Non-canonical function of SCIMP restricts JAK1 activity to counteract immune evasion in bladder cancer

Abstract

Background

Despite the fact that immunotherapy is reshaping the therapeutic landscape of bladder cancer, immune evasion remains a major obstacle to durable clinical benefit. Therefore, elucidating the mechanisms underlying immune resistance is essential for developing novel therapeutic strategies.

Methods

In this study, we performed a genome-wide CRISPR-Cas9 screen in bladder cancer cells co-cultured with CD8⁺ T cells and identified the SLP adaptor and CSK-interacting membrane protein (SCIMP) as a key gene that enhances CD8⁺ T cell antitumor immunity.

Results

SCIMP serves as a favorable prognostic biomarker in patients with bladder cancer treated with immunotherapy. Mechanistically, SCIMP restricts JAK1 phosphorylation in an adaptor-independent manner, suppressing the IFN-γ/JAK1/STAT5A/PD-L1 signaling axis. Meanwhile, YTHDF2 destabilizes the m6A-tagged SCIMP transcripts to reduce their expression. Remarkably, SCIMP synergizes with PD-1 blockade in suppressing tumor growth.

Conclusions

Collectively, these findings reveal a previously unexplored non-canonical role of SCIMP in preventing immune evasion and provide a potential therapeutic target for bladder cancer.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• CD8⁺ cytotoxic T cells play a pivotal role in antitumor immunity and are generally correlated with improved immunotherapy efficacy. However, tumor cells can inhibit CD8⁺ T-cell recruitment or accelerate their exhaustion for immune evasion. Therefore, elucidating the mechanisms underlying immune resistance is essential for developing novel therapeutic strategies that can enhance the clinical benefits of immunotherapy.

WHAT THIS STUDY ADDS

• In this study, we identified SCIMP as a key regulator of immune evasion through a genome-wide CRISPR screen. Moreover, we found that elevated SLP adaptor and CSK-interacting membrane protein (SCIMP) expression is associated with improved prognosis in patients with bladder cancer. Mechanistically, SCIMP bound to JAK1 and inhibited its phosphorylation in a non-canonical manner, leading to suppression of the IFN-γ/JAK1/STAT5A/PD-L1 signaling axis and thereby increasing tumor sensitivity to CD8⁺ T cell-mediated killing.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This study uncovers a non-canonical, adaptor-independent role of SCIMP in bladder cancer cells. In addition, our findings suggest a promising therapeutic strategy to counteract tumor immune evasion and improve treatment outcomes in bladder cancer.

Introduction

Bladder cancer is one of the most prevalent urological malignancies, accounting for an estimated 656,000 new bladder cancer cases and 234,000 deaths worldwide in 2025.^{1 2} Approximately 25% of patients are initially diagnosed with muscle-invasive bladder cancer, for which standard treatment with radical cystectomy and chemotherapy offers limited survival benefit.^{3 4} In recent years, immune checkpoint inhibitors (ICIs) have been approved by the Food and Drug Administration for advanced or metastatic bladder cancer, yet only a minority of patients achieve durable responses.^{5 6} Therefore, it is imperative to reveal the landscape of immune evasion in bladder cancer and identify effective therapeutic targets that may enhance the efficacy of ICI therapy.

The tumor microenvironment (TME) plays a critical role in immune tolerance and responses to immunotherapy.⁷ Within the TME, CD8⁺ T cells are key immune effectors that elicit antitumor immunity through antigen-specific cytotoxicity and the secretion of cytokines,^{8 9} while tumor cells can evade immune surveillance through multiple mechanisms, including upregulation of the immune checkpoint molecule programmed death-ligand 1 (PD-L1). PD-L1 expression is induced in tumor cells in response to inflammatory signals, among which interferon-γ (IFN-γ) is considered the primary stimulatory factor.¹⁰ IFN-γ triggers the Janus kinase/signal transducer and activator of transcription (JAK/STAT) signaling, thereby regulating gene transcription.^{11 12} Despite these findings, the regulatory mechanisms of PD-L1 and JAK/STAT pathway in bladder cancer remain incompletely characterized.

SLP adaptor and CSK-interacting membrane protein (SCIMP), a member of the palmitoylated transmembrane adaptor protein (pTRAP) family, functions as a transmembrane adaptor involved in cellular signal transduction.¹³ Previous studies have suggested that SCIMP was implicated in MHC class II and Dectin-1 signaling in B cells and dendritic cells, respectively.^{14 15} Recently, it has been identified as a Toll-like receptor (TLR) adaptor in macrophages, recruiting Erk1/2 to TLR4 and driving the production of pro-inflammatory cytokines.^{16 17} However, the roles and mechanisms of SCIMP in cancer cells have not been reported.

To explore the tumor immune evasion mechanisms and find novel therapeutic targets, we performed a genome-wide CRISPR-Cas9 screen under co-culture with CD8⁺ T cells and found that sgRNAs targeting SCIMP were significantly enriched in bladder cancer cells. In this study, we uncovered an unexpected function of SCIMP in restricting JAK1 kinase activity, independent of its canonical role as an adaptor protein. Specifically, SCIMP, which is negatively regulated by the N6-methyladenosine (m6A) reader YTH domain-containing family protein 2 (YTHDF2), interacts with JAK1 and inhibits its phosphorylation, thereby dampening IFN-γ/JAK1/STAT5A/PD-L1 signaling to enhance antitumor immune responses. Furthermore, SCIMP overexpression or pharmacological inhibition of JAK1 with itacitinib enhanced intratumoral CD8⁺ T cell effector function, leading to improved efficacy of ICIs. Overall, our findings elucidate a mechanism of SCIMP distinct from its adaptor function in antitumor immunity and suggest a potential therapeutic strategy for improving bladder cancer immunotherapy.

Materials and methods

Patient specimens

96 paired bladder cancer tissues and adjacent normal bladder tissues were collected from radical cystectomy specimens at the Department of Urology, Union Hospital, Wuhan, China, between 2014 and 2019. The pathological types of all tissues were independently confirmed by three experienced pathologists. The tissues were stored at −80°C after quick freezing in liquid nitrogen. Follow-up data of all patients were documented from diagnosis until death or the censoring date.

