While PTPN22 allotypes have a minor impact on T cell compartment composition, a significant impact on signal transduction in human T cells has been observed. In primary T cells, PTPN22^620R is a negative regulator of TCR (26, 28, 43, 77, 78) ( Figure 1A ) and lymphocyte function-associated antigen 1 (LFA-1) (32) signaling ( Figure 2 ) while it is a positive regulator of in vitro T regulatory cell (T_reg) induction (33). In T cells, PTPN22^620R has been shown to directly interact with Grb2 (43), VCP (44), Vav (32, 44), Zap70 (32, 44), Lck (26, 32, 44), TCRζ (44), CD3ϵ (44), c-CBL (45), CSK (32, 40, 41), EB1 (46), and the p85 subunit of PI3K (47). Studies do not agree whether PTPN22^620W is a gain-of-function or loss-of-function variant in human TCR signaling but there is compelling evidence for both views ( Figures 1B, C ) (40, 41, 47, 70, 72, 76, 79–81). PTPN22^620W is a loss-of-function variant in LFA-1 signaling ( Figure 2 ) (32). PTPN22^620W has not been studied in the context of T_reg induction in humans, however activated T_regs (aT_regs) from PTPN22^620W/W donors have a reduced capacity to inhibit IFNγ secretion from other T cells compared to those from PTPN22^620R/R donors (47).

PTPN22 is a known negative regulator of TCR signaling ( Figure 1A ) (82). To investigate the function of PTPN22 in human T cells many studies have utilized the T-cell acute lymphoblastic leukemia cell line, Jurkat (26, 28, 43). This has allowed dissection of the influence of PTPN22 on proximal TCR signaling. In Jurkat T cells, it has been shown that PTPN22 negatively regulates activation of JNK2 (26) and LCK (26), and transcriptional activity driven by NF-κB (28), CD28 response element/NF-IL2B AP-1 (43), NFAT/AP-1 (26), c-fos (26), and c-jun (26) downstream of the TCR. CRISPR/Cas9 mediated knockout of PTPN22 in Jurkat T cells revealed that PTPN22 negatively regulates TCR-driven IL-2 and CD69 expression especially in the context of weak antigen stimulation (83). CRISPR/Cas9 mediated knockout of PTPN22 in primary CD4⁺ T cells supports that PTPN22 is a negative regulator of TCR signaling (78). These studies also revealed how PTPN22 achieves negative regulation of TCR signaling. PTPN22 cooperates with CSK to inhibit initial TCR signaling ( Figure 1A ) (26). In resting T cells, PTPN22 is associated with CSK and upon TCR stimulation, this complex dissociates at a rate that parallels dephosphorylation of PTPN22 substrates (40).

While PTPN22^620R is a negative regulator of TCR signaling, the effect of the SNP on function of PTPN22^620W remains controversial. It is currently debated whether PTPN22^620W is a gain-of-function variant that reduces response to TCR stimulation ( Figure 1B ) or a loss-of-function variant that allows enhanced TCR signaling ( Figure 1C ). The most studied hypothesis is that PTPN22^620W is a gain-of-function variant that suppresses TCR signaling ( Figure 1B ) (40, 41, 47, 70, 72, 76, 79, 80). These studies have shown that PTPN22^620W reduces signaling through the TCR and is associated with significantly reduced IL-2 secretion (72, 79), calcium mobilization (72, 76, 79), and IFNγ production from CD4⁺ T cells ( Table 2 ) (79). There is also evidence that the PTPN22^620W allotype drives enhanced skewing of CD4⁺ T cells to EOMES⁺ T_H1 cells (47, 73). PTPN22^620W is also associated with reduced expression of CD25, lower proliferation, and decreased IL-10 secretion by CD4⁺ memory T cells (76, 79) ( Table 2 ). In concordance with this, in vivo TCR stimulation in the form of a trivalent influenza vaccine resulted in reduced induction of an influenza virus-specific CD4⁺ T cell response in PTPN22^620R/W subjects compared to PTPN22^620R/R subjects ( Table 2 ) (80). Another indication of reduced influenza virus-specific CD4⁺ T cell induction is the impairment of anti-flu antibody affinity maturation. Antibody affinity maturation relies on activation of CD4⁺ T follicular helper (T_FH) cells; PTPN22^620R/W subjects had reduced affinity maturation compared to PTPN22^620R/R subjects implying they had reduced activation of T_FH cells or reduced activation of anti-flu B cells (80).

Studies tracing the proximal events following TCR stimulation agree that PTPN22^620W is a gain-of-function variant in primary T cells ( Figure 1B ). Overexpression of PTPN22^620W decreases NFAT/AP-1-driven luciferase transcription more than overexpression of PTPN22^620R (72). In primary T cells from donors with PTPN22^620W/W TCR stimulation resulted in lower TCRζ-chain phosphorylation and increased ERK, AKT, and PI3K p85 activation compared to PTPN22^620R/R donor T cells ( Table 3 ) (47, 70). These studies offer a molecular mechanism for the difference in function of PTPN22^620R and PTPN22^620W centered on reduced interactions of CSK and LCK with PTPN22^620W.

As noted above, in resting T cells PTPN22^620R is associated with CSK and upon TCR stimulation this complex dissociates at a rate that parallels dephosphorylation of PTPN22 substrates ( Figure 1A ) (26, 40). Simultaneously, PTPN22 is phosphorylated at Ser⁷⁵¹ by PKCα which enhances the CSK/PTPN22 interaction and restricts PTPN22 activity to allow appropriate TCR signaling (42). PTPN22^620W interacts with CSK to a lesser extent than PTPN22^620R (immunoprecipitation of PTPN22^620R pulls down 2.9 fold more CSK than PTPN22^620W), and is more available to dephosphorylate PTPN22 substrates at the initiation of TCR signaling (28, 40, 41, 51). Both PTPN22^620R and PTPN22^620W are subject to phosphorylation at Ser⁷⁵¹ by PKCα, however this only seems to inhibit PTPN22^620R activity, by enhancing its association with CSK, while it does not inhibit PTPN22^620W or enhance PTPN22^620W/CSK interactions (42). Similarly, PTPN22^620R is associated with LCK to a greater degree than PTPN22^620W and this appears to be CSK-dependent (41). LCK phosphorylates PTPN22 on an inhibitory Y536 residue (41). PTPN22^620R has more phosphorylated Y536 residues and is less active than PTPN22^620W in Jurkat cells at rest and upon TCR stimulation (41). This may also explain why the in vitro phosphatase activity of PTPN22^620W is 50% higher compared to PTPN22^620R when the two allotypes of PTPN22 are purified from mammalian cells. When purified from insect cells, where this post-translational modification is absent, the phosphatase activity is equal among the two allotypes (41, 72). In conclusion, PTPN22^620W is a more potent inhibitor of TCR signaling than PTPN22^620R because PTPN22^620W is more available to interact with PTPN22 substrates due to reduced sequestration by CSK. Further, PTPN22^620W is more active due to reduced association with its own negative regulator, LCK, and consequent reduced phosphorylation at an inhibitory tyrosine residue ( Figure 1B ).

