 DKO:WT chimeras with anti-CD3 and then assessed the capacity for IL-2 production
399 following maximal restimulation with PMA/ionomycin. We observed that, after TCR
400 stimulation, DKO CD4+ T cells acquire a much higher capacity for IL-2 production
401 relative to WT, and this is cell-intrinsic (Figures 9B, C). This result is not due to Treg
402 deficiency in DKO compartment (Supplementary Figure 8C). Nr4a1–/– or Nr4a3–/– SKO
403 CD4+ T cells each exhibit a less robust but independent increase in capacity for IL-2
404 production relative to WT (Supplementary Figure 8D). Moreover, Il2 transcript is
405 upregulated in anergic DKO cells relative to anergic WT cells directly ex vivo (Figure
406 9D). These data suggest that the role of the NR4A family in restraining IL-2 production
407 is not completely redundant, but rather additive, and impacts naive as well as anergic
408 CD4+ T cells.
409 DKO CD8+ T cells also exhibited a much higher capacity for IL-2 production than
410 WT cells from the same mixed chimera (Supplementary Figures 8E, F). Furthermore,
411 CD8+ T cells from CD8-cre cDKO mice exhibit a nearly identical phenotype that is more
412 robust than in SKO CD8+ T cells from Nr4a3–/– or CD8-cre Nr4a1fl/fl mice
413 (Supplementary Figure 8G). These data suggest that the NR4A family negatively
18 414 regulate the IL-2 locus in peripheral CD8+ T cells in a manner that is additive and cell-
415 intrinsic, and independent of self-reactivity.
416
417 Restoring WT Treg compartment in competitive chimeras alters autoantibody
418 repertoire but does not suppress autoimmunity
419 gDKO mice exhibit spontaneous, early-onset development of autoantibodies
420 (Figures 10A, B). Indirect immunofluorescence assay (IFA) for autoantibodies revealed
421 both nuclear and cytosolic staining suggesting a widespread loss of B cell tolerance that
422 occurs with complete penetrance before 4 weeks of age, recapitulating observations in
423 Treg-deficient mice (4). It is possible that this is attributable in part to loss of T follicular
424 regulatory (Tfr) cells in gDKO as seen in other Treg-deficient mice (44). Indeed, we
425 identify a profound cell-intrinsic defect for DKO cells in Tfr (but not Tfh) compartments in
426 chimeras (Supplementary Figures 9A, B) (7, 45). Although older Nr4a3–/– (but not
427 Nr4a1–/–) mice exhibit very low titer autoantibodies with a similar pattern
428 (Supplementary Figures 9C, D), B cell tolerance is largely preserved in SKO mice. To
429 our surprise, despite reconstitution of the Treg (and Tfr) compartment in DKO:WT
430 chimeras with cells of WT origin, we nevertheless observed the development of high
431 titer autoantibodies even at early time points after reconstitution (Figures 10C, D).
432 Cytosolic staining by autoantibodies was largely eliminated, but anti-nuclear
433 autoantibodies persisted in both 1:1 and even 1:5 DKO:WT chimeras (Figures 10E, F,
434 Supplementary Figures 9E, F). This suggests that the development of autoimmunity in
435 DKO chimeras was not attributable to a residual or partial Treg defect. The titer of ANA
436 in DKO:WT chimeras increased with age (Figure 10G, Supplementary Figure 9E).
19 437 This correlated with progressive accumulation of anergic CD4+ T cells, thymic atrophy,
438 and development of polyclonal B cell activation and spontaneous GC expansion,
439 suggestive of evolving immune dysregulation in these chimeras. By contrast, mice in
440 which B cells conditionally lack both Nr4a1 and Nr4a3 (mb1-cre cDKO) did not develop
441 ANA even after 40 weeks (Supplementary Figures 9G, H). These data imply that ANA
442 in DKO:WT chimeras are not attributable to a B cell-intrinsic role for the NR4A family.
443 Rather, we propose that - although reconstituting a WT Treg compartment suppresses
444 lethal immune dysregulation in DKO chimeras - tolerance is not fully restored, and this
445 may be due to a profound defect in both negative selection and peripheral T cell
446 tolerance.
447
448 DISCUSSION
449 A vital and redundant role for NR4A factors in the Treg compartment has made it
450 challenging to isolate and dissect other functions for this family in immune tolerance and
451 homeostasis (6, 7). Unfortunately, conditional genetic strategies alone cannot
452 disentangle requirements for the NR4A family during thymic selection from their obligate
453 function in Treg. Here we used competitive BM chimeras to reconstitute a functional
454 Treg compartment of WT origin, and this enabled us to unmask additional essential
455 roles for the NR4A family in the preservation of both central and peripheral T cell
456 tolerance under homeostatic conditions.
457 We confirmed a cell-intrinsic requirement for Nr4a1 and Nr4a3 in the Treg
458 compartment as previously reported with CD4-cre conditional DKO and TKO mice (6, 7,
459 28, 42). Concurrently, we also observed expansion of DKO CD25+ FOXP3– cells both in
20 460 gDKO mice and DKO chimeras. Sekiya and colleagues propose that this compartment
461 contains highly self-reactive T cells that failed to assume Treg fate and yet escaped
462 censorship by negative selection (28). Formal fate-mapping studies will be important to
463 test this hypothesis. This population of cells may contribute to immune dysregulation in
464 DKO chimeras, but importantly, cells diverted from the Treg fate cannot account
465 numerically for excess DKO SP thymocytes that escape negative selection in mixed
466 chimeras, especially since equal or greater advantage for DKO cells relative to WT is
467 observed in CD8SP thymocytes relative to CD4SP.
468 By contrast, we showed that myeloid cell expansion in gDKO mice is not cell-
469 intrinsic because it is almost entirely suppressed in DKO chimeras, and is instead likely
470 attributable to loss of Treg as observed in other Treg-deficient mouse models (2, 3, 37).
471 In support of this hypothesis, CD4-cre TKO mice develop a similar myeloproliferative
472 disorder that is also rescued in competitive chimeras generated with mixtures of CD4-
473 cre TKO and WT donor BM (6). This result emphasizes the need to critically re-assess
474 the therapeutic potential of the NR4A family as drug targets in myeloproliferative
475 disorders.
476 One of the earliest functions identified for the NR4A family is an essential role in
477 antigen-induced cell death and during thymic negative selection, but these studies relied
478 on mis-expression of full length and truncated NR4A Tg constructs under the control of
479 the proximal Lck promoter (which is active early during the DN stage of thymic
480 development) (13, 21, 22, 33, 34). By contrast, studies of Nr4a1–/– mice have revealed
481 subtle phenotypes, consistent with redundancy among family members (23, 24, 46).
482 Here we were able to unmask the redundancy between Nr4a1 and Nr4a3 during thymic
21 483 negative selection in a physiological setting for the first time. It is estimated that six
484 times more thymocytes are negatively than positively selected in a given time frame
485 (47). Since both 1:1 and 1:5 DKO chimeras harbor 4- to 6-fold more CD4SP and
486 CD8SP DKO cells relative to WT cells when normalized to the pre-selection DP
487 compartment (Figure 2E), we propose that NR4A-dependent deletion may account for
488 most or all negative selection. Prior studies implicate the NR4A family in negative
489 selection by both ubiquitous and tissue-restricted antigens (TRA) (22, 24, 48-50).
490 Though our data does not directly distinguish between the two, the striking amplitude of
491 rescue seen in DKO chimeras suggests escape from negative selection by ubiquitous
492 self-antigens (proposed to account for 75% of all deletion) and possibly TRA as well, but
493 this remains to be determined (47, 51).
494 Caspase-3 is activated in thymocytes upon TCR stimulation and in the process of
495 negative selection (52). Reduced aCasp3 expression of in vitro stimulated DKO
496 thymocytes suggests caspase-dependent TCR-induced apoptosis is mediated, at least
497 in part, by the NR4A family. BIM/Bcl2l11, a member of Bcl-2 family that can promote
498 Caspase-3 activation, is also essential for thymic negative selection and may represent
499 a transcriptional target for Nr4a1 (24, 53). Although we find that Nr4a1 and Nr4a3
500 collectively promote Bcl2l11 transcription in naive CD4+ T cells (Figure 8A), it has also
501 been shown that NUR77 can promote apoptosis by directly binding BCL-2 in the
502 cytosol, inducing a conformational change that exposes its BH3 pro-apoptotic domain in
503 a manner that is independent of transcriptional activity of NUR77 (49, 54). It will be
504 important to define which effectors downstream of NR4As mediate negative selection in
505 vivo. It remains to be determined how additional instructional signals modulate NR4A
22 506 function to either promote Treg differentiation or, alternatively, drive deletion of self-
507 reactive thymocytes.
508 We observe the accumulation of DKO CD44hiCD8+ T cells in DKO competitive
509 chimeras, and this is eliminated in CD8-cre cDKO mice in which cre-mediated deletion
510 occurs only after thymic selection is complete (Figure 5). We also observed a marked
511 accumulation of CD4+ DKO T cells with transcriptional and functional features of anergy
512 (Figure 6, 7). We propose that these phenotypes reflect escape of self-reactive T cells
513 into the periphery due to a defect in thymic negative selection.
514 Recent work suggests that Nr4a1 is required for induction and/or maintenance of
515 CD4+ T cell anergy (17); over-expression of Nr4a1 drives upregulation of a subset of
516 anergy-related genes, whereas deletion of Nr4a1 prevents generation of functionally
517 tolerant T cells. Similarly, Nr4a TKO CAR T cells evade exhaustion and eliminate
518 tumors (9). Although DKO T cells acquire features of tolerance in chimeras, we
519 nevertheless observed the development of systemic immune dysregulation and ANA in
520 DKO chimeras despite reconstitution of a functional Treg compartment, suggesting a
521 residual defect in functional anergy. It remains to be determined if DKO T cells are also
522 resistant to Treg-mediated suppression. We identified upregulation of inflammatory
523 mediators (e.g., Il2 and Tnf) and impaired induction of negative regulators (e.g., Cblb,
524 Tnfaip3 and Bcl2l11/BIM) in naive DKO CD4+ T cells following acute TCR stimulation
525 (Figure 8). These and other transcriptional targets of the NR4A family may contribute to
526 impaired peripheral tolerance. Indeed, although suppression of IL-2 production is
527 among the most characteristic features of anergic T cells, we report enhanced capacity
528 for IL-2 production in SKO T cells (consistent with prior studies of Nr4a1–/– T cells (17)),
23 529 and much more so in DKO T cells (Figure 9). We suggest this reflects a role for the
530 NR4A family in epigenetic remodeling of the Il2 locus in response to TCR stimulation.
531 Indeed, NR4A transcription factors modulate chromatin structure in the setting of
532 chronic antigen engagement (9, 17), and interrogation of a recently published ATACseq
533 data set reveals differentially accessible regions of open chromatin (OCR) near the Il2
534 locus in Nr4a3–/– CD8+ T cells following 12 h TCR stimulation (GSE143513) (55). We
535 propose that self-reactive DKO T cells that have escaped negative selection, Treg
536 differentiation, and peripheral anergy accumulate in the periphery and drive ANA
537 production in DKO chimeras. It remains to be defined whether defective central or
538 peripheral tolerance (or both) are most relevant for the development of autoimmunity in
539 DKO chimeras, and whether specific Th subsets (such as Tfh) play a role (42).
540 Nearly complete redundancy between Nr4a1 and Nr4a3 are evident in Treg and
541 during negative selection; deletion of both family members is necessary to unmask
542 these roles. By contrast, regulation of B cell responses (20) and CD8+ T cell exhaustion
543 (9) by the NR4A family appear additive. Based on published work (17) and our
544 observations of the IL-2 module in SKO and DKO T cells, we speculate that regulation
545 of CD4+ T cell anergy is similarly additive, but this remains to be fully addressed.
