PTPN22 as a novel therapeutic target: a key intracellular checkpoint for NK cell therapy

Abstract

Background

PTPN22, a protein tyrosine phosphatase (PTP) family member, has recently emerged as an intracellular regulator in adaptive immune cells. However, its function in natural killer (NK) cells remains unclear despite its high expression in innate immune cells.

Methods

We analyzed single-cell RNA-sequencing (scRNA-seq) datasets and The Cancer Genome Atlas (TCGA) to evaluate PTPN22 expression in NK cells and its association with clinical outcomes. We established a CRISPR-mediated PTPN22 knockout platform for NK cell gene editing and evaluated PTPN22 function using both genetic deletion and pharmacological inhibition across multiple NK cell platforms. Cytotoxicity and cytokine production were assessed against hematologic and solid tumor targets, including CD19⁺ and mesothelin-positive (MSLN⁺) cancer cells. PTPN22 knockout (PTPN22^KO) anti-CD19/CAR-NK92 cells were tested in Nalm6 xenograft models.

Results

Published scRNA-seq datasets showed that PTPN22 was highly expressed in tumor-infiltrating NK cells. PTPN22^high NK cells exhibited stress and dysfunction signatures within inflammatory and immunosuppressive tumor microenvironments (TME), with elevated PTPN22 expression in TCGA associated with poorer patient survival. Genetic deletion of PTPN22 or pharmacological inhibition enhanced NK-cell cytotoxicity and pro-inflammatory cytokine release and improved the antitumor activity of chimeric antigen receptor (CAR)-NK cells against CD19⁺ and MSLN⁺ targets. PTPN22^KO preserved CAR-NK cell function under cytokine-based TME-like conditions and improved the functional persistence of primary anti-CD19 CAR-NK cells during prolonged tumor exposure. PTPN22^KO NK and CAR-NK cells showed increased or preserved signal transducer and activator of transcription 3 (STAT3) phosphorylation and reduced FAS expression under inflammatory conditions. Notably, PTPN22^KO anti-CD19/CAR-NK92 cells demonstrated superior anti-leukemic activity and significantly prolonged survival compared with non-knockout controls in vivo.

Conclusions

These findings identify PTPN22 as a novel intracellular checkpoint that limits NK and CAR-NK cell function. Targeted deletion of PTPN22 represents a promising strategy to enhance the therapeutic efficacy and persistence of both unmodified and CAR-engineered NK cells for cancer immunotherapy.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• PTPN22 is an intracellular immune regulator highly expressed in multiple immune cell subsets, including T cells and dendritic cells; however, its role in natural killer (NK) cells remains unclear.

WHAT THIS STUDY ADDS

• CRISPR/Cas9-mediated deletion of PTPN22 enhanced antitumor activity across multiple NK cell platforms, including NK92, primary human NK, and chimeric antigen receptor (CAR)-NK cells. Compared with non-edited controls, PTPN22-knockout anti-CD19/CAR-NK92 cells exhibited superior in vivo anti-leukemic efficacy and maintained functionality under cytokine-based tumor microenvironment-like conditions, associated with preserved signal transducer and activator of transcription 3 (STAT3) signaling and reduced FAS-mediated apoptotic susceptibility.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Our study establishes PTPN22 as a targetable intracellular checkpoint for NK cell-based immunotherapy, providing a clinically translatable gene-editing strategy to improve the potency and persistence of adoptive NK and CAR-NK cell therapies and support the development of next-generation immune cell therapies for overcoming inflammatory and immunosuppressive tumor microenvironments.

Introduction

Adoptive transfer therapy using natural killer (NK) cells has shown promise in treating both hematologic malignancies and solid tumors.^{1 2} To improve the specificity of NK cells toward cancer cells, chimeric antigen receptor (CAR)-engineered NK (CAR-NK) cells have been developed and are now widely tested in both preclinical and clinical settings.^{3 4} Notably, CAR-NK cells targeting CD19 have demonstrated encouraging early results in clinical trials.^5–8 However, several factors can limit the therapeutic efficacy of NK and CAR-NK cells within the tumor microenvironment (TME),^9–11 including immunosuppression, hypoxia, antigen loss, and inflammation, which is typically associated with a favorable response to immunotherapy in so-called “hot tumors”.^{11 12}

Paradoxically, excessive or chronic exposure to pro-inflammatory cytokines in the TME can suppress immune cell function and contribute to resistance to immunotherapy.^13–17 Chronic inflammation-associated cytokines such as transforming growth factor (TGF)-β, interleukin (IL)-6, and IL-10 can downregulate activating receptors (NKG2D, NKp30),^{18 19} impair mTOR pathway signaling leading to metabolic exhaustion,²⁰ and upregulate inhibitory checkpoints including PD-1, TIGIT, NKG2A, and KIRs.^{16 21} Although these extracellular and surface inhibitory pathways have been extensively studied, intracellular checkpoints that regulate NK cell function under inflammatory TME conditions remain poorly defined.

In this study, we identified PTPN22 as an intracellular checkpoint that regulates NK and CAR-NK cell function under inflammatory conditions. Members of the protein tyrosine phosphatase (PTP) family play important roles in regulating both innate and adaptive immunity.^{22 23} For example, PTPRC (CD45) is a well-established marker for lymphocytes.²⁴ Other PTPs, including PTPN1, PTPN2, PTPN6, and PTPN11, have been implicated in modulating NK cell activity.^{23 25–27} PTPN22, another member of this family, is best known for its role in autoimmunity and has recently emerged as a potential target for immunotherapy.^28–30 PTPN22 regulates T-cell receptor signaling and effector functions in T cells.^{31 32} A recent study by Ho et al introduced a small-molecule inhibitor of PTPN22 (L1) that boosted CD8⁺ T-cell activity against colon cancer, and, interestingly, NK cell numbers also increased in L1-treated mice.³⁰ However, the functional role of PTPN22 in human NK cells remains to be elucidated.

Gene editing is a powerful tool for overcoming barriers to NK cell function posed by the TME,^{3 33} and the rise of CRISPR/Cas9 technology has paved the way for the development of more potent adoptive NK and CAR-NK cell therapies.³³ Strategies to enhance NK cell efficacy now include targeting not only extracellular immune checkpoints but also intracellular regulatory mechanisms.^34–37 For example, targeting intracellular checkpoint proteins like cytokine-inducible SH2-containing protein can enhance both NK and CAR-NK cell activity.³⁷

We investigated the role of PTPN22 as a novel intracellular checkpoint regulator across multiple NK cell platforms, using both small-molecule inhibitors and CRISPR-mediated PTPN22 knockout. We used non-viral CRISPR/Cas9 gene editing with ribonucleoprotein (RNP) complexes to delete PTPN22 in NK92 cells and primary human NK cells. Our data show that PTPN22 modulates NK and CAR-NK cell functions, at least in part, in association with changes in STAT3 phosphorylation and FAS-mediated apoptotic susceptibility. Under various cytokine-based TME-like conditions, PTPN22^KO CAR-NK cells maintained their cytotoxic potential, whereas non-knockout control cells exhibited signs of dysfunction. Together, these findings identify PTPN22 as a previously unrecognized intracellular checkpoint in human NK cells and suggest that PTPN22 deletion may help preserve CAR-NK cell function in immunosuppressive TMEs.