While evidence that PTPN22^620W is a gain-of-function variant remains compelling, sufficient results exists to argue that PTPN22^620W could be a loss-of-function variant ( Figure 1C ) (70, 81). These studies have observed that T cells from healthy PTPN22^620W/W donors expand more upon TCR stimulation than those from healthy PTPN22^620R/R donors (70). Further, when CSK is co-expressed with PTPN22^620W in Jurkat T Cells, higher calcium fluxes are measured than when CSK is co-expressed with PTPN22^620R (81). A study found that the PTPN22^1858T allele enhances expression of a dominant negative isoform, PTPN22.6, that increases signaling through the TCR ( Figure 1C ) (84). The authors offered a hypothesis that reconciles human data showing that PTPN22^620W is a gain-of-function; PTPN22^620W allows chronic signaling through the TCR that drives T cell exhaustion, causing T cells from PTPN22^620R/W and PTPN22^620W/W donors to be less responsive to stimulation through the TCR—a finding reported by most studies. This is supported by evidence that expression of PD-1, a marker of T cell exhaustion, is enhanced on CD4⁺ T_eff and T_regs in healthy PTPN22^620W/W donors compared to healthy PTPN22^620R/R donors (74). Furthermore, the reduced calcium flux seen in PTPN22^620R/W donors was most notable in memory CD4⁺ T cells with no difference observed in naïve CD4⁺ T cells; this could indicate that the experienced population is exhausted (76). While it is not certain whether PTPN22^620W is a gain-of-function or loss-of-function variant in human TCR signaling, it is clear that the mouse orthologue of PTPN22^620W, PEP^619W, is a loss-of-function variant in mouse TCR signaling.

Data from mouse models support the role of PTPN22/PEP as a negative regulator of TCR signaling ( Figure 1A ). Overexpression of PEP in the mouse antigen specific T cell line, BI-141, reduced TCR-mediated phosphorylation of ZAP70, c-Cbl, and the CD3 ζ-chain (77). Overexpression of PEP also reduced IL-2 secretion from these cells (77). C57BL/6J mice with a genetic ablation of Ptpn22 (B6.Cg-Ptpn22^tm2Achn /J commonly referred to as C57BL/6-Ptpn22 ^−/− mice) as well as NOD mice with doxycycline-induced knockdown of PEP [NOD-Tg(tetO-RNAi : Ptpn22,UBC-tetR,-GFP)P2Kslr commonly referred to as NOD-Ptpn22^KD ] starting at birth have an accumulation of effector/memory CD4⁺ and CD8⁺ T cells in secondary lymphoid organs. This phenotype is thought to be a product of increased TCR signaling in the absence of PEP (85–88). Similar to humans harboring PTPN22^620W, PEP^619W knock-in C57BL/6 mice (C57BL/6-Ptpn22 ^tm1.1Kas commonly referred to as C57BL/6-PEP^619W) exhibited an expansion of CD4⁺ memory T cells compared to unaltered C57BL/6 mice that carry the PEP^619R allele (70, 71). In C57BL/6-PEP^619W mice there was also a marked expansion of the total effector/memory T cell pool and T cells from these mice exhibited increased IL-2 secretion, increased calcium mobilization, enhanced/prolonged tyrosine-phosphorylation of ZAP-70 and Lck, and increased ex vivo expansion of T cells compared to C57BL/6 mice (70, 71, 86). While the R619W conversion in PEP appears to be a loss-of-function variant with respect to TCR signaling ( Figure 1C ), controversy exists regarding the human autoimmunity risk allotype, PTPN22^620W, with regard to gain-of-function or loss-of-function TCR signaling ( Figures 1B, C ). Despite this ongoing lack of clarity for PTPN22^620W in human TCR signaling, evidence clearly supports that PTPN22^620W is a loss-of-function variant in LFA-1 signaling in T cells.

LFA-1 is fundamentally important to general leukocyte trafficking. Loss of LFA-1 causes the life-threatening disease known as leukocyte adhesion deficiency (LAD) resulting in uncontrolled microbial infections (89). LFA-1 is also critical in T cell activation and migration (90). In human T cells, PTPN22 inhibits LFA-1 signaling ( Figure 2 ) (32). T cells treated with PTPN22 targeting small interfering RNA (siRNA) exhibited increased ICAM-1 (LFA-1 ligand)-induced phosphorylation of LCK, ZAP70, ERK1/2, and Vav compared to cells treated with a non-targeting control (NTC) siRNA. There was also an increase in ICAM-1-induced motility in cells treated with the PTPN22 targeting siRNA (32). The autoimmune associated variant, PTPN22^620W, is a loss-of-function variant in LFA-1 signaling ( Table 2B ). Similar to what was observed with knockdown of PTPN22, human T cells from PTPN22^620R/W and PTPN22^620W/W donors have enhanced LFA-1 induced signaling (pERK1/2 fold change over unstimulated; PTPN22^620W/W ~35 vs. PTPN22^620R/W ~25 vs. PTPN22^620R/R ~20) and adhesion (mean # of T cells adhered to LFA-1 coated slide at 8 min under shear flow; PTPN22^620W/W ~32 vs. PTPN22^620R/R~24) compared to T cells from PTPN22^620R/R donors. At rest, PTPN22^620R and PTPN22^620W are aggregated near the plasma membrane of T cells. Upon engagement of ICAM-1 with LFA-1, PTPN22^620R leaves these aggregates, associates with CSK, and is recruited to the leading edge of migrating cells in an LCK-dependent manner where it dephosphorylates PTPN22 substrates to inhibit LFA-1 signaling ( Figure 2A ). In contrast, PTPN22^620W stays more clustered and is less recruited to the leading edge resulting in less PTPN22-mediated negative regulation of LFA-1 signaling ( Figure 2B ) (32).

As observed in human T cells, PEP negatively regulates mouse T cell responses to ICAM-1 stimulation ( Figure 2A ). T cells from C57BL/6-Ptpn22 ^−/− mice displayed enhanced ERK1/2 phosphorylation (pERK1/2 fold change over-unstimulated; Ptpn22 ^−/− ~12 vs. Ptpn22 ^+/+ ~8) after ICAM-1 stimulation and adhered better to ICAM-1 coated glass slides under shear flow (mean # of T cells adhered at 8 min; Ptpn22 ^−/− ~55 vs. Ptpn22 ^+/+ ~30) compared to Ptpn22-intact mouse T cells (32). C57BL/6-Ptpn22 ^−/− mouse T cells also had increased LFA-1 induced IFNγ secretion and were better at forming T cell-DC conjugates compared to Ptpn22-intact T cells (86). PEP and PTPN22 are both negative regulators of LFA-1 signaling in mice and humans ( Figure 2A ). PTPN22^620W is a loss-of-function variant in humans while it is not known how the PEP 619R to W conversion affects mouse LFA-1 signaling ( Figure 2B ). While the molecular mechanisms behind PTPN22’s influence on receptor-proximal signaling in T cells (i.e., activation and mobilization) are well studied, PTPN22 has also been shown to influence T_reg induction and function however the mechanism is less resolved.

PTPN22 positively regulates in vitro induced T_reg (iT_reg) differentiation in human T cells. Primary naive T cells (CD4⁺CD127⁺CD25⁻) from PTPN22^620R/R healthy donors and PTPN22^620R/R donors with T1D were subjected to PTPN22 knockdown with antisense oligonucleotides. Differentiation of iT_regs via treatment with IL-2/TGF-β1/αCD3/αCD28 was reduced with PTPN22 knockdown compared to control oligonucleotide transfected cells (% of CD4 T cells that are CD25⁺FoxP3⁺; PTPN22 knockdown resulted in ~20% iTreg vs. control ~40%) (33). No direct clinical studies have shown how PTPN22⁶²⁰ allotypes influence iT_reg differentiation; however, healthy PTPN22^620W/W donors have increased CD4⁺ T_regs compared to healthy PTPN22^620R/R donors (7.94% vs. 6.76%) implying that PTPN22^620W might potentiate iT_reg development (74, 75). Although PTPN22^620W/W donors have slightly more CD4⁺ T_regs, these T_regs exhibit a reduced capacity to inhibit IFNγ secretion from conventional T cells compared to those from PTPN22^620R/R donors (47, 76).