546 Although expression of Nr4a2 is low in the T cell lineage under steady-state conditions,
547 we also cannot exclude the possibility that Nr4a2 compensates for and partially masks
548 some immune phenotypes in DKO cells, especially in the context of inflammatory
549 stimuli.
550 We propose that Nr4a1 and Nr4a3 regulate layered T cell tolerance mechanisms
551 to preserve immune homeostasis under steady-state conditions (see model,
24 552 Supplementary Figure 10). In addition, it is likely that NR4A factors also serve to
553 counter-regulate inflammatory stimuli and promote a return to homeostasis. Indeed,
554 negative feedback by NR4A restrain responses to LPS in myeloid cells (56) and to
555 antigen stimulation in B cells (20), and suppresses inflammation in immune-mediated
556 disease models (57). Although it remains unclear whether endogenous ligands regulate
557 NR4A function in vivo, small molecule NUR77 agonist (10) and antagonist (11)
558 compounds have been reported. Agonists might be useful to suppress autoimmunity
559 and maintain transplant tolerance. Antagonizing NUR77 and perhaps other NR4A family
560 members could have applications for cancer immunotherapy (9). Since redundancy
561 among NR4A family members is important in both negative selection and in Treg,
562 selectively targeting individual NR4A family members may allow modulation of antigen-
563 specific T and B cell responses without disrupting global immune homeostasis.
564 Conversely, our studies unmask Treg-independent and redundant roles for Nr4a1 and
565 Nr4a3 in maintaining T cell tolerance under homeostatic conditions, with important
566 implications for drug design.
567
568 MATERIALS AND METHODS
569 Mice. Nr4a1–/–, Nr4a1fl/fl, and Nr4a3–/– mice were previously described (6, 20, 23).
570 Nr4a1fl/fl were previously obtained from Catherine Hedrick (La Jolla Institute for
571 Immunology, La Jolla, CA) with permission from Pierre Chambon (University of
572 Strasbourg, Strasbourg, France) (6). Nr4a1–/– mice were obtained from The Jackson
573 Laboratory and this line is used throughout the manuscript exclusively as single
574 germline knockout comparator (23), Nr4a3–/– mice were generated in our laboratory as
25 575 previously described (20). CD8-cre and mb1-cre were obtained from The Jackson
576 Laboratory (38, 58). C57BL/6 mice were from The Jackson Laboratory and CD45.1+
577 BoyJ mice were from Charles River Laboratories. To generate germline DKO Nr4a1–/–
578 Nr4a3–/– mice, we bred Nr4a3–/– and Nr4a1fl/fl mice with germline recombination of the
579 loxp-flanked locus, and confirmed loss of exon 2 both by genomic DNA PCR and
580 transcript qPCR. All strains were fully backcrossed to C57BL/6 genetic background for
581 at least 6 generations. Mice of both sexes were used for experiments between the ages
582 of 3 and 10 weeks except for BM chimeras as described below.
583
584 Antibodies and Reagents.
585 Abs for surface markers: Abs to B220, CD3, CD4, CD8, CD11b, CD11c, CD19, CD21,
586 CD23, CD25, CD44, CD45.1, CD45.2, CD62L, CD69, CD73, CD86, CD93 (AA4.1),
587 CD138, CXCR5, Fas, FR4, γδTCR, GL7, Gr1, IgD, MHC-II, NK1.1, PD-1, and pNK
588 conjugated to fluorophores were used (BioLegend, eBiosciences, BD, or Tonbo). See
589 also Supplementary Material.
590 Abs for intra-cellular staining: FOXP3 Ab conjugated to APC or FITC (clone FJK-16s,
591 Invitrogen). Anti-active Caspase-3 (aCasp3) Ab conjugated to APC (clone C92-605, BD
592 Pharmingen). Anti-NUR77 conjugated to PE (clone 12.14, Invitrogen). Anti-IL-2 Ab
593 conjugated to PE (clone JES6-5H4, Invitrogen). Anti-pERK (Phospho-p44/42 MAPK
594 (T202/Y204) (clone 197G2, Cell Signaling) Rabbit Ab. Goat Anti-Rabbit IgG (H+L)
595 conjugated to APC (Jackson ImmunoResearch).
596 Stimulatory Abs: Anti-CD3 (clone 2c11) and anti-CD28 (clone 37.51) (BioLegend).
597 Goat anti-Armenian Hamster antibody (Jackson ImmunoResearch).
26 598 ELISA reagents: 96-well, high-binding, flat-bottom, half-area, clear polystyrene Costar
599 Assay Plate (Corning). Mouse anti-dsDNA IgG-specific ELISA Kit (Alpha diagnostic.)
600 Mouse IL-2 DuoSet ELISA and DuoSet ELISA Ancillary Reagent Kit 2 (R&D Systems).
601 Anti-nuclear antibody (ANA): NOVA LiteTM HEp-2 ANA Substrate Slide and mounting
602 medium (INOVA Diagnostics, Inc, #708100); FITC Donkey Anti-Mouse IgG (Jackson
603 ImmunoResearch).
604 Culture Media: RPMI-1640 + L-glutamine (Corning-Gibco), Penicillin Streptomycin L-
605 glutamine (Life Technologies), HEPES buffer [10mM] (Life Technologies), B-
606 Mercaptoethanol [55mM] (Gibco), Sodium Pyruvate [1mM] (Life Technologies), Non-
607 essential Amino acids (Life Technologies), 10% heat inactivated FBS (Omega
608 Scientific).
609
610 Flow Cytometry. Cells were analyzed on a Fortessa and sorted on Aria (Becton
611 Dickson). Data analysis was performed using FlowJo (v9.9.6 or v10.7.1) software
612 (Becton Dickson).
613
614 Intracellular staining to detect active Caspase 3. Following in vitro stimulation, cells
615 were permeabilized and stained with APC-aCasp3, according to the manufacturer’s
616 protocol (BD Cytofix/Cytoperm kit).
617
618 FOXP3 staining. FOXP3 staining was performed utilizing a FOXP3/transcription factor
619 buffer set (eBioscience) in conjunction with APC or FITC anti-FOXP3, as per
620 manufacturer’s instructions.
27 621
622 Intracellular staining to detect IL-2. Splenocytes were stimulated with plate-bound anti-
623 CD3 Ab for 20 hours followed by a 4-hour treatment with 20 ng/ml of phorbol myristate
624 acetate (PMA) (Sigma) and 1 μM of ionomycin (Calbiochem) and protein transport
625 inhibitor cocktail (eBioscience) per manufacturer’s protocol. Following in vitro
626 stimulation, cells were permeabilized and stained, according to the manufacturer’s
627 protocol (BD Cytofix/Cytoperm kit).
628
629 Intracellular staining to detect NUR77. Following 2-hour in vitro stimulation with 20 ng/ml
630 of PMA and 1 μM of ionomycin, cells were fixed in a final concentration of 4%
631 paraformaldehyde for 10 min, permeabilized at −20 °C with 100% methanol for 30 min
632 and, following washes and rehydration, stained with primary antibody for 60 min at
633 20 °C (room temperature).
634
635 Live/dead staining. LIVE/DEAD Fixable Near-IR Dead Cell Stain kit (Invitrogen).
636 Reagent was reconstituted in DMSO as per manufacturer’s instructions, diluted 1:1000
637 in PBS, and cells were stained at a concentration of 1 × 106 cells /100 μl on ice for 15
638 minutes.
639
640 In vitro T cell culture and stimulation. Flat bottom 96 well plates were coated with
641 varying doses of anti-CD3 with or without 2 mg/ml anti-CD28 at 4°C overnight.
642 Splenocytes, lymphocytes or thymocytes were harvested into single cell suspension.
643 Splenocytes were subjected to red cell lysis using ACK buffer. Cells were plated at a
28 644 concentration of 5 × 105 cells/100 μl complete RPMI media in antibody coated flat
645 bottom 96 well plates for varying time.
646
647 Bone marrow chimeras. Host mice were irradiated with two doses of 530 rads, 4 hours
648 apart, and injected on the same day IV with a total of 2 × 106 donor BM cells at varying
649 ratios (1:1 or 1:5 or without mixture, as noted). Chimeras were sacrificed 6-14 weeks
650 after irradiation for downstream analyses.
651
652 CD4+ T cell purification. CD4+ T cell purification was performed utilizing magnetic-
653 activated cell sorting (MACS) separation, per the manufacturer’s instructions. In brief,
654 pooled spleen and/or lymph nodes were prepared utilizing the CD4+ T Cell Isolation Kit
655 (Miltenyi Biotec) and purified by negative selection through an LS column (Miltenyi
656 Biotec). Purified CD4+ T cells were then subjected to in vitro culture.
657
658 Phospho-flow. Splenocytes were rested at 37°C in serum-free RPMI for 30 minutes.
659 Cells were then stimulated with 10 μg/ml of anti-CD3 (clone 2c11) for 30 seconds
660 followed by 50 μg/ml of anti-Armenian hamster crosslinking antibody for 2 minutes, or
661 PMA for 2 minutes. Stimulated cells were fixed with 2% paraformaldehyde and
662 permeabilized with methanol at -20°C overnight. Cells were then stained with surface
663 markers and pErk at 20°C.