Materials and methods

NK cell expansion ex vivo

Blood donated by healthy donors was separated by gradient centrifugation using Lymphoprep solution (Axis-Shield, Dundee, UK) to isolate peripheral blood mononuclear cells (PBMCs). Isolated PBMCs were washed two times with 1× phosphate-buffered saline (PBS, WelGENE). NK cells were isolated from PBMCs using negative selection with the NK Cell Isolation Kit (Miltenyi Biotec, GmbH, Germany). NK cells were then cultured in NK Magnetic-Activated Cell Sorting (MACS) medium with 10% Fetal Bovine Serum (FBS), 100 U/mL penicillin, 10 µg/mL streptomycin, 4 mM L-glutamine, 1,000 U/mL recombinant human IL-2 (rhIL-2), and 20 ng/mL rhIL-15 (PeproTech, Rocky Hill, New Jersey, USA) for 7 days.

For NK cells expanded with feeder cells, isolated PBMCs were co-cultured with 100 Gy-irradiated conventional K562 feeder cells expressing membrane-bound IL-21/IL-18 in a 24-well plate using Roswell Park Memorial Institute (RPMI)-1640 medium (supplemented with 10% FBS, 100 U/mL penicillin, 10 µg/mL streptomycin, and 4 mmol/L L-glutamine). The medium was replaced every 2 days with fresh medium supplemented with 10 U/mL rhIL-2 (PeproTech). From day 7, the medium was replaced with fresh medium supplemented with 100 U/mL rhIL-2 and 5 ng/mL rhIL-15 (PeproTech).³⁸

CRISPR/Cas9-mediated knockout in primary NK and NK92 cells

PTPN22 sgRNA1 (AAGGCAATCTACCAAGTACA) and sgRNA2 (GGTTTTGTCTGCCTTGTACT) were designed and purchased from Synthego (Redwood City, California, USA) (online supplemental table S1). The sgRNA3 (GGATGTACGTTGTTACCAAG) and tracrRNA were purchased from Integrated DNA Technologies (Coralville, Iowa, USA) (online supplemental table S1). All sgRNAs were rechecked for on-target and off-target scores in Benchling. For a control, the negative control scrambled #1 from Synthego (GCACTACCAGAGCTAACTCA) was used. The crRNA and tracrRNA were mixed at a 1:1 ratio, heated at 95°C for 5 min, and then cooled for 30 min at room temperature. sgRNA and Cas9 were mixed to form an RNP complex and incubated at 25°C for 20 min. Then, 1×10⁶ NK cells were combined with RNP in P3 buffer and electroporated using the Lonza 4D-nucleofector X-Unit (Lonza, Basel, Switzerland) with electroporation programs CM137 (for primary NK cells) and CA137 (for the NK92 cell line). After 72 hours, PTPN22 expression was confirmed by western blot. The cells were then used for further experiments.

CAR construction and lentiviral particle production

CAR constructs were generated in the pCDH-MSCV vector (System Bioscience, California, USA) and consisted of an single-chain variable fragment (scFv) linked to a CD8α transmembrane, a 4–1BB co-stimulatory domain, and a CD3ζ signaling domain. The CD19-targeting CAR used the FMC63-derived scFv (BBz), while the MSLN-targeting CAR shared the same structural design but also incorporated an IL-15 expression cassette (BBz-IL15). Lentiviral particles for both constructs were produced in Lenti-X 293T cells with the pCDH transfer plasmid, the packaging plasmid pSPAX2, and the BaEV-TR envelope plasmid using Lipofectamine 2000 (Invitrogen), following the manufacturer’s instructions. Viral supernatants were collected and concentrated using a commercial lentiviral concentration reagent (LGV-1021B, Lugen Sci).³⁹

Transduction of primary NK cells

NK cell expansion and transduction were performed with minor modifications from previously described methods.^{38 39} PBMCs obtained from healthy donors were co-cultured with 100 Gy gamma-irradiated K562 feeder cells engineered to express membrane-bound IL-18/IL-21 at a 2:1 PBMC-to-feeder ratio in RPMI-1640 medium supplemented with 10 IU/mL of rhIL-2, with half-medium changes every 2–3 days. On day 5 of expansion, NK cells were transduced with concentrated lentivirus particles in the presence of 20 µg/mL Retronectin (Takara) and 10 µg/mL Vectorfusin-1 (Miltenyi Biotec). The virus-cell mixture was then incubated in RPMI-1640 medium supplemented with rhIL-2 (200 IU/mL) and rhIL-15 (5 ng/mL).

For the double gene editing, 1×10⁶ primary expanded NK cells (day 5) were first subjected to RNP delivery using CRISPR/Cas9, followed immediately by CAR transduction. The cells were then allowed to recover for at least 2 days before assessing editing efficiency and functional activity.

Mouse acute lymphoblastic leukemia model

To evaluate the in vivo efficacy of PTPN22^KO anti-CD19/CAR-NK92 cells, we used immunodeficient NOD.Cg-Prkdc^scidIL2rg^tm1Wjl/Szj mice (Jackson Laboratory, Bar Harbor, Maine, USA), which are widely used for testing CAR efficacy. 6–12 weeks old immunodeficient female NSG mice were used in this study. The mice were maintained under sterile conditions. A total of 1×10⁵ Nalm6-Luc cells were intravenously injected into the mice. After 3 days, the mice received either 5×10⁶ (low-dose) or 1×10⁷ (high-dose) Cas9 or PTPN22^KO anti-CD19/CAR-NK92 cells. 25,000 U of hIL-2 were co-administered to support the anti-CD19/CAR-NK92 cells. The control group received 1×PBS. Bioluminescence imaging was performed every 3–4 days. Before imaging, mice were injected with 150 mg/kg D-Luciferin (PerkinElmer, Waltham, Massachusetts, USA). Bioluminescence signals were captured using the NightOwl Imaging System (Berthold Technologies, Bad Wildbad, Germany). The mice were randomly assigned to groups, and all mice were included in the analysis without any exclusion.

Statistical analysis

Data were presented as single dots or mean±SEM. Data were analyzed using one-way or two-way analysis of variance (ANOVA) and paired or unpaired t-tests; p<0.05 was considered significant. Post hoc analysis using Tukey’s test was performed to confirm differences between groups identified by ANOVA. All data were analyzed using GraphPad Prism V.9 software (GraphPad Software, San Diego, California, USA).

All other materials and methods are described in detail in the online supplemental materials including online supplemental tables S1 and S2.

The schematic illustrations were generated using BioRender.

