Abstract

   Protein tyrosine phosphatase nonreceptor type 22 (PTPN22) is encoded by
   a gene strongly associated with lupus and other autoimmune diseases.
   PTPN22 regulates T cell receptor (TCR) signaling through
   dephosphorylation of the kinases lymphocyte-specific protein tyrosine
   kinase (LCK) and zeta-chain–associated protein kinase 70 (ZAP70). The
   regulation of PTPN22 remains poorly understood. Here, we identify
   PTPN22 Ser^449 as a protein kinase A phosphorylation site, which is
   triggered by TCR engagement and is hyperphosphorylated in lupus
   peripheral blood cells. PTPN22 Ser^449 phosphorylation selectively
   lowered the affinity of PTPN22 for ZAP70 versus LCK but also indirectly
   suppressed inhibitory LCK Tyr^192 phosphorylation through a ZAP70-CD45
   signaling axis. The resulting dephosphorylation of LCK Tyr^192 not only
   enhanced TCR signaling but also modulated pathway activation downstream
   the TCR. In vivo loss of PTPN22 Ser^449 phosphorylation reduced T cell
   responses and suppressed experimental lupus nephritis. These results
   suggest that PTPN22 Ser^449 phosphorylation promotes a CD45-mediated
   retrograde ZAP70-LCK feedback loop that enhances T cell responses and
   promotes autoimmunity.
     __________________________________________________________________

   PTPN22 Ser^449 phosphorylation enables ZAP70-CD45-LCK feedback loop
   that boosts T cell activation and contributes to autoimmunity.

INTRODUCTION

   Dysregulated T cell receptor (TCR) signaling is a critical contributor
   to systemic lupus erythematosus (SLE) pathophysiology and contributes
   to the breakdown of immune tolerance and the amplification of
   pathogenic immune responses ([64]1). Studies from multiple laboratories
   have established that SLE T cells display abnormal TCR signaling, which
   affects downstream T cell differentiation and contributes to the
   immunopathology of disease ([65]2, [66]3). Abnormalities in the
   composition of the TCR signaling complex, calcium signaling, and
   endosomal trafficking have been described in SLE T cells ([67]3–[68]5).

   Engagement of TCR triggers a highly regulated cascade of
   phosphorylation events via the concerted action of protein kinases and
   phosphatases ([69]6). Zeta chain–associated protein kinase 70 (ZAP70)
   and the Src family kinase lymphocyte-specific protein tyrosine kinase
   (LCK) are protein tyrosine kinases whose catalytic activities are
   central to early T cell signaling ([70]7, [71]8). Upon TCR engagement,
   with or without coreceptor involvement, activated LCK phosphorylates
   TCR CD3 and zeta chains leading to ZAP70 recruitment and its
   LCK-mediated phosphorylation on several residues, including Tyrosine
   319 (Tyr^319) ([72]9). Binding of the LCK Src homology 2 (SH2) domain
   to Tyr^319-phosphorylated ZAP70 then sustains the activities of both
   kinases in a positive feedback mechanism that promotes downstream
   signaling events ([73]10). LCK catalytic activity is controlled by its
   SH2 and SH3 and by phosphorylation at Tyr residues ([74]11).
   Phosphorylation of LCK Tyr^394 in the so-called activation loop and
   Tyr^505 near the C terminus stimulate and suppress LCK activity,
   respectively. Tyr^394 is trans-autophosphorylated by LCK and is
   dephosphorylated by the SH2 domain–containing phosphatase-1 (SHP-1) and
   protein tyrosine phosphatase nonreceptor type 22 (PTPN22) ([75]12,
   [76]13). Tyr^505 is phosphorylated by C-terminal Src kinase (CSK) and
   dephosphorylated by CD45 ([77]14). In addition, a recent study revealed
   a role for an additional phosphorylation site of LCK on Tyr^192, which
   is thought to be regulated by interleukin-2 (IL-2) inducible Tec kinase
   (ITK), enabling a negative feedback loop involving reduced access to
   the inhibitory phospho-Tyr^505 (pTyr^505) by CD45 after TCR activation
   ([78]14, [79]15). It currently remains unknown which PTP
   dephosphorylates LCK-pTyr^192.

   PTPN22 is a hematopoietic cell–specific cytosolic protein tyrosine
   phosphatase acting as a key negative regulator of early TCR signaling
   through dephosphorylation of LCK and ZAP70. PTPN22 potentially has
   additional, less well-characterized substrates in T cells ([80]16). In
   other immune cells, it regulates signaling pathways downstream of the
   type 1 interferon (IFN) and other receptors ([81]17, [82]18). PTPN22
   consists of three domains: an N-terminal, well-characterized classic
   PTP catalytic domain (amino acids 1 to 300); an interdomain (301 to
   600), which may have regulatory functions but whose mechanism of action
   remains poorly understood; and a C-terminal domain (601 to 807), which
   carries four proline-rich motifs (P1 to P4) and facilitates
   protein-protein interactions ([83]19, [84]20). A R620W polymorphism
   (C1858T) in the P1 motif of PTPN22 has emerged as a major autoimmunity
   risk factor, and a strong association between PTPN22 R620W and SLE has
   been reported by many studies ([85]21). The functional effect of the
   W620 mutation in various T cells remains to be fully clarified;
   however, studies in knock-in (KI) mice generally pointed to the
   autoimmune-associated mutant causing loss or change in function of the
   molecule ([86]16).

   Despite the critical role of PTPN22 in T cell biology and autoimmune
   diseases, its regulatory mechanisms during TCR signaling and disease
   progression remain inadequately understood. We and others have
   identified one regulatory tyrosine (Tyr^536) and two serine (Ser^35 and
   Ser^751) PTPN22 phosphorylation sites ([87]22–[88]24); however,
   phosphorylation of PTPN22 has not been assessed in autoimmunity yet.
   Here, we report that Ser^449 within the PTPN22 interdomain is an
   unrecognized protein kinase A (PKA) target site–induced post–TCR
   stimulation, and its phosphorylation is pronouncedly elevated in
   peripheral blood mononuclear cells (PBMCs) from patients with SLE. The
   biochemical characterization of the effects of phospho-Ser^449 on early
   TCR signaling, coupled with the assessment of KI mice, uncovered
   previously unreported PTPN22-controlled retrograde feedback loop, which
   is activated by ZAP70 and impairs dephosphorylation of LCK Tyr^192 by
   CD45, thus promoting TCR signaling and autoimmunity.

RESULTS

Ser^449 in the interdomain of PTPN22 is an unidentified PKA target site and
is hyperphosphorylated in SLE PBMCs

   PTPN22 phosphorylation status in immune cells has only been partially
   characterized ([89]24). To further explore PTPN22 phosphorylation in T
   cells, we took advantage of a Jurkat CD4^+ T cell line carrying PTPN22
   with a KI 3× FLAG tag [J.N22 wild-type (WT)] ([90]24). PTPN22 was
   immunoprecipitated from these cells after TCR engagement and analyzed
   by Phos-tag SDS–polyacrylamide gel electrophoresis (SDS-PAGE) followed
   by Western blotting with an anti-FLAG antibody ([91]Fig. 1A). This
   analysis revealed a range of migration rates of PTPN22 molecules,
   likely due to variable phosphorylation of multiple sites in the Jurkat
   T cells PTPN22 pool. Next, the phosphorylation sites of PTPN22 purified
   from TCR-stimulated Jurkat T cells (fig. S1, A and B) were determined
   using mass spectrometry (MS). Multiple peptides containing
   phosphorylation at residue Ser^449 (TKS + 79.66331TPFELIQQR) were
   detected in stimulated Jurkat T cells (fig. S1C). This site is located
   within the PTPN22 interdomain ([92]Fig. 1B) and is conserved among
   various species (fig. S1D). We generated a phospho-specific antibody
   recognizing PTPN22 phosphorylated on Ser^449 whose reactivity to PTPN22
   immunoprecipitated from WT cells was enhanced more than twofold by cell
   treatment with the serine/threonine phosphatase inhibitor calyculin A
   ([93]Fig. 1C) or stimulation through the TCR ([94]Fig. 1D). The
   anti–phospho-Ser^449 antibody showed no reactivity to PTPN22 mutated at
   Ser^449 to Ala ([95]Fig. 1D). In a time course (1 to 10 min) of TCR
   engagement, PTPN22 Ser^449 phosphorylation was induced at all time
   points ([96]Fig. 1E). Moreover, PTPN22 Ser^449 phosphorylation is
   present in unstimulated T cells and could be further induced in primary
   human T cells ([97]Fig. 1F). These data suggest that Ser^449 of PTPN22
   undergoes TCR-increased phosphorylation in human T cells.

Fig. 1. Ser^449 of PTPN22 is targeted by PKA and hyperphosphorylated in SLE
PBMCs.

   [98]Fig. 1.
   [99]Open in a new tab

   (A) Analysis of PTPN22 phosphorylation by Phos-tag SDS-PAGE and Western
   blotting in J.N22 KO and J.N22 WT Jurkat cells. Non–phos-tag gel served
   as control. (B) Schematic view of the PTPN22 structure with key
   catalytic residues, known phosphorylation sites including the Ser^449
   site in the interdomain and the R620W variation. (C to F) Ser^449
   phosphorylation was assessed by Western blotting following
   immunoprecipitation of PTPN22 from J.N22 WT Jurkat cells treated with
   calyculin A (C), or in J.N22 KO cells expressing FLAG-tagged WT or
   S449A PTPN22, with or without 10-min TCR stimulation (D), or in J.N22
   WT cells after 0-, 1-, 5-, and 10-min TCR stimulation (E), or in
   primary human CD4^+ effector T cells with or without 10-min TCR
   stimulation (F). (G) NetPhos hits with prediction score > 0.5, PKA
   (0.621) and RSK (0.541) were predicted Ser^449 kinases. “Generic”
   indicates overall prediction of Ser^449 as a phosphorylation site. (H
   and I) Effect of PKA Cα inhibitor [(H) H89 2HCl] or PKA Cα siRNA (I) on
   Ser^449 phosphorylation in J.N22 WT cells after TCR stimulation. (J and
   K) In vitro PKA kinase assay using purified recombinant WT or S449A
   PTPN22 from HEK293T cells as substrates. (L) Ser^449 phosphorylation in
   primary patients with SLE (n = 6) versus healthy controls (n = 4) PBMCs
   was analyzed by Western blotting following immunoprecipitation of
   PTPN22. Representative blots are shown (left). Histograms show
   comparison of Ser^449 phosphorylation levels between stimulated/treated
   and control groups (right). Data are presented as means ± SEM and are
   from three (A, C, D, I, and K) or four (E, F, and H) independent
   experiments. Statistical significance was assessed by two-tailed
   Mann-Whitney test (C, F, H, I, and L), two-way analysis of variance
   (ANOVA) followed by Bonferroni’s post hoc test (D and K), or one-way
   ANOVA followed by Dunn’s post hoc test (E). *P < 0.05, **P < 0.01,
   ***P < 0.001. (J) Created with BioRender ([100]69).

   To address the kinase(s) mediating PTPN22 Ser^449 phosphorylation, we
   first examined protein kinase C (PKC) as a candidate since PTPN22
   undergoes PKC-mediated phosphorylation at Ser^35 and Ser^751 ([101]23,
   [102]24). We used an array of PKC inhibitors; however, none influenced
   TCR-induced PTPN22 Ser^449 phosphorylation (fig. S1E). Next, we used
   the NetPhos kinase prediction software to analyze the amino acid
   sequence surrounding PTPN22 Ser^449 and found PKA and ribosomal S6
   kinase (RSK) as the top two candidate PTPN22 Ser^449 kinases ([103]Fig.
   1G). However, while an inhibitor of PKA catalytic subunits (PKA Cs)
   blocked TCR-induced phosphorylation of PTPN22 Ser^449 ([104]Fig. 1H),
   an RSK inhibitor had no effect (fig. S1F). PKA C-alpha (PKA Cα) is the
   major cytosolic PKA C in Jurkat and primary T cells ([105]25, [106]26).
   PTPN22 Ser^449 phosphorylation was dramatically reduced when PRKACA was
   knocked down by small interfering RNA (siRNA) ([107]Fig. 1I). PKA Cα
   was also coprecipitated with PTPN22 from Jurkat cells, and the
   interaction did not require the Ser^449 residue (fig. S1G). When we
   examined PKA activity in T cells by evaluating the phosphorylation of
   its key downstream mediator cyclic adenosine monophosphate–responsive
   element–binding protein (CREB) ([108]27), we found that it was markedly
   promoted by TCR stimulation (fig. S1H), with a time course compatible
   with the one observed for PTPN22-Ser^449. Last, to verify that PKA Cα
   can directly phosphorylate PTPN22 at Ser^449, purified FLAG-tagged
   PTPN22 WT and S449A proteins were incubated with PKA Cα in vitro
   ([109]Fig. 1J). Incubation of PTPN22 WT—but not the S449A mutant—with
   PKA resulted in strong signal detected by Western blotting with the
   phospho-Ser^449 antibody ([110]Fig. 1K). Because PTPN22 is a major SLE
   genetic risk factor ([111]21) and disease-activity dependent increase
   in free cytosolic PKA Cs has been observed in lupus T cells ([112]28),
   we decided to evaluate the phosphorylation of PTPN22 Ser^449 in PBMCs
   from patients with SLE with active disease (SLE disease activity index,
   SLEDAI >2) ([113]29) and healthy controls (see patient and control data
   in table S1). [114]Figure 1L shows that the phosphorylation levels of
   Ser^449 were noticeably up-regulated in PBMCs from patients with SLE.
   Together, our data indicate that Ser^449 is a TCR-inducible PKA
   phosphorylation site up-regulated in SLE immune cells.