664
665 RNA-sequencing. DKO and WT CD25–CD4+ T cells were sorted from competitive
666 chimeras to identify either naive (CD44loCD62LhiCD73loFR4lo) or anergic
667 (CD44hiCD62LloCD73hiFR4hi) populations. Cell populations were sorted directly into RLT
29 668 + 1% beta-mercaptoethanol (BME) buffer. In parallel, sorted naive CD4+ T samples
669 were stimulated ex vivo with 8 μg/ml of plate-bound anti-CD3 and 2 μg/ml of anti-CD28
670 for 3 h and lysed in RLT/BME buffer. Libraries were generated by Emory Integrated
671 Genomics Core (EIGC): RNA was isolated using the Quick-RNA MicroPrep kit
672 (Zymo,11-328M). 2000 cell equivalent of RNA was used as input to SMART-seq v4
673 Ultra Low Input cDNA Synthesis kit (Takara, 634888) and 200 pg of cDNA was used to
674 generate sequencing libraries with the NexteraXT kit (Illumina, FC-121-10300). Libraries
675 were pooled at equimolar ratios, and sequenced on the NovaSeq6000 with a PE100
676 configuration using a NovaSeq 6000 SP Reagent Kit. Fastq files were trimmed for
677 adapters and low quality base pairs using Fastp (59) then aligned to mouse genome
678 assembly mm10 using STAR (60). FeatureCounts (61) was used to obtain read count
679 data and a paired differential expression analysis comparing samples from the same
680 chimeras was performed with edgeR (62). Heatmaps and principal component analysis
681 plots were generated using the ClustVis online tool (https://biit.cs.ut.ee/clustvis/) (63).
682 Gene set enrichment analysis (GSEA) was performed with GSEA (v4.1.0) software (UC
683 San Diego and Broad institute) (64, 65). Fastq and TPM data are publicly available
684 (GEO accession number: GSE178782) and Supplementary Data 1 contains analyses.
685
686 ELISA for IL-2 detection. Purified lymph node CD4+ T cells were cultured on anti-
687 CD3/28 coated plate at 1 × 105 cells per well. Plates were spun and supernatants were
688 harvested after 24h or 48 h. IL-2 concentrations in supernatants were measured using a
689 commercial ELISA kit, per the manufacturer’s instructions (R&D Biosystems). In brief,
690 96-well plates were coated with 1 μg/ml of capture anti-IL-2 antibody. Supernatants
30 691 were diluted serially, and IL-2 was detected with detection anti-IL-2 antibody. ELISA
692 plates were developed with mixture of tetramethylbenzidine and peroxidase, then
693 stopped with 2 N sulfuric acid. Absorbance was measured at 450 nm using a
694 spectrophotometer (SpectraMax M5; Molecular Devices).
695
696 ELISA for serum anti-dsDNA. Serum was harvested from blood collected by lateral tail
697 vein sampling or cardiac puncture postmortem. Serum anti-dsDNA titer was measured
698 with a commercial ELISA kit, per the manufacturer’s instructions (Alpha diagnostic). In
699 brief, sera were added to plates coated with dsDNA. Anti-dsDNA titer was detected with
700 anti-IgG-HRP. ELISA plates were developed, and absorbance was measured as
701 described above.
702
703 Anti-nuclear antibody (ANA). Serum ANA was detected with NOVA LiteTM HEp-2 ANA
704 Substrate Slide as per manufacturer’s instructions except for using FITC-conjugated
705 donkey anti-mouse IgG secondary antibody. Images were captured with a Zeiss Axio
706 Imager M2 widefield fluorescence microscope. Images were processed with Zen Pro
707 (Zeiss). To measure titer, serum was serially diluted 2-fold from 1:40 to 1:1280. HEp-2
708 ANA slides were stained with diluted serum. Images were read by a rheumatologist in a
709 blinded manner and titer was determined as the detectable lowest dilution of each
710 sample.
711
712 Statistical analysis. Statistical analysis and graphs were generated using Prism v9
713 (GraphPad Software, Inc). Graphs show mean ± SEM unless otherwise stated.
714 Student’s unpaired or paired t-test was used to calculate the P values for all
31 715 comparisons of two groups, and correction for multiple comparisons across time points
716 or doses was then performed using the Holm–Šídák method. One-way or two-way
717 analysis of variance (ANOVA) with follow-up Tukey’s test or Dunnett's test were
718 performed when more than two groups were compared with one another. Fisher’s exact
719 test was used to compare the difference in proportions of two groups. *p<0.05,
720 **p<0.01, ***p<0.001, ****p<0.0001.
721
722 Study approval. All mice were housed in a specific pathogen-free facility at UCSF
723 according to the University and National Institutes of Health guidelines. The protocol for
724 use of mice was reviewed and approved by UCSF Institutional Animal Care Use
725 Committee (San Francisco, CA).
726
727 Author contributions
728 R.H., H.V.N., J.L.M. and J.Z. conceived of and designed the experiments. R.H., H.V.N.
729 and J.L.M performed the experiments. R.H., H.V.N., J.L.M., R.M. and J.Z. analyzed the
730 data. R.H. and J.Z. wrote the manuscript. R.H., H.V.N., J.L.M. and J.Z edited the
731 manuscript.
732
733 Acknowledgements
734 This study was supported in part by the Emory Integrated Genomics Core (EIGC),
735 which is subsidized by the Emory University School of Medicine. We thank Al Roque for
736 help with mouse husbandry. We thank Arthur Weiss, Jeremy Brooks, Trang Nguyen,
737 and Wan-Lin Lo for critical feedback.
32 738
739 References
740 1. Sakaguchi S. Naturally arising Foxp3-expressing CD25+CD4+ regulatory T cells
741 in immunological tolerance to self and non-self. Nat Immunol. 2005;6(4):345-52.
742 2. Fontenot JD, et al. Foxp3 programs the development and function of
743 CD4+CD25+ regulatory T cells. Nat Immunol. 2003;4(4):330-6.
744 3. Skuljec J, et al. Monocyte/macrophage lineage commitment and distribution are
745 affected by the lack of regulatory T cells in scurfy mice. Eur J Immunol.
746 2016;46(7):1656-68.
747 4. Aschermann S, et al. B cells are critical for autoimmune pathology in Scurfy
748 mice. Proc Natl Acad Sci U S A. 2013;110(47):19042-7.
749 5. ElTanbouly MA, and Noelle RJ. Rethinking peripheral T cell tolerance:
750 checkpoints across a T cell's journey. Nat Rev Immunol. 2021;21(4):257-67.
751 6. Sekiya T, et al. Nr4a receptors are essential for thymic regulatory T cell
752 development and immune homeostasis. Nat Immunol. 2013;14(3):230-7.
753 7. Sekiya T, et al. Suppression of Th2 and Tfh immune reactions by Nr4a receptors
754 in mature T reg cells. J Exp Med. 2015;212(10):1623-40.
755 8. Boudreaux SP, et al. Drug targeting of NR4A nuclear receptors for treatment of
756 acute myeloid leukemia. Leukemia. 2019;33(1):52-63.
757 9. Chen J, et al. NR4A transcription factors limit CAR T cell function in solid
758 tumours. Nature. 2019;567(7749):530-4.
759 10. Zhan Y, et al. Cytosporone B is an agonist for nuclear orphan receptor Nur77.
760 Nat Chem Biol. 2008;4(9):548-56.
33 761 11. Zhan YY, et al. The orphan nuclear receptor Nur77 regulates LKB1 localization
762 and activates AMPK. Nat Chem Biol. 2012;8(11):897-904.
763 12. Winoto A, and Littman DR. Nuclear hormone receptors in T lymphocytes. Cell.
764 2002;109 Suppl:S57-66.
765 13. Calnan BJ, et al. A role for the orphan steroid receptor Nur77 in apoptosis
766 accompanying antigen-induced negative selection. Immunity. 1995;3(3):273-82.
767 14. Zikherman J, et al. Endogenous antigen tunes the responsiveness of naive B
768 cells but not T cells. Nature. 2012;489(7414):160-4.
769 15. Tan C, et al. Nur77 Links Chronic Antigen Stimulation to B Cell Tolerance by
770 Restricting the Survival of Self-Reactive B Cells in the Periphery. J Immunol.
771 2019;202(10):2907-23.
772 16. Ashouri JF, et al. Reporters of TCR signaling identify arthritogenic T cells in
773 murine and human autoimmune arthritis. Proc Natl Acad Sci U S A.
774 2019;116(37):18517-27.
775 17. Liu X, et al. Genome-wide analysis identifies NR4A1 as a key mediator of T cell
776 dysfunction. Nature. 2019;567(7749):525-9.
777 18. Moran AE, et al. T cell receptor signal strength in Treg and iNKT cell
778 development demonstrated by a novel fluorescent reporter mouse. J Exp Med.
779 2011;208(6):1279-89.
780 19. Zinzow-Kramer WM, et al. Adaptation by naive CD4(+) T cells to self-antigen-
781 dependent TCR signaling induces functional heterogeneity and tolerance. Proc
782 Natl Acad Sci U S A. 2019;116(30):15160-9.
34 783 20. Tan C, et al. NR4A nuclear receptors restrain B cell responses to antigen when
784 second signals are absent or limiting. Nat Immunol. 2020;21(10):1267-79.
785 21. Zhou T, et al. Inhibition of Nur77/Nurr1 leads to inefficient clonal deletion of self-
786 reactive T cells. J Exp Med. 1996;183(4):1879-92.
787 22. Cheng LE, et al. Functional redundancy of the Nur77 and Nor-1 orphan steroid
788 receptors in T-cell apoptosis. EMBO J. 1997;16(8):1865-75.
789 23. Lee SL, et al. Unimpaired thymic and peripheral T cell death in mice lacking the
790 nuclear receptor NGFI-B (Nur77). Science. 1995;269(5223):532-5.
791 24. Fassett MS, et al. Nuclear receptor Nr4a1 modulates both regulatory T-cell
792 (Treg) differentiation and clonal deletion. Proc Natl Acad Sci U S A.
793 2012;109(10):3891-6.
794 25. Mullican SE, et al. Abrogation of nuclear receptors Nr4a3 and Nr4a1 leads to
795 development of acute myeloid leukemia. Nat Med. 2007;13(6):730-5.
796 26. Koenis DS, et al. Nur77 variants solely comprising the amino-terminal domain
797 activate hypoxia-inducible factor-1alpha and affect bone marrow homeostasis in
798 mice and humans. J Biol Chem. 2018;293(39):15070-83.
799 27. Lio CW, and Hsieh CS. A two-step process for thymic regulatory T cell
800 development. Immunity. 2008;28(1):100-11.
801 28. Sekiya T, et al. Nr4a Receptors Regulate Development and Death of Labile Treg
802 Precursors to Prevent Generation of Pathogenic Self-Reactive Cells. Cell Rep.
803 2018;24(6):1627-38 e6.
804 29. Brewer JA, et al. Thymocyte apoptosis induced by T cell activation is mediated
805 by glucocorticoids in vivo. J Immunol. 2002;169(4):1837-43.
35 806 30. Palacios EH, and Weiss A. Distinct roles for Syk and ZAP-70 during early
807 thymocyte development. J Exp Med. 2007;204(7):1703-15.
808 31. Taghon T, et al. Developmental and molecular characterization of emerging beta-
809 and gammadelta-selected pre-T cells in the adult mouse thymus. Immunity.
810 2006;24(1):53-64.
811 32. Baldwin TA, and Hogquist KA. Transcriptional analysis of clonal deletion in vivo.