Results

PTPN22 expression in tumor-infiltrating NK cells is associated with worse patient outcomes

To characterize PTPN22 expression across immune cell types, we analyzed published single-cell RNA sequencing (scRNA-seq) datasets using the TISCH2 platform.⁴⁰ PTPN22 was most prominent in exhausted CD8⁺ T cells and NK cells, while it remained relatively low in monocytes and macrophages (online supplemental figure S1A). Among PTP family members, PTPN22, along with PTPN7 and PTPN2, was the most highly expressed in tumor-infiltrating NK cells (online supplemental figure S1B).

We further examined patient data from The Cancer Genome Atlas to determine whether high PTPN22 expression affects patients with a high NK cell signature. High PTPN22 expression was associated with worse survival in patients with primary tumors (p=0.0013; figure 1A). In a subgroup analysis with high NK cell abundance (NK^High), patients with an NK^HighPTPN22^High profile had poorer survival outcomes compared with the NK^HighPTPN22^Low group (figure 1B). Further analysis of the TME revealed that PTPN22 expression correlated with elevated expression of most immune checkpoint genes, suggesting that high PTPN22 expression is associated with an exhausted immune microenvironment (figure 1C).

Figure 1. PTPN22 expression in tumor-infiltrating NK cells is associated with worse patient outcomes. Data from the GDC Pan-Cancer project of The Cancer Genome Atlas on patients with primary tumors were used to evaluate associations among patient survival, NK levels, and PTPN22 expression. (A) Survival of patients with low and high expression of PTPN22. (B) Survival of patients with low and high expression of PTPN22 in the NK^High signature. (C) Correlation of PTPN22 with other PTP family members and immune checkpoints in patients with NK^High signature. (D) Hallmark pathway analysis between PTPN22^High and PTPN22^Low in patients with NK^High signature. Significance was calculated by the log-rank test (A and B). The correlation in (C) was calculated by the Pearson test. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. GSEA, Gene Set Enrichment Analysis; NK, natural killer; PTP, protein tyrosine phosphatase.

We then performed differential gene expression analysis between NK^HighPTPN22^High and NK^HighPTPN22^Low groups to identify differences in gene signatures that could explain patient outcomes. The NK^HighPTPN22^High group showed upregulation of hallmark pathways related to inflammation (inflammatory response, interferon (IFN)-α response, IFN-γ response, IL-6–Janus kinase (JAK)–signal transducer and activator of transcription 3 (STAT3) signaling, and tumor necrosis factor (TNF)-α signaling via nulclear factor kappa B (NF-κB), and apoptosis (figure 1D). In contrast, the NK^HighPTPN22^Low group was enriched for pathways related to NK cell functionality, such as oxidative phosphorylation and glycolysis (figure 1D). Consistent with the correlation between PTPN22 and inflammatory pathways, analysis of the publicly available cytokine stimulation dataset by Cui et al⁴¹ revealed that inflammatory cytokines, including TNF-α, IL-1β, and IL-1α, upregulated PTPN22 expression (online supplemental figure S1C).

Given the relationship between PTPN22 and inflammatory pathways, we further analyzed a human acute myeloid leukemia (AML) scRNA-seq dataset, which represents a highly inflammatory TME.¹⁴ PTPN22 was upregulated in NK cells from AML patients under high-inflammatory TME conditions (online supplemental figure S1D). Among unsupervised NK-cell clusters, cluster 5 showed the highest PTPN22 expression together with increased TNF gene expression, reduced cytotoxicity genes, and elevated inflammation-response and stress-response genes (online supplemental figure S1E,F).

We further analyzed solid tumor-infiltrating NK cell scRNA-seq datasets from Lozada et al across multiple immunosuppressive solid tumor types.⁴² Across these cohorts, PTPN22^high NK-cell clusters consistently showed reduced expression of cytotoxicity-related genes, including PRF1, GZMB, GZMM, and GZMH, together with increased exhaustion-associated and stress-associated markers (online supplemental figure S2A–G). Several PTPN22^high clusters also exhibited elevated TNF expression, further linking PTPN22 to inflammatory and dysfunctional NK-cell states within the solid tumor TME. Together, these findings suggest that elevated PTPN22 expression defines a dysfunctional NK-cell state across multiple tumor types.

PTPN22 inhibition enhances NK92 antitumor functions

To further assess the role of PTPN22, we used two potent commercially available PTPN22 inhibitors, PTPN22-IN-1 and LTV-1, to block PTPN22 signaling in NK92 cells.^{30 43} NK92 cells were treated with each inhibitor for 24 hours, before performing functional assays (online supplemental figure S3A). Neither inhibitor decreased PTPN22 protein expression levels (online supplemental figure S3B), consistent with previous reports that PTPN22-IN-1 and LTV-1 inhibit PTPN22 enzymatic activity.^{30 43} Both inhibitors showed maximal enhancement of NK-92 cell function at 10 and 20 µM, respectively (online supplemental figure S3C,D); therefore, these concentrations were used throughout the study. Inhibition of PTPN22 signaling using PTPN22-IN-1 significantly enhanced NK92 cytolysis against both hematologic (Raji, Ramos, K562 and Nalm6) and solid malignancies (SK-OV-3, Huh7, HepG2, MDA-MB-231) (online supplemental figure S3E and figure 2A). This inhibition also increased IFN-γ secretion by NK92 cells in response to cancer cell lines (figure 2B).

Figure 2. PTPN22 inhibition enhances NK92 antitumor functions. (A and B) NK92 cells were pretreated with the PTPN22-IN-1 inhibitor (10 or 20 µM) for 24 hours before assaying their antitumor functions with various solid cancer cell lines. Control cells were treated with DMSO. (A) Cytotoxicity of PTPN22-inhibited NK92 cells. (B) Release of IFN-γ by PTPN22-inhibited NK92 cells. (C) Editing efficiency of CRISPR/Cas9-mediated PTPN22 knockout in NK92 cells assessed by western blot. Signal intensities were measured using ImageJ, and fold changes were calculated relative to the Cas9 group. (D) Cytolytic activity of PTPN22^KO NK92 cells against leukemic K562 cells at E:T ratios of 1:1 and 5:1. Cas9, control NK92 cells. (E) IFN-γ release by PTPN22^KO NK92 cells against K562 cells. Significance was calculated by two-way ANOVA with Tukey post hoc test (A and D); one-way ANOVA with Tukey post hoc test (B and E); and unpaired t-test (C). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; DMSO, dimethyl sulfoxide; E:T, Effector-to-Target; IFN, interferon.

Since both PTPN22-IN-1 and LTV-1 target the PTPN22 catalytic domain and may also affect other PTP family members, we used non-viral CRISPR/Cas9 RNP editing to specifically knock out PTPN22 (online supplemental figure S4A), which successfully eliminated most PTPN22 protein expression (figure 2C). Despite previous reports suggesting that PTPN22 variants affect NK cell proliferation,⁴⁴ we observed no change in NK92 expansion post-knockout (online supplemental figure S4B). Compared with Cas9 control cells, NK92 with the PTPN22 knockout (PTPN22^KO) exhibited significantly improved cytotoxicity against various cancer cell lines (figure 2D and online supplemental figure S4C) and increased cytokine secretion in response to cancer cells (figure 2E and online supplemental figure S4D).