Ser^449 phosphorylation inhibits PTPN22 functions and enhances TCR signaling

   To understand the potential effects of Ser^449 phosphorylation on the
   function of PTPN22 in T cells, we leveraged a nuclear factor of
   activated T cells and activator protein-1 (NFAT/AP-1) luciferase
   reporter activity, which is commonly used to quantitate signaling
   downstream TCR activation ([115]30). Expression of S449A PTPN22 led to
   notably reduced TCR-induced luciferase reporter activation compared to
   similar levels of WT PTPN22 ([116]Fig. 2A) suggesting that S449A PTPN22
   behaves as a gain-of-function mutant. The R620W genetic variation in
   the P1 motif of PTPN22 impairs the interaction of PTPN22 with CSK,
   tumor necrosis factor receptor-associated factor 3 (TRAF3), and other
   TCR effectors; however, we found no substantial impairment in Ser^449
   phosphorylation between R620 and W620 PTPN22 upon TCR stimulation (fig.
   S2A). Further suppression of NFAT/AP-1 luciferase activity was also
   observed in cells expressing the double W620 S449A mutant compared to
   W620, indicating that Ser^449 phosphorylation inhibits the function of
   both R620 and W620 PTPN22 variants (fig. S2B). Moreover, diminished
   expression of CSK had no impact on Ser^449 phosphorylation of PTPN22
   (fig. S2C), overall suggesting that the P1 domain of PTPN22 is not
   involved in the regulation of Ser^449 phosphorylation.

Fig. 2. Ser^449 phosphorylation inhibits PTPN22 functions during TCR
signaling.

   [117]Fig. 2.
   [118]Open in a new tab

   (A) Evaluation of TCR signaling–dependent PTPN22 activity by
   dual-luciferase reporter assay in J.N22 KO cells. Luciferase activity
   was measured (left, with pooled replicates) and normalized to the
   amount of PTPN22 relative to that of glyceraldehyde-3-phosphate
   dehydrogenase (GAPDH) as assessed by Western blotting (right). (B)
   Real-time RT-PCR analysis of IL2 and TNFA mRNA expression in J.N22 WT
   and J.N22 449 cells. (C) TCR-induced CD69 activation in J.N22 WT and
   J.N22 449 cells analyzed by flow cytometry and evaluated by median
   fluorescence intensity (MFI) of cells. (D) Protein phosphorylation
   induced by TCR stimulation in J.N22 WT and J.N22 449 cells analyzed by
   Western blotting. (E and F) TCR stimulation–induced SLP-76 Tyr^128
   phosphorylation in J.N22 WT and J.N22 449 cells assessed by flow
   cytometry (E) or Western blotting (F) at time points between 0 and 10
   min. (G) Endogenous Rap1 activity in J.N22 WT and J. N22 449 cells
   assessed through RAS-like guanine nucleotide dissociation stimulator
   (Ral GDS) pull down was detected by immunoblotting and normalized to WT
   at 0 min. (H) The percentage of attached J.N22 WT or J.N22 449 cells to
   immobilized ICAM-1 was assessed following stimulation with or without
   antibodies against human CD3/CD28 and with plate-coated ICAM-1. (I)
   Integrin-mediated outside-in signaling in J.N22 WT and J.N22 449 cells
   was assessed by measuring protein phosphorylation as shown after
   stimulation with ICAM-1 for 0, 10, and 20 min. Representative blots are
   shown. Histogram shows quantification of phosphorylated protein
   normalized to total protein (D, F, and I). Data are presented as
   means ± SEM and are from three (A, E, and F), six (B and C), four (D,
   G, and I), or eight (H) independent experiments. Statistical
   significance was assessed by one-way ANOVA followed by Dunn’s post hoc
   test (A), Kolmogorov-Smirnov test (B), and two-way ANOVA followed by
   Bonferroni’s post hoc test (C to I). *P < 0.05, **P < 0.01,
   ***P < 0.001.

   To avoid the confounding effects of endogenous WT PTPN22, we generated
   a Jurkat PTPN22 S449A KI line by CRISPR-Cas9–mediated genome editing
   (J.N22 449; fig. S2D). The phosphorylation-null mutation was confirmed
   by various methods (table S2 and fig. S2, E to H). TCR-induced IL-2 and
   tumor necrosis factor–α (TNF-α) mRNA expression and activation of CD69
   were decreased in J.N22 449 cells ([119]Fig. 2, B and C, and fig. S2I),
   suggesting attenuated downstream T cell activation. We next assessed
   early phosphorylation events after TCR engagement and found markedly
   diminished TCR-triggered tyrosine phosphorylation of ZAP70 Tyr^319, of
   ZAP70 substrate SH2 domain–containing leukocyte protein of 76 kDa
   (SLP-76) Tyr^128 and ZAP70 downstream adaptor phospholipase C ϒ (PLCϒ)
   Tyr^783 in J.N22 449 cells compared to WT cells. However,
   phosphorylation of another well-established PTPN22 substrate, LCK
   Tyr^394, remained unaffected ([120]Fig. 2, D to F, and fig. S2J).

   PTPN22 has been implicated with integrin lymphocyte function–associated
   antigen 1 (LFA-1) ([121]31). TCR triggers rapid integrin-mediated
   adhesion of T cells involving “inside-out” signals and “outside-in”
   signaling mediated by ZAP70 ([122]32, [123]33). TCR-induced activation
   of RAS-associated protein 1 (Rap1) was strongly inhibited ([124]Fig.
   2G), and decreased cell adhesion to the LFA-1 ligand intercellular
   adhesion molecule 1 (ICAM-1) was observed in J.N22 449 cells compared
   to WT or knock-out (KO) cells ([125]Fig. 2H). Furthermore, J.N22 449
   cells showed reduced phosphorylation of ZAP70 upon ICAM-1 binding,
   pointing to impaired integrin outside-in signaling in these cells
   ([126]Fig. 2I). Together, these results indicate that Ser^449
   phosphorylation reduces the ability of PTPN22 to dephosphorylate ZAP70,
   which promotes TCR signaling downstream ZAP70.

Ser^449 phosphorylation impairs PTPN22 binding to ZAP70

   We next sought to determine whether Ser^449 phosphorylation inhibits
   PTPN22’s phosphatase activity and found that the activity of PTPN22
   immunoprecipitated from TCR-stimulated S449A cells was unchanged
   compared to WT when using 6,8-difluoro-4-methylumbelliferyl phosphate
   (DiFMUP) as a substrate (fig. S3A). Previously, we reported that PTPN22
   Ser^751 phosphorylation in the C-terminal domain extends PTPN22
   half-life in T cells ([127]24), prompting us to determine whether
   Ser^449 phosphorylation also regulates PTPN22’s half-life. However, the
   abundance of S449A PTPN22 remained akin to WT in time courses after
   cycloheximide treatment (fig. S3B).

   As TCR-induced phosphorylation of ZAP70 but not LCK was notably reduced
   in J.N22 449 cells, we next explored whether PTPN22-pSer^449
   differentially regulates the ability of PTPN22 to dephosphorylate its
   two main early TCR signaling substrates. After confirming that PTPN22
   Ser^449 is highly phosphorylated in COS-7 cells (fig. S3C), we assessed
   dephosphorylation of LCK in COS-7 cells coexpressing LCK with WT or
   S449A PTPN22 or with the catalytically inactive D195A/C227S (DACS)
   PTPN22 mutants ([128]Fig. 3A). Phosphorylation of LCK Tyr^394 decreased
   similarly in cells expressing WT or S449A PTPN22 compared to vector
   alone, while the DACS WT or DACS S449A PTPN22 mutants had no effect
   ([129]Fig. 3B). Similar results were obtained using human embryonic
   kidney (HEK) 293T cells (fig. S3D). [130]Figure 3C shows that
   phosphorylation of ZAP70 Tyr^319 and Tyr^493 was dramatically reduced
   in COS-7 cells coexpressing S449A PTPN22 compared with WT, and this
   difference was abolished in cells expressing DACS mutants of WT or
   S449A PTPN22. Consistent with the high stoichiometry of Ser^449
   phosphorylation observed in COS-7 cells, phospho-tyrosine mimetic
   PTPN22 S449E or S449D variants did not cause further decreases of ZAP70
   phosphorylation in this system compared to WT PTPN22 (fig. S3E).

Fig. 3. Ser^449 phosphorylation impairs PTPN22 binding to ZAP70.

   [131]Fig. 3.
   [132]Open in a new tab

   (A) Schematic illustration of PTPN22 variants and dephosphorylation of
   substrates LCK and ZAP70. (B) Immunoblotting analysis of COS-7 cells
   cotransfected with untagged LCK WT and 3× FLAG tagged WT, S449A, DACS
   WT, or DACS S449A PTPN22. Histogram shows quantification of the
   phospho-Tyr^394 LCK /total LCK ratio (right). (C) Dephosphorylation of
   ZAP70 in COS-7 cells cotransfected with hemagglutinin (HA)–tagged ZAP70
   and 3× FLAG tagged WT, S449A, DACS WT, or DACS S449A PTPN22 assessed by
   immunoblotting. Histograms show quantification of the phospho-Tyr^319
   or Tyr^493 ZAP70 /total ZAP70 ratio. (D) PLA in J.N22 KO, WT and 449
   cells stimulated by plate-bound anti-CD3 and soluble anti-CD28
   antibodies for 30 min. (E) PLA was performed between PTPN22 and ZAP70
   or LCK using mouse anti-FLAG and rabbit anti-ZAP70 or -LCK antibodies.
   Images of PLA (red, left) and quantification of PLA are shown (right).
   Scale bars, 10 μm. (F) Schematic model of PTPN22 intracellular FRET.
   (G) FRET measurements of PTPN22 conformation dynamics by flow
   cytometry. PTPN22 Ser^449 phosphorylation and expression of PTPN22 FRET
   fusion proteins was evaluated by immunoprecipitation and Western blot
   analysis (left). FRET was assessed by adjusting the gates to identify
   cells that are negative in transfected with CFP-PTPN22 only, PTPN22-YFP
   only, CFP-PTPN22, and PTPN22-YFP. Histogram shows quantification of
   FRET % (right). Data are presented as means ± SEM and from three to
   five (B, C, and E) or six (G) independent experiments. Representative
   blots are shown (left). Statistical significance was assessed by
   one-way ANOVA followed by Dunnett’s post hoc test (B, C, and E), or
   two-tailed Mann-Whitney test (G). *P < 0.05, **P < 0.01, ***P < 0.001.
   (A) and (D) created with BioRender ([133]69).

   Distinct effects of PTPN22 Ser^449 phosphorylation on LCK versus ZAP70
   suggested that Ser^449 phosphorylation differentially modulates
   PTPN22-binding affinity to these two substrates. Thus, PTPN22-LCK and
   -ZAP70 proximity ligation assays (PLAs) in J.N22 KO, J.N22 WT, and
   J.N22 449 Jurkat cells were performed, and revealed a substantial
   increase in the molecular colocalization of S449A PTPN22 with ZAP70
   after T cell activation, but not with LCK ([134]Fig. 3, D and E). To
   investigate whether the interaction between PTPN22 and ZAP70 is
   similarly regulated by S449A PTPN22 downstream ICAM-1 stimulation, PLA
   assay was also performed. The results demonstrated an increased
   molecular colocalization between PTPN22 and ZAP70 in J.N22 449 Jurkat
   cells (fig. S3, F and G). Considering the location of Ser^449, which is
   situated in the more loosely ordered interdomain region of PTPN22
   ([135]16), one potential explanation for the above observation is that
   Ser^449 phosphorylation contributes to a conformational change in
   PTPN22 after TCR engagement. To explore this, we used intramolecular
   Förster resonance energy transfer (FRET) detected by flow cytometry to
   monitor the dynamic conformation of PTPN22 in HEK293T transfected cells
   ([136]Fig. 3, F and G). We observed enhanced FRET signal in cells
   expressing S449A PTPN22 compared with WT ([137]Fig. 3G), suggesting
   that Ser^449 phosphorylation affects the intracellular conformation of
   PTPN22. We conclude that PTPN22 phospho-Ser^449 selectively promotes
   TCR-induced proximity to and dephosphorylation of ZAP70 versus LCK (and
   perhaps other PTPN22 substrates in T cells), likely through an effect
   on the interdomain-regulated PTPN22 conformation.

