Vero, SKOV3 (human ovarian), Panc-1, and Capan-2 cell lines were obtained from ATCC. Luciferase-expressing lines (Capan-2-luc, Panc-1-luc) were created via lentiviral transduction followed by puromycin selection (3 µg/mL) and clonal isolation. Panc-1-luc-MSLN cells were similarly engineered to express human mesothelin and validated by flow cytometry. All lines tested negative for mycoplasma and were STR-authenticated. Culture conditions: Capan-2-luc and SKOV3 in McCoy’s 5A + 10% FBS (with puromycin for Capan-2-luc); Panc-1-luc in DMEM + 10% FBS (with puromycin); Vero in MEM + 10% FBS; incubated at 37 °C and 5% CO₂. Oncolytic HSV-2 vectors (oHSV-2-GFP and oHSV-2-IL-12) derived from HG52 strain with ICP34.5 deletion and IL-12 insertion were produced, purified, and titered by plaque assay by Binhui Biotech (Wuhan, China) (16).

A third-generation, four-plasmid lentiviral system (pMDLg/pRRE, pRSV-Rev, pCMV-VSV-G, pRRLSIN backbone) was used. The backbone promoter was switched to EF-1α. Chimeric cDNA sequences containing SS1-TM ICOS–ICOS-BBz were custom synthesized (Nanjing GenScript Biotechnology Co., Ltd, China), digested with BamHI and SalI and ligated into the pRRLSIN-EF-1a-MCS. Plasmids were sequence-verified. Lentivirus production and titration were performed in 293T cells by Huaxia Yingtai (Beijing).

PBMCs were isolated from healthy donor apheresis samples (Shanghai Miaoshun Biotechnology Co., Ltd.). CD4⁺ and CD8⁺ T cells were purified using negative selection (Miltenyi, Cat# 130-096-535), activated with TransAct (Miltenyi Biotec; Cat# 130-111-160) at a density of 1 × 10⁶ cells per 10 μL reagent for 24 hours (24 h), and cultured in TexMACS + IL-15 (84 U/mL) and IL-7 (500 U/mL). After 24 h, cells were transduced at multiplicities of infection (MOI) = 3 and maintained with cytokine support. CAR-T cells were harvested after 12 days.

Target cells (Capan-2-luc, Panc-1-luc, Panc-1-luc-MSLN) were plated in 96-well format at respective densities in McCoy’s or DMEM + 10% FBS. After 24 h, CAR-T cells were added at various E:T ratios. Triton X-100 (1%) served as a positive lysis control. After 30 h co-culture, Bright-Glo luciferase reagent (Promega, Cat# E2620) was added and luminescence measured using a GloMax^® 96 Microplate Luminometer (Promega) with a 1-second integration time at 560 nm. Percent cytotoxicity was calculated as:

X = experimental RLU, Max = effector-untreated control, Min = Triton-X 100-treated control.

Capan-2-luc cells (5×10⁴/well) were seeded in xCELLigence E-plates (Agilent, Cat# 300600890), then infected with oHSV-2-IL-12 at MOIs of 0.05, 0.5, or 5. After 2 h adsorption, medium with 2.5% FBS was added. Cell index (CI) was monitored every 10 min for 96 h. In the same real-time cytolysis assay, KT₈₀ (time to reach 80% cytolysis) was also calculated.

In vitro, Capan-2-luc cells (5×10⁵–6×10⁵/well) were infected with oHSV-2-IL-12 (MOI 0.1 or 0.5). Cell and supernatant samples were collected at 24 and 48 h post-infection. IL-12 p70 levels were measured using RayBio^® Human IL-12 p70 ELISA Kit (Cat# ELH-IL12P70-1). In vivo, tumor-bearing B-NDG mice received three intratumoral doses of oHSV-2-IL-12. Tumors were harvested on day 9 post-first dose, homogenized, lysed, and IL-12 quantified via Abcam p70 kit Human IL-12 p70 SimpleStep ELISA^® Kit (abcam, ab223592), normalized to tissue weight.

Immunodeficient B-NDG (NOD.Cg-PrkdcscidIl2rgtm1/Bcgen) female mice (18–21 g, SPF) from Biocytogen Pharmaceuticals (Beijing) Co., Ltd were housed at NIFDC vivarium. All procedures were IACUC-approved.

2×10⁶ Capan-2-luc cells were injected subcutaneously (s.c.). Tumor volume (V) was measured by calipers:. V = 1/2 × L × W × W. Endpoint: volume ≥ 1,000 mm³.