PTPN22 suppression enhances anti-leukemia functions of anti-CD19/CAR-NK92 cells

To evaluate the impact of PTPN22 on the effectiveness of CAR-NK cells, we generated anti-CD19 CAR-NK92 (anti-CD19/CAR-NK92) cells that express a receptor comprising the anti-CD19 FMC63 clone, a CD8α transmembrane domain, a 4–1BB (CD137) co-stimulatory domain, and a CD3ζ signaling domain.³⁹ Treatment with PTPN22-IN-1 modestly enhanced the cytotoxicity of these cells against CD19-sensitive Raji and Ramos cells and K562 cells overexpressing CD19 (CD19-OE) (online supplemental figure S5A,B). Notably, PTPN22 inhibition of anti-CD19/CAR-NK92 cells significantly improved the lysis of resistant target Nalm6 cells (online supplemental figure S5B) and substantially enhanced pro-inflammatory cytokine release (TNF-α and IFN-γ) (online supplemental figure S5C,D).

To further assess long-term cytotoxic potential, we pretreated anti-CD19/CAR-NK92 cells with either 10 µM or 20 µM PTPN22-IN-1 and co-cultured them with K562-CD19-OE cells. As maximal cytolysis was observed by 9 hours (data not shown), fresh tumor cells were added every 9 hours to simulate rechallenge. PTPN22 inhibition consistently improved the killing ability of anti-CD19/CAR-NK92 cells across all rechallenge time points (online supplemental figure S5E).

We next generated PTPN22^KO anti-CD19/CAR-NK92 cells using three different sgRNAs (online supplemental table S1 and figure 3A), with both sgRNA1 and sgRNA3 eliminating most PTPN22 protein expression (figure 3B). Importantly, none of the sgRNAs affected the proliferation of anti-CD19/CAR-NK92 cells (online supplemental figure S6A). PTPN22^KO enhanced NK-92 cell-mediated lysis across multiple target cell lines, whereas CD19 CAR expression selectively increased cytotoxicity against CD19⁺ targets but not CD19-negative K562 cells. Interestingly, PTPN22 knockout further enhanced CAR-NK92-mediated lysis of the CD19-positive Ramos, K562-CD19-OE, and Nalm6 target cells (figure 3C).

Figure 3. PTPN22 knockout enhances anti-leukemia functions of anti-CD19/CAR-NK92 cells. (A) Schematic of PTPN22 deletion using CRISPR/Cas9. Three sgRNAs were designed to target exon 2 and exon 8 of the PTPN22 gene. (B) Editing efficiency of CRISPR/Cas9-mediated PTPN22 knockout in anti-CD19/CAR-NK92 cells assessed by western blot. (C) PTPN22^KO anti-CD19/CAR-NK92 cell cytotoxicity was tested against hematologic cancer cells. (D and E) Assessment of potential off-target effects of PTPN22 inhibitors on the cytotoxicity of anti-CD19/CAR-NK92 cells against K562-CD19-OE (D) and Nalm6 (E) cells. (F and G) IFN-γ release (F) and TNF-α release (G) by PTPN22^KO anti-CD19/CAR-NK92 cells after co-culture with CD19⁺ hematologic cancer cells. Significance was calculated by two-way ANOVA with Tukey post hoc test (C, D and E) and unpaired t-test (F and G). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; CAR, chimeric antigen receptor; E:T, Effector-to-Target; IFN, interferon; NK, natural killer; TNF, tumor necrosis factor.

PTPN22^KO anti-CD19/CAR-NK92 cells demonstrated cytolytic activity against K562-CD19-OE and Nalm6 cells that was comparable to that of PTPN22-IN-1-treated anti-CD19/CAR-NK92 cells (figure 3D,E). Treatment of PTPN22^KO CAR-NK92 cells with PTPN22-IN-1 did not further improve their cytolytic activity. In contrast, LTV-1 enhanced the cytotoxicity of PTPN22^KO CAR-NK92 cells against both K562-CD19-OE and Nalm6 cells, resulting in the highest cytotoxicity among all groups. These findings raise the possibility of off-target effects of LTV-1 on other PTPs, including PTPN1 and PTPN2.⁴³ Furthermore, PTPN22 knockout in anti-CD19/CAR-NK92 cells led to increased secretion of the pro-inflammatory cytokines IFN-γ and TNF-α in response to all tested hematologic cancer cell lines (figure 3F,G).

Using dual sgRNAs (sgRNA1+sgRNA2) improved editing efficiency, comparable to that achieved with sgRNA1 or sgRNA3 (online supplemental figure S6B). As expected, both single-sgRNA and dual-sgRNA-mediated targeting of PTPN22 significantly enhanced the anti-CD19/CAR-NK92 cells’ antitumor function against Nalm6 cancer cells (online supplemental figure S6C–E). Additionally, we tested the Cas9 HiFi variant, which reportedly reduces off-target effects when compared with wild-type (WT) Cas9.⁴⁵ PTPN22 knockout using Cas9 HiFi also enhanced anti-CD19/CAR-NK92 cytotoxicity and degranulation compared with the Cas9 control, consistent with observations using WT Cas9 (online supplemental figure S6F,G). These results support that the enhanced effector functions observed after PTPN22 deletion are mainly attributable to loss of PTPN22 rather than sgRNA-related or Cas9-related off-target effects.

Targeting PTPN22 signaling enhances anti-MSLN CAR-IL-15-NK92 effector functions against ovarian cancers

We next evaluated PTPN22 inhibition in a solid tumor-targeting CAR-NK model, using anti-MSLN/CAR-IL-15-NK92 cells. Inhibition of PTPN22 signaling using either PTPN22-IN-1 or LTV-1 significantly enhanced the cytolytic activity of anti-MSLN/CAR-IL-15-NK92 cells against MSLN⁺ SK-OV-3 cells (online supplemental figure S7A,B). Moreover, anti-MSLN/CAR-IL-15-NK92 cells in combination with a PTPN22 inhibitor exhibited increased IFN-γ release compared with the dimethyl sulfoxide (DMSO) control (online supplemental figure S7C). We also knocked out PTPN22 in anti-MSLN/CAR-IL-15-NK92 cells. The PTPN22^KO cells enhanced SK-OV-3 cell lysis when compared with the Cas9 control anti-MSLN/CAR-IL-15-NK92 cells (online supplemental figure S7D). These findings suggest that the functional benefits of PTPN22 inhibition are not restricted to anti-CD19/CAR-NK92 cells but extend to other CAR-NK platforms.