As observed in humans, PEP also influences T_reg development in mice; C57BL/6-Ptpn22 ^−/− mice and NOD-Ptpn22^KD mice had increased numbers of T_regs (87, 88). Data from C57BL/6-Ptpn22 ^−/− mice and NOD-Ptpn22^KD mice provided evidence that deficiency of PEP reduces the TCR signal strength required for in vitro induction of iT_regs (91, 92). The iT_reg induction can be accomplished by stimulating naïve FoxP3-CD4⁺ T cells with a combination of agonistic anti-CD3 and anti-CD28 targeting antibodies in the presence of TGF-β (87). Lower levels of stimulation with reduced concentrations of anti-CD3 antibodies increased in vitro iT_reg induction in Ptpn22 ^−/− cells compared to Ptpn22-intact cells. Increased concentrations of anti-CD3 resulted in elevated stimulation and decreased iT_reg induction in Ptpn22 ^−/− cells compared to Ptpn22-intact cells. At levels of TCR-stimulation that drive optimal in vitro iT_reg induction in parental C57BL/6 mice, C57BL/6-Ptpn22 ^−/− had reduced iT_reg induction (87). Much like PTPN22^620W humans, aged C57BL/6-PEP^619W mice had increased T_regs compared to C57BL/6 mice. However, young C57BL/6-PEP^619W mice exhibited no increase in T_regs. T_regs from young C57BL/6-PEP^619W mice exhibited no differences in suppressive activity when compared to C57BL/6 mice, however T_regs from aged mice were not assessed. This difference may be due to the age of the mice, however, it remains to be seen if T_regs from older C57BL/6-PEP^619W mice exhibit the same defect in suppression as human T_regs (71). It is clear that PTPN22 plays multiple roles in human T cells and that the diabetogenic allotype of PTPN22, PTPN22^620W, alters these roles; how might the altered function of PTPN22^620W in T cells impact T1D development?

T1D is generally considered a T cell mediated disease where CD8⁺ T cells are the major islet infiltrating immune cells (93, 94). The SNP in PTPN22, rs2476601, is associated with increased risk for T1D, reduced age at onset (95), and reduced residual β cell function at diagnosis (96). This SNP affects T cell function. PTPN22 is a negative regulator of TCR (26, 28, 43, 77) and LFA-1 (32) signaling and influences aT_reg suppressive capacity ( Figures 1 and 2 ) (47). In T cells, the effect of the T1D-risk variant PTPN22^620W on TCR-induced signaling is currently unresolved with data supporting both gain-of-function and loss-of-function hypotheses (40, 41, 47, 70, 72, 76, 79, 80). In contrast, PTPN22^620W has been characterized as a loss-of-function variant in LFA-1-induced signaling because it is not available to interact with its substrates (32). Adaptive T_regs (aT_regs) from PTPN22^620W/W donors have reduced capacity to suppress IFNγ secretion from conventional T cells (47). The enhanced LFA-1-induced signaling and motility, and the reduced capacity of aT_regs to suppress IFNγ secretion from conventional T cells seen in PTPN22^620R/W and PTPN22^620W/W humans could help explain why rs2476601 is associated with increased overall risk of T1D development. The seemingly small magnitudes of reported biochemical, phenotypic, and functional effects of PTPN22^620W in human T cells are surprising for a genetic variation that ranks near the top of the list for T1D genetic risk. We ask ourselves, “How could such minor fluctuations contribute to a life-threatening pathology?” The answer might lie in the thymus - the immune tissue where developing thymocytes (soon to be T cells) are exquisitely sensitive to the strength and duration of nascent TCR signaling. If PTPN22^620W is a gain-of-function variant in TCR signaling, the PTPN22^620W variant might impair the process of negative selection whereby autoreactive thymocytes are normally eliminated upon strong TCR signaling. Thus, effectively blinded to the fact that a given TCR is recognizing a self-antigen (e.g., insulin), autoreactive T cells might survive and escape into the periphery (72). More autoreactive T cells in the periphery would lead to increased autoreactive T cells surveying tissues, including the pancreas, and more opportunities for an autoimmune reaction to occur.

The alternate scenario postulates that thymic selection is more or less unaffected, and that the biologic effects of PTPN22^620W manifest in the periphery. If PTPN22^620W is a loss-of-function variant in TCR signaling, circulating T cells would be more sensitive to TCR ligation and this could explain the genesis of autoreactive T cell activation and thus autoimmunity. Both intrathymic and peripheral scenarios would be complicated by enhanced LFA-1-induced signaling (enhancing T cell migration) and reduced capacity of aT_regs to suppress IFNγ secretion from activated T cells that could result in enhanced T cell infiltration into tissues (i.e., islets of Langerhans) as well as secretion of more IFNγ, thus creating a more inflammatory local environment. For T cells, additional new work will be needed to understand how thymic development and intra-islet T cell function is modulated by PTPN22 variants. Is there a single dominant mechanism at fault for autoimmune risk, or is this a case of death by a thousand cuts—multiple subtle effects which alone appear innocuous but together add up to complete destruction of a vital tissue? If the story weren’t complicated enough, T cells alone might not be the culprit of T1D. Autoantibodies produced by B cells are a prevalent feature and remain the gold standard biomarker of T1D progression. While it is hypothesized that autoantibodies are not pathogenic in human T1D, B cells are thought to play an important role as antigen specific APCs. It is known that depletion of B cells with Rituximab can delay disease progression (97). Additionally, many of the other rs2476601-asocciated autoimmune diseases are characterized by production of autoantibodies (e.g., RA, SLE, etc.). As such, many studies have focused on the effect of the PTPN22^620R versus PTPN22^620W in human B cells.

PTPN22 functions to dampen BCR signaling as well as BCR-induced apoptosis ( Figure 3B ). PTPN22 is overexpressed in primary chronic B lymphocytic leukemia (CLL) cells (34). CLL cells express functional BCRs and have been characterized for ligand-dependent signaling. PTPN22 depletion in CLL cells increased soluble-αIgM (simulated strong BCR signaling) induced apoptosis (34). Knockdown of PTPN22 also resulted in increased soluble αIgM-induced phosphorylation of LYN, SYK, BLNK, PKCδ, ERK, JNK, and p38 MAPK and reduced soluble-αIgM-induced phosphorylation of AKT, GSK3, and FOXO (34). PTPN22^620W is a gain-of-function variant that acts to further blunt BCR signaling ( Figure 3B ). In heterozygous PTPN22^620R/W donors there is reduced BCR-induced calcium flux compared to PTPN22^620R/R donors (27, 76, 98). Heterozygous donors also had reduced phosphorylation of the BCR-proximal signaling components, SYK, PLCγ2 (MFI phospho-PLCγ2-Y759; PTPN22^620R/W ~700 vs. PTPN22^620R/R ~950), and AKT compared to PTPN22^620R/R donors (27, 76, 98). In PTPN22^620R/W donors there is also reduced total phosphorylated tyrosine in resting (% of CD27⁺ B cells that are phospho-tyrosine⁺; PTPN22^620R/W ~4% vs. PTPN22^620R/R ~8%) and BCR-activated memory B cells compared to PTPN22^620R/R donors (27). Inhibition of PTPN22 in B cells of PTPN22^620R/W donors increased SYK, PLCγ2, and AKT phosphorylation to levels equivalent to those of B cells from PTPN22^620R/R donors (27, 98). Signaling through the BCR can also induce B cell expansion. However, it is not clear how PTPN22^620W affects BCR-induced expansion of B cells; different studies have shown conflicting results (27, 70). Overall, PTPN22^620W is more effective at regulating BCR signaling ( Figure 3B ).