Mutation of PTPN22 Ser^449 elevates LCK Tyr^192 phosphorylation

   Intrigued by the selective effect of the PTPN22 S449A mutation on ZAP70
   versus LCK-activating phosphorylation sites, we next turned our
   attention to other LCK phosphorylation sites in J.N22 449 cells. We
   thus assessed the three major LCK Tyr phosphorylation sites in an
   independent set of experiments. While we confirmed that phospho-Tyr^394
   was unaffected, we found that LCK Tyr^192 and Tyr^505 phosphorylation
   were up-regulated in resting as well as stimulated J.N22 449 cells
   ([138]Fig. 4A). A potent and selective ITK inhibitor ([139]34) did not
   block LCK Tyr^192 phosphorylation (fig. S4A), overall ruling out ITK as
   a major LCK-Tyr^192 kinase in our cells. Since ZAP70 had also been
   previously excluded as an LCK-Tyr^192 kinase ([140]35), we next
   hypothesized that Tyr^192 could be an autophosphorylation site of LCK.
   We expressed LCK mutated at its various Tyr phosphorylation sites in
   HEK293T cells—which lack ITK—and in J.CaM1.6 cells, a derivative of
   Jurkat that lacks LCK ([141]36) but expresses ITK. LCK Tyr^192
   phosphorylation was nearly absent in HEK293T cells expressing the
   inactive LCK Y394F mutant but was substantially enhanced in cells
   expressing the hyperactive LCK Y505F mutant ([142]Fig. 4B), suggesting
   that LCK Tyr^192 can be phosphorylated by active LCK. In J.CaM1.6
   cells, LCK Tyr^192 phosphorylation was again substantially reduced by
   the LCK Y394F mutation; however, the LCK Y505F mutant also displayed
   decreased Tyr^192 phosphorylation when compared to WT LCK ([143]Fig.
   4C). Comparing these results, we suspected that LCK Tyr^192
   phosphorylation is under control of a T cell–expressed kinase or
   phosphatase that is absent in HEK293T cells and respectively inhibited
   or induced by TCR signaling. We decided to initially test this
   hypothesis by blocking signaling downstream ZAP70 and observed that
   Tyr^192 phosphorylation was elevated in J.N22 WT cells subjected to
   knockdown of ZAP70 ([144]Fig. 4D). Knockdown of ZAP70 also
   substantially increased LCK Tyr^192 phosphorylation in JCaM 1.6 cells
   expressing LCK Y505F ([145]Fig. 4E), while knockdown of CSK—which
   enhances TCR signaling—decreased LCK Tyr^192 phosphorylation in J.N22
   WT cells ([146]Fig. 4F). These data suggest that TCR-induced ZAP-70
   activity inhibits LCK Tyr^192 phosphorylation through a retrograde
   feedback mechanism. We next assessed LCK phospho-Tyr^192 in P116 T
   cells, a derivative of Jurkat lacking ZAP70 ([147]37), and observed
   that it was highly enhanced in resting or TCR-stimulated conditions.
   Notably, phosphorylation of LCK Tyr^394 and Tyr^505 were also increased
   in P116 cells ([148]Fig. 4G) ([149]37). These data rule out ZAP70 and
   ITK as major LCK Tyr^192 kinases and suggest that LCK activity and a T
   cell–expressed phosphatase induced by TCR signaling are necessary,
   respectively, for LCK Tyr^192 phosphorylation and its down-regulation
   induced by low or absent ZAP70 activity.

Fig. 4. Mutation of PTPN22 Ser^449 elevates LCK Tyr^192 phosphorylation.

   [150]Fig. 4.
   [151]Open in a new tab

   (A) Immunoblotting analysis of phospho-proteins in total lysates of TCR
   stimulated J.N22 WT and J.N22 449 cells. Blots show LCK phosphorylation
   (Tyr^192, Tyr^394, or Tyr^505) induced by TCR stimulation. (B and C)
   LCK phosphorylation in HEK293T (B) or J.CaM1.6 (C) cells transfected
   with LCK variants. (D) TCR-induced LCK Tyr^192 phosphorylation in ZAP70
   knockdown WT cells. (E) LCK phosphorylation in J.CaM1.6 cells
   transfected with LCK Y505F variant and subjected to ZAP70 knockdown
   after 10-min TCR stimulation. (F) LCK Tyr^192 and ZAP70 Tyr^319
   phosphorylation in CSK knockdown WT cells. (G) Immunoblotting analysis
   of LCK phosphorylation in P116 cells induced by TCR stimulation. Data
   are presented as means ± SEM and are from four (A, E, and F) or three
   (B to D, and G) independent experiments. Representative blots are shown
   (left), and histograms show quantification of phosphorylated LCK
   normalized to total protein (right). Statistical significance was
   assessed by two-way ANOVA followed by Bonferroni’s post hoc test (A and
   D to G), and one-way ANOVA followed by Dunnett’s post hoc test (B and
   C). *P < 0.05, **P < 0.01. Schematics in (B) to (G) created with
   BioRender ([152]69).

ZAP70-CD45 axis regulates LCK Tyr^192 phosphorylation

   Initially, we hypothesized that PTPN22 might be the LCK Tyr^192
   phosphatase. However, deletion of PTPN22 enhances Tyr^394
   phosphorylation ([153]17), and we found reduced Tyr^192 phosphorylation
   in J.N22 KO T cells (fig. S4B). Overexpression of SH2-containing SHP-1
   and SHP-2, which both have roles in TCR signaling ([154]38) did not
   affect LCK Tyr^192 phosphorylation in WT cells (fig. S4C). We next
   hypothesized that CD45 could serve as an LCK Tyr^192 phosphatase and
   assessed LCK Tyr^192 phosphorylation in J45.01 cells, a Jurkat
   derivative lacking CD45 ([155]39). As CD45 is also known to
   dephosphorylate LCK Tyr^505 and Tyr^394, a regulation of CD45 function
   by ZAP70 could explain the triple-site hyperphosphorylation of LCK
   observed in P116 cells ([156]Fig. 4G). However, Tyr^192 phosphorylation
   was drastically reduced in J45.01 cells compared to J.N22 WT cells upon
   TCR stimulation ([157]Fig. 5A). To rule out that the profound
   obliteration of LCK activity caused by total lack of CD45 in J45.01 had
   prevented the autophosphorylation of Tyr^192, we next assessed LCK
   Tyr^192 phosphorylation in WT cells in which CD45 was only partially
   knocked down by siRNA. In a notable reversal of the result obtained in
   J45.01 cells, LCK Tyr^192 phosphorylation was markedly increased in WT
   cells when CD45 was knocked down ([158]Fig. 5B). We confirmed that CD45
   expression was equal in J.N22 WT and J.N22 449 cells (fig. S4D). As
   CD45 knockdown tends to lower LCK activity, these data suggest a
   potential dominant role for CD45 in control of LCK Tyr^192
   phosphorylation. Consistent with this hypothesis, when we expressed LCK
   Y505F in JCaM 1.6 cells subjected to CD45 knockdown, we observed a
   strong induction of LCK Tyr^192 phosphorylation (fig. S4E). As
   phosphorylation at Tyr^505 can be reduced by inhibiting CSK activity
   ([159]40, [160]41), we also assessed LCK Tyr^192 phosphorylation in
   J.Csk^AS/CD45 cells, which lack CD45 and can be subjected to inhibition
   of CSK-analog-sensitive (Csk^AS) activity to treatment with the PP1
   analog compound 3-IB-PP1 ([161]40). LCK Tyr^192 phosphorylation was
   strikingly induced by 3-IB-PP1 in J.Csk^AS/CD45 cells ([162]Fig. 5C),
   ruling out CSK as a major LCK Tyr^192 kinase and showing that
   restoration of LCK activity (as indicated by enhanced ZAP70 Tyr^319 and
   PLCϒ Tyr^783 phosphorylation; [163]Fig. 5C) is sufficient to enhance
   LCK Tyr^192 phosphorylation in the absence of CD45.

Fig. 5. ZAP70-CD45 axis regulates LCK Tyr^192 phosphorylation.

   [164]Fig. 5.
   [165]Open in a new tab

   (A) TCR stimulation–induced LCK phosphorylation in J45.01 cells. (B)
   TCR stimulation–induced LCK phosphorylation in CD45 knockdown WT cells.
   (C) Phosphorylation of LCK Tyr^192 after CSK inhibition in CD45
   deficient cells. J.Csk^AS/CD45 cells were treated with antibodies
   against CD3/CD28 or the Csk^AS inhibitor (5 μM 3-IB-PP1) for 2 min
   (left). (D and E) LCK Tyr^192 phosphorylation in CD45 knockdown P116
   cells (D) or in ZAP70 knockdown J45.01 cells (E) after TCR stimulation
   for 0, 5, and 10 min. Representative blots are shown (left). Histograms
   show quantification of phospho-Tyr^192 LCK to total protein ratio
   (right). (F) PLA assay in J.N22 KO, WT, and 449 cells stimulated with
   antibodies against human CD3/CD28 for 30 min. PLA was performed between
   CD45 and LCK using mouse anti-LCK and rabbit anti-CD45 antibodies.
   Representative images of PLA (red, left) and quantification of PLA are
   shown (right). Scale bars, 10 μm. Data are presented as means ± SEM and
   are from three (A to F) independent experiments. Statistical
   significance was assessed by two-way ANOVA followed by Bonferroni’s
   post hoc test (A, B, D, and E) and by one-way ANOVA followed by
   Dunnett’s post hoc test (C and F). *P < 0.05, **P < 0.01, ***P < 0.001.
   Schematics in (A) to (E) created with BioRender ([166]69).

   We next reasoned that if LCK Tyr^192 is an LCK autophosphorylation site
   and is dephosphorylated by CD45, the increased phosphorylation of
   Tyr^192 observed in J.N22 449 cells could be due to a ZAP70-CD45
   retrograde feedback loop that inhibits LCK Tyr^192 phosphorylation in T
   cells. We explored the effects of ZAP70 on LCK Tyr^192 phosphorylation
   in cells lacking CD45 by knocking down ZAP70 in J45.01 cells or CD45 in
   P116 cells. CD45 knockdown strongly increased LCK Tyr^192
   phosphorylation in P116 cells ([167]Fig. 5D), while ZAP70 knockdown did
   not enhance LCK Tyr^192 phosphorylation in J45.01 cells ([168]Fig. 5E).
   We reasoned that decreased LCK Tyr^192 phosphorylation in cells lacking
   PTPN22 might be due to constitutive activation of the retrograde
   ZAP70-CD45 feedback loop (see also the effect of CSK knockdown in
   [169]Fig. 4F) and thus decided to assess CD45 recruitment to LCK by PLA
   assay in TCR-stimulated J.N22 KO, J.N22 WT and J.N22 449 cells. As
   shown in [170]Fig. 4F, a substantial increase in the molecular
   colocalization of CD45 with LCK was observed in KO cells versus WT
   cells. In contrast, PTPN22 S449A induced a further decrease in the
   molecular colocalization of CD45 with LCK compared with WT cells after
   T cell activation ([171]Fig. 5F). Together, these data support CD45 as
   a key LCK Tyr^192 phosphatase and that ZAP70 promotes this function of
   CD45 downstream TCR engagement by inducing proximity between CD45 and
   LCK. Phosphorylation of PTPN22 Ser^449 enhances such CD45-mediated
   negative regulation of LCK Tyr^192 by selectively impairing PTPN22
   recruitment to and dephosphorylation of ZAP70. Our data are consistent
   with the reported model that LCK Tyr^192 works upstream LCK Tyr^505 to
   inhibit LCK ([172]14).