PTPN22 knockout potentiates anti-CD19/CAR-NK92 activity in an acute lymphoblastic leukemia model

We next evaluated the in vivo efficacy of PTPN22^KO anti-CD19/CAR-NK92 cells in an intractable Nalm6 leukemia model, using luciferase-expressing Nalm6 cells (Nalm6-Luc) and in vivo imaging (figure 4A). NSG mice were injected with the Nalm6-Luc cells before receiving three intravenous injections of 5×10⁶ PTPN22^KO or Cas9 control anti-CD19/CAR-NK92 cells, both with approximately 70% CAR expression (online supplemental figure S8A). The PTPN22^KO anti-CD19/CAR-NK92 group demonstrated enhanced control of Nalm6 tumor growth compared with PBS-treated groups (online supplemental figure S8B,C). Furthermore, PTPN22^KO CAR-NK92 cells showed improved effector activity in lung, liver, and spleen tissues relative to the PBS group, whereas Cas9 control CAR-NK92 cells failed to demonstrate such activity (online supplemental figure S8D). However, PTPN22^KO did not enhance CAR-NK92 cell function in challenging-to-infiltrate tissues such as the brain or bone marrow, which may explain the eventual tumor escape (online supplemental figure S8D). Importantly, this improved tumor control, achieved with a relatively low dose of PTPN22^KO anti-CD19/CAR-NK92 cells, translated into prolonged mouse survival compared with both PBS and Cas9 controls (online supplemental figure S8E).

Figure 4. PTPN22 knockout potentiates anti-CD19/CAR-NK92 activity in an acute lymphoblastic leukemia model. (A) Experimental design of an in vivo study. NSG mice were injected with 1×10⁵ Nalm6-luciferase cells, followed by either three intravenous injections of 5×10⁶ anti-CD19/CAR-NK92 cells (Scheme 1, low-dose) or two injections of 1×10⁷ anti-CD19/CAR-NK92 cells (Scheme 2, high-dose), with bioluminescence imaging by IVIS. Both groups received IL-2 (25,000 U) with the injection of the Cas9 or PTPN22^KO anti-CD19/CAR-NK92 cells. (B) Bioluminescence imaging of tumor burden, (C) quantification of tumor bioluminescence, and (D) Kaplan-Meier survival curves of high-dose groups (PBS n=2, Cas9 n=7, PTPN22^KO n=8). (E) Combined survival analysis from both experimental batches. (F) Body weight of mice receiving a high-dose treatment with CAR-NK92 cells. Significance was calculated by two-way ANOVA with Tukey post hoc test (C and F). Survival significance was analyzed using the log-rank test (D and E). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; CAR, chimeric antigen receptor; IL, interleukin; IVIS, in vivo imaging system; NK, natural killer; PBS, phosphate-buffered saline.

We also tested a high-dose approach, using two doses of 1×10⁷ PTPN22^KO anti-CD19/CAR-NK92 cells. Consistent with the low-dose results, high-dose PTPN22^KO anti-CD19/CAR-NK92 cells exhibited superior anti-leukemic activity relative to Cas9 controls (figure 4B,C), again resulting in significantly extended mouse survival (figure 4D). Across both the low-dose and high-dose studies, PTPN22^KO anti-CD19/CAR-NK92 cells consistently prolonged survival compared with control groups, and body weight remained stable throughout the treatment period, supporting treatment tolerability under the tested conditions (figure 4E,F and online supplemental figure S8F).

STAT3-associated signaling contributes to the functional effects of PTPN22 targeting in NK and CAR-NK cells

We further explored the relationship between PTPN22 knockout and NK92 cell activation. PTPN22 knockout did not alter the expression of NK92 activation markers (online supplemental figure S9A). As STAT3, STAT5, and ERK are among the best-characterized signaling pathways involved in NK-cell activation, we investigated whether PTPN22 targeting was associated with changes in these pathways. PTPN22^KO NK-92 cells exhibited increased STAT3 phosphorylation (pSTAT3) compared with Cas9 controls (figure 5A), whereas no differences were observed in STAT5 or ERK phosphorylation (online supplemental figure S9B). This increase in STAT3 activation was not limited to CRISPR/Cas9-mediated gene editing, as both PTPN22-IN-1-treated and PTPN22^KO NK-92 cells exhibited elevated pSTAT3 levels compared with their respective controls (figure 5B). Following stimulation with HepG2 cancer cells, PTPN22-IN-1-treated NK-92 cells also exhibited increased pSTAT3 levels compared with the DMSO control (figure 5C).

Figure 5. PTPN22 is an intracellular checkpoint that limits CAR-NK cell function under cytokine-induced inflammatory TME-like conditions. (A) Western blot analysis of pSTAT3 in PTPN22^KO and Cas9 NK92 cells. (B) Flow cytometric analysis of pSTAT3 in PTPN22-inhibited or PTPN22^KO NK92 cells. (C) Western blot analysis of pSTAT3 in DMSO-treated or PTPN22-IN-1-treated NK92 cells after co-culture with HepG2 cells. (D and E) Anti-CD19/CAR-NK92 cells were cultured with K562-CD19-OE cells (D) or Nalm6 cells (E) for 4 or 24 hours before analysis by flow cytometry for pSTAT3 levels. (F) Assaying the effects of inflammatory cytokines on anti-CD19/CAR-NK92. PTPN22^KO and Cas9 NK cells were stimulated with different inflammation-related cytokines (10 ng/mL) for 24 hours in the absence of IL-2. (G and H) pSTAT3 expression in Cas9 control and PTPN22^KO anti-CD19/CAR-NK92 cells after incubation with inflammation-related cytokines (10 ng/mL) for 24 hours. Representative histograms are shown in (G), and quantified results are shown in (H). (I) Cytotoxicity assays using TNF-α-treated PTPN22^KO anti-CD19/CAR-NK92 against K562-CD19-OE cells. (J) PTPN22^KO and Cas9 anti-CD19/CAR-NK92 cells were incubated with different inflammation-related cytokines (10 ng/mL) for 24 hours. Cytotoxicity assays were then performed against K562-CD19-OE. (K) Assaying the effects of inflammatory cytokines in cell supernatants. Anti-CD19/CAR-NK92 cells were co-cultured with K562-CD19-OE cells for 24 hours. Supernatants were harvested and then incubated with PTPN22^KO or Cas9 anti-CD19/CAR-NK92 for 24 hours. Cytotoxicity assays were then performed against K562-CD19-OE cells. Significance was calculated by two-way ANOVA with Tukey post hoc test (D, E, H, I, J and K) and by one-way ANOVA with Tukey post hoc test (B). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; CAR, chimeric antigen receptor; DMSO, dimethyl sulfoxide; E:T, Effector-to-Target; IL, interleukin; MFI, Mean Fluorescence Intensity; NK, natural killer; pSTAT3, phosphorylation signal transducer and activator of transcription 3; TGF, transforming growth factor; TME, tumor microenvironment; TNF, tumor necrosis factor.

The STAT3 inhibitor Stattic⁴⁶ reduced both pSTAT3 and total STAT3 levels in Cas9 control and PTPN22^KO NK-92 cells (online supplemental figure S9C). Stattic suppressed NK-92 cell cytotoxicity and abolished the functional enhancement induced by PTPN22^KO and PTPN22-IN-1, but only modestly reduced the effect of LTV-1 (online supplemental figure S9D), potentially reflecting the broader inhibitory activity of LTV-1 against PTPN1 and PTPN2.⁴³ These results suggest that STAT3 signaling contributes to the enhanced NK-92 cell function induced by PTPN22 deletion.