Mice received intraperitoneal D-luciferin (150 mg/kg; PerkinElmer, Cat#122799) and were imaged using IVIS Spectrum Imaging System (PerkinElmer)10 min post-injection. Bioluminescence signals were quantified using Living Image software (PerkinElmer) and expressed as total flux (photons/second) within a region of interest drawn around the tumor site.

In vivo xenograft tumors were established by s.c. injection of 2 × 10⁶ Capan-2 cells in PBS. On day 10 (tumor size: 50–100 mm³), mice received 1×10⁷ CAR-T cells intravenously in CAR-T monotherapy experiments. In combination therapy, mice received 1×10⁷ CAR-T cells intravenously and intratumoral injections of oHSV-2-IL-12 or oHSV-2-GFP (1×10⁵ CCID₅₀ in 50 µL PBS) every 3 days for 3 doses. For OV monotherapy dose titration, oHSV-2-IL-12 was administered at 1×10⁶, 1×10⁵, or 1×10⁴ CCID₅₀. Except for CAR-T monotherapy, where Capan-2 is treated 10 days after inoculation, OV therapies or OV combination with CAR-T therapies are administered 14 days after Capan-2 inoculation. Peripheral blood was obtained from retro-orbital bleeding. Blood and spleen were sampled for CAR-T persistence (human CD45⁺CD3⁺CAR⁺) by flow cytometry.

Pancreatic cancer–xenografted B-NDG mice that achieved complete remission (≥ 14 days tumor-free) were re-challenged s.c. with 2×10⁶ Capan-2-luc (right flank) and 2×10⁶ Panc-1-luc (left flank). Tumor growth and bioluminescence were monitored weekly.

Blood: 70 μL collected via retro-orbital bleed on days 7, 9, and every 3 days until day 27 post-treatment.7Tumors: harvested on treatment day 9 post-first oHSV-2-IL-12 dose. Homogenized (Shanghai Jingxin Industrial Development Co., Ltd., Shanghai, China) in PBS-lysis buffer (absin, Cat# abs9225) with protease inhibitors (absin, Cat# abs9162); lysates centrifuged and supernatants retained.Spleens: harvested, processed through 70 μm strainers, RBC-lysed with BioLegend buffer (BioLegend, Cat# 420302), and used for flow cytometry.

The following fluorescent reagents were used for staining: MSLN-PE (ACRO, Cat# MSN-HP2H5) for CAR, CD4-AmCyan (BD, Cat# 562970), and CD8-APC-Cy7 (BD, Cat# 557834), each diluted at 1:100; CD3-FITC (BD, Cat# 555332) diluted at 1:10; and 7-AAD (BD, Cat# 559925).

CAR-T Phenotyping: Cells were stained in PBS + 2% FBS with antibodies against CD3, CD4, CD8, viability dye 7-AAD, and CAR detected via PE-labeled mesothelin. 7-AAD unstained live cells were gated, and CD3⁺ cells were identified. Within the CD3⁺ population, CAR⁺ (PE postive) T cells were further gated, and the proportions of CD4⁺ and CD8⁺ cells within the CAR⁺ T cell population were analyzed.

CAR+ T cell monitoring In Vivo: Blood and spleen cells stained similarly, blood was processed with BD FACS™ Lysing Solution. Tumor homogenates were stained with human CD45-BV650, MSLN-PE, CD3, and 7-AAD. Analysis was performed on an LSRFortessa cytometer (Becton Dickinson) using FlowJo. Live CD45+ cells were gated, the proportion of CAR⁺ cells within the CD3⁺ population was analyzed.

After 24 h recovery, SS1-ICOSBBZ-CAR-T cells were co-cultured with Capan-2 target cells at E:T = 2:1 for 48 h. CD4⁺ and CD8⁺ subsets were separated using the human CD8 MicroBeads (Miltenyi), stained with AF647-conjugated antibodies (IsoPlexis,STAIN-1002–1 and STAIN-1003-1), loaded onto IsoLight chips (IsoPlexis), and analyzed using IsoSpeak v3.0.1.

Analyses conducted in GraphPad Prism 9.0. Data are mean ± SD. Two-group comparisons were by two-tailed Student’s t−test; multiple groups by one-way ANOVA with Tukey’s range test. Tumor growth was analyzed by two-way repeated measures ANOVA with variance normalization if necessary.