Cell culture

The human bladder cancer cell lines T24, UMUC3, and human embryonic kidney 293T cells were purchased from ATCC. The murine bladder cancer cell line MB49 was obtained from Sigma-Aldrich. T24 cells were cultured in McCoy’s 5A medium (Gibco). UMUC3, HEK293T, and MB49 cells were cultured in DMEM medium (Gibco). Purified CD8⁺ T cells were cultured in RPMI-1640 medium (Gibco). Cells were cultured at 37°C, 5% CO₂. All cell lines were authenticated using short tandem repeat profiling and were free of Mycoplasma contamination.

Lentivirus production

293T cells were seeded in 100 mm dishes and transfected at 70%–90% confluency using Lipofectamine 2000 (Invitrogen). For lentivirus packaging, 8 µg transfer plasmid, 6 µg psPAX2 (Addgene, 12260), and 4 µg pMD2.G (Addgene, 12259) were mixed in 750 µL of OPTI-MEM medium (Gibco). The transfection mix was replaced with complete DMEM 8 hours later. The lentiviral supernatant was harvested at 48 and 72 hours after transfection, filtered through 0.45 µm filters, and concentrated by ultracentrifugation at 100,000×g for 2 hours at 4°C. The concentrated viruses were then resuspended in 100 µL PBS, aliquoted, and subsequently used for cell infection or stored at −80°C until use.

Genome-wide CRISPR-Cas9 screen

The human GeCKO V.2 library was used to screen for genes preventing immune escape in bladder cancer.¹⁸ The GeCKO V.2 library was delivered into T24 cells at a multiplicity of infection of 0.3. After puromycin selection for 1 week, the cells were expanded and co-cultured with CD8⁺ T cells at an effector-to-target ratio of 3:1 for 72 hours. Following co-culture, T cells were removed, and a minimum of 3×10⁷ viable tumor cells were collected for genomic DNA extraction to ensure a coverage of 400× or greater. The sgRNA sequences were amplified using 2× PCR Master Mix (NEB, M0541) and submitted to Novogene (Beijing, China) for sequencing. The analysis of sgRNA read count and hit calling was performed by the MAGeCK V.0.5.7 algorithm.¹⁹

Plasmids construction

The sgRNAs sequences targeting human SCIMP and mouse Scimp (online supplemental table S1) were synthesized by TSINGKE and cloned into the lentiCRISPR V.2 vector. The human SCIMP and PD-L1, as well as mouse Scimp and PD-L1 cDNAs, were synthesized by TSINGKE and cloned into the pcDNA3.1−3× Flag vector. Truncations of SCIMP and JAK1 were cloned into the pcDNA3.1-3× Flag vector and the pLVX-HA vector, respectively. The PD-L1 promoter reporter was inserted into the pGL3-Basic vector. Wild-type and mutant SCIMP variants synthesized by TSINGKE were cloned into the pmirGLO vectors. Short hairpin RNAs targeting YTHDF2 (online supplemental table S1) were synthesized by TSINGKE and cloned into the pLKO.1 vector.

RNA preparation and qRT-PCR

In brief, total RNA was isolated from bladder tissues and cells using TRIzol (Invitrogen). cDNA was synthesized using HiScript IV SuperMix (Vazyme). Quantitative RT-PCR was performed with SYBR Green Master Mix (Vazyme) on StepOnePlus Real-Time PCR System (Applied Biosystems). The primers are listed in online supplemental table S1.

Western blotting

Cells were lysed on ice in RIPA buffer (Invitrogen) supplemented with protease inhibitor cocktail (MCE), phosphatase inhibitor cocktail (MCE), and PMSF (MCE) for 30 min. Following quantification with a BCA assay kit (HYcezmbio), protein samples were denatured in 5× loading buffer at 95°C for 10 min. Proteins were electrophoresed on 8%–15% SDS-PAGE gels and transferred onto PVDF membranes (Millipore). The membranes were blocked in 5% skim milk for 1 hour and incubated with primary antibodies overnight at 4°C. The following day, membranes were washed three times with TBST and incubated with HRP-conjugated goat anti-mouse/rabbit secondary antibodies (Proteintech, SA00001-1/SA00001-2) for 1 hour at room temperature. After the final TBST washes, immunoreactive bands were visualized using the ChemiDoc-It Imaging System (UVP). The following primary antibodies were used: mouse anti-GAPDH (Proteintech, 60004-1-Ig), mouse anti-SCIMP (Invitrogen, MA5-28633), mouse anti-JAK1 (Proteintech, 66466-1-Ig), rabbit anti-Flag (Abcam, ab205606), rabbit anti-HA (Abcam, ab9110), rabbit anti-p-JAK1 (Invitrogen, 44–422G), rabbit anti-STAT2 (CST, 72604), rabbit anti-p-STAT2 (CST, 23078), rabbit anti-STAT4 (CST, 2653), rabbit anti-p-STAT4 (Abcam, ab28815), mouse anti-STAT5A (Invitrogen, 13–3600), rabbit anti-p-STAT5A (Abcam, ab128896), mouse anti-PD-L1 (Proteintech, 66248-1-Ig), rabbit anti-YTHDF2 (Proteintech, 24744-1-AP), rabbit anti-IGF2BP3 (Proteintech, 14642-1-AP), rabbit anti-YTHDF3 (Proteintech, 25537-1-AP), and rabbit anti-YTHDC1 (Proteintech, 14392-1-AP).

Cell viability assays

Cell viability was determined by Cell Counting Kit-8 (Dojindo). Briefly, cells were seeded into 96-well plates at a density of 3000 cells per well in 100 µL of medium. At the indicated time points (0, 24, 48, 72, and 96 hours), 10 µL CCK-8 reagent was added to each well, followed by incubation for 2 hours at 37°C in the dark. The absorbance was measured at 450 nm using a microplate reader (Synergy4; BioTek, Winooski, VT, USA).

T-cell isolation, activation, and culture

Human peripheral blood was obtained from healthy volunteers and diluted with an equal volume of PBS. The PBS-diluted blood was gently layered onto an equal volume of Ficoll (Solarbio, P8900) and subjected to centrifugation at 850×g for 20 min at 20°C with minimal acceleration and breaking. The peripheral blood mononuclear cells (PBMCs) layer was gently collected and subjected to lysis using ACK lysis buffer (Solarbio, R1013) on ice for 5 min, after which the reaction was terminated with PBS.