PTPN22 Ser^449 phosphorylation enhances T cell activation and modulates
effector functions by promoting dephosphorylation of LCK Tyr^192

   Next, we sought to elucidate the role of LCK Tyr^192 phosphorylation in
   the mechanism of action of phospho-Ser^449 PTPN22. Through potentially
   deeper inhibition of early TCR signaling and/or modulation of the ratio
   of LCK/ZAP70 activation, the retrograde inhibition of LCK by selective
   ZAP70 dephosphorylation observed in J.N22 449 cells could result in
   qualitative differences in the pathways that are ultimately activated
   by TCR engagement. We decided to take a multiomic approach to this
   problem by expressing monomeric enhanced green fluorescent protein
   (mEGFP) WT or S449A PTPN22 in J.CaM 1.6 cells engineered to reexpress
   LCK WT (J.Lck WT) or LCK Y192F (J.Lck Y192F) ([173]14), which we
   confirmed to be devoid of LCK-Tyr^192 phosphorylation (fig. S5A), and
   assessing downstream signaling in these cells through cellular indexing
   of transcriptomes and epitopes sequencing (CITE-seq) by 3′ GEM (Gel
   Bead-in-EMulsion) single-cell sequencing (scRNA-seq, 10× Genomics). To
   assess the intensity of TCR signal transduction at the single-cell
   level, we added adapters conjugated with an antibody against the
   TCR-induced molecule CD69, and to allow quantification of the PTPN22
   constructs via 3′ GEM scRNA-seq methods, 450–base pair (bp) nucleotides
   sequences at the C-terminal open reading frame were optimized via
   silent mutagenesis (table S3 and fig. S5A). We confirmed that
   mEGFP-tagged PTPN22 S449A retained a higher ability to inhibit the
   NFAT/AP-1 luciferase reporter activity compared with WT PTPN22 (fig.
   S5B).

   We next subjected the four TCR-stimulated cell groups (J.Lck WT N22 WT,
   J.Lck WT N22 449, J.Lck Y192F N22 WT, and J.Lck Y192F N22 449), plus a
   group of unstimulated cells expressing WT LCK and PTPN22 (J.Lck WT N22
   WT Unstim), to sorting for mEGFP^+ cells followed by CITE-seq (fig.
   S5A). We selected cells expressing optimized PTPN22 mRNA sequences for
   analysis after sequencing, which revealed three distinct cell clusters
   ([174]Fig. 6A) corresponding to J.Lck WT N22 WT Unstim (cluster 1,
   total 600), J.Lck WT (cluster 2, total 1955), and J.Lck Y192F (cluster
   3, total 1163) cells. The distribution of optimized PTPN22 expression
   was comparable among the cell groups ([175]Fig. 6B). Next, we split the
   cells expressing optimized PTPN22 at the median of PTPN22 expression
   and performed differential expression analysis with below the median
   cells, which showed more narrowly and homogeneously distributed PTPN22
   expression (rectangles in [176]Fig. 6B). Using these results, Gene Set
   Enrichment Analysis (GSEA) was performed with the Gene Ontology (GO)
   biological process database terms and normalized enrichment scores
   (NES) and significance values were calculated. Several pathways related
   to T cell responses were down-regulated (NES < −1.4, P-adj < 0.05) in
   cells expressing S449A PTPN22, which is consistent with the
   gain-of-function phenotype of PTPN22 S449A ([177]Fig. 6C), but we
   noticed that pathways down-regulated in cells expressing WT versus
   Y192F LCK were only partially overlapping. Pathways regulating antigen
   processing and presentation, T cell differentiation, molecules involved
   in T cell activation, adaptive immune responses, and regulation of cell
   adhesion were enriched in all cells expressing S449A PTPN22 (purple
   rows in [178]Fig. 6C). However, pathways related to antigen processing
   and presentation of endogenous antigen, regulation of T cell–mediated
   cytotoxicity, and response to IL-4 and TCR signaling were selectively
   enriched in cells expressing LCK WT (orange rows in [179]Fig. 6C),
   while pathways related to T cell activation involved in immune
   responses, IL-4 production, T cell–mediated immunity, and lymphocyte
   activation were enriched in cells expressing the Y192F mutant (blue
   rows in [180]Fig. 6C). Notably, even if the TCR activation gene pathway
   was inhibited in cells expressing either LCK WT or the Y192F mutant
   (black line, [181]Fig. 6C), the pattern of genes whose expression was
   inhibited by S449A PTPN22 depended on LCK Tyr^192 phosphorylation
   ([182]Fig. 6D). To assess whether the functional interaction between
   PTPN22 Ser^449 and LCK Tyr^192 phosphorylation depends on the intensity
   of T cell activation (which in turn affects the stoichiometry of
   Ser^449 phosphorylation), we leveraged the CD69 data obtained by
   CITE-seq. At the population level, CD69 expression was up-regulated in
   stimulated versus unstimulated cells ([183]Fig. 6E); however,
   consistent with PTPN22 S449A causing inhibition of TCR signaling
   pathway only in the presence of WT Lck (green line, [184]Fig. 6C), the
   expression of PTPN22 S449A correlated with lower expression of CD69 in
   J.Lck WT cells but not in J.Lck Y192F cells. In highly stimulated T
   cells (CD69 high, expression level over median value), LCK Tyr^192
   phosphorylation was required for S449A PTPN22 to down-regulate pathways
   involved in antigen processing and presentation (orange rows in
   [185]Fig. 6F) but inhibited the down-regulation of pathways associated
   with T cell activation, regulation of phosphoprotein phosphatase
   activity, IL-4 production, and TCR signaling (blue rows in [186]Fig.
   6F). In less stimulated T cells (CD69 low, expression level below the
   median value), several pathways including T cell–mediated cytotoxicity,
   T cell differentiation, and adaptive immune response were
   down-regulated by S449A PTPN22 independent of LCK phosphorylation
   (purple rows in [187]Fig. 6G). LCK Tyr^192 phosphorylation was required
   for S449A PTPN22 to down-regulate pathways related to antigen
   processing, TCR signaling, innate immune response, and antigen
   receptor–mediated signaling (orange rows in [188]Fig. 6G), and
   inhibited the down-regulation of pathways related to IL-8 production, T
   cell activation, lymphocyte activation, and cell adhesion caused by
   S449A PTPN22 (blue rows in [189]Fig. 6G). Together, these data confirm
   the gain-of-function nature of S449A PTPN22. LCK Tyr^192
   dephosphorylation quantitatively and qualitatively modulates the
   promotion of early TCR signaling operated by PTPN22 Ser^449
   phosphorylation over the full range of TCR activation, and the outcome
   of such feedback interaction is influenced by the intensity of TCR
   stimulation.

Fig. 6. PTPN22 Ser^449 phosphorylation enhances T cell activation and
modulates effector functions by promoting dephosphorylation of LCK Tyr^192.

   [190]Fig. 6.
   [191]Open in a new tab

   (A) Uniform Manifold Approximation and Projection (UMAP) of J.Lck cell
   clusters expressing optimized PTPN22 gene sequences. (B) Violin plots
   showing expression of optimized PTPN22 WT and S449A genes in single
   J.Lck WT or J.Lck Y192F cells; each dot represents a single cell;
   horizontal bars show median expression value. (C) Dot plot of GSEA GO
   pathway enrichment analysis of the differentially expressed genes
   (DEGs) in the optimized S449A versus WT populations of J.Lck WT and
   Y192F cells (P-adjust <0.05). NES, normalized enrichment score. (D)
   Heatmap of genes related to T cell activation enriched between
   optimized S449A and WT groups in J.Lck WT and Y192F cells (P < 0.05).
   (E) Violin plots showing the distribution of CD69 activation by
   CITE-seq; horizontal bars show median expression value. (F) Dot plot of
   GSEA GO pathway enrichment analysis of the DEGs in the optimized S449A
   versus WT populations of highly activated (as estimated by CD69
   expression over median value of 0.65) J.Lck WT or Y192F cells. (G) Dot
   plot of GSEA GO pathway enrichment analysis of the DEGs in the
   optimized S449A versus WT populations of less activated J.Lck WT or
   Y192F cells (CD69 expression of all expressing cells below median
   0.65). Specific enriched pathways are labeled in different colors
   (orange, purple, or blue).

Phosphorylation of PTPN22 Ser^449 facilitates T cell recognition of peptide
antigens presented by MHC

   PTPN22 is known to preferentially inhibit TCR responses to weaker
   ligands ([192]42). To assess the phospho-Ser^449–based PTPN22
   regulation mechanism across a range of ligand strengths, we studied
   J.OT1.hCD8^+ Jurkat cells ([193]43) stimulated with TCR ligands of
   varying affinity presented by T2-K^b cells ([194]44). We generated
   PTPN22 KO (J.OT1 N22 KO) and 3× FLAG tag KI PTPN22 (3× FLAG N22 J.OT1)
   J.OT1.hCD8^+ Jurkat cells via CRISPR-Cas9 editing (fig. S6, A and B).
   Next, we assessed PTPN22 Ser^449 phosphorylation in 3× FLAG N22 J.OT1
   cells stimulated by T2-K^b cells pulsed with ovalbumin (OVA) peptide
   (residues 257 to 264) and found it dramatically increased at 30 min
   (fig. S6C). We then repeated the experiment in 3× FLAG N22 J.OT1 cells
   stimulated by T2-K^b cells pulsed with an array of varying affinity
   peptides [OVA, Q4R7, T4, Q4H7, G4, beta catenin (Catnb), and vesicular
   stomatitis virus (VSV)] for 30 min (fig. S6D). Ser^449 phosphorylation
   was increasingly induced by higher affinity peptides (OVA, Q4R7, T4,
   and Q4H7), while lower affinity, self or control peptides (G4, Catnb,
   and VSV) induced similar lower levels of phosphorylation (fig. S6E). We
   next generated J.OT1 N22 449 (PTPN22 S449A) cells via CRISPR-Cas9
   editing (fig. S6, F and G) and stimulated these cells with T2-K^b
   pulsed with the abovementioned OVA peptides ([195]Fig. 7A).
   Phosphorylation of SLP-76 and extracellular signal–regulated kinase
   (ERK) were reduced in J.OT1 N22 449 cells with differences proportional
   to the affinity of the stimulating peptide and gradually growing from
   weak to strong peptides ([196]Fig. 7, B and C, and fig. S6H). We
   observed a similar pattern of affinity-correlated and enhanced versus
   reduced CD69 induction in J.OT1 N22 WT versus J.OT1 N22 449 cells,
   respectively; however, differences were notably absent in cells
   stimulated with the strongest ligand OVA ([197]Fig. 7D and fig. S6H).
   These data support the idea that PTPN22 Ser^449 phosphorylation
   contributes to the ability of cells to discriminate among ligands over
   a wide range of affinities and potencies. High stoichiometry
   phosphorylation and inhibition of PTPN22 by strong ligands might be
   consistent with the reported preferential enhancement of weak ligand
   signaling in PTPN22 KO cells ([198]42).

Fig. 7. Phosphorylation of PTPN22 Ser^449 facilitates T cell recognition of
peptide antigens presented by MHC.

   [199]Fig. 7.
   [200]Open in a new tab

   (A) Schematic illustration of the interaction of J.OT1N22 and J.OT1N22
   449 cells with antigen-presenting cells pulsed by multiple peptides. (B
   to D) Flow cytometry analysis of SLP-76, extracellular signal–regulated
   kinase (ERK) phosphorylation, and CD69 activation in J.OT1 N22 WT or
   J.OT1 N22 449 cells stimulated with or without 1 nM OVA, Q4R7, T4,
   Q4H7, G4, Catnb, or VSV control peptide–pulsed T2-K^b cells for the
   indicated time. Percentage of SLP-76 [Tyr^128 (B)] and ERK1/2
   [Thr^202/Tyr^204 (C)] phosphorylation was measured after stimulation
   for 1 and 5 min, respectively. The activation of CD69 (as % CD69^+
   cells) was assessed after 16 hours of stimulation (D). Data in
   histogram are presented as means ± SEM and are from at least three (B
   to D) independent experiments. Statistical significance was assessed by
   two-tailed Mann-Whitney test. *P < 0.05, **P < 0.01, ***P < 0.001. (A)
   created with BioRender ([201]69).