Consistent with data from human B cells, PEP is also a negative regulator of BCR signaling in mice ( Figure 3 ). Silencing of PEP via doxycycline-induced expression of a Ptpn22-targeting siRNA in NOD mice (the NOD-Ptpn22^KD mice) increased B cell response to anti-IgM/anti-CD40 stimulation (88). Additionally, silencing of PEP resulted in the increased proliferation, robust expression of CD25 and CD69, and elevated phosphorylation of PLCγ2 (88). These results have been replicated via Ptpn22 knockout in other mouse strains (81, 101, 102). Unlike PTPN22^620W humans, C57BL/6-PEP^619W mice had a phenotype similar to C57BL/6-Ptpn22 ^−/− mice, with increased anti-IgM induced B cell activation, increased anti-IgM induced proliferation, and increased phosphorylation of PLCγ2 compared to C57BL/6 mice (70, 71, 103). While PTPN22^620W is a gain-of-function variant in human BCR signaling, PEP^619W is a loss-of-function variant with respect to BCR signaling in mice ( Figure 3B ).

PTPN22^620W alters gene expression and immune receptor levels in B cells. Naïve B cells from both PTPN22^620R/W and PTPN22^620W/W donors had significantly upregulated IL4R, IL13R, IL17R, and IL21R mRNA expression (genes involved in B cell proliferation/differentiation) and significantly upregulated genes in the BCR, CD40, and TLR activating pathways compared to those from PTPN22^620R/R donors (100). PTPN22^620W differentially affects expression of other genes with SNPs associated with T1D and other autoimmune diseases (BLK, PTPN2, CD40, TRAF1, CD19, SLAM, IRF5) (100). The surface expression of BAFFR, CD40, and SLAMF6 was enhanced in PTPN22^620R/W and PTPN22^620W/W donors compared to PTPN22^620R/R donors ( Table 4 ) (100, 103). Naïve B cells from PTPN22^620R/W and PTPN22^620W/W donors were more responsive to CD40L stimulation with an increased percent of B cells expressing CD69 and CD25 than those from PTPN22^620R/R donors (100). CpG stimulation of PBMC for 4 days resulted in greater expansion of IgM+ memory B cells (CD19+CD27+IgM+) and IgM- Plasma cells (CD19+CD27^hiIgM-) in PTPN22^620R/W patients with T1D compared to PTPN22^620R/R patients with T1D and in healthy control PTPN22^620R/W donors compared to healthy control PTPN22^620R/R donors (99). The combination of increased BAFFR, CD40, and SLAMF6 surface levels and the increased expression of IL4R, IL13R, IL17R, IL21R, as well as genes belonging to the CD40, TLR, and BCR activation pathways may explain the enhanced CpG-induced expansion of IgM+ memory B cells and IgM- Plasma cells in PBMCs seen in PTPN22^620R/W and PTPN22^620W/W donors compared to PTPN22^620R/R donors. Importantly, this phenomenon is present in both PTPN22^620R/W and PTPN22^620W/W donors implying that the effects of PTPN22^620W are either dominant or co-dominant.

Unlike humans, there was decreased CD40 and BAFFR surface expression on total splenocytes with decreased CD40 and BAFFR on immature B cells and increased CD40 on T2 B cells of C57BL/6-PEP^619W mice compared to C57BL/6 mice (103). Tnfrsf13c (BAFFR) mRNA levels were enhanced in immature B cells and Cd40 mRNA levels were enhanced in T2 B cells of C57BL/6-PEP^619W mice compared to C57BL/6 mice (103). Taken together we see that PTPN22 and PEP affect expression of costimulatory molecules in B cells of both humans and mice however the effects of the R to W conversion are not consistent when comparing humans to mice.

PTPN22^620W alters the central and peripheral B cell tolerance checkpoints as well as the composition of the B cell compartment in humans (76, 98–100). Central B cell tolerance is mediated via clonal deletion or receptor editing to remove autoreactive or polyreactive B cells from the bone marrow before they enter the periphery (e.g., spleen, blood, lymph nodes, tissues) (104). Central tolerance results in a large reduction of polyreactive and autoreactive B cells and is readily apparent when comparing the bone marrow to the spleen and blood; 40%–70% of early immature B cells are polyreactive and 50%–75% are autoreactive in the bone marrow while 5%–10% of transitional B cells are polyreactive and 30%–50% are autoreactive in the periphery (104, 105). A common method for determining if B cells are autoreactive is to assess their response to human epithelial type 2 (HEp-2) cells. HEp-2 cells express a large array of self-antigens and HEp-2 reactive B cells are considered autoreactive (106). Healthy PTPN22^620R/W and PTPN22^620W/W donors had an increased proportion of polyreactive and HEp-2-reactive new emigrant/transitional B cells (CD20⁺CD10⁺CD21^loIgM^hiCD27⁻: 25%–30% of new emigrant/transitional B cells were polyreactive and ~50% were HEp-2-reactive) compared to healthy PTPN22^620R/R donors (8%–10% of new emigrant/transitional B cells were polyreactive and ~30% were HEp-2 reactive) (100, 107). Most studies agreed that transitional B cells (CD19⁺CD27⁻CD24^hiCD38^hi) were increased in healthy PTPN22^620R/W and PTPN22^620W/W donors compared to healthy PTPN22^620R/R donors (percentage of total B cells that are transitional; PTPN22^620W/W and PTPN22^620R/W ~5% vs. PTPN22^620R/R ~2.5%), although not all studies observe this effect (98, 99, 108). The increased numbers of transitional B cells and polyreactive/HEp-2-reactive new emigrant/transitional B cells in healthy PTPN22^620R/W and PTPN22^620W/W donors indicates that the central B cell tolerance checkpoint is altered by PTPN22^620W. Ergo, the autoimmune-linked allotype allows more polyreactive and autoreactive B cells to escape central tolerance and proceed into the periphery. B cells that enter the periphery will go through another round of selection to remove or inactivate autoreactive cells.