For CD8⁺ T cell sorting, approximately 5×10⁶ PBMCs in 100 µL of PBS were incubated with 10 µL of human CD8 microbeads (Miltenyi Biotec) for 30 min at 4°C. CD8⁺ T cells were subsequently purified using an LS column and MidiMACS Separator (Miltenyi Biotec) according to the manufacturer’s instructions. The isolated CD8⁺ T cells were cultured in 24-well plates at a density of 2×10⁶ cells/mL in RPMI-1640 medium supplemented with 5 µg/mL anti-human CD3 (BioLegend), 5 µg/mL anti-human CD28 (BioLegend), and 20 ng/mL recombinant human IL-2 (PeproTech).

Spleens from C57BL/6 mice were isolated and gently ground in PBS using a 10 mL syringe, then filtered through a 70 µm cell strainer to obtain a single-cell suspension. Following centrifugation (500×g, 5 min) and red blood cell lysis, CD8⁺ T cells were purified using Ficoll (Solarbio, P8860) and mouse CD8 microbeads (Miltenyi Biotec). The purified T cells were then activated with anti-mouse CD3/CD28 antibodies (BioLegend) and mouse IL-2 (PeproTech).

Co-culture assay

T24 and UMUC3 cells were co-cultured with activated human CD8⁺ T cells from healthy donors at an effector-to-target ratio of 3:1 for 8–72 hours. Similarly, MB49 cells were subjected to co-culture with activated CD8⁺ T cells from C57BL/6 mice.

Animal experiments

All animal experiments were approved by the Animal Care Committee of Tongji Medical College. Six- to eight-week-old C57BL/6 and BALB/c nude mice were used in this study. MB49 cells (5×10⁵ cells per mouse) stably transfected with either pcDNA3.1-vector or pcDNA3.1-SCIMP were subcutaneously injected into the right flank of the mice. Once palpable tumors developed, tumor size was measured every 3–4 days using a caliper, and the volume was calculated according to the formula: (length×width²)/2. When tumors reached approximately 100 mm³, mice received intraperitoneal injections of 100 µg control IgG (Bioxcell, BE0089) or anti-PD-1 antibody (Bioxcell, BE0146) two times a week for 3 weeks. For itacitinib treatment, mice were administered DMSO or itacitinib (60 mg/kg, two times a day) for 3 weeks.

Immunofluorescence staining

Cells seeded into confocal dishes were fixed with 4% paraformaldehyde (Beyotime) for 20 min and permeabilized with 0.1% Triton X-100 on ice for 10 min. After permeabilization, the cells were blocked with 1% BSA for 1 hour at room temperature and then incubated overnight at 4°C with anti-SCIMP (Invitrogen, MA5-28633) and anti-JAK1 (ABclonal, A18323) antibodies. The next day, cells were incubated with the appropriate secondary antibodies, CoraLite488 (SA00013-1, Proteintech) and CoraLite594 (SA00013-4, Proteintech), for 1 hour at room temperature before DAPI staining. Images were acquired using a Nikon A1Si Laser Scanning Confocal Microscope (Nikon Instruments, Japan).

Co-immunoprecipitation

Cells were lysed on ice in NP-40 lysis buffer (Beyotime) supplemented with protease inhibitor cocktail and PMSF for 1 hour. The lysates were centrifuged at 12,000 rpm for 5 min at 4°C, and 5% of the supernatant was collected as the input control. The remaining supernatant was divided equally and incubated overnight at 4°C with either IgG or the target antibody. Protein A/G beads (MCE) were then added for an additional 2 hours of incubation at 4°C with rotation. The protein-bead complexes were washed three times with PBS containing 0.5% Tween-20 and eluted by boiling in SDS-PAGE loading buffer at 95°C for 10 min.

Silver staining and mass spectrometry analysis

Silver staining of the immunoprecipitated samples was performed according to the protocol of the Silver Staining Kit (Solarbio, G7210). Mass spectrometry analysis was conducted by Novogene (Beijing, China) and processed using Proteome Discoverer software V.1.4 (Thermo Fisher Scientific).

Phosphorylation antibody microarray

JAK/STAT II Phospho Antibody Array (PJS202, Full Moon Biosystems), which features 205 highly specific antibodies related to the Jak/Stat pathway, was used in this study. The array images acquired using the GenePix 4000B scanner were processed with GenePix Pro 6.0.

RNA-seq

After 24 hours of co-culture with CD8⁺ T cells, bladder cancer cells transfected with a vector or an SCIMP were collected. Total RNA was extracted using TRIzol (Invitrogen) for RNA-seq analysis.

Chromatin immunoprecipitation

Cells were crosslinked with 1% formaldehyde at room temperature for 10 min, and the reaction was quenched by adding 125 mM glycine. The crosslinked chromatin was then sonicated into 200–1000 bp fragments, which were incubated overnight at 4°C with either anti-STAT5A or control IgG antibody and then added to Protein A/G magnetic beads for 2 hours. The enrichment of STAT5A bound to fragments of the PD-L1 promoter region was analyzed by PCR. Chromatin immunoprecipitation (ChIP) primers are listed in online supplemental table S1.

Luciferase reporter assay

The reporter plasmid and Renilla control plasmid were co-transfected into bladder cancer cells. Luciferase activities were measured using the Dual Luciferase Reporter Gene Assay Kit (Beyotime), and the Firefly luciferase activity was normalized to Renilla luciferase activity.

RNA pull-down and RNA immunoprecipitation

Biotin-labeled RNAs were synthesized in vitro using the MEGAscript T7 Transcription Kit (Invitrogen, AM1333) and the Pierce RNA 3′ End Desthiobiotinylation Kit (Thermo Fisher Scientific, 20163). RNA pull-down assays were performed with the Pierce Magnetic RNA-Protein Pull-Down Kit (Thermo Fisher Scientific, 20164) according to the manufacturer’s instructions. For RNA immunoprecipitation (RIP) assay, we employed the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, 17-700).