Phosphorylation of Ptpn22 Ser^452 promotes TCR signaling and adaptive
responses to immunization

   Next, to determine whether phosphorylation on Ser^449 (Ser^452 in mouse
   Ptpn22) regulates the function of Ptpn22 in vivo, we generated a mouse
   model carrying a Ptpn22^S452A mutation (fig. S7, A and B). We noticed
   that phosphorylation of Ptpn22 at Ser^452 selectively decreases the
   binding efficiency of a monoclonal PTPN22 antibody raised against a
   PTPN22 peptide surrounding Pro^451 (Cell Signaling Technology), while
   it does not affect the binding of a polyclonal Ptpn22 antibody raised
   against the P1 motif (amino acids 588 to 654) (fig. S7C, top)
   ([202]17). Using these two antibodies, we could confirm that Ser^452
   phosphorylation was eliminated by the Ptpn22^S452A mutation without
   alterations in Ptpn22 expression (fig. S7C, bottom). We then proceeded
   to assess TCR signaling in CD4^+ effector T cells from homozygous
   Ptpn22^S452A versus Ptpn22^WT mice. Consistent with the observations in
   human cells, the phosphorylation of Zap70 Tyr^319 and its downstream
   substrate Plcϒ Tyr^783 in CD4^+ effector T cells of Ptpn22^S452A mice
   were inhibited compared to cells from Ptpn22^WT mice ([203]Fig. 8A),
   while phosphorylation of Lck Tyr^394 was unaffected. Lck Tyr^192
   phosphorylation was up-regulated in Ptpn22^S452A mice at time 0 and
   then down-regulated by TCR stimulation ([204]Fig. 8A). TCR-induced Cd69
   expression in the Ptpn22^S452A mice was notably inhibited after 4 hours
   of stimulation ([205]Fig. 8B). PTPN22 is known to inhibit TCR-induced T
   cell proliferation ([206]17); thus, we further evaluated proliferation
   of CD4^+ effector T cells from Ptpn22^WT or Ptpn22^S452A mice by
   CellTrace Violet staining. As shown in [207]Fig. 8C, TCR-induced CD4^+
   T cell proliferation was considerably reduced in Ptpn22^S452A mice.
   Consistent with the finding in Jurkat cells ([208]Fig. 2G), we also
   found Rap1 activity to be reduced in CD4^+ effector T cells from
   Ptpn22^S452A mice when compared to Ptpn22^WT T cells after TCR
   stimulation ([209]Fig. 8D). In response to stimulation with OVA peptide
   (residues 329 to 336, specific for OT-II^+ TCR) pulsed splenocytes,
   OT-II^+ CD4^+ T cells from Ptpn22^S452A mice displayed blunted
   induction of Cd69, and a similar Ptpn22 gain-of-function phenotype was
   also observed when T cells from OT-II^+ CD4^+ T cells from Ptpn22^S452A
   mice were stimulated with E336Q peptide (a partial agonist of OVA
   peptide) pulsed splenocytes. In these assays, cells stimulated with
   self-peptide Class II-associated invariant chain peptide (CLIP) pulsed
   splenocytes did not induce any Cd69 activation (fig. S7, D to F).

Fig. 8. Phosphorylation of Ptpn22 Ser^452 promotes TCR signaling and adaptive
responses to immunization.

   [210]Fig. 8.
   [211]Open in a new tab

   (A) Protein phosphorylation induced by TCR stimulation in CD4^+
   effector T cells from Ptpn22^WT (n = 3) or Ptpn22^S452A mice (n = 3)
   analyzed by Western blotting (left) and quantification of
   phosphorylated Lck, Zap70, and Plcϒ normalized to total proteins
   (right). (B) Flow cytometry analysis of Cd69 activation in CD4^+
   effector T cells from Ptpn22^WT (n = 4) or Ptpn22^S452A (n = 4) mice
   stimulated with or without antibodies against mouse CD3/CD28. (C) Flow
   cytometry analysis of proliferation of CD4^+ effector T cells from
   Ptpn22^WT (n = 4) or Ptpn22^S452A (n = 4) mice. (D) Endogenous Rap1
   activity in CD4^+ effector T cells from Ptpn22^WT (n = 3) or
   Ptpn22^S452A (n = 3) mice was assessed through Ral GDS pull down and
   detected by immunoblotting, and quantification of the precipitated Rap1
   was normalized to WT at time 0. (E) Changes in the absolute numbers of
   CD4^+ and CD8^+ T cells with naïve (CD44^lo CD62L^hi), effector
   (CD44^hi CD62L^lo), and central memory (CD44^hi CD62L^hi) phenotypes in
   spleens of aged (over 6 months) Ptpn22^WT (n = 11) or Ptpn22^S452A
   (n = 9) mice. (F) An illustration of OVA immunization of Ptpn22^WT or
   Ptpn22^S452A mice. (G) OVA-specific IgGs were quantified in 8-week
   Ptpn22^WT (n = 9) or Ptpn22^S452A (n = 14) mice immunized by
   subcutaneous injection at day 10 (priming) and 21 (boosting). (H)
   Numbers of Tfh (Tcrβ^+ CD4^+ PD-1^+ Bcl6^+), GC B cells (Tcrβ^− CD19^+
   GL7^+ Bcl6^+), and T[reg] cells (Tcrβ^+ CD4^+ CD25^+ FoxP3^+) in
   draining lymph nodes of OVA-immunized Ptpn22^WT (n = 7) or Ptpn22^S452A
   (n = 12) mice were analyzed by flow cytometry at 21 days after primary
   immunization. Representative blots and flow gating are shown (A to D,
   E, and H). Data represent means ± SEM, and statistical significance was
   assessed by two-way ANOVA followed by Bonferroni’s post hoc test (A, B,
   and D) and two-tailed Mann-Whitney test (C, E, G, and H). *P < 0.05,
   **P < 0.01, ***P < 0.001.

   Increased positive thymic selection has been reported in Ptpn22 KO mice
   ([212]17), and Tyr^192 phosphorylation impairs thymic reconstitution of
   retrogenic mice ([213]14). However, when we analyzed the thymus of
   4-week-old Ptpn22^WT or Ptpn22^S452A mice, we did not find alterations
   of double-negative, CD4 or CD8 single-positive, or double-positive
   thymocyte populations in Ptpn22^S452A mice (fig. S8A). KO of Ptpn22 in
   mice enhances responses to immunization and leads to a marked expansion
   of effector memory CD4 and CD8 T cells in aged (>6 months old) animals
   ([214]16, [215]17). We immunophenotyped T cells in spleen and lymph
   nodes of >6-month-old Ptpn22^S452A versus littermate WT mice and found
   that aged Ptpn22^S452A mice displayed reduced numbers of CD4^+ and
   CD8^+ central memory T cells in spleens ([216]Fig. 8E), a phenotype
   that was not evident in lymph nodes (fig. S8, B and C). Next, we
   immunized Ptpn22^WT or Ptpn22^S452A mice with OVA ([217]Fig. 8F) and
   observed decreased titers of anti-OVA immunoglobulin G (IgG) antibodies
   in Ptpn22^S452A mice after both priming and boosting injections
   ([218]Fig. 8G). Immunophenotyping of draining lymph nodes after
   boosting showed that CD4^+ T follicular helper (Tfh) and germinal
   center (GC) B cell frequencies were highly reduced in the draining
   lymph nodes of Ptpn22^S452A mice, while there was no change in
   regulatory T cell (T[reg]) numbers ([219]Fig. 8H and fig. S8D).
   Together, these results support the notion that Ptpn22 phospho-Ser^452
   reduces the inhibitory function of Ptpn22 on TCR signaling in vivo.

Ptpn22 Ser^452 phosphorylation promotes disease severity in a T
cell–dependent lupus mouse model

   Given that PTPN22 phospho-Ser^449 functionally regulates PTPN22 to
   promote T cell responses and PTPN22 has a strong genetic association
   with SLE, we next sought to collect data to support a role of Ser^449
   phosphorylation in SLE pathogenesis by subjecting Ptpn22^S452A mice to
   a T cell–dependent inducible model of SLE ([220]45, [221]46). Repeated
   topical treatment of C57BL/6 mice with resiquimod (R848) induces T cell
   activation, splenomegaly, B cell expansion with hypergammaglobulinemia,
   IgG deposition in kidney, and overt proteinuria (see outline in
   [222]Fig. 9A). When subjected to the R848 lupus model, Ptpn22^S452A
   mice displayed markedly attenuated proteinuria level ([223]Fig. 9B)
   correlating with reduced deposition of IgG in kidney glomeruli,
   pointing to protection against lupus nephritis ([224]Fig. 9C). The
   Ptpn22^S452A mutation did not attenuate R848-induced splenomegaly
   ([225]Fig. 9D). Immunophenotyping of the spleen showed comparable
   numbers of splenocytes in Ptpn22^S452A versus Ptpn22^WT mice ([226]Fig.
   9D), and there were no changes in the total number of B cells (Tcrβ^−
   CD19^+) or GC B cells (Tcrβ^− CD19^+ GL7^+ Bcl6^+) ([227]Fig. 9E and
   fig. S8E). Numbers of total splenic T cells, CD4^+ (Tcrβ^+ CD4^+) T
   cells, CD4^+ naïve (CD44^lo CD62L^hi) T cells, and CD4^+ effector
   (CD44^hi CD62L^lo) T cells were also comparable between Ptpn22^S452A
   versus Ptpn22^WT mice ([228]Fig. 9F and fig. S8F); however, Tfh (Tcrβ^+
   CD4^+ PD-1^+ Bcl6^+) and T[reg] cells (Tcrβ^+ CD4^+ CD25^+ Foxp3^+)
   were strongly reduced in the spleen of Ptpn22^S452A mice ([229]Fig. 9G
   and fig. S8G). Together, these data suggest that Ptpn22 Ser^452
   phosphorylation enhances T cell responses to promote SLE pathogenesis.

Fig. 9. Ptpn22 Ser^452 phosphorylation promotes disease severity in a T
cell–dependent lupus mouse model.

   [230]Fig. 9.
   [231]Open in a new tab

   (A) Schematic illustration of the 28-day R848 treatment protocol
   performed on Ptpn22^WT (n = 8) or Ptpn22^S452A (n = 8) mice. (B)
   Proteinuria score of Ptpn22^WT or Ptpn22^S452A mice treated with R848
   for 28 days and measured every 7 days. (C) Periodic acid–Schiff (PAS)
   staining of kidney sections and immunohistochemical staining of IgGs in
   kidney glomeruli from R848-treated Ptpn22^WT or Ptpn22^S452A mice.
   Images were taken at 10× magnifications. Scale bars, 50 μm. IgG
   intensity in the glomeruli was quantified in each mouse (right). (D)
   Spleen/body weight ratio and total splenocytes comparison between
   Ptpn22^WT and Ptpn22^S452A R848-treated mice. (E to G) Splenocytes
   analysis by flow cytometry was performed on cells collected at day 28
   from Ptpn22^WT or Ptpn22^S452A mice, and total numbers of target cell
   populations were quantified. Numbers of total B cells (Tcrβ^− CD19^+),
   GC B cells (Tcrβ^− CD19^+ GL7^+ Bcl6^+) (E); total T cells (Tcrβ^+),
   CD4^+ T cells (Tcrβ^+ CD4^+), CD4^+ CD8^+ T cells (Tcrβ^+ CD4^+CD8^+),
   CD4^+ naïve (Tcrβ^+ CD4^+CD44^lo CD62L^hi), CD4^+ effector (Tcrβ^+
   CD4^+CD44^hi CD62L^lo) T cells (F), Tfh (Tcrβ^+ CD4^+ PD-1^+ Bcl6^+),
   and T[reg] cells (Tcrβ^+ CD4^+ CD25^+ FoxP3^+) (G) were quantified.
   Data are presented as means ± SEM, and statistical significance was
   assessed by two-way ANOVA followed by Bonferroni’s post hoc test (B)
   and two-tailed Mann-Whitney test (C to G). *P < 0.05, **P < 0.01.

DISCUSSION

   PTPN22, a key negative regulator of TCR signaling, is encoded by a gene
   whose R620W variation is strongly associated with SLE ([232]47,
   [233]48). Here, through phosphoproteomics analysis, we identified
   Ser^449 as a major TCR-induced PKA phosphorylation site of PTPN22 in T
   cells. PKA contributes to modulating signaling downstream TCR
   engagement in response to a variety of concurrent stimuli ([234]49),
   and we show that PKA activity is induced in T cells by early TCR
   engagement. We find that PKA Cα can directly phosphorylate PTPN22 on
   Ser^449 and is required for TCR-induced PTPN22 phosphorylation. We also
   show that PBMCs from patients with SLE and active disease display
   elevated Ser^449 phosphorylation. Our data about the function of
   Ser^449 phosphorylation in TCR signaling and the phenotype of
   Ptpn22^S452A mice suggest that the hyperphosphorylation of Ser^449
   found in SLE might contribute to the pathogenesis of T cell
   hyperactivation reported in SLE and in disease mechanism ([235]50). The
   protective effect of gain-of-function Ptpn22^S452A against experimental
   lupus is also consistent with various reports suggesting that the
   lupus-associated PTPN22-R620W variation alters or reduces the function
   of PTPN22 ([236]16). The hyperphosphorylation of PTPN22 observed in
   lupus PBMCs could be secondary to the intrinsic abnormalities of PKAs
   and alterations of the TCR signaling complex that have been reported in
   SLE T cells ([237]2, [238]28). However, cytokine signaling
   dysregulation due to disease activity could also play a role and, for
   example, IFN-γ can activate PKA signaling and is increased in active
   SLE ([239]51, [240]52). Further studies are needed to fully clarify
   regulation of PTPN22 Ser^449 by PKA in healthy and lupus cells, whether
   the hyperphosphorylation of Ser^449 is limited to lupus T cells and its
   relationship with disease severity in SLE.