Inducible T-cell co-stimulator (ICOS), a member of the CD28/CTLA-4 family, plays a pivotal role in modulating immune responses and T-cell differentiation. To harness this property, we engineered third-generation CAR-T cells by transducing peripheral blood T cells from healthy donors with a lentiviral vector encoding the SS1-ICOSBBZ-CAR. This CAR comprises an SS1-derived single-chain variable fragment (scFv) targeting mesothelin, the transmembrane domain of ICOS, and intracellular signaling domains from ICOS, 4-1BB, and CD3ζ. The design of this CAR construct is illustrated in Figure 1A, as previously described by Guedan et al. (4). The synthesized CAR gene was cloned into a lentiviral vector, packaged using 293T cells, and subsequently used to transduce peripheral blood T cells, generated SS1-ICOSBBZ-CAR-T cells.

After completing the manufacturing of the CAR-T cells, we evaluated the final product’s quality by assessing the CAR expression (CAR⁺ percentage) and determining the proportions of CD4⁺ and CD8⁺ T cells. Flow cytometric analysis revealed that the CAR expressing rate was 23.4% (Figure 1B, left), among CD3⁺ T cells, 37.6% were CD4⁺ and 61.5% were CD8⁺ (Figure 1B, right), resulting in a CD4/CD8 ratio of 0.61.

To evaluate the cytotoxic function of SS1-ICOSBBZ-CAR-T cells, in vitro co-culture assays with pancreatic cancer cell lines revealed potent, dose-dependent and antigen specific killing of MSLN -positive targets but not MSLN-negative controls (Figure 1C). Specifically, the CAR-T cells lysed ~30% of MSLN⁺ Capan-2-luc cells at a 1:10 effector-to-target (E:T) ratio, increasing to 50% at 1:1. similarly, they achieved 20% lysis against MSLN⁺ Panc-1-luc-MSLN cells at 2:1 E:T, escalating to 80% at 8:1 within 24 hours. Critically, no significant cytotoxicity (<5% lysis) was observed against MSLN^- Panc-1-luc cells. These results confirm the potent and antigen-specific antitumor activity of SS1-ICOSBBZ-CAR-T cells in vitro (Figure 1C).

We next evaluated the antitumor efficacy in an immune-deficient B-NDG mouse model bearing subcutaneous pancreatic tumors in vivo. Capan-2-luc cells (2×10⁶) were subcutaneously inoculated into the right flank of B-NDG mice to establish tumors. On day 10 post-engraftment, mice received 1×10⁷ SS1-ICOSBBZ-CAR-T cells via tail vein injection (experimental timeline, Figure 1D). CAR-T treatment significantly suppressed tumor growth (p<0.01, Figures 1E, F), though complete eradication was not achieved. Tumor regrowth emerged from day 28 onward, indicating transient efficacy. Notably, no significant body weight loss occurred during treatment (Figure 1G), suggesting minimal systemic toxicity of SS1-ICOSBBZ-CAR-T cells.

OVs can infect and lyse tumor cells, and OVs engineered to express cytokines can modulate immune activity—these are crucial antitumor mechanisms of OV. Accordingly, we verified the capability of the oHSV-2 to infect and lyse pancreatic cancer cell lines both in vitro and in vivo, and the expression of IL-12 within tumors. OHSV-2-IL-12 has been reported to exhibit antitumor activity in the murine CT26 colon cancer model (16). In this study, oHSV-2-IL-12 demonstrated potent, dose-dependent antitumor activity against Capan-2-luc pancreatic cancer cells, successfully infecting cells at MOIs of 0.1 and 0.5 and inducing evident cytopathic effects (CPE) within 24 to 48 hours post-infection (Figure 2A). RTCA confirmed this dose-dependent cytolytic activity, showing that higher MOIs caused a faster and more pronounced decrease in Cell Index (CI) (Figure 2B, left). The cytotoxicity kinetics, measured as the time required to achieve 80% cell lysis (KT₈₀), exhibited a clear inverse relationship with MOI. KT₈₀ was approximately 54 hours at an MOI of 0.05, decreased to 34 hours at an MOI of 0.5, and reached about 12 hours at a higher MOI of 5 (Figure 2B, right), conclusively demonstrating that increased viral doses accelerate and enhance tumor cell killing.