Because PTPN22^KO enhanced STAT3 signaling in NK92 cells, we next asked whether similar signaling changes were also present in CAR-NK92 cells. A similar trend was observed in CAR-NK92 cells, as PTPN22 deficiency did not affect activation marker expression (online supplemental figure S9E). To investigate the molecular mechanisms by which PTPN22 regulates CAR-NK cell function, bulk RNA sequencing (RNA-seq) on PTPN22^KO anti-CD19/CAR-NK92 cells was performed. Among the genes upregulated in PTPN22^KO cells were key activation and inflammation-related markers, including CD69 and IFNG (online supplemental figure S9F). Pathway enrichment analysis further showed upregulation of pro-inflammatory pathways, including toll-like receptor signaling and cytokine–cytokine receptor interaction, together with activation of the JAK–STAT signaling pathway (online supplemental figure S9G,H). In addition, functional transcriptional scoring showed a reduced cellular stress-response signature in PTPN22^KO CAR-NK92 cells, which was further supported by reduced expression of stress-related genes compared with Cas9 controls (online supplemental figure S9I,J).

To further define whether PTPN22^KO sustains STAT3 signaling during CAR-NK/tumor-cell interaction, anti-CD19/CAR-NK92 cells were co-cultured with CD19⁺ cancer cells, and pSTAT3 expression was evaluated. PTPN22^KO enhanced the capacity of CAR-NK cells to suppress cancer cell proliferation (online supplemental figure S9K) and more effectively maintained pSTAT3 signaling compared with the Cas9 control group (figure 5D,E). Consistent with the observations in anti-CD19/CAR-NK92 cells, anti-MSLN/CAR-NK92 cells also showed increased pSTAT3 signaling following PTPN22 deletion (online supplemental figure S9L). These results further support that PTPN22 deletion was associated with changes in STAT3 signaling in CAR-NK92 cells.

PTPN22 is an intracellular checkpoint that limits CAR-NK cell function under cytokine-induced inflammatory TME-like conditions

Given the association between PTPN22 expression and TNF-driven inflammatory signatures in tumor-infiltrating NK cells, and consistent with previous reports in chronic inflammation TME demonstrating that elevated TNF levels suppress NK cell function,¹⁶ we next investigated whether PTPN22 knockout could preserve CAR-NK cell activity under inflammatory conditions.

We first examined STAT3 and STAT5 activation in PTPN22^KO anti-CD19/CAR-NK92 cells under inflammatory cytokine stimulation. PTPN22^KO preserved pSTAT3, but not pSTAT5, levels under most inflammatory conditions (figure 5F–H; online supplemental figure S10A,B). Under TNF-α inflammatory conditions, anti-CD19/CAR-NK92 cells showed markedly reduced cytolytic activity, whereas PTPN22^KO CAR-NK92 cells preserved their cytotoxic function against K562-CD19-OE cells (figure 5I). Even after prolonged TNF-α exposure, PTPN22^KO CAR-NK92 cells maintained cytotoxicity, in contrast to Cas9 controls (online supplemental figure S10C,D). Notably, pharmacological inhibition of STAT3 signaling markedly reduced the cytotoxic activity of both Cas9 and PTPN22^KO anti-CD19/CAR-NK92 cells against K562-CD19-OE cells (online supplemental figure S10E). We further evaluated PTPN22^KO CAR-NK92 cells under various inflammation-related cytokine conditions (IL-2, IL-1β, IL-6, and TGF-β). Notably, PTPN22^KO preserved the cytotoxic function of anti-CD19/CAR-NK92 under various inflammatory conditions, including immunosuppressive cytokines such as IL-6 and TGF-β (figure 5J).

As CAR-NK92 and tumor-cell co-culture induced high levels of inflammatory cytokines in vitro (figure 3F,G), we asked whether this cytokine milieu could, in turn, impair CAR-NK92 cell function. Anti-CD19/CAR-NK92 cells were co-cultured with cancer cells at a 1:1 effector-to-target ratio for 24 hours. The resulting supernatant, containing inflammation-induced cytokines, was harvested and used to precondition either PTPN22^KO or Cas9 anti-CD19/CAR-NK92 cells for 24 hours (figure 5K). PTPN22^KO CAR-NK92 cells retained significantly better cytolytic activity under these conditions, suggesting that PTPN22 deletion helps preserve CAR-NK92 cell function under inflammation-induced stress (figure 5K).

We further examined the phenotypes of PTPN22^KO anti-CD19 CAR-NK92 cells under inflammatory cytokine conditions. TNF-α stimulation significantly increased surface FAS expression in Cas9 control CAR-NK92 cells, whereas FAS expression remained lower in PTPN22^KO cells regardless of TNF-α stimulation (online supplemental figure S11A). A similar protective effect was observed following IL-1β stimulation. Consistent with the reduced FAS expression, PTPN22^KO CAR-NK92 cells exhibited lower Annexin V positivity than Cas9 controls following exposure to conditioned medium or TNF-α (online supplemental figure S11B,C).

To clarify the contribution of FAS signaling, we generated FAS^KO and PTPN22^KOFAS^KO CAR-NK92 cells. Both groups maintained low FAS expression following TNF-α stimulation (online supplemental figure S11D), whereas neither genetic modification affected FASL expression (online supplemental figure S11E). Under TNF-α-induced inflammatory conditions, FAS^KO CAR-NK92 cells exhibited cytotoxic activity comparable to that of PTPN22^KO CAR-NK92 cells. Notably, PTPN22^KOFAS^KO double-knockout cells showed the highest cytolytic activity among all tested groups (online supplemental figure S11F).

Finally, Stattic markedly reduced cytotoxicity across all groups and largely diminished the enhanced activity induced by PTPN22^KO and FAS^KO, whereas PTPN22^KOFAS^KO retained the highest cytolytic activity under STAT3 inhibition (online supplemental figure S11G). These findings suggest that PTPN22 deletion helps preserve CAR-NK cell function under inflammatory stress, in association with maintained STAT3 phosphorylation and reduced FAS-mediated apoptotic susceptibility.

PTPN22 knockout enhances antitumor activity of primary human NK cells

To assess the role of PTPN22 in primary human PBMCs, we co-cultured freshly isolated PBMCs from healthy donors with PTPN22-IN-1 inhibitor for 24 hours and tested their cytotoxicity against major histocompatibility complex (MHC)-I-low K562 cells (figure 6A). PTPN22-IN-1-treated PBMCs showed increased cytolysis against K562 cells (figure 6B).