812 J Immunol. 2007;179(2):837-44.
813 33. Liu ZG, et al. Apoptotic signals delivered through the T-cell receptor of a T-cell
814 hybrid require the immediate-early gene nur77. Nature. 1994;367(6460):281-4.
815 34. Woronicz JD, et al. Requirement for the orphan steroid receptor Nur77 in
816 apoptosis of T-cell hybridomas. Nature. 1994;367(6460):277-81.
817 35. Nowyhed HN, et al. The nuclear receptor nr4a1 controls CD8 T cell development
818 through transcriptional suppression of runx3. Sci Rep. 2015;5:9059.
819 36. Ramirez-Herrick AM, et al. Reduced NR4A gene dosage leads to mixed
820 myelodysplastic/myeloproliferative neoplasms in mice. Blood. 2011;117(9):2681-
821 90.
822 37. Clark LB, et al. Cellular and molecular characterization of the scurfy mouse
823 mutant. J Immunol. 1999;162(5):2546-54.
824 38. Maekawa Y, et al. Notch2 integrates signaling by the transcription factors RBP-J
825 and CREB1 to promote T cell cytotoxicity. Nat Immunol. 2008;9(10):1140-7.
826 39. Kalekar LA, et al. CD4(+) T cell anergy prevents autoimmunity and generates
827 regulatory T cell precursors. Nat Immunol. 2016;17(3):304-14.
36 828 40. Fathman CG, and Lineberry NB. Molecular mechanisms of CD4+ T-cell anergy.
829 Nat Rev Immunol. 2007;7(8):599-609.
830 41. Zheng Y, et al. Molecular regulation of T-cell anergy. EMBO Rep. 2008;9(1):50-5.
831 42. Sekiya T, et al. Regulation of peripheral Th/Treg differentiation and suppression
832 of airway inflammation by Nr4a transcription factors. iScience.
833 2021;24(3):102166.
834 43. Jennings E, et al. Nr4a1 and Nr4a3 Reporter Mice Are Differentially Sensitive to
835 T Cell Receptor Signal Strength and Duration. Cell Rep. 2020;33(5):108328.
836 44. Vinuesa CG, et al. Follicular Helper T Cells. Annu Rev Immunol. 2016;34:335-68.
837 45. Ma W, et al. Nuclear receptors of the NR4a family are not required for the
838 development and function of follicular T helper cells. Int Immunopharmacol.
839 2015;28(2):841-5.
840 46. Hu QN, and Baldwin TA. Differential roles for Bim and Nur77 in thymocyte clonal
841 deletion induced by ubiquitous self-antigen. J Immunol. 2015;194(6):2643-53.
842 47. Stritesky GL, et al. Murine thymic selection quantified using a unique method to
843 capture deleted T cells. Proc Natl Acad Sci U S A. 2013;110(12):4679-84.
844 48. Suen AY, and Baldwin TA. Proapoptotic protein Bim is differentially required
845 during thymic clonal deletion to ubiquitous versus tissue-restricted antigens. Proc
846 Natl Acad Sci U S A. 2012;109(3):893-8.
847 49. Thompson J, and Winoto A. During negative selection, Nur77 family proteins
848 translocate to mitochondria where they associate with Bcl-2 and expose its
849 proapoptotic BH3 domain. J Exp Med. 2008;205(5):1029-36.
37 850 50. Odagiu L, et al. Role of the Orphan Nuclear Receptor NR4A Family in T-Cell
851 Biology. Front Endocrinol (Lausanne). 2020;11:624122.
852 51. Klein L, et al. Central CD4(+) T cell tolerance: deletion versus regulatory T cell
853 differentiation. Nat Rev Immunol. 2019;19(1):7-18.
854 52. Alam A, et al. Specific activation of the cysteine protease CPP32 during the
855 negative selection of T cells in the thymus. J Exp Med. 1997;186(9):1503-12.
856 53. Bouillet P, et al. BH3-only Bcl-2 family member Bim is required for apoptosis of
857 autoreactive thymocytes. Nature. 2002;415(6874):922-6.
858 54. Lin B, et al. Conversion of Bcl-2 from protector to killer by interaction with nuclear
859 orphan receptor Nur77/TR3. Cell. 2004;116(4):527-40.
860 55. Odagiu L, et al. Early programming of CD8(+) T cell response by the orphan
861 nuclear receptor NR4A3. Proc Natl Acad Sci U S A. 2020;117(39):24392-402.
862 56. McEvoy C, et al. NR4A Receptors Differentially Regulate NF-kappaB Signaling in
863 Myeloid Cells. Front Immunol. 2017;8:7.
864 57. Liebmann M, et al. Nur77 serves as a molecular brake of the metabolic switch
865 during T cell activation to restrict autoimmunity. Proc Natl Acad Sci U S A.
866 2018;115(34):E8017-E26.
867 58. Hobeika E, et al. Testing gene function early in the B cell lineage in mb1-cre
868 mice. Proc Natl Acad Sci U S A. 2006;103(37):13789-94.
869 59. Chen S, et al. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics.
870 2018;34(17):i884-i90.
871 60. Dobin A, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics.
872 2013;29(1):15-21.
38 873 61. Liao Y, et al. featureCounts: an efficient general purpose program for assigning
874 sequence reads to genomic features. Bioinformatics. 2014;30(7):923-30.
875 62. Robinson MD, et al. edgeR: a Bioconductor package for differential expression
876 analysis of digital gene expression data. Bioinformatics. 2010;26(1):139-40.
877 63. Metsalu T, and Vilo J. ClustVis: a web tool for visualizing clustering of
878 multivariate data using Principal Component Analysis and heatmap. Nucleic
879 Acids Res. 2015;43(W1):W566-70.
880 64. Subramanian A, et al. Gene set enrichment analysis: a knowledge-based
881 approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U
882 S A. 2005;102(43):15545-50.
883 65. Mootha VK, et al. PGC-1alpha-responsive genes involved in oxidative
884 phosphorylation are coordinately downregulated in human diabetes. Nat Genet.
885 2003;34(3):267-73.
886
887
888
889
890
891
892
893
894
895
39
FIGURE LEGENDS
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 to healthy littermate, 4 w;
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, 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. Schematics depict competitive BM chimera design.
H, I. Flow plots show thymic CD4SP (H) or splenic CD4+ T cell (I) sub-populations 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, L) or splenic (N, O) Treg, CD25+FOXP3– and
CD25–FOXP3– cells in 1:1 chimeras, normalized to DP thymocytes (n 3 (J-L) or 6 (M-
O), pooled from 2 sets of independently generated chimeras 6-10 w post≧ -transplant.
P, Q. Flow plots show thymic CD4SP (P) or splenic CD4+ T cell (Q) sub-populations in
1:5 DKO:WT chimera. Representative of 3 chimeras from one chimera setup.
Graphs depict mean +/- SEM. Statistical significance≧ was assessed by one-way ANOVA with Tukey’s test (C, D, K, L, N, O) or two-tailed unpaired Student’s t-test (J, M). *P <
0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. NS, not significant.
40
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 Supp. Fig. 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 6-7 w post-transplant chimeras pooled from 3 sets of independentl≧ y 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 h. 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 % aCasp3+ cells among DP (G, H) or CD4SP (J,
K) in 1:1 DKO:WT (G, J) or 1:1 WT:WT (H, K) chimeras (n = 3 from one chimera setup).
Graphs depict mean +/- SEM. Statistical significance was assessed by one-way (B) or two-way (C, E) ANOVA with Tukey’s test or two-tailed unpaired Student’s t-test with the
Holm–Sidak method (G, H, J, K). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
NS, not significant.
41
Figure 3. Myeloproliferative disorder in DKO mice is a non-cell-autonomous effect of NR4A deficiency
A-D. Lymph nodes cells (A, B) and thymocytes (C, D) from WT and gDKO mice (A, C) or 1:1 DKO:WT chimeras (B, D) were stained to detect CD11b and Gr1 expression.
Shown are representative plots of 5 mice.
E-H. Quantification of CD11b+Gr1–≧ cells in lymph nodes (E, F) and thymocytes (G, H) from WT, Nr4a3–/–, Nr4a1–/– and gDKO mice (E, 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, 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 one-way ANOVA with Tukey’s test (E, G), two-tailed unpaired Student’s t-test with (I) or without (F, H) the
Holm–Sidak method. **P < 0.01; ****P < 0.0001. NS, not significant.
42
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, 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 one chimera setup).
D, G. Representative flow plots show FAShiGL7+ GC B cells pre-gated on B220+IgDlo 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+IgDhi 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, 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 Nr4a1fl/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 one-way ANOVA with Tukey’s test (B) or Dunnett's test (E, F, H, I), or two-tailed unpaired Student’s t-test with (C) or without (K, M) the Holm–Sidak method. *P < 0.05; **P < 0.01; ****P <
0.0001. NS, not significant.
43
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 CD44hiCD8+ 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 CD44hiCD8+ T≧ cells from chimeras as gated in C at varied time points post-transplant (n 3).
E. Ratio of CD45.2 to CD45.1/2≧ for CD8+CD44hi population as gated in C, normalized to naive CD8+CD44loCD62Lhi gate (n 3). Data in C-E pooled from 2 sets of independently generated chimeras.≧
F, G. Thymocytes and splenocytes from CD8-cre and CD8-cre Nr4a1fl/fl Nr4a3–/– (cDKO) mice were stimulated with PMA and ionomycin (PMA/Io) for 2 h. Flow plots show intra- cellular 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+CD44hi T cells from CD8-cre, Nr4a3–/–, CD8-cre
Nr4a1fl/fl and CD8-cre cDKO mice (n = 3 mice/genotype).
44 Graphs depict mean +/- SEM. Statistical significance was assessed by one-way ANOVA with Tukey’s test (B, H) or two-way ANOVA with Dunnett's test (D, E). *P < 0.05; ***P <
0.001; ****P < 0.0001. NS, not significant.
45
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+ CD44hiCD62Llo 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+ CD44hiCD62Llo T cells as gated in Supp. Fig.
6A from competitive chimeras at indicated time points post-transplant (n 3, pooled from 2 sets of independently generated chimeras). ≧
D. Splenocytes from 12 weeks post-transplant DKO:WT = 1:5 chimera were stained to detect anergic T cell subsets. Flow plots depict CD73hiFR4hi (anergic) T cells within
CD44loCD62Lhi (naive), CD44hiCD62Lhi and CD44hiCD62Llo (memory) compartments of
CD4+FOXP3– cells of each donor genotype. Representative of 7 chimeras, generated in one set.
E-G. Ratio of CD45.2 to CD45.1/2 within CD73hiFR4hi gate among naive (E),
CD44hiCD62Lhi (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 post-transplant (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 one-way (B) ANOVA with Tukey’s test, two-way ANOVA with Dunnett's test (C) or two-tailed unpaired Student’s t-test with the Holm–Sidak method (E-G). *P < 0.05; **P <
0.01; ****P < 0.0001. NS, not significant.