Peripheral B cell tolerance results in anergy or clonal deletion via apoptosis that is dependent on caspase-3 activation and is triggered by strong signaling though the BCR (98). This results in the reduction of autoreactive peripheral B cells. There are more autoreactive transitional B cells than autoreactive naïve mature B cells due to the peripheral B cell tolerance checkpoint; 30%–50% of transitional B cells are autoreactive while 10%–30% of naïve mature B cells are autoreactive (105). To simulate strong BCR signaling in naïve B cells, anti-IgM is used to crosslink the BCRs; this is similar to encountering a multivalent self-antigen during peripheral B cell tolerance and will cause some naïve B cells to undergo apoptosis. After 12 h of anti-IgM treatment, significantly fewer naïve B cells from PTPN22^620R/W donors had begun the process of apoptosis by cleaving/activating caspase-3 when compared to PTPN22^620R/R donors (% of naïve B cells with cleaved/active caspase-3; PTPN22^620R/W ~10% vs. PTPN22^620R/R ~18%) (98). Basal levels of the anti-apoptotic protein, Bcl-2, were higher in transitional B cells from PTPN22^620R/W donors compared to PTPN22^620R/R donors (Normalized BCL-2 MFI; PTPN22^620R/W ~20 vs. PTPN22^620R/R ~12) with no alteration in the pro-apoptotic protein, Bim (98). Healthy PTPN22^620W/W and PTPN22^620R/W donors had increased frequencies of polyreactive and HEp-2-reactive mature naïve B cells (CD20⁺CD10⁻CD21⁺IgM⁺CD27⁻). In these donors ~30% of mature naïve B cells were polyreactive and ~45% were HEp-2-reactive. In contrast, healthy PTPN22^620R/R donors had ~10% polyreactive mature naïve B cells were and ~20% HEp-2-reactive (100). A unique subset of autoreactive anergic B cells (naïve IgD ⁺ B cells [B_ND]: CD19⁺CD27⁻IgD⁺IgM⁻) are cells in the periphery thought to be anergic due to low chronic antigen stimulation through the BCR (109). B_ND cells were increased in healthy PTPN22^620R/W donors compared to healthy PTPN22^620R/R donors (% of CD19⁺ B cells that are B_ND cells; PTPN22^620R/W ~3% vs. PTPN22^620R/R ~2%) (98). PTPN22^620R/W donors had a lower percentage of memory B cells compared to PTPN22^620R/R donors (% of CD19⁺ B cells that are CD27⁺; PTPN22^620R/W ~35% vs. PTPN22^620R/R ~45%) (76). The reduced caspase-3 activation, increased levels of Bcl-2, increased frequencies of B_ND cells, HEp-2-reactive mature naïve B cells, and polyreactive mature naïve B cells, and decreased frequency of mature B cells found in PTPN22^620R/W and PTPN22^620W/W donors indicates that PTPN22^620W alters the peripheral B cell tolerance checkpoint (98, 100, 107). The increase in autoreactive/polyreactive new emigrant/transition B cells, all transitional B cells, B_ND cells, and decrease in memory B cells was also seen when comparing T1D donors regardless of genotype to healthy PTPN22^620R/R donors and this may represent a common B cell phenotype present in T1D patients (98, 100). Currently, it is thought that the blunting of BCR signaling by the gain-of-function PTPN22^620W allotype leads to reduced negative selection and is responsible for the alterations seen in central and peripheral B cell tolerance mechanisms (76, 98, 100, 107). These B cell phenotypes are observed in both patients with autoimmunity and healthy controls that encode PTPN22^620W.

C57BL/6-Ptpn22 ^−/− as well as other strains of Ptpn22 knockout mice exhibit an altered B cell compartment. Deletion of Ptpn22 increased age-associated B cells (ABCs), plasma cells, autoantibodies, as well as germinal center activity and size when compared to Ptpn22-intact mice. However, germinal center size and activity appears to be partially dependent on an alteration in T follicular helper cells (81, 101, 102). Unlike humans harboring PTPN22^620W, alterations in the B cell compartment of the loss-of-function PEP^619W variant in mice is attributed to altered positive B cell selection due to enhanced BCR signaling (103). C57BL/6-PEP^619W mice have increased splenic transitional 1 B cells, increased age-dependent B cells (ABCs), increased class-switched B cells, increased germinal center B cells, and less mature recirculating B cells when compared to C57BL/6 mice (103). Like humans however, the enhanced positive selection leads to increased self-reactive B cells, increased autoantibody titers, and reduced apoptosis of T1 B cells in C57BL/6-PEP^619W when compared to C57BL/6 mice (70, 71, 103). The similarities between the B cell compartments of Ptpn22 ^−/− mouse strains and C57BL/6-PEP^619W mice implies that PEP^619W is a loss-of-function variant in mice with respect to its effects on B cell positive selection while PTPN22^620W decreases human B cell negative selection.

While PEP^619W mice do not display the same central B cell tolerance phenotype as humans heterozygous or homozygous for PTPN22^620W, immunodeficient NOD.Cg-Prkdc^scid.Il2rg^tm1Wjl (NSG) mice engrafted with human CD34⁺ hematopoietic stem cells (HSCs) from either PTPN22^620R/W, PTPN22^620W/W donors, or with HSCs overexpressing PTPN22^620W phenocopy humans that are heterozygous or homozygous for PTPN22^620W. These PTPN22^620W HSC engrafted NSG mice display an increased proportion of polyreactive and HEp-2-reactive new emigrant/transitional B cells when compared to NSG mice engrafted with HSCs from PTPN22^620R/R donors or HSCs overexpressing PTPN22^620R (100, 107). Importantly, inhibition of PTPN22 in NSG mice engrafted with PTPN22^620W HSCs reduced polyreactive and HEp2-reactive new emigrant B cells to the same levels as NSG mice engrafted with PTPN22^620R HSCs indicating that PTPN22 is the main driver of this difference (107). The increased numbers of transitional B cells and polyreactive/HEp-2-reactive new emigrant/transitional B cells in healthy PTPN22^620R/W and PTPN22^620W/W donors and in PTPN22^620W HSC engrafted NSG mice indicates that the central B cell tolerance checkpoint is altered by PTPN22^620W. This alteration allows more polyreactive and autoreactive B cells to escape central tolerance and proceed into the periphery. Overall, PEP^619W is a loss-of-function variant in mice with respect to its effects on B cell positive selection while PTPN22^620W decreases human B cell negative selection; both of these alterations result in more autoreactive B cells with increased autoantibody titers.

Autoantibodies produced by B cells are a prevalent feature of T1D and remain the gold standard biomarker of islet autoimmunity and T1D progression (110). The SNP in PTPN22, rs2476601, is associated with increased risk of persistent islet autoimmunity (i.e., autoantibodies directed against insulin, GAD65, or IA-2) (111). While the role of pathogenesis of human T1D remains controversial, the importance of B cells has been demonstrated in preclinical models and clinical trials. Depletion of B cells pauses the loss of β cell function in some patients with recent onset T1D and can prevent or reverse disease in NOD mice (97, 112). B cells are not only capable of producing antibodies, they also act as APCs to present antigen to T cells in a process called linked recognition (113). In linked recognition, B cells uptake antigen recognized by the BCR, process it, and load peptides derived from the antigen on MHC-II to present to CD4⁺ T cells (113). These responding CD4⁺ T cells must have already encountered antigen and been activated by other APCs in the periphery before they can provide T cell help to the B cells. The T cell help initiates class-switching in germinal centers, while the B cells provide co-stimulatory signals to the T cells capable of enhancing in-progress T cell responses (114). In NOD mice, it is thought that B cells primarily enhance autoreactive T cell function as APCs and through the production of pro-inflammatory cytokines (115). PTPN22^620R/W and PTPN22^620W/W donors have increased B cell surface expression of CD40, SLAMF6, and BAFFR ( Table 4 ), as well as B cell mRNA expression of IL4R, IL13R, IL17R, IL21R compared to PTPN22^620R/R donors. PTPN22^620R is a negative regulator of BCR signaling and PTPN22^620W is a gain-of-function variant that reduces signaling through the BCR. This reduction in BCR signaling alters central and peripheral B cell tolerance allowing more autoreactive and polyreactive B cells into the periphery. The increased surface expression of CD40, SLAMF6, and BAFFR ( Table 4 ), as well as B cell mRNA expression of IL4R, IL13R, IL17R, IL21R could enhance clonal expansion of B cells, differentiation into plasma cells, class switching, and cell survival in PTPN22^620R/W and PTPN22^620W/W humans (116–121). Increased SLAMF6 and CD40 expression on B cells could also enhance/prolong B cell-T cell interactions leading to more T cell and B cell activation in PTPN22^620R/W and PTPN22^620W/W humans. The combination of these phenotypes could lead to increased class switching of autoreactive B cells and increased survival of autoreactive and polyreactive B cells. These autoreactive/polyreactive B cells could go on to increase or simply sustain activation of autoreactive T cells. The increased/sustained activation of autoreactive T cells by autoreactive/polyreactive B cells could explain why rs2476601 is associated with increased risk of persistent islet autoimmunity (111) and why treatment with a B cell depleting therapy (rituximab) can delay loss of, but not restore, the c-peptide response in patients with recent onset T1D (97).