Flow cytometry analysis of tumor-infiltrating lymphocytes

Tumors harvested from mice were minced and digested in RPMI-1640 containing collagenase IV, hyaluronidase, and DNase I (Sigma-Aldrich) at 37°C for 1 hour. For intracellular cytokine staining, cells were cultured in complete RPMI-1640 medium supplemented with Leukocyte Activation Cocktail (BD Pharmingen, 550583) for 6 hours. Live/dead staining (BD Pharmingen, 564997) was performed, followed by Fc blocking (BD Pharmingen, 553141) and surface staining with APC-Cy7 anti-mouse CD45 (BD Pharmingen, 557659), BV421 anti-mouse CD3 (BD Pharmingen, 562600), and BV711 anti-mouse CD8 (BD Pharmingen, 563046). Cells were then fixed in fixation/permeabilization solution at 4°C for 30 min, washed with Perm/Wash buffer (BD Pharmingen, 554714), and stained with APC anti-mouse TNF-α (BD Pharmingen, 554420) and PE anti-mouse IFN-γ (BD Pharmingen, 554412) for 30 min. The stained cells were resuspended in FACS buffer and analyzed on a BD FACSymphony A1 flow cytometer.

Statistical analysis

Data are presented as mean±SD. Survival differences were analyzed using Kaplan-Meier curves and the log-rank test, while Student’s t-test and analysis of variance were used for the determination of statistically significant differences between groups. Statistical analyses were performed using GraphPad Prism V.9, and p value <0.05 was considered statistically significant.

Results

SCIMP suppresses bladder cancer growth in a CD8⁺ T cell-dependent manner

To identify genes restraining bladder cancer immune evasion, we performed a genome-wide CRISPR/Cas9 screen on T24 cells in a co-culture system with human CD8⁺ T cells (figure 1A). High-throughput sequencing analysis of sgRNA abundance in tumor cells revealed that sgRNAs targeting SCIMP were among the most highly enriched in the gene list (figure 1B). Consistently, the survival analysis across both pan-cancer and bladder cancer cohorts from the KMplot database²⁰ showed that SCIMP was associated with improved prognosis in patients receiving immunotherapy (figure 1C). Moreover, SCIMP was significantly downregulated in bladder cancer tissues compared with paired normal tissues (figure 1D), and Kaplan-Meier survival analysis showed that patients with bladder cancer with high SCIMP expression had a better prognosis (figure 1E). In addition, according to the KMplot database, elevated SCIMP expression was associated with improved survival in stage III–IV patients with bladder cancer with high CD8⁺ T cell infiltration, whereas no significant correlation was observed in patients with low CD8⁺ T cell infiltration (online supplemental figure S1A). To interrogate the effect of SCIMP on cancer cells, we generated SCIMP knockout (KO) cells and KO cells re-expressing SCIMP in T24, MB49, and UMUC3 cell lines (figure 1F; online supplemental figure S1B). We found that neither SCIMP KO nor its re-expression significantly altered the viability of bladder cancer cells (online supplemental figure S1C). In contrast, SCIMP KO significantly enhanced the viability of bladder cancer cells on co-culture with CD8⁺ T cells, while exogenous SCIMP expression abrogated the increased viability induced by SCIMP deficiency (figure 1G; online supplemental figure S1D). To determine the effect of SCIMP in vivo, SCIMP-overexpressing (SCIMP-OE) and control MB49 cells were subcutaneously injected into nude mice and C57BL/6 mice. Notably, SCIMP significantly suppressed tumor growth in immunocompetent C57BL/6 mice, whereas this effect was absent in nude mice and CD8⁺ T cell-depleted C57BL/6 mice (figure 1H; online supplemental figure S1E-H). Collectively, these in vitro and in vivo results suggest that the antitumor effect of SCIMP is dependent on CD8⁺ T cell-mediated immunity.

Figure 1. Genome-wide CRISPR/Cas9 screens identify SCIMP as a modulator of tumor immunity. (A) Schematic diagram of CRISPR-Cas9 screening strategy. (B) Rank plot showed SCIMP was identified as the most enriched gene in the positive screen. (C) Kaplan-Meier curves showed overall survival of immunotherapy-treated pan-cancer and patients with bladder cancer from KMplot database stratified by SCIMP expression. (D) Relative SCIMP mRNA levels measured by qRT-PCR in paired human bladder cancer and adjacent normal tissues (n=96). (E) Kaplan-Meier curves showing overall survival of patients with bladder cancer (n=96) stratified by SCIMP expression. (F) Western blot analysis of SCIMP protein levels in T24 and MB49 cells with SCIMP knockout or SCIMP re-expression. GAPDH was used as a loading control. (G) CCK-8 assay revealed the cell viability of T24 and MB49 cells (SCIMP knockout or SCIMP re-expression) co-cultured with CD8⁺ T cells for 24 hours. (H) Schematic diagram and tumor growth in BALB/c nude (n=6) and C57BL/6 mice (n=6) subcutaneously injected with MB49 cells stably expressing empty a vector or an SCIMP. KO, knockout; MOI, multiplicity of infection; SCIMP, SLP adaptor and CSK-interacting membrane protein.

SCIMP interacts with JAK1 at the plasma membrane of bladder cancer cells

Although SCIMP has been characterized as a transmembrane adaptor protein in immune cells,¹⁶ its subcellular localization in tumor cells remains unclear. Therefore, we conducted immunofluorescence assays and found that SCIMP was localized to the plasma membrane in bladder cancer cells (figure 2A; online supplemental figure S2A). Subsequently, co-immunoprecipitation (Co-IP) and mass spectrometry were used to identify potential binding partners of SCIMP (figure 2B; online supplemental table S2). Unexpectedly, we did not find an association between SCIMP and receptor complexes, implying a non-canonical function of SCIMP in bladder cancer separate from its role as an adaptor. Through a systematic analysis pipeline, JAK1 was identified as physically interacting with SCIMP (figure 2C). In addition, Co-IP assays confirmed the interaction between SCIMP and JAK1 (figure 2D,E; online supplemental figure S2B, C), which was further supported by their colocalization at the plasma membrane of bladder cancer cells (figure 2F; online supplemental figure S2D). To delineate the structural determinants of the interaction between SCIMP and JAK1, the different functional domains of SCIMP and JAK1 were constructed based on the SMART database and previous literature^{11 17} (figure 2G). The results showed that amino acids 40–103 of SCIMP interacted with amino acids 583–855 of JAK1 (figure 2H,I; online supplemental figure S2E, F), indicating that SCIMP interacts with the JAK1 pseudokinase domain at the plasma membrane, which may be implicated in the regulation of JAK1 activity.