46
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 crosslinking antibody for 2 minutes, or alternatively with
PMA for 2 minutes. Cells were fixed, permeabilized, and then stained to detect surface markers, FOXP3, and pErk. Representative histograms showing intra-cellular pErk expression in non-anergic (CD73loFR4lo; NA), int-anergic (CD73intFR4int ; IA) or anergic
(CD73hiFR4hi; A) among naive (CD44loCD62Lhi) or memory (CD44hiCD62Llo) CD4+ T cells gated as Supp Fig. 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 one chimera setup). Graphs depict mean +/- SEM.
Statistical significance was assessed by two-tailed unpaired Student’s t-test with the
Holm–Sidak method. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
C, D. Naive or anergic CD4+ T cells from CD45.1 (WT) or CD45.2 (DKO) cells gated as in Supp Fig. 7E were sorted directly into RLT buffer for RNA sequencing. ClustVis heatmaps depict expression of selected genes associated with anergy (C) or Treg (D).
E, F. GSEA plots for the genes downregulated (E) or upregulated (F) by Nr4a1 (Nature.
2019;567(7749):525-9.) against differentially expressed genes (DEGs) in DKO and WT anergic cells. DEGs were defined as genes upregulated in DKO compared to WT anergic cells with p < 0.05. NES = normalized enrichment score, FDR = false discovery rate.
G. Heatmap shows expression of selected inflammatory mediators.
47
Figure 8. Transcriptional targets of NR4A family in acutely stimulated naive CD4+
T cells
A. ClustVis heatmap shows overlap primary response genes (PRGs) and differentially expressed genes (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 Fig. 7.
B, C. GSEA plots for the genes downregulated (B) or upregulated (C) by Nr4a1 (Nature.
2019;567(7749):525-9.) against DEGs in DKO and WT stimulated naive CD4+ T cells.
NES = normalized enrichment score, FDR = false discovery rate.
48
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 h (left) or 48 h (right).
IL-2 concentration in supernatant was measured with ELISA (n = 3 biological replicates).
B. Lymph node cells from 10 weeks post-transplant DKO:WT = 1:1 chimera were cultured in plates coated with indicated doses of anti-CD3 for 20 h. Then cells were re- stimulated with PMA, ionomycin and brefeldin for additional 4 h. Representative histograms of 3 mice showing intracellular IL-2 in CD4+ cells of each donor genotype.
C. Quantification of %IL-2+ as described for C above (n = 3 biological replicates from one 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 two-way ANOVA with Tukey’s test (A), two-tailed unpaired Student’s t-test with the Holm–Sidak 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.
49
Figure 10. Restoring Treg compartment in competitive chimeras alters autoantibody repertoire but does not restore tolerance
A-D. Anti-nuclear antibody (ANA) immunofluorescence images. 1:40 diluted serum of indicated mice were applied to Hep-2 substrate slides, washed, and stained with FITC- anti-mouse IgG. Images are representative of biological replicates as quantified below in E, 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 includes analysis of serum from 2 sets of independently generated chimeras 6-12 w post-transplant. 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 two-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 post-transplant stained as in A-D.
50 Supplementary Figure 1 A B C 2 6000 Source: Immgen.org **** **** 5000 Nr4a1 *** *** **** ) 4000 ) **** 7 Nr4a2 7 6
15 0 3000 Nr4a3 0 2000 1000 10 4 500 400 5 2 300
200 0 0 Splenoytes cell # (x1 Thymocytes cell # (x1 100 –/– –/– –/– –/– WT WT 0 gDKO gDKO Nr4a3 Nr4a1 Nr4a3 Nr4a1 Expression value normalized by DESeq DN3 DN4 DP CD4SP CD8SP Naive Foxp3+ Naive CD4 Treg CD8
D CD4SP E F G WT Nr4a3–/– 5 5 10 3.5 10 6.0
4 4 10 10 ****
) **
*** ) 5 P **** 5 3 3 ** 0 * 0 **** 10 10 8 * 40 *** 4 2 2 10 10 0 2.0 0 3.3 6 CD4S 30 3
2 3 4 5 2 3 4 5 of 0 10 10 10 10 0 10 10 10 10 –
–/– cell # (x1 cell # (x1 – Nr4a1 gDKO 4 20 – 5 5 2 10 7.7 10 2.4 Thymic Splenic
4 4 Foxp3
10 10 +
Foxp3 2 10 Foxp3
+ 1 3 3 + 10 10
2 2 10 10 0 CD25 0 0
0 0 CD25 CD25 3.5 7.9 –/– –/– % –/– –/–
Foxp3-APC –/– –/– 2 3 4 5 2 3 4 5 0 10 10 10 10 0 10 10 10 10 WT WT WT gDKO gDKO gDKO CD25-PE Nr4a3 Nr4a1 Nr4a3 Nr4a1 Nr4a3 Nr4a1
H I J K ) ) 5 Thymocytes 5 P = 0.067 0 Splenocytes 0 20 P = 0.063 30 NS 2 2 10 * 10 ** 1:5 chimera 15 8 8 CD45.2 DKO 20 * 6 *** CD45.1/2 WT 10 6 ** 10 4 4 5 2 2 Thymic Treg cell # (x1 0 Splenic Treg cell # (x1 0
Ratio CD45.2:CD45.1/ 0 Ratio CD45.2:CD45.1/ 0 CD45.2 WT DKO DKO CD45.2 WT DKO DKO + – – + – – CD45.1/2 WT WT WT CD45.1/2 WT WT WT
Foxp3+ Foxp3– Foxp3 Foxp3+ Foxp3– Foxp3 Chimera 1:1 1:5 Chimera 1:1 1:5
CD25 CD25 CD25 CD25
L M N O ) ) ) 15,000 15,000 1,500 1,500 **** NS **** 10,000 10,000 1,000 1,000 reg CD25 (MFI reg Foxp3 (MFI NS T
T 5,000
5,000 500 500
0 0 Splenic Treg Foxp3 (MFI) Splenic Treg CD25 (MFI Splenic 0 0 Splenic CD45.1/2 CD45.2 CD45.1/2 CD45.2 CD45.1/2 CD45.2 CD45.1/2 CD45.2 WT WT WT DKO WT WT WT DKO 1:1 chimera 1:1 chimera 1:1 chimera 1:1 chimera
Supplementary Figure 1. Supporting data for Figure 1 A. Expression of Nr4a family mRNA in T cell subsets quantified via RNA-seq. Data are exported from Immgen database (http://rstats.immgen.org/Skyline/skyline.html). B, C. Quantification of viable thymocyte (B) and splenocyte (C) cell number from WT, Nr4a3–/–, Nr4a1–/– and Nr4a1–/– Nr4a3–/– (gDKO) mice. (Data in C-G include n ≥ 5 biological replicates/genotype, 3 to 4-week-old gDKO and 5 to 6-week-old mice with the other genotypes were analyzed). D. Representative flow plots showing thymic Treg gate in WT, Nr4a3–/–, Nr4a1–/– and Nr4a1–/– Nr4a3–/– (gDKO) mice, as determined by FOXP3 and CD25 expression. Plots are representative of at least 5 mice/genotype. E, F. Quantification of cell number (E) and frequency (F) of thymic CD4SP CD25+FOXP3– cells as gated in (D). G. Quantification of splenic CD4+CD25+FOXP3– cell number as gated in Fig. 1B (n ≥ 5 biological replicates, 3 to 4-week-old gDKO and 5 to 6-week-old mice with the other genotypes were analyzed). H, I. Ratio of CD45.2 to CD45.1/2 for Treg, CD25+FOXP3– cells and CD25–FOXP3– cells among thymic CD4SP (H) or splenic CD4+ (I) in DKO:WT = 1:5 chimera (n = 3 biological replicates from one chimera setup). Ratios were normalized to DP thymocytes. J, K. Quantification of thymic (J) and splenic (K) Treg cell number in WT:WT = 1:1, DKO:WT = 1:1 and 1:5 chimera at 6 weeks post-transplant (n = 3 biological replicates pooled from 2 sets of individually generated chimeras). L-O. Quantification of FOXP3 (L, M) or CD25 (N, O) expression (MFI) on splenic Treg cells of each donor genotype from WT:WT = 1:1 chimeras (L, N) and DKO:WT = 1:1 (M, O) determined via flow staining. Lines connect donor genotypes within an individual chimera (n = at least 6 biological replicates from one chimera setup).Graphs depict mean +/- SEM. Statistical significance was assessed by one-way ANOVA with Tukey’s test (B, C, E-K) or two-tailed paired Student’s t-test (L-O). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. NS, not significant. Supplementary Figure 2
A B ** NS
) NS 7 20 *** 1:1 chimera
0 – ** CD45.1/2 Thymocytes Size-gate Dump DN3 CD45.2 250K 5 5 **** 10 10 DN1 DN2 15 WT WT 2.1 5.0 50 200K 4 4 10 10 40 DKO WT 150K 3 3 30 10 10 10 1:5 chimera 100K 20 2 2 DN3a DN3b 10 10 50K DN4 DN3 10 80 20 CD45.2 CD45.1/2 79 0 3.0 0 73 18 5 SSC 0 Dump CD44-PB # Cells 0 2 3 4 5 DKO WT 0 50K 100K 150K 200K 250K 0 50K 100K 150K 200K 250K 0 10 10 10 10 0 50K 100K 150K 200K 250K FSC SSC CD25-APC FSC
Thymocyte cell # (x1 0 6w 10 - 12w
C Thymocytes DP D 5 5 10 Post-sel. 2 10 14 77 4 2.2 4 10 10 3 10 3 10
2 2 10 2.0 Post-sel. 1 1:1 chimera 0 4.6 0 89 7.2 NS CD4-PE 3 3 4 5 3 4 5 CD45.2 CD45.1/2 -10 0 10 10 10 0 10 10 10 1.5 NS Pre-sel. WT WT CD8-PB CD69-PE-Cy7 DKO WT 1.0 Pre-sel. Post-sel. 1 Post-sel. 2 1:5 chimera 5 5 5 0.5 10 12 89 10 13 86 10 13 86 CD45.2 CD45.1/2 4 4 4 10 10 10 DKO WT