Figure 6. PTPN22 knockout enhances antitumor activity of primary human NK cells. (A) PBMCs were isolated from healthy donor blood and treated with PTPN22-IN-1 for 24 hours. (B) Cytotoxicity of PTPN22-IN-1-treated PBMCs against MHC-I-low K562 cells. (C) Editing efficiency of CRISPR/Cas9-mediated PTPN22 knockout in cytokine-expanded primary NK cells assessed by western blot (n=5 donors). (D) Cytotoxicity of PTPN22^KO primary NK cells against K562 cells (E:T 1:1) (n=2 donors). (E) The pSTAT3 and pSTAT5 expression in PTPN22^KO and Cas9 primary NK cells. (F to H) PTPN22^KO primary NK cells were assayed for degranulation (F), TNF-α release (G) and IFN-γ release (H) against solid cancer cells (n=2–4 donors). (I) Design scheme for delivery of the CRISPR/Cas9 RNP complex to primary NK cells expanded by exposure to gamma-irradiated K562 feeder cells. The feeder cells expressed membrane-bound IL-18 and IL-21. (J) Cytolytic activity of PTPN22^KO expanded NK cells against different cell lines. Significance was calculated by a paired t-test (C, D and F); by two-way ANOVA with a Tukey post hoc test (B and J); and by one-way ANOVA with a Tukey post hoc test (G and H). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; DMSO, dimethyl sulfoxide; E:T, Effector-to-Target; IFN, interferon; IL, interleukin; MFI, Mean Fluorescence Intensity; MHC, major histocompatibility complex; NK, natural killer; PBMC, peripheral blood mononuclear cell; pSTAT, phosphorylation signal transducer and activator of transcription 3; RNP, ribonucleoprotein; TNF, tumor necrosis factor.

We then isolated fresh NK cells and expanded them with high-dose IL-2 and IL-15. To verify the role of PTPN22 in primary NK cells, we knocked out PTPN22 in cytokine-expanded primary NK cells (online supplemental figure S12A), achieving efficient knockout across all tested donor cells (figure 6C). Consistent with the NK92 cell results, PTPN22^KO did not affect NK-cell purity, proliferation, or phenotype (online supplemental figure S12B–D), but enhanced cytotoxicity against K562 cells (figure 6D). Consistently, PTPN22^KO primary NK cells also exhibited elevated pSTAT3 expression levels (figure 6E). We further tested PTPN22^KO NK cells against the hard-to-treat solid cancer cell lines Hep3B and MDA-MB-231. PTPN22^KO NK cells showed increased degranulation against these solid tumor cell lines (figure 6F) and elevated pro-inflammatory cytokine production (figure 6G,H). Feeder cell-expanded NK cells have been used clinically due to their strong expansion capacity.⁴⁷ In this study, we expanded NK cells using mbIL-18/IL-21 K562 feeder cells, which improved NK-cell expansion and effector function compared with conventional K562 feeder cells³⁸ (figure 6I). PTPN22^KO further enhanced the function of expanded NK cells against different cancer cell types (figure 6J), indicating that PTPN22 deletion improves NK-cell function regardless of the expansion platform.

PTPN22 knockout enhances the functional persistence of primary CAR-NK cells

We next evaluated the impact of PTPN22 inhibition in primary CAR-NK cells. PTPN22-IN-1-treated anti-CD19/CAR primary NK cells, generated from feeder cell-expanded NK cells, also enhanced degranulation against both K562-CD19-OE and Nalm6 cells (online supplemental figure S13A). Additionally, increased cytotoxicity was observed across all tested donors (online supplemental figure S13B).

Dual engineering through CAR transduction and CRISPR-based gene knockout represents an essential strategy for generating gene-edited NK cells capable of overcoming the suppressive conditions of the TME. We successfully achieved double editing by transducing anti-CD19/CAR and performing PTPN22 knockout simultaneously on day 5 after NK cell expansion using K562 feeder cells expressing membrane-bound IL-18/IL-21 (figure 7A). Double editing did not affect CAR transduction efficiency in any donor (online supplemental figure S13C). Despite the donor variation, PTPN22^KO anti-CD19/CAR primary NK cells consistently showed enhanced cytolysis and cytokine release against K562-CD19-OE and Nalm6 cancer cells (online supplemental figure S13D–G).

Figure 7. PTPN22 knockout enhances the functional persistence of primary CAR-NK cells. (A) Schematic of the generation of PTPN22^KO anti-CD19/CAR-NK cells. Primary NK cells were expanded using mbIL-18/IL-21-expressing K562 feeder cells. After 5 days, the NK cells were simultaneously transduced with anti-CD19/CAR and subjected to PTPN22 knockout. The modified cells were subsequently used for functional assays. (B to D) Cytotoxicity (B), TNF-α release (C), and IFN-γ release (D) of PTPN22^KO primary anti-CD19/CAR-NK cells against K562, K562-CD19-OE, and Nalm6 cells, as indicated. (E) Persistence of primary anti-CD19/CAR-NK cells after 48 hours of co-culture with K562, K562-CD19-OE, or Nalm6 cells. (F) Cytotoxicity assays using TNF-α-treated PTPN22^KO primary anti-CD19/CAR-NK cells against K562-CD19-OE and Nalm6 cells. (G) pSTAT3 expression in primary anti-CD19/CAR-NK cells. (H) FAS expression in primary anti-CD19/CAR-NK cells. Significance was calculated by one-way ANOVA with Tukey post hoc test (C, D and E), two-way ANOVA with Tukey post hoc test (B and F), and paired t-test (G and H). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; CAR, chimeric antigen receptor; E:T, Effector-to-Target; IFN, interferon; IL, interleukin; NK, natural killer; PBMC, peripheral blood mononuclear cell; pSTAT, phosphorylation STAT; RNP, ribonucleoprotein; TNF, tumor necrosis factor.

We next evaluated whether PTPN22^KO enhances the function of primary NK and anti-CD19 CAR-NK cells under different levels of CD19 antigen exposure (online supplemental figure S14A). PTPN22^KO improved primary NK-cell cytotoxicity across the tested target cell lines, whereas anti-CD19 CAR expression enhanced NK-cell cytotoxicity only against CD19-expressing targets, including K562-CD19 OE and Nalm6 cells (figure 7B). Notably, combining PTPN22^KO with anti-CD19 CAR expression further increased cytotoxicity compared with anti-CD19 CAR-NK cells alone, indicating that PTPN22^KO cooperates with CAR-mediated antigen recognition to enhance primary CAR-NK cell function. Consistent with the enhanced cytotoxicity, PTPN22^KO anti-CD19 CAR-NK cells also produced higher levels of IFN-γ and TNF-α after co-culture with CD19-expressing target cells (figure 7C,D).

Furthermore, during prolonged 48 hours co-culture, PTPN22^KO anti-CD19 CAR-NK cells showed improved functional persistence, as reflected by reduced residual tumor-cell burden and increased recovery of NK cells after long-term tumor exposure (online supplemental figure S14B,C and figure 7E).

Consistently, PTPN22^KO anti-CD19/CAR primary NK cells maintained their cytolytic function under TNF-α-induced inflammatory conditions, whereas Cas9 control cells exhibited reduced lytic capacity (figure 7F). This maintained function was associated with preserved pSTAT3 expression and reduced FAS expression, but not with pSTAT5 expression (figure 7G,H and online supplemental figure S14D). These findings further support the role of PTPN22 in regulating NK cell function by extending its effects to primary NK cells, including CAR-engineered cells.