While adaptive immune cells are integral for targeting and destroying β cells, they are not the only cells implicated in development of T1D. The innate arm of the immune system is generally required to initiate antigen-specific responses by T and B cells. Monocytes, macrophages, and dendritic cells (DCs) are all APCs capable of initiating these potent immune responses in inflammatory contexts.

PTPN22^620R associates with TRAF3 following LPS stimulation and promotes T1-IFN production while PTPN22^620W does not ( Figure 4A ) (35). This effect is not limited to TLR4 stimulation, plasmacytoid dendritic cells (pDCs) from PTPN22^620W/W and PTPN22^620R/W patients with SLE have reduced IFNα production following R848 (TLR7/8 agonist) stimulation compared to PTPN22^620R/R patients (PTPN22^620R/W+PTPN22^620W/W; ~35% pDCs IFNα2+ with gMFI of ~250 vs. PTPN22^620R/R; 45% pDCs IFNα2+ with gMFI of ~500) (122). STAT1 phosphorylation, a marker of interferon receptor signaling, is significantly reduced by about 50% in PBMCs from PTPN22^620R/W donors after LPS treatment when compared to PTPN22^620R/R donors. T1-IFN-inducible genes (IRF7, MX1, and ISG15) were also significantly reduced by about 50% in PBMC-derived DCs from PTPN22^620R/W donors compared to PTPN22^620R/R donors, probably due to reduced production of T1-IFNs. TRAF3 is an adaptor protein that links TLR4 and TLR7/8 signaling to induction of T1-IFNs. PTPN22 co-immunoprecipitated TRAF3 from human monocyte derived DCs (moDCs). In transgenic C57BL/6-Ptpn22 ^−/− mice expressing either human PTPN22^620R or PTPN22^620W, PTPN22^620R associated with TRAF3 and promoted its poly-ubiquitination and subsequent induction of Ifnb1 while PTPN22^620W did not. C57BL/6-Ptpn22 ^−/− mice expressing human PTPN22^620W had reduced LPS-induced T1-IFN production [~50% of Ifnb1 from bone marrow-derived dendritic cells (BMDCs), and ~50% of Ifnb1/Ifna4 from bone marrow-derived macrophages (BMMΦ)] compared to those expressing human PTPN22^620R (35).

Like PTPN22^620W in humans and transgenic mice, BMMΦ from C57BL/6-Ptpn22 ^−/− mice had impaired TLR4-induced T1-IFN (Ifnb1 and Ifna4 mRNA production were ~50% less) and decreased TLR4- and TLR3-induced IFN-β production (~60% less) compared to WT BMMΦ ( Figure 4A ). BMMΦ from C57BL/6 mice reconstituted with PEP^227S, a phosphatase-inactive mutant, restored TLR-induced Ifnb1 expression indicating that the phosphatase activity of PEP is not required in this process. C57BL/6-Ptpn22 ^−/− BMMΦ have reduced K63-linked polyubiquitination of TRAF3 following LPS stimulation compared to WT BMMΦ. These data are not confined to mouse BMMΦ, pDCs from C57BL/6-Ptpn22 ^−/− mice and BXSB/MpJ-Ptpn22 ^−/− mice had fewer pDCs making IFNα (~50% reduction) and the pDCs that were making IFNα made less than pDCs from WT mice (again ~50% reduction) (102). Also like PTPN22^620W humans, C57BL/6-PEP^619W mice had significantly reduced TLR-7-driven T1-IFN serum levels following injection of R848 compared to C57BL/6 mice (~3 ng/ml in C57BL/6-PEP^619W mice vs. 5 ng/ml in C57BL/6 mice) (122). The combined data from mice and humans shows that both PTPN22^620W and PEP^619W are loss-of-function variants with respect to TLR-induced T1-IFN resulting in reduced T1-IFN following TLR stimulation ( Figure 4A ) (35). TLR stimulation does not only induce T1-IFN, it is also capable of priming the NLRP3 inflammasome for subsequent activation following an inflammatory stimulus such as murmamyldipeptide (MDP), an aganoist of nucleotide-binding oligomerization domain-containing protein (NOD2) that is a component of bacterial cell walls. The role of PEP/PTPN22 allotypes in NLRP3 inflammasome may also impact autoimmunity.

PTPN22^620R positively regulates activation of NLRP3 and subsequent release of IL-1β ( Figure 4B ). PTPN22^620W is a gain-of-function variant that potentiates NLRP3 activity ( Figure 4B ) (36, 37). PTPN22 dephosphorylates NLPR3 at Y861 which prevents it from being sequestered into phagophores and degraded via autophagy (36, 37). PTPN22 knockdown in THP-1 macrophages primed with ultrapure LPS (upLPS) led to increased NLRP3 phosphorylation and increased NLRP3 sequestration in autophagosomes, with a concomitant reduction in IL-1β secretion ranging from about 50% with MDP treatment and up to 80% with monosodium urate (MSU) treatment (36, 37). In support of this, inhibiting autophagy restored IL-1β secretion from PTPN22 knockdown THP-1 cells (37). PTPN22^620W is a gain-of-function variant and is better able to dephosphorylate NLRP3 and prevents its sequestration into phagophores and subsequent degradation ( Figure 4B ). PTPN22^620W has an enhanced capacity to dephosphorylate NLRP3 in a cell free system compared to PTPN22^620R (36). When moDCs from PTPN22^620R/W donors were primed with ultrapure LPS and treated with monosodium urate (MSU) cleaved caspase-1 was increase by 500% and produced 300% more mature IL-1β compared to PTPN22^620R/R donors (36).

Much like THP-1 cells with PTPN22 knockdown, C57BL/6-Ptpn22 ^−/− mice exhibited a 50% reduction in MDP-, MSU-, and ATP-induced IL-1β secretion from BMDCs compared to those of Ptpn22-competent mice and this effect was abrogated by inhibition of autophagy ( Figure 4B ) (37). This is due to the catalytic activity of PTPN22. In C57BL/6-Ptpn22 ^−/− BMDCs expressing the catalytically dead human PTPN22 (PTPN22^263Q) the same effect was observed. Similar to moDCs from PTPN22^620R/W donors, BMDCs from C57BL/6-PEP^619W mice have less NLRP3 in autophagosomes upon upLPS/MSU treatment and over 50% increased IL-1β secretion compared to C57BL/6 mice (36, 37). The same was seen when comparing BMMΦ from C57BL/6-PEP^619W mice with C57BL/6 mice (36). Taken together, these data demonstrate that PTPN22^620W and PEP^619W are gain-of-function variants with respect to NLRP3 dephosphorylation and enhance NLRP3-inflammasome activation and mature IL-1β release. While signaling via NOD2 is capable of activating the NLRP3 inflammasome following priming with LPS, it also induces autophagy and cytokine secretion.