Figure 2. SLP adaptor and CSK-interacting membrane protein (SCIMP) interacts with JAK1 pseudokinase domain at the plasma membrane. (A) Representative immunofluorescent images of SCIMP (green) and DAPI (blue) in T24 and UMUC3 cells. Scale bar, 10 µm. (B) Silver staining showed proteins pulled down by SCIMP from the lysates of T24 (left) and UMUC3 (right) cells. Arrowhead indicated the major differential band. (C) Venn diagram showed overlapping proteins identified by MS assay in T24 and UMUC3 cells. Analysis pipeline was performed to identify SCIMP-binding proteins; (i) 48 proteins were identified after overlapping; (ii) nine proteins with molecular masses of 130–180 kDa were then selected as candidates based on the differential band observed in silver staining; (iii) JAK1 was selected as it was the only membrane-associated protein. (D and E) Co-immunoprecipitation (Co-IP) analysis of the interaction between SCIMP and JAK1 in T24 and UMUC3 cells. (F) Immunofluorescence staining showed co-localization of SCIMP (green) and JAK1 (red) in T24 and UMUC3 cells. Nuclei were stained with DAPI (blue). Scale bar, 10 µm. (G) Schematic illustration of the functional domains of JAK1 and SCIMP, as well as the truncation domains. (H) Co-IP analysis of the interaction between SCIMP and full-length or truncations of HA-tagged recombinant JAK1 in the T24 cells. (I) Co-IP analysis of the interaction between JAK1 and full-length or truncations of Flag-tagged recombinant SCIMP in the T24 cells.

SCIMP mediates negative regulation of IFN-γ/JAK1/STAT5A signaling through inhibiting JAK1 phosphorylation

JAK1, a member of the non-receptor tyrosine kinase family, transduces extracellular signals on receptor activation through autophosphorylation, thereby initiating downstream signaling cascades.¹² Given the interaction between SCIMP and JAK1, we hypothesized that SCIMP might act on JAK1 to regulate its activity. Indeed, SCIMP overexpression significantly reduced JAK1 phosphorylation without altering the total protein level in tumor cells when co-cultured with CD8⁺ T cells (figure 3A; online supplemental figure S3A, B). In addition, the inhibitory effect of SCIMP on tumor cell viability in the co-culture system was abolished by treatment with the JAK1 inhibitor itacitinib (figure 3B; online supplemental figure S3C), suggesting that SCIMP promotes antitumor immune responses mainly through inhibition of JAK1 activity.

Figure 3. SLP adaptor and CSK-interacting membrane protein (SCIMP) inhibits IFN-γ/JAK1/STAT5A signaling. (A) Western blot analysis of SCIMP, p-JAK1, and JAK1 protein levels in T24 and MB49 cells transfected with a vector or an SCIMP expression plasmid and co-cultured with or without CD8⁺ T cells for 8 hours. GAPDH was used as a loading control. (B) CCK-8 assay showed the viability of T24 and MB49 cells treated with DMSO or itacitinib and co-cultured with CD8⁺ T cells for 24 hours. Cells were transfected with a vector or an SCIMP expression plasmid. (C) Western blot analysis of SCIMP, p-JAK1, and JAK1 protein levels in T24 and MB49 cells transfected with a vector or an SCIMP expression plasmid and treated with or without IFN-γ stimulation for 1 hour. GAPDH was used as a loading control. (D) Western blot analysis of SCIMP, p-JAK1, and JAK1 protein levels in T24 cells co-cultured with or without CD8⁺ T cells in the presence or absence of anti-IFN-γ antibody for 8 hours. Cells were transfected with a vector or an SCIMP expression plasmid. GAPDH was used as a loading control. (E) Western blot analysis of SCIMP, p-JAK1, and JAK1 protein levels in MB49 cells co-cultured with or without CD8⁺ T cells in the presence or absence of anti-IFN-γ antibody for 8 hours. Cells were transfected with a vector or an SCIMP expression plasmid. GAPDH was used as a loading control. (F) JAK/STAT II Phospho Antibody Array depicting the differentially phosphorylated proteins between vector and SCIMP-overexpressing T24 cells co-cultured with CD8⁺ T cells for 8 hours. (G) Western blot analysis of SCIMP, p-JAK1, JAK1, p-STAT5A, STAT5A, p-STAT2, STAT2, p-STAT4, and STAT4 protein levels in T24 and MB49 cells transfected with a vector or an SCIMP expression plasmid and co-cultured with CD8⁺ T cells for 8 hours. GAPDH was used as a loading control. (H) Western blot analysis of SCIMP, p-JAK1, JAK1, p-STAT5A, and STAT5A protein levels in T24 cells co-cultured with or without CD8⁺ T cells in the presence or absence of anti-IFN-γ antibody for 8 hours. Cells were transfected with a vector or an SCIMP expression plasmid. GAPDH was used as a loading control. (I) Western blot analysis of SCIMP, p-JAK1, JAK1, p-STAT5A, and STAT5A protein levels in MB49 cells co-cultured with or without CD8⁺ T cells in the presence or absence of anti-IFN-γ antibody for 8 hours. Cells were transfected with a vector or an SCIMP expression plasmid. GAPDH was used as a loading control.

As a key effector of CD8⁺ T cell function, IFN-γ is responsible for activating JAK1 in tumor cells.^{11 21} Notably, SCIMP markedly inhibited the increase in JAK1 phosphorylation induced by IFN-γ treatment (figure 3C; online supplemental figure S3D). Moreover, under CD8⁺ T cell co-culture conditions, the suppressive effect of SCIMP on JAK1 phosphorylation is dependent on IFN-γ signaling and is not observed when IFN-γ is neutralized (figure 3D,E; online supplemental figure S3E), demonstrating that SCIMP acts as a negative regulator of IFN-γ-mediated JAK1 activation.