   To explore effects of PTPN22 Ser^449 phosphorylation on TCR signaling,
   we generated a PTPN22 S449A Jurkat KI line by CRISPR-Cas9–mediated
   genome editing. Analysis of these cells compared to cells carrying WT
   PTPN22 showed a reduction in TCR-induced phosphorylation of activating
   residues of proximal TCR signaling mediators ZAP70, SLP-76, and PLCϒ,
   and reduced expression of CD69, IL-2, and TNF-α, suggesting that PTPN22
   Ser^449 phosphorylation enhances early TCR signaling and T cell
   activation. To more deeply characterize the role of PTPN22 Ser^449
   phosphorylation in regulation of T cell function, we also generated a
   KI mouse model carrying the Ptpn22 S452A mutation. Concordant with the
   effect of Ser^449 mutation in human T cells, loss of this
   phosphorylation site in mice led to reduced TCR signaling, decreased
   numbers of CD4^+ and CD8^+ central memory T cells in spleens of aged
   mice, and reduced antigen-specific IgG antibodies following priming and
   boost immunizations, suggesting diminished peripheral T cell responses.
   These phenotypes are mostly opposite to the ones reported for mice
   carrying deletion or R619W mutation of Ptpn22, with the exception of
   thymocyte counts that were not affected in Ptpn22 S452A mice. The
   latter could be due to potential repertoire-based or other compensation
   mechanisms in the context of PTPN22 gain-of-function and is consistent
   with the previously reported phenotype of mice overexpressing human
   PTPN22 ([241]53). Ptpn22^S452A mice also showed decreased disease
   progression in the R848-induced model of lupus when compared to WT
   mice. The splenic T cell phenotypes from this lupus model, combined
   with the abovementioned immunophenotype of healthy Ptpn22^S452A mice
   suggest that reduced TCR signaling underlies at least in part the
   observed protection from disease. However, our KI mouse model causes a
   global S452A mutation of PTPN22, thus a contribution of other immune
   cells cannot be ruled out at the moment. As Ptpn22 also regulates IFN-α
   signaling ([242]54), additional studies are also warranted to clarify
   whether Ser^449 affects additional signaling pathways that are relevant
   to the pathogenesis of lupus in humans and in R848-induced lupus-like
   disease.

   We observed increased phosphorylation of PTPN22 Ser^449 in J.OT1 cells
   stimulated with higher affinity OVA-derived peptide ligands, suggesting
   that induction of phosphorylation at this site is dependent on the
   strength of TCR signaling. This may in part explain the observed
   greater effect of PTPN22 Ser^449 phosphorylation on OVA-derived peptide
   ligand–induced phosphorylation of early TCR signaling mediators. In
   mouse OT-II–expressing CD4^+ T cells, Ptpn22 S452A suppressed signaling
   to a greater extent than WT Ptpn22 when cells were stimulated with
   full-strength OVA peptide, while in human OT-I Jurkat T cells, there
   was no effect of PTPN22 S449A on TCR signaling in response to OVA
   peptide. This may be due to a higher affinity of OVA for OT-I versus
   OT-II TCRs, causing a greater strength of stimulation that overcame the
   effect of PTPN22 phosphorylation in OVA-stimulated OT-I cells. Thus, in
   vivo Ser^449 phosphorylation might contribute to the reported role of
   PTPN22 in discriminating strong versus weak agonists and enhancing
   signaling induced by stimuli of weak and intermediate strength.

   Since PTPN22 S449A cells exhibit impaired TCR signaling and downstream
   T cell activation, we initially assumed that phosphorylation on this
   site directly inhibited the catalytic activity of PTPN22. However, we
   had noted that S449A cells display equivalent TCR-induced
   phosphorylation of LCK Tyr^394 and PTPN22 immunoprecipitated from
   TCR-stimulated S449A cells showed equivalent phosphatase activity to
   PTPN22 immunoprecipitated from WT cells, suggesting that
   phosphorylation affects the ability of PTPN22 to bind and/or
   dephosphorylate ZAP70. Compared to WT PTPN22, PTPN22 S449A showed
   enhanced colocalization in TCR-stimulated cells with ZAP70 but not LCK.
   Furthermore, in intracellular dephosphorylation assays in COS-7 cells,
   expression of PTPN22 S449A had the same effect as WT PTPN22 on the
   phosphorylation of coexpressed LCK Tyr^394 yet led to reduced
   phosphorylation of coexpressed ZAP70 on Tyr^319 and Tyr^493. These
   results are likely due to a conformational change in phosphorylated
   PTPN22, as PTPN22 S449A showed enhanced FRET compared to WT PTPN22 when
   fused at the N and C termini to cyan fluorescent protein and yellow
   fluorescent protein, respectively. The location of the Ser^449
   phosphorylation site within the PTPN22 interdomain may facilitate an
   altered PTPN22 structure upon phosphorylation that lowers the affinity
   for ZAP70 while keeping the structural determinants needed for LCK
   binding intact. We do not exclude the possibility that Ser^449
   phosphorylation operates through a more indirect mechanism or affects
   recognition of other PTPN22 substrates.

   In our examinations of the effect of PTPN22 phosphorylation on LCK
   activation state, we uncovered the unexpected finding that LCK Tyr^192
   is more phosphorylated in T cells where PTPN22 Ser^449 phosphorylation
   is abolished. Although LCK Tyr^192 has been reported to be
   phosphorylated by ITK ([243]15), our data suggest that in our T cell
   system, LCK Tyr^192 is primarily an autophosphorylation site, as
   Tyr^192 phosphorylation is unaffected by cell treatment with an ITK
   inhibitor, reduced on the inactive Y394F LCK mutant, and strongly
   enhanced on the hyperactive Y505F LCK mutant. However, it is possible
   that phosphorylation of Tyr^192 is under ITK control in specific T cell
   populations or conditions and that a similar TCR-driven retrograde loop
   could be promoted by an LCK-ITK mechanism rather than LCK alone in
   different contexts. Our data also point to a role for both ZAP70 and
   CD45 in the negative regulation of LCK Tyr^192 phosphorylation state.
   Knockout or knockdown of ZAP70 and knockdown of CD45, enhances Tyr^192
   phosphorylation. Cells carrying PTPN22 S449A also showed reduced
   colocalization of CD45 with LCK. Together, these data suggest a model
   by which during TCR stimulation, LCK Tyr^192 is autophosphorylated,
   which would depress TCR signaling. However, CD45 acts as a key negative
   regulator of LCK Tyr^192 phosphorylation—possibly through direct
   dephosphorylation—through a mechanism that is enhanced by ZAP70
   activation. Phosphorylation of PTPN22 Ser^449 impairs its ZAP70
   dephosphorylation ability, which results in enhanced CD45-mediated
   inhibition of LCK Tyr^192 phosphorylation, and promotes strength of TCR
   signaling and T cell activation in a retrograde positive feedback loop.

   A question that remains open is how the enhanced activation of ZAP70
   upon PTPN22 Ser^449 phosphorylation leads to increased
   dephosphorylation of LCK Tyr^192 by CD45. We demonstrate through PLA
   that CD45 colocalization with LCK is regulated by PTPN22 Ser^449
   phosphorylation, suggesting that ZAP70 activation induces recruitment
   of CD45 to LCK. Whether this occurs through direct phosphorylation of
   CD45, CD45 complexing with phosphorylated adaptor proteins such as
   lymphocyte phosphatase–associated phosphoprotein ([244]55) and/or other
   mechanisms remains to be clarified.

   Our proposed retrograde feedback mechanism prompts the question of how
   relevant is the dephosphorylation of an inhibitory site on LCK induced
   by the activation of its downstream effector. The retrograde loop
   likely underlies the complete absence of phosphorylation of LCK Tyr^192
   in PTPN22 KO cells. However, interpreting the interaction between LCK
   Y192F and PTPN22 KO would be complicated by the considerable LCK
   hyperactivation caused by PTPN22 KO, which might obliterate the
   modulatory role of the CD45-mediated loop. We thus opted to study the
   interaction between LCK Y192F and PTPN22 S449A at the single-cell level
   using CITE-seq of cells expressing combination of mutants. The results
   of our analysis support the gain-of-function nature of S449A and
   previous data about the role of LCK Tyr^192 in TCR signaling ([245]14).
   They also suggest that LCK Tyr^192 phosphorylation not only is
   important to sustain PTPN22 Ser^449–phosphorylation-mediated promotion
   of TCR signaling but also modulate its outcome in terms of which
   pathways are preferentially promoted by Ser^449 phosphorylation. We
   also show that, as expected, such differences are further modulated by
   the strength of TCR signaling, likely in part through control of
   Ser^449 phosphorylation stoichiometry. Further studies are needed to
   understand whether such qualitative differences in signaling induced by
   the cross-talk between LCK Tyr^192 and PTPN22 Ser^449 phosphorylation
   have physiological relevance and whether they depend on overall
   intensity of TCR signaling downstream ZAP70 or the ratio of LCK/ZAP70
   activation plays a role.

   We believe that our data are sufficient to conclude that TCR-induced
   autophosphorylation of LCK Tyr^192 balanced by PTPN22
   phospho-Ser^499–promoted dephosphorylation of the same site by CD45
   allows for fine feedback regulation of LCK versus ZAP70 activation in
   response to stimuli of various strength, as well as a variety of
   additional variables. Our study also highlights the “linchpin role” of
   CD45 in early TCR signaling as the key phosphatase that regulates all
   three major phosphorylation sites of LCK in response to various cues.
   The opposite phenotypes we observed in CD45 KO versus knock-down cells
   is consistent with previous studies highlighting that CD45 complex
   function in TCR signaling can hardly be recapitulated by studies in
   cells completely lacking CD45 ([246]40).

   In summary, through a combination of phosphoproteomic, cell-based, in
   vivo, and multiomic approaches, we report Ser^449 phosphorylation of
   PTPN22 as a key mechanism by which T cells enhance and modulate the
   outcome of TCR-induced signals through activation of a ZAP70-CD45-LCK
   signaling loop. Our model ([247]Fig. 10, A and B) is that PKA-mediated
   phosphorylation of PTPN22 on Ser^449 selectively inhibits PTPN22
   binding to and dephosphorylation of ZAP70 on Tyr^319 and Tyr^493,
   resulting in dephosphorylation of LCK Tyr^192 by CD45 and enhanced
   activation T cells. PTPN22 Ser^449 phosphorylation contributes to
   support proximal TCR signaling, peripheral T cell activation, and
   promotion of SLE pathogenesis.

Fig. 10. Current working model.

   [248]Fig. 10.
   [249]Open in a new tab

   (A) PTPN22 Ser^449 is phosphorylated by PKA and induces a PTPN22
   conformational change that selectively lowers the affinity of the
   phosphatase for phosphorylated ZAP70. The consequent
   hyperphosphorylation of ZAP70 promotes a ZAP70-CD45 feedback loop,
   which leads to dephosphorylation of LCK Tyr^192, which has been shown
   to facilitate dephosphorylation of inhibitory Tyr^505. The overall
   effect of PTPN22 phospho-Ser^449 is to promote TCR signaling and
   autoimmunity, and it is abolished by the S449A mutation (B), which
   causes gain-of-function inhibition of TCR signaling and protects mice
   from autoimmunity. Schematic created with BioRender ([250]69).

   In addition to PTPN22, several other protein phosphatases including
   PP2A ([251]56) and DUSP22 ([252]57) have recently emerged as potential
   pathogenic players in SLE T cells. Further investigation into the
   regulation of PTPN22 and these other phosphatases in T cell is
   warranted to understand the complex intricacies of how T cells are
   activated and how TCR signaling pathways can be exploited to develop
   targeted immunotherapies for autoimmune diseases.

MATERIALS AND METHODS

Cells

   Human PBMCs were isolated from peripheral blood [obtained from La Jolla
   Institute for Immunology's in-house Normal Blood Donor Program (NBDP),
   with EDTA] using SepMate tubes with Lymphoprep (STEMCELL Technologies)
   according to the manufacturer’s instructions. The human CD4^+ T cells
   were enriched using the EasySep Human CD4^+ T Cell Isolation Kit
   (STEMCELL Technologies) and expanded using Dynabeads Human T-Activator
   CD3/CD28 (Thermo Fisher Scientific) in the presence of recombinant
   human IL-2 (20 ng/ml; R&D Systems) for 5 days. J.OT1.hCD8^+ cells,
   J.Csk^AS/CD45, J.Lck WT, and J.Lck Y192F were gifts from A. Weiss
   (University of California, San Francisco) ([253]14, [254]40, [255]58).
   These and the other Jurkat-derived cell lines were maintained in RPMI
   1640 medium supplemented with 10% fetal bovine serum (FBS), 1 mM sodium
   pyruvate, 2 mM l-glutamine, d-glucose (2.5 mg/ml), 10 mM Hepes,
   penicillin (100 units/ml), and streptomycin (100 μg/ml). All the Jurkat
   and its derivative cell lines are summarized in table S4. Mouse CD4^+ T
   cells were isolated from pooled spleens and lymph nodes by EasySep
   Mouse CD4^+ T cell isolation kit (STEMCELL Technologies) and cultured
   with Dynabeads mouse T-Activator CD3/CD28 (Thermo Fisher Scientific) in
   complete RPMI 1640 medium supplemented with recombinant mouse Il-2 (10
   ng/ml) for 3 days. HEK293T and COS-7 cell lines were purchased from
   American Type Culture Collection and cultured in Dulbecco’s modified
   Eagle’s medium supplemented with 10% FBS. All cells were kept under 5%
   CO[2] at 37°C.