PTPN22 is a negative regulator of NOD2-induced autophagy ( Figure 5A ). Knockdown of PTPN22 via shRNA in THP-1 monocytes enhanced NOD2-induced LC3B-II, a cleaved and activated form of LC3B indicative of autophagosome formation. There was also a decrease in p62 protein levels consistent with enhanced autolysosome activity. Knockdown of PTPN22 via shRNA in THP-1 monocytes also led to enhanced JNK, p38, NF-κB-p65, and NF-κB-p50, activation downstream of NOD2 while reducing ERK activation. Enhanced NOD2-induced IL-6 and TNF mRNA expression and IL-6, IL-8, and TNF secretion were also seen with PTPN22 knockdown (36, 39). In addition, the reduction in PTPN22 resulted in decreased NOD2-induced ICAM1, NOD2, T-bet, and IFN-γ mRNA expression as well as reduced IFN-γ secretion (39). Interestingly, the variant in PTPN22 is associated with reduced risk of Crohn’s disease while loss-of-function mutations in NOD2 are associated with increased risk of Crohn’s disease (123, 124). This could indicate that the T1D-risk allotype (PTPN22^620W) enhances NOD2 activity to suppress gastrointestinal pathology, however, more studies are necessary to clarify how PTPN22^620W alters NOD2 response compared to PTPN22^620R. These PTPN22 knockdown studies indicate that PTPN22 negatively regulates NOD2-induced autophagy, IL-6, IL-8, and TNF production while positively regulating NOD2-induced ICAM1, NOD2, and IFN-γ production.

Like PTPN22 knockdown in THP-1 cells, C57BL/6-Ptpn22 ^−/− mice demonstrate that PEP is a negative regulator of NOD2-induced cytokine secretion in BMDCs of mice ( Figure 5A ). BMDCs from Ptpn22 ^−/− mice treated with MDP had increased p38, NF-κB p65, and NF-κB p50 phosphorylation, and decreased ERK phosphorylation compared to Ptpn22-competent BMDCs. MDP-treated BMDCs from Ptpn22 ^−/− mice had increased levels of IL6 and TNF but decreased levels of NOD2, ICAM-1, and IFNγ mRNA compared to Ptpn22 competent BMDCs (39). MDP-treated Ptpn22 ^−/− BMDCs had enhanced IL-6, IL-8, and TNF secretion compared to Ptpn22-intact BMDCs (39). These data closely mirror data from PTPN22 knockdown THP-1 cells and demonstrate that PTPN22^620R and PEP^619R are negative regulators of NOD2-induced autophagy and cytokine secretion ( Figure 5A ). PTPN22 does not only influence signaling downstream of TLRs and other pattern recognition receptors in monocytes, macrophages, and DCs, it also influences cytokine secretion and signaling in response to IFNγ.

PTPN22 regulates IFN-γ receptor (IFNγR) signaling in human monocytes ( Figure 5B ). PTPN22 knockdown in THP-1 monocytes followed by treatment with IFNγ induced increased SOCS1 phosphorylation and activity and reduced protein levels of SOCS3 compared to control siRNA transfected cells. PTPN22 pulls down with SOCS1, suggesting that PTPN22 may be responsible for dephosphorylating and inactivating SOCS1 when it is present. In agreement with this, PTPN22 knockdown reduced activation (phosphorylation) of known SOCS1 targets, Jak1, STAT1, and STAT3 in response to IFNγ. It also reduced subsequent production of ICAM1 (~70% reduced), NOD2 (~15% reduced), and T-bet mRNA (~40% reduced) when compared to control siRNA transfected cells. Knockdown of PTPN22 also decreased IFNγ-induced MCP-1 (~70% reduced), IL-8 (~50% less), and IL12p40 (~75% reduced) secretion (29). These data indicate that PTPN22 is a positive regulator of STAT1 and STAT3 activation following IFNγ treatment. Activation of STAT1 and subsequent gene induction is the most well characterized portion of IFNγR signaling, however, the signaling cascade activated by the IFNγR includes many other signaling molecules. Treatment with IFNγ also induces signaling via p38 MAPK and Src. Upon knockdown of PTPN22 in THP-1 monocytes, IFNγ-induced p38 MAPK activation and subsequent IL-6 mRNA expression and protein production were enhanced compared to control siRNA transfected cells. This suggests that PTPN22 is negatively regulating p38 MAPK activation downstream of the IFNγR. It is unknown how PTPN22 regulates p38 MAPK activation downstream of the IFNγR, however, there are several plausible targets. Current literature indicates that p38 MAPK is activated by the IFNγR via a signaling cascade involving JAK2, Pyk2, MEKK4, MEK6, and finally p38 MAPK (125, 126). Pyk2, MEKK4, and p38 MAPK are attractive potential targets of PTPN22 because they are all activated by phosphorylation on a tyrosine residue. At this time, more targeted research is necessary to define the PTPN22 target(s) in this pathway. Similarly, PTPN22 knockdown induced basal Src phosphorylation that increased after IFNγ treatment; however, in control siRNA transfected cells there was no basal Src phosphorylation nor was there IFNγ-induced Src phosphorylation. This indicates that PTPN22 negatively regulates basal Src activation and IFNγR-induced Src activation (29). While PTPN22 influences response to IFNγ treatment alone it also influences macrophage cytokine secretion following polarization in response to IFNγ/LPS or IL-4/IL-13 treatment.

In primary MDMs, PTPN22 is a negative regulator of IL-12 and IL-23 production following M1 polarization ( Figure 6A ) and a positive regulator of IL-10 production following M2 polarization ( Figure 6B ). PTPN22 knockdown in MDMs led to increased IL-23 (~60% more) and IL-12 (~30% more) secretion upon IFNγ/LPS treatment (M1 polarization) and decreased IL-10 expression (~50% less) following IL-4/IL-13 treatment (M2 polarization). PTPN22^620W appears to be a gain-of-function negative regulator of IL-12 and IL-23 production following M1 polarization ( Figure 6A ). M1 polarized macrophages from PTPN22^620W/W donors expressed significantly less IL-12, IL-1β, and IL-6 than those from PTPN22^620R/R donors. It is thought that this gain-of-function phenotype is due to enhanced expression of PTPN22^620W upon M1 polarization. M1 polarized macrophages from PTPN22^620W/W donors expressed significantly more PTPN22 than those from PTPN22^620R/R donors. PTPN22^620W and PTPN22^620R are comparable positive regulators of M2 polarization with no differences in IL-10 expression following IL-4/IL-13 treatment ( Figure 6B ) (38).

Like PTPN22 knockdown in human MDMs, splenic macrophages from C57BL/6-Ptpn22 ^−/− mice had increased expression of IL-23 (~200%) and IL-12 (~250%) following M1 polarization ( Figure 6A ) and decreased expression of IL-10 (~50%) following M2 polarization ( Figure 6B ) compared to those from Ptpn22-intact mice (38). These Ptpn22 ^−/− splenic macrophages had increased NF-κB activity (~200%) compared to Ptpn22-intact macrophages and this could explain the increase in LPS/IFNγ-induced IL-12 and IL-23. Splenic macrophages from C57BL/6-Ptpn22 ^−/− mice reconstituted in vitro with PEP^619R or PEP^619W and then polarized to M1 or M2 macrophages had no difference in gene expression. If the level of PEP expression is important in mouse macrophages like the level of PTPN22 expression is in human MDMs, then reconstituting macrophages with the same amount or PEP^619R and PEP^619W would not capture the effects seen in human MDMs where PTPN22^620W and PTPN22^620R expression levels are different (38). Like human PTPN22^620W M1 macrophages, M1 peritoneal macrophages from C57BL/6-PEP^619W mice had lower mRNA levels for the M1 genes, iNOS (~50 fold less) and TNF (~2 fold less), than those from WT mice (127). Overall, these data indicate that PTPN22^620W and PEP^619W are gain-of-function negative regulators of macrophage cytokine secretion following M1 polarization due to increased PTPN22 expression ( Figure 6A ). PTPN22 has multiple roles in macrophage polarization and in fact influences diverse functions in monocytes, macrophages, and DCs. The T1D-associated variant of PTPN22, PTPN22^620W, influences a large number of these functions and these cellular phenotypes could contribute to the pathogenesis of T1D.