It is well established that phosphorylated JAK1 activates STAT transcription factors.^{11 22} Accordingly, we performed a phospho-antibody array comprising site-specific and phospho-specific antibodies related to the JAK/STAT pathway. The array analysis validated the SCIMP-induced reduction in JAK1 phosphorylation, but also revealed a significant decrease in the phosphorylation levels of STAT2 (Tyr690), STAT4 (Tyr693), and STAT5A (Ser726) in SCIMP-OE cells co-cultured with CD8⁺ T cells (figure 3F). Next, we used immunoblotting to validate whether SCIMP could impede these phosphorylation events under CD8⁺ T cell co-culture conditions, and found that only STAT5A phosphorylation at Ser726 was regulated by SCIMP (figure 3G; online supplemental figure S3F, G). Furthermore, the anti-IFN-γ antibody completely abolished CD8⁺ T cell-induced activation of the JAK1/STAT5A pathway (figure 3H,I; online supplemental figure S3H). Importantly, the inhibitory effect of SCIMP on STAT5A phosphorylation was lost under co-culture conditions following IFN-γ blockade (figure 3H,I; online supplemental figure S3H), suggesting that the inhibition of JAK1/STAT5A caused by SCIMP is dependent on IFN-γ elicited by CD8⁺ T cells. These findings support the notion that SCIMP is crucial for repressing the IFN-γ/JAK1/STAT5A signaling pathway.

SCIMP prevents tumor immune evasion by repressing PD-L1 transcription

To delve deeper into the underlying molecular mechanism, vector control and SCIMP-OE T24 cells were harvested for transcriptome analysis following co-culture with CD8⁺ T cells. In line with the Gene Ontology analysis linking SCIMP to T cell costimulation (figure 4A), we found that SCIMP significantly downregulated PD-L1 mRNA levels (figure 4B,C; online supplemental figure S4A; online supplemental table S3). Furthermore, SCIMP overexpression decreased the luciferase activity of the PD-L1 promoter (online supplemental figure S4B). Therefore, we speculated that SCIMP represses PD-L1 transcription by inhibiting IFN-γ/JAK1 signaling. As expected, treatment with either anti-IFN-γ antibody or itacitinib could eliminate the decline in PD-L1 mRNA and protein levels caused by SCIMP overexpression (figure 4D–G; online supplemental figure S4C-F). In addition, the impact of SCIMP on bladder cancer cell viability was compromised in the presence of exogenous PD-L1 expression, supporting that SCIMP counteracts tumor immune evasion from CD8⁺ T cell cytotoxicity via reducing PD-L1 expression (figure 4H,I; online supplemental figure S4G, H). Consistent with the role of STAT5A as a transcription factor, ChIP experiments demonstrated that STAT5A binds to the PD-L1 promoter region, while SCIMP overexpression decreased the binding of STAT5A to the PD-L1 promoter in co-culture with CD8⁺ T cells (figure 4J,K; online supplemental figure S4I, J). Thereby, SCIMP attenuates IFN-γ/JAK1/STAT5A signaling in bladder cancer cells, leading to PD-L1 downregulation and the subsequent inhibition of tumor immune escape.

Figure 4. SLP adaptor and CSK-interacting membrane protein (SCIMP) regulates programmed death-ligand 1 (PD-L1) expression to counteract immune evasion. (A and B) Gene Ontology enrichment analysis (A) and Volcano plot (B) of differentially expressed genes between vector and SCIMP-overexpressing T24 cells following 24 hours of co-culture with CD8⁺ T cells. (C) Relative PD-L1 mRNA levels measured by qRT-PCR in T24 and MB49 cells transfected with a vector or an SCIMP expression plasmid and co-cultured with CD8⁺ T cells for 24 hours. (D) Relative PD-L1 mRNA levels measured by qRT-PCR in T24 and MB49 cells treated with IgG or anti-IFN-γ antibody and co-cultured with CD8⁺ T cells for 24 hours. Cells were transfected with a vector or an SCIMP expression plasmid. (E) Western blot analysis of SCIMP and PD-L1 protein levels in T24 and MB49 cells co-cultured with CD8⁺ T cells in the presence or absence of anti-IFN-γ antibody for 24 hours. Cells were transfected with a vector or an SCIMP expression plasmid. GAPDH was used as a loading control. (F) Relative PD-L1 mRNA levels measured by qRT-PCR in T24 and MB49 cells treated with DMSO or itacitinib and co-cultured with CD8⁺ T cells for 24 hours. Cells were transfected with a vector or an SCIMP expression plasmid. (G) Western blot analysis of SCIMP and PD-L1 protein levels in T24 and MB49 cells treated with DMSO or itacitinib and co-cultured with CD8⁺ T cells for 24 hours. Cells were transfected with a vector or an SCIMP expression plasmid. GAPDH was used as a loading control. (H) Western blot analysis of SCIMP and PD-L1 protein levels in T24 and MB49 cells transfected with SCIMP or PD-L1 expression plasmids and co-cultured with CD8⁺ T cells for 24 hours. GAPDH was used as a loading control. (I) CCK-8 assay showed the viability of T24 and MB49 cells transfected with SCIMP or PD-L1 expression plasmid and co-cultured with CD8⁺ T cells for 24 hours. (J) ChIP analysis of STAT5A enrichment at the indicated regions of the PD-L1 promoter in T24 and MB49 cells. (K) STAT5A-specific ChIP followed by RT-qPCR showed binding of STAT5A to PD-L1 promoter in T24 and MB49 cells transfected with the vector or SCIMP expression plasmid and co-cultured with or without CD8⁺ T cells for 8 hours. The relative binding of STAT5A to the PD-L1 promoter was calculated by first determining the STAT5A/input ratio for each group and then normalizing the SCIMP-OE groups to the vector group for comparison.