Ratio CD45.2:CD45.1/ 0.0 3 3 3 10 10 10 Pre-sel. Post-sel. 1 Post-sel. 2 0 0 0 3 3 3 -10 -10 -10 CD45.2-PCPCy5.5 3 3 4 5 3 3 4 5 3 3 4 5 -10 0 10 10 10 -10 0 10 10 10 -10 0 10 10 10 CD45.1-APC
E F 50 NS ) 40 WT CD45.1 (co-cultured)
) NS NS NS NS WT CD45.2 40 NS NS 30 NS NS –/– NS NS Nr4a3 CD45.2 30 –/– NS NS NS NS NS NS Nr4a1 CD45.2 of DP (%
+ 20 NS
20 of CD4SP (% + 10 10 aCasp3
0 aCasp3 0 0 0.1 1 0 0.1 1 Anti-CD3 (µg/ml) Anti-CD3 (µg/ml)
G
2 4 1:1 chimera 3 CD45.2 CD45.1/2 WT WT –/– 2 Nr4a3 WT
1
Ratio CD45.2:CD45.1/ 0 DP CD4SP CD8SP
Supplementary Figure 2. Supporting data for Figure 2 A. Quantification of thymocyte cell number in WT:WT = 1:1, DKO:WT = 1:1 and 1:5 chimera at indicated time points post-transplant (n ≥ 3 biological replicates pooled from 3 sets of individually generated chimeras). B. Representative flow plots show gating strategy to identify pre- and post- b selection subsets DN3a/b in 1:1 DKO:WT chimera. Dump staining includes C. Representative flow plots show gating strategy to detect pre- and post-positive selection DP subsets in DKO:WT = 1:5 chimera. Plots are representative of 3 chimeras. D. Ratio of CD45.2 to CD45.1/2 of pre-selection, post-selection 1 and post-selection 2 DP thymocytes (as gated in C) in chimeras, normalized to pre-selection DP (n = 3 biological replicates pooled from 2 sets of individually generated chimeras). E, F. Thymocytes from CD45.2 WT, Nr4a3–/– and Nr4a1–/– were mixed with CD45.1 WT thymocytes in 1:1 ratio. Cells were co-cultured, stimulated and stained + cells among DP (E) or CD4SP (F) with indicated dose of anti-CD3 (n = 3 biological replicates). G. Ratio of CD45.2 to CD45.1/2 thymic subsets in WT:WT = 1:1 chimera and Nr4a3–/–:WT = 1:1 chimera, normalized to DP thymocytes (n = 2 biological replicates from one chimera setup). Graphs depict mean +/- SEM. Statistical significance was assessed by two-way ANOVA with Tukey’s test (A, D) or two-tailed unpaired Student’s t-test with the Holm–Šídák method (E, F). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. NS, not significant. Supplementary Figure 3
A Splenocytes B C Bone marrow WT gDKO DKO:WT = 1:1 WT gDKO 250K 250K 250K 250K 250K
200K 200K 200K 200K 200K
150K 150K 150K 150K 150K
100K 100K 100K 100K 100K
50K 50K 50K 50K 50K
SSC 0 0
SSC 0 SSC 0 0 0 50K 100K 150K 200K 250K 0 50K 100K 150K 200K 250K 0 50K 100K 150K 200K 250K 0 50K 100K 150K 200K 250K 0 50K 100K 150K 200K 250K FSC FSC FSC Live cells Live cells
5 5 5 5 10 10 105 10 10 2.6 1.5 2.1 21 11 4 4 4 4 10 10 104 10 10
3 3 3 3 10 10 103 10 10
2 2 2 2 10 10 102 10 10 0 6.9 0 13 0 5.7 0 13 0 51 Gr1-PE Gr1-PE Gr1-PE 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 0 10 10 10 10 0 10 10 10 10 0 102 103 104 105 0 10 10 10 10 0 10 10 10 10 CD11b-FITC CD11b-FITC CD11b-FITC
D E F
**** 50 *** 50 80 **** ** ****
*** – – 40 40 – 60 Gr1 Gr1 Gr1 + +
30 30 + 40 20 20 CD11b
%splenocyte NS CD11b CD11b %splenocyte 10 20
10 %bone marrow 0 0 0
–/– –/– CD45.2 WT DKO –/– –/– WT CD45.1/2 WT WT WT gDKO gDKO Nr4a3 Nr4a1 1:1 chimera Nr4a3 Nr4a1
Supplementary Figure 3. Expansion of myeloid cells in spleen and bone marrow of gDKO mice A-C. Splenocytes (A) and BM cells (C) from WT and gDKO mice, as well as splenocytes from DKO:WT = 1:1 chimera (B) were stained as described for Figure 3A-D. Shown are representative plots of at least 5 mice. D, F. Quantification of CD11b+Gr1– cell frequency in spleen (D) and bone marrow (F) from WT, Nr4a3–/–, Nr4a1–/– and gDKO mice (n ≥ 5 biological replicates, 3 to 4-week-old gDKO and 5 to 6-week-old mice with the other genotypes were analyzed). E. Quantification of CD11b+Gr1– cell frequency in spleen from WT:WT = 1:1 and DKO:WT = 1:1 chimera (n ≥ 3 biological replicates pooled from 2 sets of individually generated chimeras). Graphs depict mean +/- SEM. Statistical significance was assessed by one-way ANOVA with Tukey’s test (D, F) or two-tailed unpaired Student’s t-test (E). **P < 0.01; ***P < 0.001; ****P < 0.0001. NS, not significant. Supplementary Figure 4 A B C D + lo
NS s Splenic B220 IgD 2.0 800 40 **** NS ****
WT gDKO s **** ) 5 5 1.5 30 10 10 600
4 0.72 4 10 10 7.8 1.0 400 20 3 3 10 10 of total B cell
2 2 hi
10 10 % GC B cells CD86 (MFI of total B cell 0.5 10 0 0 200 FAS-PE-Cy7
2 3 4 5 2 3 4 5 0 10 10 10 10 0 10 10 10 10 0.0 0 %FAS 0 GL7-PCP-Cy5.5 –/– –/– –/– –/– –/– –/– WT WT WT gDKO gDKO Nr4a3 Nr4a1 gDKO Nr4a3 Nr4a1 Nr4a3 Nr4a1
E Splenocytes F G WT gDKO 0.06 * 2 2.5 ** 1:1 chimera
5 5 + 10 0.26 10 0.60 2.0 CD45.2 CD45.1/2 4 4 10 10 0.04 NS NS WT WT
3 3 /B220 1.5 10 10 + DKO WT 8
2 2 10 10 1.0 0 0 0.02 CD138-PE 2 3 4 5 2 3 4 5 CD13 0 10 10 10 10 0 10 10 10 10 0.5 B220-PB
0.00 Ratio CD45.2:CD45.1/ 0.0 T1 T2/3 Fo MZ –/– –/– WT gDKO Nr4a3 Nr4a1 H I ** *** 500 NS NS 100,000 NS NS )
) 400 80,000 300 60,000 200 40,000
CD86 (MFI 100 MHC II (MFI 20,000 0 0 CD45.2 WT DKO CD45.2 WT DKO CD45.1/2 WT WT CD45.1/2 WT WT 1:1 chimera 1:1 chimera
J Splenic K L B220+IgDlo GC B cells
5 5 10 10 46 51 7.7 4 4 NS 10 10 2.5 1.5
3 3 10 10 mb1-cre:WT s 2.0 * 2 2 chimera 10 10 1.0 0 0 1.5
2 3 4 5 2 3 4 5 0 10 10 10 10 0 10 10 10 10 5 5 1.0 10 10 30 67 0.5 7.7 4 4 (GC B cells) 10 10 %GC B cells of total B cell 0.5
3 3 10 10 mb1-cre cDKO:WT
0.0 Ratio CD45.2:CD45.1/2 0.0 2 2 10 10 chimera 0 0 CD45.2 mb1-cre cDKO CD45.2 mb1-cre cDKO FAS-PE-Cy7 CD45.2-PB 2 3 4 5 2 3 4 5 0 10 10 10 10 0 10 10 10 10 CD45.1/2 WT WT CD45.1/2 WT WT GL7-PCP-Cy5.5 CD45.1-FITC 1:1 chimera 1:1 chimera Supplementary Figure 4. Supporting data for Figure 4 A. Representative flow plots show FAShiGL7+ GC B cells from WT and gDKO mice as described for Figure 4D. Plots are representative of 6 mice. B. Quantification of frequency of GC B cells, as gated in (A) above, among total B cells from WT, Nr4a3–/–, Nr4a1–/– and gDKO mice (n ≥ 5 biological replicates). C. Quantification of CD86 MFI on splenic B cells from WT, Nr4a3–/–, Nr4a1–/– and gDKO mice (n ≥ 4 biological replicates). D. Quantification of FAShi cells among B220+ cells in spleen from WT, Nr4a3–/–, Nr4a1–/– and gDKO mice (n ≥ 5 biological replicates). E. Representative flow plots show CD138+ cells in splenocytes from WT and gDKO mice. Plots are representative of 4 mice. F. Quantification of ratio of CD138+ to B220+ cells, as gated in (E) above, from WT, Nr4a3–/–, Nr4a1–/– and gDKO mice (n ≥ 4 biological replicates). 3 to 4-week-old gDKO and 5 to 6-week-old mice with the other genotypes were analyzed (B, C, D, F). G. The ratio of CD45.2 to CD45.1/2 for transitional1 (T1), T2/3, follicular (Fo) and marginal zone (MZ) B cell subsets (gated on the basis of B220, CD21, CD23, and CD93) from WT:WT and DKO:WT = 1:1 chimera, normalized to T1 subset (n = 6 biological replicates from one chimera setup). H, I. Quantification of CD86 (H) and MHC II (I) MFI on splenic B cells of each donor genotype in in WT:WT and DKO:WT = 1:1 chimera (n = 3 biological replicates from one chimera setup). J. Representative flow plots showing FAShiGL7+ GC B cells in spleen from mb1-cre:WT = 1:1 chimera and mb1-cre Nr4a1fl/fl Nr4a3–/– (cDKO):WT = 1:1 chimera on the left. Right-hand plots depict CD45.2 cDKO donor and CD45.1/2 WT donor populations. Plots are representative of n = 4 biological replicates. K. Quantification of frequency of GC B cells among total B cells as described for J above. L. Ratio of CD45.2 to CD45.1/2 GC B cells in spleen from chimeras described in (J) above, normalized to B220+IgDhi cells (data in K, L represent n = 4 biological replicates from n = 2 independent experiments with one chimera setup). Graphs depict mean +/- SEM. Statistical significance was assessed by Brown-Forsythe ANOVA (B, F), one-way ANOVA with Tukey’s test (C, D) or two-tailed unpaired Student’s t-test with (G, H, I) or without (K, L) the Holm–Šídák method. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. NS, not significant. Supplementary Figure 5 A B
**** + **** 100 **** 8 **** 80 **** + **** 8 80 of CD 60 lo 60 40 40 CD62L hi 20 20 % naive of CD 0 0 %CD44 –/– –/– –/– –/– WT WT gDKO gDKO Nr4a3 Nr4a1 Nr4a3 Nr4a1
C D DKO:WT = 1:5 chimera CD8+ T cells Naive CD44hi CD45.1/2 WT