Discussion

The role of PTP family members in NK cells has remained largely underexplored, despite extensive studies highlighting their regulatory functions in T cell-based therapies.^{22 23} Here, we identified a novel role for PTPN22 as an intracellular checkpoint that modulates NK cell effector functions. Consistent with this concept, a recent study showed that the autoimmunity-associated PTPN22 allele PEP-R619W enhances NK-cell maturation, IFN-γ production, and antiviral function during acute COVID-19 infection, further supporting a context-dependent role for PTPN22 in regulating NK-cell immunity.⁴⁸ Using non-viral CRISPR/Cas9-mediated knockout approaches, we demonstrated that suppression of PTPN22 significantly enhances NK cell activity. Notably, pharmacological inhibition partially recapitulated the functional enhancement observed with genetic PTPN22 deletion. However, given the potential off-target effects of PTPN22 inhibitors, particularly LTV-1, further validation with more selective PTPN22 inhibitors will be needed to fully define the therapeutic potential of pharmacological PTPN22 targeting in NK-cell therapy.

In clinical trials, anti-CD19/CAR-NK cells are emerging as a promising therapeutic modality for hematologic malignancies.^5–8 We showed that disrupting PTPN22 signaling augments the effector functions of anti-CD19/CAR-NK92 and primary anti-CD19/CAR-NK cells against CD19⁺ leukemic targets in vitro, while the principal in vivo efficacy study was evaluated only using anti-CD19/CAR-NK92 cells. Moreover, PTPN22 deletion enhances the cytotoxic activity of anti-MSLN/CAR-NK cells against MSLN⁺ ovarian tumor cells, supporting its applicability beyond hematologic malignancies. Importantly, PTPN22^KO also improved the functional persistence of primary anti-CD19 CAR-NK cells during prolonged tumor exposure, as shown by reduced residual tumor-cell burden and increased NK-cell recovery after long-term co-culture. This finding is particularly relevant because limited persistence remains a major barrier for NK and CAR-NK cell therapy.^{2 3 9} These findings highlight a critical role for PTPN22 in regulating CAR-NK cell function, particularly under immunosuppressive conditions.

Although inflammation typically facilitates immune activation,^{11 12} increasing evidence indicates that chronic inflammatory cues can exert immunosuppressive effects,^{13 14 17} particularly within the context of NK-based and CAR-based therapies.^{13 15 16} Pro-inflammatory cytokines such as IFN-γ and TNF-α, once regarded as supportive factors in the TME, have now been implicated in promoting exhaustion and apoptosis in CAR-T cells.^{13 49} We identified a PTPN22^high NK cell cluster from publicly available scRNA-seq datasets, which exhibited features of exhaustion and showed a notable correlation between PTPN22 expression and TNF levels in the TME. Notably, this cluster was observed across both hematologic malignancies (AML) and multiple solid tumors. Furthermore, patients with a NK^highPTPN22^high transcriptional signature displayed upregulation of inflammatory response pathways and downregulation of key effector pathways, including glycolysis and oxidative phosphorylation, and this signature was associated with poorer patient survival.

To experimentally evaluate the effects of inflammation-related cytokines on CAR-NK cells, we tested TNF-α, IL-1β, IL-6, and TGF-β. Disruption of PTPN22 preserved CAR-NK-cell function under cytokine-based TME-like conditions, accompanied by preserved STAT3 phosphorylation. Consistently, bulk RNA-seq analysis of PTPN22^KO anti-CD19/CAR-NK92 cells showed enrichment of JAK–STAT-related gene signatures, which may be associated with their improved function under cytokine-based TME-like conditions. In contrast, stress-related gene signatures were relatively enriched in Cas9 control cells, suggesting that PTPN22 deficiency may be associated with reduced cellular stress responses.

FAS-mediated apoptotic signaling may also contribute to the enhanced function of PTPN22-deficient CAR-NK cells. Recent studies have identified FAS/FASL-mediated apoptosis as a mechanism limiting CAR-T and CAR-NK cell persistence in the TME.⁵⁰ In our study, FAS^KO partially recapitulated the functional benefit of PTPN22^KO, whereas PTPN22^KOFAS^KO further enhanced cytolytic activity. These findings suggest that reduced FAS-mediated apoptotic susceptibility contributes to the cytoprotective effect of PTPN22 deletion, while additional pathways are also involved. One such pathway may involve STAT3, a critical regulator of NK-cell metabolism and function.⁵¹ We found that either genetic disruption or pharmacological inhibition of PTPN22 enhanced STAT3 signaling in NK92 cells. PTPN22^KO anti-CD19/CAR-NK and anti-MSLN/CAR-NK cells retained STAT3 signaling more robustly than Cas9 controls, suggesting an association between PTPN22 deletion and STAT3 signaling in CAR-NK cells. Although the role of STAT3 in NK cells remains debated,^52–54 our findings suggest that STAT3-associated signaling may contribute to sustaining the effector function of both native and CAR-modified NK cells under cytokine-induced inflammatory and immunosuppressive conditions.

A previous study reported increased NK cell numbers in mice following pharmacological inhibition or global knockout of PTPN22.³⁰ Although we determined that PTPN22^KO anti-CD19/CAR-NK92 cells more effectively controlled tumor growth in a mouse leukemia model when compared with Cas9 control, our study did not specifically explore the role of PTPN22 in murine NK cells, leaving open questions regarding how PTPN22 may regulate NK cell maturation or proliferation in vivo. Future studies using NK-cell-specific PTPN22 conditional knockout mouse models will be valuable to further validate the intrinsic role of PTPN22 in regulating NK-cell development, persistence, and antitumor function in vivo. Another limitation is that in vivo efficacy was evaluated using PTPN22^KO anti-CD19/CAR-NK92 cells, warranting further validation in primary CAR-NK cells.

We focused on clinically relevant human NK cell platforms to evaluate the therapeutic potential of PTPN22 targeting. Our findings reveal that PTPN22 knockout enhances NK cell effector function across different NK platforms, including NK92, primary NK cells, and multiple types of CAR-NK cells. This approach offers a compelling strategy to enhance the durability and efficacy of CAR therapy within the TME,^{55 56} particularly under chronic inflammatory conditions.^{16 49} However, because our current in vitro systems do not fully recapitulate the complexity of a TME, we were unable to evaluate PTPN22^KO CAR-NK cells under authentic inflammatory or immunosuppressive TME conditions. Future studies should investigate the antitumor activity of PTPN22^KO CAR-NK cells against patient-derived tumor cells using more physiologically relevant immunosuppressive TME models.

In conclusion, PTPN22 represents a promising therapeutic target for enhancing NK cell-based immunotherapy, particularly within immunosuppressive TMEs. Our findings support genetic deletion of PTPN22 as a potential strategy for enhancing NK cell function, while pharmacological inhibition provides complementary evidence that will require further validation with more selective PTPN22 inhibitors before clinical translation.

The funders had no role in the design of the study; collection, analysis, or interpretation of data; writing of the manuscript; or the decision to submit the manuscript for publication.

Data availability statement

Data are available in a public, open access repository. Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.