YTHDF2 contributes to SCIMP downregulation via m6A modification

The mechanisms leading to aberrant SCIMP expression in bladder cancer remain unclear. RNA modifications play critical roles in gene expression; we wondered whether m6A modification contributes to SCIMP downregulation. Using the SRAMP online tool,²³ two potential m6A modification sites with very high confidence (Site1 and Site2) were predicted in the 3′-UTR region of SCIMP mRNA (figure 5A). Subsequently, MeRIP-qPCR revealed abundant m6A enrichment in the SCIMP fragments corresponding to Site2 (figure 5B). To further validate the m6A modification at Site2, we inserted the wild-type or m6A-site-mutant SCIMP 3′-UTR sequences into luciferase reporter plasmids. Compared with the wild-type sequence, mutation of Site2 resulted in diminished m6A levels (figure 5C), indicating that Site2 is essential for the m6A modification. The biological effects of m6A modification depend on its recognition by specific reader proteins.²⁴ Accordingly, a comprehensive analysis integrating catRAPID omics, RBPsuite, and the RM2Target database predicted four reader proteins potentially associated with SCIMP, including IGF2BP3, YTHDC1, YTHDF2, and YTHDF3 (figure 5D; online supplemental table S4). In order to verify these candidates, we employed RNA pull-down assays, which revealed that only YTHDF2 specifically interacted with biotinylated SCIMP (figure 5E). Consistently, SCIMP mRNA was also enriched in YTHDF2 RIP assays (figure 5F). Moreover, mutation of the m6A site could disrupt the interaction between SCIMP 3′-UTR and YTHDF2 (figure 5G,H), implicating that YTHDF2 recognizes the m6A modification of SCIMP. We next observed that the increase in luciferase activity induced by YTHDF2 knockdown was abolished on mutation of the m6A site (figure 5I). YTHDF2 is widely known for specifically regulating the stability of m6A-modified mRNAs.²⁵ In line with this, YTHDF2 knockdown enhanced SCIMP mRNA stability, leading to increased SCIMP mRNA and protein expression (figure 5J,K; online supplemental figure S5A). These results confirm that YTHDF2 accelerates SCIMP mRNA decay through the m6A modification within its 3′-UTR.

Figure 5. YTHDF2 degrades SLP adaptor and CSK-interacting membrane protein (SCIMP) mRNA in an N6-methyladenosine (m6A)-dependent manner. (A) The potential m6A sites of SCIMP predicted by SRAMP software. (B) MeRIP-qPCR assays of Site1 and Site2 regions of SCIMP mRNA in T24 and UMUC3 cells using m6A and IgG antibodies. (C) Top: schematic diagram of SCIMP-3′-UTR-WT and SCIMP-3′-UTR-Mut luciferase reporters. Down: MeRIP-qPCR assays showed the enrichment of SCIMP on m6A and IgG antibodies in T24 and UMUC3 cells transfected with SCIMP-3'-UTR-WT or SCIMP-3'-UTR-Mut reporters. (D) Venn diagram showed the potential m6A readers targeting SCIMP, as predicted by catRAPID omics, RBPsuite, and the RM2 target database. (E) The interaction between YTHDF2 protein and SCIMP RNA in T24 and UMUC3 cells was verified by RNA pull-down assay. (F) RNA immunoprecipitation (RIP) assays in T24 and UMUC3 cells using YTHDF2 and IgG antibodies. The precipitate was subjected to Western blot with the antibody against YTHDF2. The YTHDF2-enriched SCIMP RNA was measured by qRT-PCR. (G and H) RIP-qPCR assays showed the enrichment of SCIMP on YTHDF2 and IgG antibodies in T24 and UMUC3 cells transfected with SCIMP-3'UTR-WT or SCIMP-3'UTR-Mut reporters. (I) The relative luciferase activity of SCIMP-3'UTR-WT or SCIMP-3'UTR-Mut reporters in T24 and UMUC3 cells transfected with scramble, sh-YTHDF2#1 or sh-YTHDF2#2. (J) The relative remaining levels of SCIMP mRNA in T24 and UMUC3 cells transfected with scramble, sh-YTHDF2#1 or sh-YTHDF2#2 were analyzed by qRT-PCR after treatment with 5 µg/µL actinomycin D. (K) Western blot analysis of YTHDF2 and SCIMP protein expression levels in T24 and UMUC3 cells transfected with scramble, sh-YTHDF2#1, or sh-YTHDF2#2. GAPDH was used as a loading control.

SCIMP enhances the efficacy of tumor immunotherapy in bladder cancer

Based on our in vitro findings, we first sought to determine whether SCIMP similarly regulates PD-L1 expression in vivo. Consistent with in vitro data, SCIMP-OE tumors displayed reduced PD-L1 levels (online supplemental figure S6A,B). To investigate the impact of SCIMP overexpression on the therapeutic efficacy of PD-1 blockade, syngeneic C57BL/6 mice bearing MB49 tumors were treated with IgG or anti-PD-1 antibody. Although either SCIMP overexpression or anti-PD-1 antibody suppressed tumor growth, the combination treatment exhibited a synergistic effect in reducing tumor burden (figure 6A–C). In addition, flow cytometry demonstrated that the combination therapy significantly promoted CD8⁺ T cell infiltration and antitumor cytokine production (figure 6D–G; online supplemental figure S6C). Given that SCIMP enhances antitumor immune responses by inhibiting JAK1 activation, we employed the subcutaneous tumor model to evaluate the efficacy of JAK1 inhibitor itacitinib and PD-1 blockade. As shown in figure 6H–J, an efficient delay in tumor growth was observed in mice treated with itacitinib in combination with anti-PD-1 therapy. Immune infiltration analysis further revealed that the proportion and activity of CD8⁺ T cells were markedly augmented following the combined administration of itacitinib and PD-1 blockade (figure 6K–N). Taken together, these data demonstrate a potent synergy between SCIMP and PD-1 blockade in the treatment of bladder cancer.

Figure 6. Combination therapy of SLP adaptor and CSK-interacting membrane protein (SCIMP) and PD-1 blockade in bladder cancer. (A–C) Tumor growth and tumor weight in C57BL/6 mice injected subcutaneously with MB49 cells stably expressing a vector or an SCIMP and treated with anti-PD-1 or isotype control (n=6). (D–G) Flow cytometry assay revealed the tumor-infiltrating CD8⁺ T cells, IFN-γ+ CD8⁺ T cells, and TNF-α+ CD8⁺ T cells in MB49 tumors of the indicated groups. (H–J) Tumor growth and tumor weight in C57BL/6 mice injected subcutaneously with MB49 cells administered with itacitinib and treated with anti-PD-1 or isotype control (n=6). (K–N) Flow cytometry assay revealed the tumor-infiltrating CD8⁺ T cells, IFN-γ+ CD8⁺ T cells, and TNF-α+ CD8⁺ T cells in MB49 tumors of the indicated groups.