Subsequent investigations demonstrated that oHSV-2-IL-12 drives functional IL-12 expression both in vitro and in vivo. In vitro, infecting Capan-2-luc cells at MOIs of 0.1 and 0.5 triggered a dose-dependent increase in IL-12 secretion, with ELISA quantifying approximately 1.5 ng/mL in culture supernatants by 48 hours post-infection (Figure 2C, left). In vivo, intratumoral administration of oHSV-2-IL-12 (three times, days 0, 3 and 6) into Capan-2-luc xenograft-bearing B-NDG mice led to pronounced IL-12 expression in tumor tissues. ELISA analysis of tumor homogenates 9 days post-first treatment revealed IL-12 concentrations of ~100 pg/mg tissue in two mice and ~400 pg/mg in the third (Figure 2C, right). Following intratumoral injection of oHSV-2-IL-12, IL-12 was undetectable in serum samples. This confirms that IL-12 expression mediated by the OV was localized specifically to the injected tumor site and sites of viral infection/spread, with no systemic leakage. Collectively, these findings demonstrate that oHSV-2-IL-12 achieves tumor-restricted delivery of IL-12, minimizing systemic exposure and the associated toxicity risks. This highly localized expression profile underscores the favorable safety profile of this approach for targeted immunotherapeutic applications.

We evaluated the in vivo antitumor efficacy of oHSV-2-IL-12 by intratumoral administration at three dose levels (1×10⁶, 1×10⁵, 1×10⁴ CCID₅₀/mouse) and as shown in Figure 2D. BLI results showed that between days 8 and 14, high- and medium-dose groups displayed significantly reduced tumor signal compared to untreated controls, whereas the low-dose group did not (Figure 2E). Consistent with these findings, tumor volumes in all treated groups were significantly lower than controls (Figure 2F); however, there were no statistically significant differences among the three dose groups. Body weights remained unchanged across all groups (Figure 2G), indicating that oHSV-2-IL-12 was well tolerated without detectable systemic toxicity.

Taken together, these data demonstrate that oHSV−2−IL−12 not only infects tumor cells and tissues but also drives IL−12 gene expression within infected cells. Compared to untreated controls, oHSV−2−IL−12 exhibited significant antitumor activity in vivo, as evidenced by both imaging and tumor growth measurements. However, although treatment inhibited tumor progression, it did not induce tumor regression or complete remission, indicating that oHSV−2−IL−12 monotherapy is insufficient for robust antitumor efficacy.

We hypothesized that oHSV-2-IL-12 plus MSLN targeted CAR-T cells could overcome the incomplete tumor clearance observed with each monotherapy. To evaluate whether combining OV therapy with CAR-T cells enhances tumor killing, we therefore used B-NDG mouse Capan-2-luc xenograft models, where 2×10⁶ tumor cells were implanted subcutaneously. On day 14 post-inoculation, medium-dose oHSV-2-IL-12 (1×10⁵ CCID₅₀) was administered intratumorally and repeated every three days for a total of three injections, SS1-ICOSBBZ-CAR-T cells (1×10⁷ cells/mouse) were delivered two days after the first OV injection, the experimental design is illustrated in Figure 3A. oHSV-2-GFP served as the control virus.

As hypothesized, the combination therapy significantly outperformed either CAR-T or OV monotherapy (Figures 3B, C). Complete tumor elimination occurred on days 14 and 21 post-treatment in the combination groups (Figure 3B). Statistical analysis of tumor volumes revealed that OV + CAR-T treatment showed significantly reduced tumor size compared to monotherapy groups, namely, the tumor volume in group 5 (oHSV-2-GFP + CAR-T) was significantly smaller than that in groups 2 (oHSV-2-GFP) (p < 0.0001), group 3 (oHSV-2-IL-12) (p < 0.0001), and group 4 (CAR-T) (p < 0.001), and tumor size in group 6 (oHSV-2-IL-12 + CAR-T) also significantly smaller than those monotherapy groups (Figure 3C). However, no significant difference in antitumor efficacy was observed between the two combination groups (5 vs. 6; p > 0.05).

Among three monotherapy groups, CAR-T appeared most potent as shown by bioluminescence imaging (Figure 3B); but tumor volume analysis showed no significant difference between CAR-T and oHSV-2-IL-12 groups (group 4 vs. 3, p > 0.05) (Figure 3C). In contrast, CAR-T treatment significantly reduced tumor size compared to the oHSV-2-GFP control (group 4 vs. 2, p < 0.0001), likely because oHSV-2-GFP lacks the capacity to fully activate host immune cell mediated cytotoxicity in immunodeficient B-NDG mice. Importantly, tumor volumes were also significantly less in the oHSV-2-IL-12 group than in oHSV-2-GFP (group 3 vs. 2, p < 0.001), demonstrating enhanced antitumor efficacy conferred by the IL-12-armed OV.