Mice

   Mice were maintained in a specific pathogen–free facility, and all
   experiments were carried out in accordance with the Institutional
   Animal Care and Use Committee–approved protocols at University of
   California, San Diego (no. S16098 and no. [256]S17160) or Cedars Sinai
   Medical Center (CSMC, No. 010724). Ptpn22^S452A KI mice on the C57BL/6
   genetic background were generated using CRISPR-Cas9 technology at the
   Transgenic Mouse Model Core Facility of the National Core Facility for
   Biopharmaceuticals, National Science and Technology Council, and the
   Animal Resources Laboratory of National Taiwan University Centers of
   Genomic and Precision Medicine (Taiwan). Briefly, the single guide RNA
   (sgRNA) sequence with a protospacer adjacent motif (PAM) site (NGG)
   (5′-ctctgctgaatcagttcaaa-3′) was selected following the CRISPick
   ([257]59) and the Cas-OFFinder online resources ([258]60). The sequence
   for the repairing template (including 5′ and 3′ homologous sequences
   and the codon of the S452A mutation -TCA > GCT, serine to alanine) was
   5′-tcgtgtcctagtgctctgcccataaacacagcggacaggtatcacaattcaaaggggccggtaaaacg
   gaccaaagctactcccttcgaattgattcagcagagaaaaacaaatgacttggccgtgg-3′, (GCT
   was bold italicized), including an artificially inserted sequence for
   the BstBI restriction enzyme site (TTCGAA) and the synonymously mutated
   TTG codon for leucine at amino acid residue 457 (CTG > TTG) to
   facilitate recombination efficiency and future genotyping (Integrated
   DNA Technologies Inc.). The sgRNA and the recombinant Cas9 protein were
   purchased from Synthego Corporation. Electroporation was performed with
   premixed sgRNA and Cas9 added to the repairing DNA oligonucleotide
   template and fertilized eggs from C57BL/6J mice, which were then
   electroporated on a BEX electroporation apparatus (CUY21EDIT2) set at
   pulse three times; current, 165 mA; voltage, 30 V; and resistance, 160
   kilohm. The primers for genotyping polymerase chain reaction (PCR)
   amplification and BstBI digestion are as follows: VF1:
   5′-ATGTGAAAACGACAAACCAGCA-3′, VR1: 5′-CAGCACTATTCGGAGGGGATGA-3′. To
   minimize off-target effects, the Ptpn22^S452A mice were backcrossed
   with C57BL/6J mice for six generations before being used for
   experimentation. OT II mice [JAX 004194, B6.Cg-Tg(TcraTcrb)425Cbn/J]
   and RAG2 KO mice (JAX 008449, B6.Cg-Rag2tm1.1Cgn/J) were obtained from
   the Jackson Laboratory. Genotyping of mice was performed by Transnetyx
   (Cordova, TN).

Patient samples

   All patients with SLE (n = 6) [as per the American College of
   Rheumatology diagnostic criteria ([259]61)] and healthy controls (table
   S1) were recruited at CSMC. All participants provided informed written
   consent, and the study was approved in advance by the CSMC
   institutional review board (protocol no. 19627). SLEDAI scores ranged
   from 2 to 9 at the time of enrollment.

Plasmids, antibodies, reagents and primers

   C-terminal triple FLAG tagged (3× FLAG) full-length PTPN22 WT plasmid
   has been described ([260]24) and its derived mutants were constructed
   by site-directed mutagenesis or overlap extension PCR using the primers
   in table S2. Custom-made PTPN22 FRET, mEGFP-tagged codon-optimized WT
   or S449A PTPN22 plasmids were purchased from GenScript (Piscataway,
   NJ). pD1321-AP WT Cas9 plasmid with red fluorescent protein (RFP)
   selection marker containing the guide RNA (gRNA) sequences was
   purchased from DNA 2.0 (ATUM). NFAT/AP-1 firefly luciferase and Renilla
   luciferase reporters were purchased from Promega (Madison, WI). All
   plasmids were confirmed by DNA sequencing.

   The polyclonal rabbit anti-human phospho-PTPN22 Ser^449–specific
   antibody was custom-made by Pacific Immunology (Ramona, CA). Rabbit
   anti-mouse Ptpn22 P1 domain antibody was from Genentech (San Francisco,
   CA). Goat anti-human PTPN22 polyclonal antibody (AF3428) was purchased
   from R&D systems (Minneapolis, MN). Rabbit anti-PTPN22 (D6D1H),
   anti–phospho-Src family (Tyr^416) (D49G4), anti-LCK (73A5, D88),
   anti–phospho-ZAP70 (Tyr^493) (2704), anti–phospho-ZAP70 (Tyr^319)
   (65E4), anti-ZAP70 (D1C10E), anti–phospho–PLC-γ (Tyr^783) (D6M9S),
   anti–PLC-γ (D9H10), anti–phospho-CREB (Ser^133) (87G3), anti-CREB
   (48H2), anti-PKA Cα (4782), anti-CSK (C74C1), and
   anti–glyceraldehyde-3-phosphate dehydrogenase (D16H11) antibodies were
   from Cell Signaling Technology (Boston, MA). Mouse anti-FLAG M2 (F3165)
   and anti–β Actin (AC15) antibodies were purchased from Millipore Sigma
   (Burlington, MA). Rabbit anti-p56lck/LCK (phospho-Tyr^192) (LS-C199194)
   antibody was from Lifespan Biosciences. Ultra-LEAF purified human CD3
   antibody (clone, OKT3), Ultra-LEAF purified human CD28 antibody (clone,
   CD28.2), Ultra-LEAF purified anti-mouse CD3ε (clone, 145-2C11), and
   Ultra-LEAF purified anti-mouse CD28 (clone, 37.51) were purchased from
   BioLegend (San Diego, CA). The polyclonal goat anti-mouse Ig (553998)
   and mouse anti-Syrian/Armenian Hamster IgG (550637) used for
   cross-linking were from BD Biosciences (Carlsbad, CA). Horseradish
   peroxidase–conjugated anti-rabbit IgG (95017-556), anti-mouse IgG
   (76835-688), and anti-goat IgG (103262-184) were purchased from VWR
   (Radnor, PA).

   Anti-FLAG M2 magnetic beads, dimethyl sulfoxide (DMSO),
   β-mercaptoethanol and recombinant human PKA Cα protein were purchased
   from Millipore Sigma (Burlington, MA). Pierce protease inhibitor
   tablets (EDTA-free), Dynabeads human or mouse T-Activator CD3/CD28
   magnetic beads, DiFMUP, SuperScript III First-Strand Synthesis
   SuperMix, Platinum PCR SuperMix, and Phos-tag Acrylamide were purchased
   from Thermo Fisher Scientific (Rockford, IL). 2× Laemmli sample buffer
   was from Bio-Rad (Hercules, CA). Protein G Sepharose 4 Fast Flow was
   purchased from GE Healthcare (Chicago, IL). The dual-luciferase
   reporter assay system was purchased from Promega (Madison, WI).
   Endofree plasmid Maxi kit, QIAquick Gel Extraction kit, RNeasy Plus
   Micro Kit, and DNeasy Blood & Tissue kit were purchased from QIAGEN
   (Redwood City, CA). Mouse OVA sIgG (OVA-specific IgG) enzyme-linked
   immunosorbent assay kit was purchased from MyBioSource (San Diego, CA).
   OVA and relative mutated or control peptides were ordered or
   synthesized from GenScript (Piscataway, NJ). Human PKA, ZAP70, CSK,
   CD45, and nontargeted siRNA were purchased from Horizon Discovery
   (Lafayette, CO). Recombinant human or mouse cytokines were from R&D
   systems (Minneapolis, MN). All primers used in the study were
   synthesized by Integrated DNA Technologies.

Kinase prediction

   Kinase prediction was performed using NetPhos 3.1 software, with a
   prediction score of 0.5 or higher indicating a positive hit ([261]62,
   [262]63).

Inhibitors and cell treatment

   PKA (H89 2HCl), PKC (Sotrastaurin, GÖ 6983, Staurosporine, Ro 31-8220),
   RSK (LJH685), and ITK (BMS-509744) inhibitors were from SelleckChem
   (Houston, TX). For cell treatments, the cells were pretreated with the
   inhibitors for 30 min at 37°C and then stimulated with or without
   antibodies (1 μg/ml) against human CD3/CD28 in the presence of
   inhibitors or DMSO for the indicated times.

CRISPR-Cas9 strategies for generation of PTPN22 S449A KI cell line

   The 3× FLAG KI PTPN22 or PTPN22 KO Jurkat cell lines have been
   described ([263]24). The 3× FLAG KI PTPN22 or PTPN22 KO in J.OT1.hCD8^+
   cells were constructed as previously described ([264]24). These were
   used as a parent line to generate PTPN22 S449A KI Jurkat or
   J.OT1.hCD8^+ cells as follows: gRNA targeting PTPN22 genomic DNA
   nucleotides sequences around Ser^449 in Exon 13 was designed and cloned
   into pD1321-AP WT Cas9 plasmid. Plasmids containing about 834-bp
   homology-directed repair (HDR) arms with S449A mutation were ordered
   from GenScript (Piscataway, NJ). The HDR fragments were amplified by
   PCR using Platimum PCR SuperMix and purified by QIAquick Gel Extraction
   kit. The double-stranded DNA fragments were concentrated and
   quantitated by Nanodrop. Ten micrograms of Cas9 plasmid and 10 μg of
   HDR fragments carrying the S449A mutation were electroporated into WT
   or J.OT1 WT Jurkat cells for 24 hours. RFP^+ cells were sorted into
   96-well round-bottom plates using a SONY SH800 cell sorter; genomic DNA
   was extracted by DNeasy Blood& Tissue kit and used as template to
   screen positive clones by allele-specific PCR amplification. The
   mutation on mRNA was further confirmed by sequencing DNA synthesized by
   SuperScript III First-Strand Synthesis SuperMix using RNA extracted
   from positive candidate clones (RNeasy Plus Micro Kit) as template. The
   primers for HDR arms sequences amplification, positive candidates
   screening (allele specific primers), and CspCI digestion were
   summarized in table S2.

In vitro phosphorylation and phosphatase activity assays

   In vitro kinase phosphorylation assay was performed as previously
   described ([265]24). For phosphatase activity assays, PTPN22 from
   lysates of KO, WT, or S449A cells was immobilized on anti-FLAG M2
   magnetic beads. Beads bound to target proteins were suspended in
   kinetics buffer [50 mM tris (pH 7.2), 1 mM dithiothreitol, and 0.01%
   Triton X-100], and combined with an equal volume of 1 mM DiFMUP
   dissolved in kinetics buffer containing 8% DMSO in round-bottom black
   96-well polystyrene microplates (Corning). Phosphatase activity was
   then measured by continuously monitoring the fluorescence signal as
   described ([266]64). Levels of immunoprecipitated PTPN22 were evaluated
   by Western blotting. For intracellular dephosphorylation assay, 3× FLAG
   PTPN22 WT, S449A, D195A/C227S (DACS PTPN22 WT), or D195A/C227S/S449A
   (DACS PTPN22 S449A) PTPN22 were cotransfected with hemagglutinin-tagged
   ZAP70 (HA-ZAP70) or untagged LCK into HEK293T or COS-7 cells. The cells
   were washed twice with cold phosphate-buffered saline (PBS) and
   harvested for Western blot analysis at 48 hours.

Samples preparation for phospho-MS

   KO or WT Jurkat cell lines were starved for 48 hours and then
   stimulated for different purposes. For mapping PTPN22 phosphorylation
   sites, KO or WT Jurkat cells were stimulated with antibodies (1 μg/ml)
   against human CD3 and CD28 for 5 min and harvested, and PTPN22 protein
   was affinity-purified using anti-FLAG M2 magnetic beads
   (Sigma-Aldrich). The immunoprecipitated (IPed) proteins were
   dissociated from beads by incubation with FLAG peptide (1 mg/ml) in 50
   mM tris base (pH 7.4) and 150 mM NaCl buffer at 4°C for 1 hour with
   rotation. The eluted proteins were collected and confirmed by protein
   gel electrophoresis, Coomassie blue staining, and Western blot.