Discussion

Despite the promising results of ICIs in bladder cancer therapy, tumors frequently develop resistance by evading CD8⁺ T cell immune surveillance.^{7 26–28} Beyond T cell receptor-mediated cytotoxicity, a substantial proportion of CD8⁺ T cells within the TME are not tumor antigen-specific and can exert bystander cytotoxic activity independent of tumor antigen-specific recognition.²⁹ To gain broader insights into tumor immune evasion in the context of bystander T-cell responses, we performed a genome-wide CRISPR screen on bladder cancer cells followed by co-culture with CD8⁺ T cells, and identified SCIMP as a critical regulator that counteracts tumor immune evasion. Mechanistically, SCIMP inhibited the induction of PD-L1 expression in response to IFN-γ secreted by CD8⁺ T cells via the JAK1/STAT5A axis, thereby promoting CD8⁺ T cell antitumor activity. Furthermore, combining PD-1 blockade with either SCIMP overexpression or pharmacological JAK1 inhibition promoted CD8⁺ T-cell infiltration and activity within the TME, thereby suppressing bladder cancer growth in vivo. In conclusion, we unveiled a non-canonical, adaptor-independent role of SCIMP in bladder cancer through direct interaction with and inhibition of JAK1, which may provide a promising strategy for improving bladder cancer immunotherapy.

The pTRAP family member SCIMP, lacking intrinsic enzymatic activity, modulates specific receptor pathways in myeloid and lymphoid cells by recruiting effector molecules to the plasma membrane.³⁰ Intriguingly, the function of SCIMP in bladder cancer appeared to differ from its classical adaptor role, since receptor complexes, including those previously reported as associated with SCIMP,^{14 17} were not detected in our mass spectrometry data. Instead, our analysis revealed that SCIMP physically interacts with JAK1 in cancer cells. Contrary to previous reports documenting that SCIMP scaffolds kinases to facilitate their phosphorylation and propagate downstream signaling,¹⁶ we found that SCIMP unexpectedly inhibits JAK1 phosphorylation on IFN-γ stimulation in bladder cancer. This inhibitory effect may result from SCIMP binding to the pseudokinase domain of JAK1, which potentially induces a conformational change in JAK1 that restricts its kinase activity.³¹ Alternatively, SCIMP might indirectly decrease JAK1 phosphorylation via an as yet unidentified intermediary. Although the potential roles of SCIMP in other tumors warrant further investigation, our findings expand the current perception regarding SCIMP function. In addition, SCIMP overexpression improved the therapeutic effect of PD-1 blockade in vivo, and patients with bladder cancer with high SCIMP expression derived greater benefit from immunotherapy, suggesting that SCIMP may serve as a predictive biomarker for immunotherapy response and a potential candidate for counteracting immune evasion. However, targeted restoration of SCIMP expression in tumor tissues in vivo remains challenging. Nonetheless, our study highlights the potential clinical relevance of SCIMP in bladder cancer immunotherapy.

JAK1 has been implicated in tumor immune evasion,³² but the regulatory mechanisms governing JAK1 in cancer remain to be fully elucidated. Here, we demonstrated that SCIMP functions as a novel JAK1 phosphorylation inhibitor that specifically targets the pseudokinase domain in bladder cancer. As recently revealed, targeting the pseudokinase domain could induce JAK1 allosteric inhibition and improve selectivity within the JAK family.^{33 34} Our findings further provide the rationale for the development of JAK1 allosteric inhibitors. Additionally, consistent with observations in other malignancies,^{35 36} we found that JAK1 inhibition reduced PD-L1 expression and enhanced the activity of tumor-infiltrating CD8⁺ T cells, effectively limiting immune evasion in bladder cancer. Notably, our unbiased JAK/STAT phospho-antibody array screening revealed that SCIMP-mediated reduction in JAK1 activity predominantly inhibited STAT5A phosphorylation, rather than JAK1/STAT1 signaling.³⁵ We speculate that cellular and environmental contexts, as well as compensatory regulatory mechanisms, may contribute to the selective inhibition of JAK1/STAT5A. In contrast to STAT1, which plays paradoxical roles in tumor development by driving the expression of immunosuppressive molecules while simultaneously mediating the antitumor effects on cancer cells, STAT5A predominantly functions as a driver of immune evasion.^{37 38} Therefore, selective inhibition of the JAK1/STAT5A branch by SCIMP raises the possibility that SCIMP may offer a superior therapeutic strategy compared with conventional JAK1 inhibitors by circumventing the broad-spectrum inhibition of STATs. However, elucidating the conformational basis of the SCIMP-JAK1 interaction remains an essential step toward the development of SCIMP-based mimetics.

As the most prevalent modification of mRNA, m6A modification is associated with cancers.³⁹ Herein, we identified an m6A site within the 3′-UTR of SCIMP, which could be recognized by the reader YTHDF2. It has been shown that YTHDF2 promotes UBXN1 mRNA decay via binding to m6A modification sites, which in turn drives the progression of glioma.⁴⁰ Similarly, YTHDF2 facilitates bladder cancer immune evasion by degrading the m6A-modified RIG-I mRNA.⁴¹ In this study, YTHDF2 targets SCIMP mRNA for degradation in an m6A-dependent manner, which may partially account for SCIMP downregulation in bladder cancer. Although the specific writer complex responsible for m6A modification of SCIMP mRNA has yet to be determined, our results provide evidence that m6A modification is involved in the post-transcriptional regulation of SCIMP expression.

In summary, this study provides a perspective on immune evasion in bladder cancer and identifies SCIMP as a key regulator of tumor immunity. SCIMP exerts a non-canonical function in restricting JAK1 phosphorylation, thereby inhibiting the IFN-γ/JAK1/STAT5A/PD-L1 signaling pathway to potentiate CD8⁺ T cell responses and counteract tumor immune evasion. Hence, SCIMP may represent a promising therapeutic strategy to enhance immunotherapy efficacy for patients with bladder cancer.

Data availability statement

Data are available upon reasonable request.