OHSV-2-IL-12 monotherapy demonstrated significantly superior antitumor activity over oHSV-2-GFP monotherapy in our study. However, we observed no significant difference in antitumor efficacy between the combination therapy group (CAR-T cells + oHSV-2-IL-12) and the control group (CAR-T cells + oHSV-2-GFP) in the xenograft tumor model. To elucidate whether oHSV-2-IL-12 enhances CAR-T cell therapy, we analyzed CAR-T cell proliferation and persistence in the peripheral blood and spleens of the mice. Peripheral blood was collected every three days starting one week post-treatment from five mice per group. Flow cytometry analysis of CD3⁺CAR⁺ T cells revealed that these cells were detected only in the oHSV-2-IL-12 + CAR-T combination group (Figure 3E). Within this group, SS1-ICOSBBZ-CAR-T cells were first detectable on day 13, peaked by day 16, and subsequently declined, becoming nearly undetectable by day 22. Substantial inter-animal heterogeneity in CAR-T cell numbers was observed. Figure 3E presents data from a single representative mouse. The exclusive detection of CAR⁺ T cells in the oHSV-2-IL-12 + CAR-T combination group suggests oHSV-2-IL-12 drives SS1-ICOSBBZ-CAR-T cell expansion and persistent.

Consistent with the peripheral blood findings, oHSV-2-IL-12 also increased the number of CAR-T cells in the spleen. At the study endpoint, spleens were harvested, dissociated, and analyzed for CAR-T cells within single-cell suspensions by flow cytometry (Figure 3F). On day 11 post-treatment, CD3⁺CAR⁺ cells were detectable in the spleens of all groups receiving CAR-T cell therapy. While oHSV-2-GFP combination therapy resulted in slightly higher CAR-T cell frequencies compared to CAR-T cell monotherapy, the oHSV-2-IL-12 + ICOSBBZ-CAR-T group exhibited the highest CAR-T frequency. These findings suggest that IL-12 augments both the expansion and splenic homing of SS1-ICOSBBZ-CAR-T cells. Thus, anti-MSLN SS1-ICOSBBZ-CAR-T cells expanded in the periphery and migrate to or persist within the spleen more effectively when supported by oHSV-2-IL-12.

In the xenograft tumor model, the combination therapy achieved complete tumor clearance. Therefore, we sought to determine whether CAR-T treatment could induce durable immunological protection by tumor rechallenge experiment.

As depicted in Figure 3, on day 28 after the combined treatment with 1×10⁵ CCID₅₀ oHSV-2-IL-12 and 1×10⁷ SS1-ICOSBBz-CAR-T cells, maintained a tumor-free for two weeks. At this time point, a re-challenge experiment was performed in the same Capan-2-luc xenograft mouse model. The schematic diagram of the re-challenge experimental is shown as Figure 4A. Simultaneously, oHSV-2-IL-12 + ICOSBBZ-CAR-T group and the blank group, which had neither received any tumor implantation nor treatment previously, was similarly injected with Capan-2-luc cells on the right side and Panc-1-luc cells on the left side. Tumor development was monitored using BLI and tumor volume was measured on days 7 and 14 post tumor cell injection. The results demonstrated that in the combination therapy group, the reintroduced Capan-2-luc cells failed to form new tumors, whereas tumors developed at the Panc-1-luc cells planted side. In the untreated blank group, tumors developed on both sides, as shown in Figures 4B, C. These results indicate that the combination therapy maintained tumor-free status for at least 5 weeks from the initial CAR-T administration and prevented tumor formation for over 2 weeks after rechallenge (Figure 4D). These results suggest that the combined therapy induced a durable, tumor-specific immune response capable of preventing tumor recurrence following the combination therapy of oHSV-2-IL-12 and SS1-ICOSBBz-CAR-T cells.