Monocytes, macrophages, and DCs are APCs that are all capable of initiating and enhancing adaptive immune responses. The precipitating events that lead to loss of tolerance and the development of T1D are unknown; be it physiological β cell death, viral infection, bacterial infection, or some other initiating event, monocytes, macrophages, and DCs are the cells most likely to sense β cell death/inflammation and initiate the adaptive immune response. After APCs trigger the adaptive immune response, these cells enhance and support the ongoing immune response against β cells. In APCs, PTPN22^620R plays a role in signaling downstream of many PRRs [i.e., TLR4 (35), TLR7/8 (122), NOD2 (36, 37, 39)], cytokine receptors [i.e., IL-4R/IL-13R (38), and IFNγR (29, 128)]. PTPN22^620W enhances NLRP3 activation and subsequent IL-1β release following priming via TLR4 (LPS) and treatment with a NOD2 agonists (MDP) while dampening the T1-IFN response following TLR4/7/8 stimulation. The combination of these phenotypes renders APCs from PTPN22^620R/W and PTPN22^620W/W humans more sensitive to NLRP3 activation while dampening their ability to produce T1-IFNs in response to PRR signaling. IL-1β enhances naïve and memory CD4 T cell expansion and this could in turn exacerbate activation of autoreactive CD4 T cells during the initiation of T1D (129). T1-IFNs enhance CD8 T cell activation and support activated T cell survival and are considered a major feature of the diabetic islet microenvironment where they enhance expression of MHC-I on β cells and expression of T cell chemoattractants (e.g., CXCL10) (130–132). Importantly, the T1-IFN phenotype results in a reduction of T1-IFN and not a complete loss. This might reduce the induction of MHC-I and T cell chemoattractants, however, it would not ablate them and in a genetically predisposed individual this may still be more than sufficient to help initiate and sustain T1D especially in combination with enhanced IL-1β production.

PTPN22 is a negative regulator of protein citrullination and NETosis and PTPN22^620W is a loss-of-function variant (30) ( Figure 7A ). Neutrophils from heterozygous PTPN22^620R/W donors displayed a hypercitrullinated protein profile (~4 fold more in PTPN22^620R/W neutrophils), they had enhanced citrullination of histone H3, a marker of NETosis (~5 fold more in PTPN22^620R/W neutrophils), and they were more prone to NETosis (3%–15% of PTPN22^620R/W neutrophils vs. ~2% of PTPN22^620R/R neutrophils) compared to those from PTPN22^620R/R donors (30, 136). PAD4 co-immunoprecipitated PTPN22 in human neutrophils and PTPN22 allotype influences this interaction; there is a significantly decreased amount of PTPN22 co-immunoprecipitated with PAD4 in heterozygous PTPN22^620R/W donors when compared to PTPN22^620R/R donors (~66% decreased). The total PTPN22 protein level was the same between donors implying that PTPN22^620R interacts with PAD4 more than PTPN22^620W. In C57BL/6-Ptpn22 ^−/− mouse macrophages transfected with human PTPN22^620R or PTPN22^620W expressing constructs, PTPN22^620R but not PTPN22^620W reduced protein citrullination and co-immunoprecipitated with PAD4 further supporting the lack of association of PTPN22^620W with PAD4 (30).

Much like in human neutrophils, PEP in C57BL/6 mouse neutrophils interacts with PAD-4. PEP co-immunoprecipitated with PAD-4. The absence of PEP in C57BL/6 mice enhanced protein citrullination by approximately 100%; however, the enhanced protein citrullination was abrogated in the presence of a catalytically dead PEP indicating that the catalytic activity of PEP is not involved in this process. Unlike in humans, PEP does not specifically impact histone H3 citrullination or NETosis in mouse neutrophils (30). Taken together, these data indicate that PTPN22^620R is a negative regulator of protein citrullination and NETosis in human neutrophils and PTPN22^620W is a loss-of-function variant ( Figure 7A ).

PTPN22 plays a role in transmigration across inflamed endothelium, as well as the response to fMLP, a highly chemotactic n-formylated oligopeptide actively released by invading bacteria or passively released by mitochondria of dying host cells (31, 137, 138). Significantly more neutrophils from PTPN22^620R/W donors transmigrate across inflamed (TNF treated) endothelium over 2 min than those from PTPN22^620R/R donors (PTPN22^620R/W = 43 ± 9% vs. PTPN22^620R/R = 24 ± 4%). Stimulation of neutrophils from healthy PTPN22^620R/W donors with fMLP resulted in increased calcium flux compared to neutrophils from healthy PTPN22^620R/R donors (PTPN22^620R/W = 0.28 ± 0.02 vs. PTPN22^620R/R = 0.24 ± 0.02 Indo-1 ratio). Priming of neutrophils from healthy PTPN22^620R/W donors with TNF followed by stimulation with fMLP resulted in significantly increased ROS production (4-fold increase) compared to PTPN22^620R/R donors ( Figure 7B ) (31).

Unlike in humans, PEP does not appear to play a role in transmigration across inflamed endothelium or the response to fMLP in C57BL/6 mice. Ptpn22 ^−/− and Ptpn22-intact mouse neutrophils migrated across TNF-treated endothelium at the same rate (139). Ptpn22 ^−/− and PTPN22-intact neutrophils produce similar amounts of ROS in response to fMLF (also called fMLP) and PMA stimulation, however, they were not primed with TNF like the human neutrophils which may explain why there was no difference in ROS production. Neutrophils from C57BL/6-Ptpn22 ^−/− mice did however exhibit decreased ROS production (~50% reduced) and degranulation (~25% reduced) in response to FcγR and integrin stimulation compared to neutrophils from C57BL/6 mice. These pathways have not been investigated in the context of PTPN22 in humans (139). In human neutrophils, PTPN22^620W enhances transmigration across inflamed endothelium, calcium flux in response to fMLP stimulation, and ROS production in response to TNF priming followed by fMLP stimulation.

Current data indicates that neutrophils most likely do not play a direct role in the pathogenesis of T1D in humans (133, 134) and it is apparent that they do not influence pathogenesis in NOD mice; depletion of neutrophils starting at 4 weeks of age does not impact development of T1D in NOD mice (134). While data indicate that neutrophils do not play a role in human T1D pathogenesis, many neutrophil products (e.g., ROS, NETs, cytokines) are capable of damaging tissues, including pancreatic β cells (140). Neutrophils from PTPN22^620R/W donors had enhanced calcium flux and ROS production in response TNF priming followed by treatment with fMLP compared to those from PTPN22^620R/R donors. These PTPN22^620R/W neutrophils also transmigrated across TNF-inflamed epithelium faster than their PTPN22^620R/R counterparts, displayed enhanced protein citrullination, and were more prone to NETosis (30, 31, 136). The combined effects of these phenotypes mean that PTPN22^620R/W and PTPN22^620W/W patients with T1D could display enhanced neutrophil accumulation in the exocrine pancreas due to enhanced transmigration across inflamed epithelium and increased frequency of these infiltrating neutrophils releasing NETs and producing high amount of ROS. More studies need to be undertaken to understand if neutrophils participate in the pathogenesis of human T1D and if the influence of PTPN22⁶²⁰ allotype effects their participation. Overall, it is clear that PTPN22 plays diverse roles in many cell types that have the potential to influence the pathogenesis of T1D.