MS-based data acquisition

   Samples were loaded into a Thermo Scientific Easy-nLC 1000 coupled to
   an Acclaim PepMap c18 column and an EasySpray NG (Thermo Fisher
   Scientific). Mass spectrometer acquisition methods are detailed in
   Lapek et al. ([267]65). Briefly, samples loaded onto column were
   subjected to increasing % organic solvent over the course of 180-min
   gradient method. Data were acquired using a Thermo Fisher Scientific
   Orbitrap Fusion and uploaded to ProteomeXchange.

IP, pull-down, and Western blot

   Immunoprecipitation (IP) and Western blot were performed as described
   ([268]24). Briefly, the cells were lysed in 1× TNE lysis buffer [50 mM
   tris base (pH 7.4), 150 mM NaCl, and 1 mM EDTA] supplied with
   proteinase inhibitors and 1 mM phenylmethylsulfonyl fluoride. After
   removing the cell debris by centrifuging at 15,000 rpm for 10 min at
   4°C, the supernatants were collected and subjected to IP or pull-down
   using the indicated antibodies. The IPed proteins were separated by
   SDS-PAGE and transferred to nitrocellulose membranes. Target proteins
   were probed with specific antibodies and visualized by
   chemiluminescence detection system.

Dual-luciferase reporter assay

   The dual-luciferase reporter assay was performed as previously
   described ([269]24).

Proximity ligation assay

   The Duolink PLA was performed by using Duolink In Situ Red Starter Kit
   mouse/rabbit according to Duolink In Situ Fluorescence protocol
   (Sigma-Aldrich). Briefly, coverslips were precoated with anti-human CD3
   antibody (1 μg/ml) or ICAM-1 (3 μg/ml) at 4°C overnight. Jurkat cells
   (3× 10^6) were added onto the coverslips and stimulated at 37°C for
   indicated time. For TCR stimulation–induced PLA, Jurkat cells were
   stimulated together with soluble anti-human CD28 antibody (1 μg/ml) for
   30 min. For ICAM-1–induced PLA, the cells were stimulated for 20 min.
   The cells were then fixed with 4% paraformaldehyde in PBS for 15 min at
   room temperature, followed by permeabilization with 0.2% Triton X-100
   for 10 min. After three washes in cold PBS, PLA assay was performed
   following the standard protocol using anti-FLAG M2 (mouse), anti-ZAP70
   (rabbit), anti-LCK (rabbit or mouse), or anti-CD45 (rabbit) antibodies.
   Images of PLA signal (red dots) were acquired using an Echo Revolve
   fluorescence microscope, and quantification was performed using ImageJ.

siRNA knockdown assay

   siRNA knockdown assay was performed as previously described ([270]24).
   In brief, ON-TARGETplus SMARTpool siRNA-targeting human PRKACA, CD45,
   ZAP70, CSK, or ON-TARGETplus Nontargeting Pool siRNA (Dharmacon) at a
   concentration of 50 nM were nucleofected into 20 × 10^6 Jurkat cells in
   six-well plates using the Human T Cell Nucleofector Kit (Lonza)
   according to the manufacturer’s instructions using program T-23 on an
   Amaxa Nucleofector II. Transfected cells were harvested for Western
   blot analysis after 48 hours, and expression of target proteins was
   calculated with ImageJ.

Flow cytometry and cell sorting

   Flow cytometry and cell sorting were performed following described
   protocols ([271]66). For surface markers staining, single-cell
   suspensions were prepared and washed twice with fluorescence-activated
   cell sorting buffer (2.5% FBS and 1 mM EDTA), blocked with human or
   mouse Fc blocker (BD Biosciences) for 15 min at room temperature and
   stained with fluorochrome-conjugated antibodies: CD3 (17A2), CD19
   (6D5), CD44 (IM7), and CD69 (FN50, H1.2F3) antibodies were purchased
   from BioLegend. PD-1 (RMP1-30), F4/80 (BM8), GL7 (GL-7), CD25 (PC61.5),
   CD62L (MEL-14), TCR-β (H57-597), CD8 (53-6.7), and CD4 (RM4-5)
   antibodies were purchased from Thermo Fisher Scientific. Intracellular
   staining was performed with the intracellular fixation buffer,
   permeabilization buffer, and the Foxp3/Transcription Factor Staining
   Buffer Set (Thermo Fisher Scientific). For intracellular staining of
   transcription factors, antibodies recognizing FoxP3 (FJK-16 s) and Bcl6
   (BCL-DWN) were obtained from Thermo Fisher Scientific. Dead cells were
   excluded from analysis by staining with Fixable Viability dye from
   Thermo Fisher Scientific. Cells were assessed by a ZE5 (Bio-Rad) or
   ID7000 (SONY) flow cytometer. Results were analyzed by FlowJo software
   (Treestar, Ashland, OR). Targeted cell populations were sorted by a
   SONY SH800 or a BD Symphony S6 Cell Sorter. For thymic
   immunophenotyping, thymocytes from either 4-week-old Ptpn22^WT or
   Ptpn22^S452A mice were collected, and single-cell suspensions were used
   to analyze the percentage of CD4^−CD8^− double-negative, CD4^+
   single-positive, CD8^+ single-positive, and CD4^+CD8^+ double-positive
   populations.

   For CD4^+ T cell proliferation, CD4^+ T cells were isolated from
   spleens of either 8-week-old Ptpn22^WT or Ptpn22^S452A female mice
   using the EasySep Mouse CD4^+ T cell isolation kit (STEMCELL
   Technologies) as previously described ([272]66). Isolated CD4^+ T cells
   were stimulated with Dynabeads mouse T-Activator CD3/CD28 in the
   presence of Il-2 (10 ng/ml) for 3 days. Next, the total cell numbers
   were counted, and T cell proliferation was assessed by staining equal
   amounts of T cells with CellTrace Violet dye (Thermo Fisher Scientific)
   for 3 to 4 days together with mouse CD3/CD28 Dynabeads and Il-2 (10
   ng/ml). After removal of Dynabeads, the cells were harvested, washed,
   stained with Fixable Viability Dye 780 (Thermo Fisher Scientific), and
   analyzed by flow cytometry.

   For the CD69 activation assay, mouse CD4^+ T cells were isolated and
   stimulated as described above. The cells (1 × 10^6) were cultured in
   plates uncoated or coated with anti-CD3 (0.5 μg/ml) antibodies in
   complete RPMI 1640 media containing soluble anti-CD28 (1 μg/ml)
   antibodies, (R&D Systems) for 4 hours. The cells were then collected,
   stained with fluorescein isothiocyanate anti-mouse CD69 antibody and
   Fixable Viability Dye 780 and analyzed by flow cytometry.

Optimized scRNA sequencing

   For optimized scRNA sequencing, J.Lck WT or J.Lck Y192F cells were
   electroporated with plasmids encoding mEGFP-tagged and C-terminal amino
   acid codon sequence optimized WT or S449A PTPN22, respectively, for
   24 hours. The cells were stimulated with anti-human CD3/CD28 antibodies
   (0.5 μg/ml) for 12 hours. mEGFP^+ cells were sorted with a BD Symphony
   S6 Cell Sorter and subjected to 3′ single-cell and CITE-seq by using
   TotalSeq-B0146 anti-human CD69 Antibody (BioLegend). Next, mEGFP^+
   cells were subjected to library preparation and sequencing under
   illumina Novaseq X plus system (UCLA). Libraries were constructed using
   the Chromium Next GEM Single Cell 3′ Kit v3.1 with the 3′ Feature
   Barcode Kit (10x Genomics) by loading sorted cells to target
   11,000 ~ 15,000 cells. 3′ GEX and Feature Barcode libraries were
   sequenced on the Illumina NovaSeq X Plus platform (2× 100 bp) at a 4:1
   ratio.

   We quantified mRNA and antibody-derived tag unique molecular identifier
   (UMI) counts using Cell Ranger v6.1.2 ([273]67). Reads were aligned to
   a custom GRCh38 human reference genome with the addition of the
   optimized PTPN22 WT and S449A sequences (table S3). The Seurat package
   was used for quality control and downstream processing of single-cell
   CITE-seq data. Cells with less than 500 genes or more than 5000 genes
   expressed were removed. In addition, cells with more than 30,000 UMI
   were removed along with those that had a mitochondrial content higher
   than 7%. Further, doublets were removed using DoubletFinder version
   2.0.3 ([274]68). After QC filtering, a total of 17,677 cells remained
   across all five conditions. Data were merged, log normalized, and
   scaled using the ScaleData function with the linear model before
   dimension reduction with RunPCA. Clusters were identified with the
   FindNeighbors function constructing a KNN graph based on the Euclidean
   distance in principal components analysis space (based on the first 14
   principal components) with edge weight between cells based on their
   Jaccard similarity. Following this, modularity optimization with the
   Louvain algorithm, to iteratively group cells together, was performed
   using the FindClusters function with a resolution of 0.1. The Uniform
   Manifold Approximation and Projection embeddings were then calculated
   and plotted. Differential expression analysis to compare between
   genotypes was conducted using the Findmarkers function.

Phospho-ERK, phospho-SLP-76, and CD69 activation assay

   The J.OT1.hCD8^+ or its derivatives with PTPN22 S449A variant (J.OT1
   N22 WT or J.OT1 N22 449) were incubated with peptide-pulsed T2-K^b
   cells as previously described with minor modifications ([275]44).
   Briefly, 1 × 10^6 T2-K^b cells were pulsed with a series of titrated
   concentrations of OVA or altered peptide ligand (APL) peptides and
   incubated at 37°C for 1 hour in flat-bottom 96-well plates. The J.OT1
   N22 WT or J.OT1 N22 449 cells were added into each well of the 96-well
   plate that contained the peptide-pulsed T2-K^b cells. The plates were
   incubated at 37°C for the indicated times. OT-II^+ CD4^+ T cells
   isolated from mice expressing Ptpn22^WT or Ptpn22^S452A were cultured
   with 1 × 10^6 irradiated (35 Gy) splenocytes from RAG2 KO mice as
   antigen-presenting cells and pulsed with titrated concentrations of OVA
   or APL peptides at 37°C for 16 hours. The phosphorylation of ERK
   (Thr^202/Tyr^204), SLP-76 (Tyr^128), and CD69 activation were analyzed
   by flow cytometry.

OVA immunization and tissue processing

   Complete Freund’s Adjuvant (CFA) + OVA (EK-0301, lot: 125) and
   Incomplete Freund’s Adjuvant (IFA) + OVA (EK-0311, lot: 125) were
   purchased from Hooke Laboratories. On day 0, two distinct 50-μl
   subcutaneous injections of OVA/CFA emulsion were administered in the
   back of 8-week-old Ptpn22^WT or Ptpn22^S452A mice. On day 10, blood
   samples were collected by retro-orbital bleeding, and the mice received
   a subcutaneous booster immunization of 100 μl of OVA/IFA emulsion in
   the back. On day 21, the mice were bled retro-orbitally and euthanized
   to isolate the serum for specific IgG testing and to collect axillary
   lymph nodes for flow cytometry analysis.

Resiquimod (R848) lupus model and tissue processing

   The resiquimod lupus model was performed following a previously
   published protocol ([276]46). Briefly, 8 to 12 weeks old Ptpn22^WT or
   Ptpn22^S452A littermate mice were treated with resiquimod for four
   weeks. Proteinuria was monitored by using Albustix urinalysis reagent
   test strips (Siemens Healthineers) every week. After mouse euthanasia,
   the spleens, kidneys, and peripheral blood were collected and processed
   for further experiments. Spleen and body weights were measured to
   calculate the spleen/body weight ratio. Splenocytes were isolated and
   counted for flow cytometry analysis. Mouse kidneys were preserved in
   10% buffered formalin and embedded into paraffin after tissue
   processing. The sections were stained with periodic acid–Schiff and
   imaged using an Echo Revolve microscope. IgG deposition in the kidney
   was assessed by using Alexa Fluor 647–conjugated anti-mouse IgG
   antibody (Thermo Fisher Scientific, A-31571). Fluorescent images were
   taken using the fluorescence microscope (Echo Revolve), and intensity
   was quantified by ImageJ in a blinded fashion.

Statistical analysis

   Statistical analyses were performed using GraphPad Prism 9.2.0 software
   by applying one-way and two-way analysis of variance (ANOVA), unpaired
   t test, Kruskal-Wallis test, Kolmogorov-Smirnov test, and Mann-Whitney
   test as indicated in the figure legends. Error bars represent
   means ± SEM from at least three biological repeats. Significant
   differences were defined as *P < 0.05, **P < 0.01, ***P < 0.001.

Acknowledgments