We observed that oHSV-2-IL-12 enhanced CAR-T cell proliferation and persistence in B-NDG pancreatic cells xenograft tumor model, but the combination therapy of oHSV-2-IL-12 with SS1-ICOSBBZ-CAR-T cells did not demonstrate greater efficacy than the oHSV-2-GFP and SS1-ICOSBBZ-CAR-T combination group as shown in BLI and tumor volume (Figures 3B, C). We speculate that it may be due to the rapid tumor clearance within approximately 7 days after treatment, which likely limited viral replication and limited IL-12 expression. As a result, the amount and duration of IL-12 exposure were probably insufficient to further enhance CAR-T proliferation under this high-dose condition. To better evaluate the therapeutic advantage of combination with oHSV-2-IL-12, we reduced the SS1-ICOSBBZ-CAR-T cell to one-twentieth of the original dose (from 1×10⁷ to 5×10⁵ cells). The schematic diagram of the experimental procedure was shown in Figure 5A. Briefly, following established combination therapy procedures (including tumor inoculation, oHSV-2 intra-tumor injections 1 × 10⁵ CCID₅₀ every 3 days) as previously described, a reduced dose of 5 × 10⁵ SS1-ICOSBBZ-CAR-T cells was administered intravenously one day after the initial oHSV-2 injection. BLI results indicated that the control virus and SS1-ICOSBBZ-CAR-T cells combination group did not achieve complete tumor eradication. In contrast, treatment with oHSV-2-IL-12 and CAR-T cells exhibited tumor regression in some tumor-bearing mice as early as day 14. By day 21, all five mice in oHSV-2-IL-12 combination with CAR-T group showed complete tumor clearance. These findings aligned with previous studies demonstrating that oHSV-2-IL-12 enhances the antitumor efficacy of SS1-ICOSBBZ-CAR-T cells. Consistent with these findings, a significant difference in tumor volumes treated by oHSV-2-IL-12 and CAR-T was observed between the two groups at treatment 21 day and 27 day, as shown in Figure 5C. In all groups, there were no significant differences in mouse body weight, indicating that the combination therapy is safer when CAR-T used lower dose (Figure 5D). As previously observed, only the treatment group receiving oHSV-2-IL-12 in combination with CAR-T cells—rather than the group treated with oHSV-2-GFP and CAR-T cells—exhibited detectable CAR⁺ T cells in peripheral blood (Figure 5E).

This result allowed us to more clearly observe the enhanced efficacy of the oHSV-2-IL-12 and SS1-ICOSBBZ-CAR-T combination therapy over the oHSV-2-GFP and SS1-ICOSBBZ regimen.

Compared with Figure1E, we observed that administering CAR-T cells alone at a dose of 1×10⁷ cells was insufficient to completely eradicate tumors. However, when combined with oHSV-2-IL-12, even a significantly reduced dose of CAR-T cells (one-twentieth of the original amount) achieved complete tumor clearance. This finding underscores the synergistic effect of oHSV-2-IL-12 in enhancing the antitumor efficacy of CAR-T cell therapy.

Occasional mouse deaths were observed in the oHSV-2 plus CAR-T cell combination therapy groups(Figures 3B, 5B), whereas no mortality was detected in the oHSV-2 alone or CAR-T monotherapy groups, it suggests there is a risk in combination OV with CAR-T, while, reduced CAR-T cell dose might be safer.

Polyfunctional T cells are recently reported for their association with long-term immune responses in the clinical settings (19).The SS1-ICOSBBZ-CAR-T cells used in our study exhibited moderate tumor-killing activity, the proliferation and persistence in vivo were limited. We analyzed their polyfunctional cytokine secretion profiles. Here, we used the 32-plex panel that included the key immune elements of T cells. The experimental results of single-cell secretion of multiple cytokines by CAR-T cells showed approximately 4% of CD8+ cells secreted two or more cytokines, while only 0.4% of CD4+ CAR-T cells were multifunctional, secreting two cytokines, almost no CD4+ CAR-T cells secreted more than two cytokines. As shown in Figure 6A, it indicated relatively low multifunctionality in CD4+ cells of SS1-ICOSBBZ-CART. The polyfunctional strength index (PSI) values are defined by cytokine function to highlight the contribution of each group to the overall polyfunctionality of the sample. As shown in Figure 6B, CD8+ SS1-ICOSBBZ-CAR-T cells showed 10 times higher than CD4+ CAR T PSI when co-cultured with Canpan-2-luc cells, indicating CD8+ CAR T cells were more polyfunctional than CD4+ CAR T cells. Heatmap analysis, as shown in Figure 6C, the most frequently secreted cytokines included Granzyme B, perforin, IFN-γ, IP-10(CXCL11), sCD137(41BB), and GM-CSF. CD8+ cells exhibited a higher secretion frequency of these cytokines. In contrast, CD4+ cells had a lower proportion secreting IL-12, IL-15, and TNF-β in addition to the aforementioned cytokines. These findings further highlight the functional differences between CD8+ and CD4+ SS1-ICOSBBZ-CAR-T cells. The poor proliferation of SS1-ICOSBBZ-CAR-T in vivo may be attributed to the low percent multifunctional CD4+ of SS1-ICOSBBZ-CAR-T.