5 250K 250K 10 38 CD45.2 DKO 200K 200K 4 10 150K 150K 3 CD45.1/2 15 10 100K 0.34 100K 3.8 WT 2 ** 10 60 50K 50K 02 -10 Naive 0 0 + 3 3 4 5 3 4 5 3 4 5 10 -10 0 10 10 10 0 10 10 10 0 10 10 10
5 250K 250K 10 71
200K 200K %PD-1 4 10 5 150K 150K 3 CD45.2 10 NS 100K 0.55 100K 10 2 DKO 10 50K 50K 0 02 26
-10 SSC 0 0 CD44-BV421 hi 3 3 4 5 3 4 5 3 4 5 Naive CD44 -10 0 10 10 10 0 10 10 10 0 10 10 10 CD62L-BV711 PD-1-PCPCy5.5
E CD8+ T cells CD8-cre CD8-cre CD8-cre Nr4a3–/– Nr4a1fl/fl cDKO
5 5 5 5 10 7.4 16 10 10 16 10 11 9.2 10 6.3 19
4 4 4 4 10 10 10 10
3 3 3 3 10 10 10 10
0 0 0 0 3 3 3 3 -10 71 -10 59 -10 46 -10 64 CD44-PB 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 CD62L-PE
Supplementary Figure 5. Supporting data for Figure 5 A, B. Quantification of frequency of CD8+ naive CD44loCD62Lhi (A) and CD44hiCD62Llo (B) cells from WT, Nr4a3–/–, Nr4a1–/– and gDKO mice as gated in Fig. 5A (n ≥ 5 biological replicates, 3 to 4-week-old gDKO and 5 to 6-week-old mice with the other genotypes were analyzed). C. Splenocytes from DKO:WT = 1:5 chimera was stained to detect PD-1 expression on naive CD44loCD62Lhi and on CD44hi CD8+ cells among each donor genotype. Plots are representative of 6 mice. D. Quantification of %PD-1+ of naive and CD44hi CD8+ cells as gated in C above. Lines connect donor genotypes within an individual chimera (n = 6 biological replicates from one chimera setup). E. Splenocytes from CD8-cre, Nr4a3–/–, CD8-cre Nr4a1fl/fl and CD8-cre Nr4a1fl/fl Nr4a3–/– (cDKO) mice were stained as described for Fig. 5A to detect CD8+ T cell subsets. Plots are representative of 3 mice and correspond to quantification in Fig. 5H. Graphs depict mean +/- SEM. Statistical significance was assessed by one-way ANOVA with Tukey’s test (A, B) or two-tailed unpaired Student’s t-test with the Holm–Šídák method (D). **P < 0.01; ****P < 0.0001. NS, not significant. Supplementary Figure 6
A Splenic CD4+ Foxp3– Naive Memory 5 5 5 5 10 10 10 10 2.0 20 4 4 4 10 10 10 4 10 3 3 3 10 10 10 3 2 2 WT:WT = 1:1 chimera 2 10 14 10 10 10 0 0 0 1 10 3 75 2 2 88 -10 -10 -10 3 3 4 5 3 3 4 5 3 4 5 3 4 5 -10 0 10 10 10 -10 0 10 10 10 0 10 10 10 0 10 10 10
5 5 5 5 10 10 10 10 34 85 4 4 4 10 10 10 4 10 3 3 3 10 10 10 3 DKO:WT = 1:1 chimera 2 2 2 10 10 30 10 10 0 0 0 1 10 80 3 35 2 2 Foxp3-FITC CD44-BV421 -10 -10 -10 FR4-PE-Cy7 3 3 4 5 3 3 4 5 3 4 5 3 4 5 -10 0 10 10 10 -10 0 10 10 10 0 10 10 10 0 10 10 10 CD25-APC CD62L-BV711 CD73-BV605
B Splenic CD4+ Foxp3– Naive Memory 5 5 5 10 5 10 10 1.4 10 22 4 4 4 10 10 4 10 10 3 3 3 10 10 10 3 2 10 10 2 WT 2 11 10 10 0 76 0 0 1 2 10 3 -10 2 90 -10 -10 3 3 4 5 3 3 4 5 3 4 5 3 4 5 -10 0 10 10 10 -10 0 10 10 10 0 10 10 10 0 10 10 10
5 5 5 10 5 10 10 4.1 10 33 4 4 4 10 10 4 10 10 3 3 3 10 10 10 –/– 3 2 10 10 2 Nr4a3 2 8.7 10 10 0 75 0 0 1 2 10 3 -10 2 90 -10 -10 3 3 4 5 3 3 4 5 3 4 5 3 4 5 -10 0 10 10 10 -10 0 10 10 10 0 10 10 10 0 10 10 10
5 5 5 10 5 10 10 2.4 10 20 4 4 4 10 10 4 10 10 3 3 3 10 10 10 –/– 3 2 Nr4a1 10 10 2 2 9.1 10 10 0 80 0 0 1 2 10 3 -10 2 90 -10 -10 Foxp3-FITC CD44-BV421 FR4-PE-Cy7 3 3 4 5 3 3 4 5 3 4 5 3 4 5 -10 0 10 10 10 -10 0 10 10 10 0 10 10 10 0 10 10 10 CD25-APC CD62L-BV711 CD73-BV605
C D E 1:1 chimera CD45.2 CD45.1/2 100 100 100 WT WT **** + + hi – 4 80 DKO WT hi 80 80 cells hi FR4 FR4 hi 60 60 hi 60 Foxp3 +
** FR4 4 40 *** 40 hi 40 NS of naive CD4 %CD73 20 %CD73 20 20 NS of CD
**** of memory CD 0 0 %CD73 0 6w 10 - 12w 6w 10 - 12w –/– –/– WT Nr4a3Nr4a1 F G H ) ) ) 5 5 40 5 40 40 0 0 *
30 30 – 30 ** cell# (x10 cell# (x1 ve
i cell# (x1 hi hi 20 20 Foxp3 hi 20 +
Na NS Memory FR4 FR4 * FR4 NS
10 hi 10 hi CD4 hi 10 NS 0 0 0 CD73 CD73 CD73 6w 10 - 12w 6w 10 - 12w –/– –/– WT Nr4a3Nr4a1
Supplementary Figure 6. Supporting data for Figure 6 A. Splenocytes from 10 weeks post-transplant WT:WT = 1:1 and DKO:WT = 1:1 chimeras were stained as described for Fig. 6D to identify CD73hiFR4hi (anergic) T cells. Plots are representative of 3 mice. B. Splenocytes from WT, Nr4a3–/– and Nr4a1–/– mice were stained as described for Fig. 6D. Plots are representative of 3 mice. C, D, F, G. Quantification of frequency (C, D) and cell number (F, G) of CD73hiFR4hi T cells among naive (C, F) or memory (D, G) CD4+ cells in WT:WT = 1:1 chimera and DKO:WT = 1:1 chimera at indicated time points post-transplant (n ≥ 3 biological replicates from one chimera setup). E, H. Quantification of frequency (E) and cell number (H) of CD73hiFR4hi T cells among total CD4+FOXP3– T cells from WT, Nr4a3–/– and Nr4a1–/– mice (n = 3 biological replicates). Graphs depict mean +/- SEM. Statistical significance was assessed by two-tailed unpaired Student’s t-test with the Holm–Šídák method (C, D, F, G) or one-way ANOVA with Dunnett’s test (E, H). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. NS, not significant. Supplementary Figure 7 A B DKO:WT = 1:5 chimera + CD4+Foxp3– Naive CD44hiCD62Lhi Memory CD8 5 5 5 5 5 10 44 10 Memory 10 1.1 10 17 10 30 29 4 4 25 4 4 4 10 10 10 37 10 53 10 48 3 3 3 3 3 10 10 CD45.1/2 WT 10 10 10
2 2 2 10 10 10 0 36 0 Naive 27 0 56 0 24 0 17 Naive 3 3 4 5 3 3 4 5 3 3 4 5 3 3 4 5 3 3 4 5 -10 0 10 10 10 -10 0 10 10 10 -10 0 10 10 10 -10 0 10 10 10 -10 0 10 10 10
5 5 5 5 5 5 10 10 10 10 53 10 Anergic 10 55 39 21 9.0 35 4 4 4 4 4 10 Int-anergic 4 10 10 55 10 46 10 36 10 3 3 3 3 3 10 10 3 CD45.2 DKO 10 10 10 10 2 2 2 2 Non- 10 10 10 10 0 0 26 0 30 0 14 0 6.8 0 anergic 29 FR4-PE-Cy7 CD44-BV421 CD44-BV421 3 3 4 5 3 3 4 5 3 3 4 5 3 3 4 5 3 3 4 5 3 3 4 5 -10 0 10 10 10 -10 0 10 10 10 -10 0 10 10 10 -10 0 10 10 10 -10 0 10 10 10 -10 0 10 10 10 CD62L-BV711 CD73-BV605 CD62L-BV711 C D F WT CD45.1/2 4 Naive CD44hi 3 DKO CD45.2 Lif 2 Unstim. 100 ** 1 80 0 -1
+ **** Anti-CD3 60 40 %pErk PMA 20 0 3 3 4 5 3 3 4 5 -10 0 10 10 10 -10 0 10 10 10 Naive CD44hi pErk-APC WT CD45.1/2 DKO CD45.2
E Live CD4+ T cells CD25– CD44loCD62Lhi CD44hiCD62Llo 105 105 105 105 11 Anergic 4 4 4 4 10 10 10 10 26 3 10 103 3 CD45.1 WT 103 10 0 72 2 -103 10 0 Naive 0 0 71 59 3 3 4 5 3 4 5 3 4 5 3 4 5 -10 0 10 10 10 0 10 10 10 0 10 10 10 0 10 10 10
5 5 5 10 10 10 105 15 Anergic 4 4 4 4 CD45.2 DKO 10 10 10 10 21 3 10 3 3 103 10 10 0 96 2 -103 10 0 Naive 0
0 60 FR4 32 CD25 CD44 3 3 4 5 3 4 5 3 4 5 -10 0 10 10 10 0 10 10 10 0 10 10 10 0 103 104 105 CD4 CD62L CD73
Il4 G H I CD45.1/2 WT CD45.1/2 WT CD45.2 DKO CD45.2 DKO
20 DKO anergic 80 * 3000 10 * ) 60 + 2000 0 WT or DKO 40 PC2 (13.1%) -10 %PD-1 1000 NS naive+/-stim. LAG3 (MFI 20 -20 WT anergic NS 0 0 -20 0 20 Naive Memory Naive Memory PC1 (75.8%) anergic anergic
DKO WT DKO WT DKO or WT Stimulated Anergic Naive ex vivo naive Supplementary Figure 7. Supporting data for Figure 7 A. Gating strategy corresponding to Fig. 7A to identify splenic CD4+ T cell sub-populations for phosflow assay. Plots are representative of 3 mice. B, C. Splenocytes from DKO:WT = 1:5 chimera were stimulated and stained as described in Fig. 7A. (B) Representative plots show naive and CD44hiCD8+ T cell gates. (C) Representative histograms show intracellular pErk expression in naive and CD44hiCD8+ T cells. Vertical dashed line shows the threshold of positive gate. Plots are representative of 6 mice. D. Quantification of %pErk+ as described in C above (n = 3 biological replicates, representative of n = 2 independent experiments from one chimera setup). E. Gating strategy for sorting naive and anergic CD4+ T cells from competitive chimeras. Splenocytes and lymph node cells from DKO:WT = 1:5 chimera were pooled, and CD4+ T cells were negatively isolated with MACS. Enriched CD4+ T cells were stained and pre-gated on live CD4+CD45.1+ (WT) or live CD4+CD45.2+ (DKO). CD25–CD44loCD62LhiCD73loFR4lo cells were sorted as “naive”, and CD25–CD44hiCD62LloCD73hiFR4hi cells were sorted as “anergic”. F. Heatmap showing expression levels of anergy-related genes in each subset detected with RNA-seq as described in Fig. 7. Anergy-associated genes were identified as genes upregulated in WT anergic cells relative to WT naive cells identified with FDR < 0.05, CPM > 2, and fold change > 4. G. Principal component analysis (PCA) plot of sorted cells assessed for the expression of anergy-related genes as described for F above. F, G. H, I. Splenocytes from DKO:WT = 1:5 chimera 12 (I) or 13 weeks (H) post-transplant were stained to detect PD-1 expression on naive (FOXP3–CD44loCD62LhiCD73loFR4lo) and memory anergic (FOXP3–CD44hiCD62LloCD73hiFR4hi) CD4+ cells of each donor genotype. Shown are quantification of %PD-1+ (H) or MFI of LAG3 (I) of naive and memory anergic. Lines connect donor genotypes within an individual chimera (n = 3 biological replicates from one chimera setup). Graphs depict mean +/- SEM. Statistical significance was assessed by two-tailed unpaired (D) or paired (H, I) Student’s t-test with the Holm–Šídák method. *P < 0.05; **P < 0.01; ****P < 0.0001. NS, not significant. A C Supplementary Figure 8 CD4+Foxp3– 100 WT CD45.1/2 80 DKO CD45.2 2 60 40 1 20
% of Max 0 2 3 4 5 0 0 10 10 10 10 IL-2-PE -1
-2 D CD4+ WT Nr4a3–/– Nr4a1–/– 100 100 100 WT CD45.1 80 80 80 60 60 60 CD45.2 40 40 40 20 20 20
% of Max 0 0 0 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 IL-2-PE
E CD8+ Anti-CD3 0 0.1 1 100 100 100 WT CD45.1/2 80 80 80 60 60 60 DKO CD45.2 40 40 40 20 20 20
% of Max 0 0 0 2 3 4 5 2 3 4 5 2 3 4 5 0 10 10 10 10 0 10 10 10 10 0 10 10 10 10 IL-2-PE F
80 **** 60 WT CD45.1/2 + **** DKO CD45.2 40
%IL-2 **** 20
WT or DKO WT DKO 0 Naive ex vivo Stimulated 0 0.1 1 naive Anti-CD3 (µg/ml) G B + 20 CD8 DKO naive+stim. WT naive DKO naive CD8-cre CD8-cre 10 WT naive+stim. CD8-cre Nr4a3–/– Nr4a1fl/fl cDKO 100 100 100 100 DKO naive+stim. WT CD45.1 0 80 80 80 80 WT or DKO 60 60 60 60 CD45.2 PC2 (8%) -10 naive 40 40 40 40 20 20 20 20
% of Max 0 0 0 0 -20 3 4 5 3 4 5 3 4 5 3 4 5 WT naive+stim. 0 10 10 10 0 10 10 10 0 10 10 10 0 10 10 10
-30 -20 -10 0 10 20 IL-2-PE PC1 (86.4%)
Supplementary Figure 8. Supporting data for Figure 8 A. Heatmap showing expression levels of all primary response genes (PRGs) in sorted naive CD4+ T cells with or without TCR stimulation in vitro. PRGs were defined as described in Fig. 8A. B. Principal component analysis (PCA) plot of naive CD4+ T cells with or without TCR stimulation in vitro assessed for the expression of PRGs as described for A above. Heatmap and PCA plot were generated using the ClustVis online tool (https://biit.cs.ut.ee/clustvis/). stained with CD4, CD8, CD45.1 and CD45.2, followed by permeabilization and intracellular staining for FOXP3 and IL-2. Shown is a representative histogram pre-gated on CD4+FOXP3– of each genotype. D. Lymph node cells from WT, Nr4a3–/– and Nr4a1–/– stained as described for Fig. 8E. Plots are representative of 3 mice from one chimera setup. E, F. Lymph node cells from 10 weeks post-transplant DKO:WT = 1:1 chimera were stimulated and stained as described in Fig. 8E. (E) Representative histograms show intracellular IL-2 in CD8+ T cells of each genotype. (F) Quantification of %IL-2+ in CD8+ T cells (data in E, F represent n = 3 biological replicates from one chimera setup). Graphs depict mean +/- SEM. Statistical significance was assessed by two-tailed unpaired Student’s t-test with the Holm–Šídák method. ****P < 0.0001. G. Lymph node cells from CD8-cre, Nr4a3–/–, CD8-cre Nr4a1fl/fl or CD8-cre Nr4a1fl/fl Nr4a3–/– of plate-bound anti-CD3, and then treated and stained as described for Fig. 8E. Representative histograms show intracellular IL-2 in CD8+ T cells of each genotype. Plots are representative of 2 independent experiments. Supplementary Figure 9
A CD44hiCD62Llo CD4+ CD44hiCD62Llo Naive CXCR5hiPD-1hi Tfh Tfr Naive
5 20 5 5 5 5 5 5 10 10 18 10 0.041 10 10 10 10 84 4 4 4 Tfr 2.1 37 10 10 4 4 8.9 4 63 4 49 10 10 17 10 10 10 3 3 3 3 3 3 3 10 10 10 10 10 10 10 56 2 2 10 0 0 10 Tfh 0 0 0 0 2 40 02 40 Foxp3-FITC -10 CD45.2-PE-Cy7 PD-1-PCPCy5.5 -10 82 CD44-BV421 3 3 4 5 4 5 4 5 3 3 4 5 3 4 5 3 4 3 4 5 -10 0 10 10 10 0 10 10 0 10 10 -10 0 10 10 10 0 10 10 10 0 10 10 0 10 10 10 CD62L-BV711 CXCR5-BV605 CD25-APC CD45.1-PE
B C Nr4a1–/– D Nr4a3–/–
2 15 ** 1:5 chimera CD45.2 DKO 10 CD45.1/2 WT
5
Ratio CD45.2:CD45.1/ 0 Tfh Tfr E DKO:WT = 1:5 6w 10w 12w 14w
F G mb1-cre H mb1-cre cDKO DKO:WT = 1:5
Nuclear+cytosol
12 Nuclear Negative
Supplementary Figure 9. Supporting data for Figure 9 A. Splenocytes from DKO:WT = 1:5 chimera were harvested and stained with CD4, CD25, CD44, CD62L, CXCR5 and PD-1, then permeabilized and stained with FOXP3. Representative plots show FOXP3– follicular helper T cells (Tfh) and FOXP3+ follicular regulatory T cells (Tfr) within CD4+CD44hiCD62LloCXCR5hiPD-1hi gate. Naive CD4+ T cell gate is shown for reference to define Tfh/Tfr gate. Donor genotype gates among each population are shown. B. Ratio of CD45.2 to CD45.1/2 for Tfh or Tfr, normalized to naive CD4+ as gated in A above (n = 3 biological replicates from one chimera setup). Statistical significance was assessed by two-tailed unpaired Student’s t-test. **P < 0.01. C, D, E, G, H. Anti-nuclear antibody (ANA) immunofluorescence images as described for Fig. 9A-D. Serum was collected from 12-week-old Nr4a1–/– (C), Nr4a3–/– (D) mice, from DKO:WT = 1:5 chimera at 6, 10, 12, or 14 weeks post-transplant (E), and from mb1-cre Nr4a1fl/fl Nr4a3–/– (mb1cre-cDKO) (G) or mb1-cre (H) non-competitive chimeras 40 weeks post-transplant. Images are representative of n = 3 (mb1-cre chimera and DKO:WT = 1:5 chimera), n = 5 (mb1-cre cDKO chimera) or n = 10 (Nr4a1–/– and Nr4a3–/–) biological replicates. F. Graph depict frequency of Hep-2 staining patterns in DKO:WT = 1:5 chimera from one chimera setup. A WT Thymus Periphery Supplementary Figure 10
Nr4a1/3 Negative selection
Nr4a1/3 Foxp3 Foxp3
Treg Treg
Naive CD4+T Positive selection CD4SP CD4+T Nr4a1/3 Self-reactivity of T cell receptor
Anergic T
B gDKO Thymus Periphery
Nr4a1/3 Autoreactive T cells Negative selection CD4+T
Nr4a1/3 Foxp3 Foxp3
Treg Treg Polyclonal Autoantibody against B cell activation broad self-antigens (nuclear + cytosol)
Positive selection CD4SP +
Self-reactivity of T cell receptor CD4 T
C DKO:WT = 1:1 chimera Thymus Periphery
Nr4a1/3 Nr4a1/3 Autoreactive Negative selection T cells CD4+T CD73hiFR4hi Defect in anergy
Nr4a1/3 Foxp3 Foxp3
DKO Treg WT Treg Polyclonal B cell activation Autoantibody specific to nuclear antigen Nr4a1/3 Positive selection CD4SP CD4+T CD73hiFR4hi Self-reactivity of T cell receptor Defect in anergy
T cell receptor
Supplementary Figure 10. Model. Roles for NR4A family in T cell tolerance and immune homeostasis A. In WT, central and peripheral tolerance are intact. Nr4a1 and Nr4a3 mediate thymic negative selection and Treg homeostasis. Nr4a1 and Nr4a3 also play a role on induction and maintenance of anergy in periphery. B. In Nr4a1–/– Nr4a3–/– (gDKO) mice, both thymic negative selection and Treg homeostasis are impaired, and highly autoreactive T cells escape to periphery. Autoreactive CD4+ T cells are activated in part because of Treg deficiency. Polyclonal B cells are activated by autoreactive T cells and produce autoantibody against a broad spectrum of self-antigens as detected by both cytoplasmic and nuclear Hep-2 staining pattern. C. Treg compartment is reconstituted by WT donor BM in DKO:WT competitive chimeras. However, negative selection of self-reactive DKO thymocytes is still impaired. Self-reactive DKO T cells that escaped negative selection encounter self-antigens in the periphery, acquire features of anergy including expression of CD73 and FR4, yet exhibit a defect in peripheral tolerance. WT-origin Treg and cell-intrinsic peripheral tolerance mechanisms are insufficient to completely suppress autoreactive DKO T cells, which in turn drive anti-nuclear autoantibody production. Although Treg compartment restores some immune homeostasis (and tolerance to cytosolic antigens as detected by Hep-2 staining), autoimmunity is not suppressed due to role of NR4A family in other T cell-intrinsic tolerance mechanisms.