The structures of CFRs and characteristics of the CFR T cells
Three distinct CFRs were generated using different extracellular regions of human FcγR subtypes, namely 16s3, 32-8a, and 64-8a, respectively (Fig. 1a). The expression levels of CFRs and GFP of the transduced T cells (CFR T cells) were determined by flow cytometry on Day 5 post-infection (Fig. 1b, c). The expression rates of CD16, CD32, CD64 on the 16s3, 32-8a, 64-8a CFR T cells were 73.98–86.33%, 74.81–85.01%, 70.66–82.99%, respectively. The expression rates of CD16, CD32, CD64 on the T cells transduced with a mock construct lacking extracellular sequences (mock T cells) were 0.012–0.073%, 0–0.036%, 0–0.063%, respectively. The CFR T cells were further characterized for their proliferation potential. The cell proliferation after transduction was monitored by cell number at different days (Day 3, 5, 7, 9, 12, 15, 18, 21). No significant differences in proliferation were observed among various CFR T cells and the mock T cells (Fig. 1d). The exhaustion (PD-1, LAG3, TIGIT, TIM3), activation (CD25, CD69), and subpulations [CD8, CD4, naïve T (Tn), central memory T (Tcm), effector memory T (Tem), terminal effector T (Te)] during culture were also compared. Overall, no significant differences were observed despite variations between different samples (Fig. S1–3).
The CFR T cells exhibited RTX-mediated specific cytotoxicity against CD20+ cell lines
The CD20-expressing cell lines (Raji-wt and K562-CD20+) were used as target cells (Fig. S4a). Both 16s3 and 32-8a CFR T cells exhibited superior cytotoxicity at RTX concentrations ≥ 1 μg/ml, whereas a RTX concentration of ≥ 0.1 μg/ml was optimal for 64-8a CFR T cells (Fig. S4b). Therefore, the 1 μg/ml concentration of RTX was chosen for subsequent investigations.
CD107a, also known as lysosomal-associated membrane protein 1 (LAMP-1), is a marker for T cell degranulation/activation. The CD107a level assay was employed to quantify degranulation and characterize the production of cytotoxic granules from the CFR T cells. Upon co-culturing the CFR T cells with Raji-wt and K562-CD20+ cells in the presence of RTX, a significant increase in the percentages of CD107a expression was observed in the CFR T cells when compared to those in the mock T cells (Fig. 2a–c). Similarly, the specific lysis of Raji-wt and K562-CD20+ cells by the CFR T cells in the presence of RTX was significantly increased (Fig. 2d–f). Furthermore, the cytokine levels of IFN-γ, TNF-α, IL-2 and IL-6 secreted by the CFR T cells increased dramatically when the CFR T cells and CD20+ target cells were co-cultured in the presence of RTX (Fig. 2g, h), with the exception for IL-2 secreted by 32-8a CFR T cells co-cultured with K562-CD20+ cells (Fig. 2h IL-2 panel). For the control groups treated with a non-specific MoAb (herceptin groups) or normal saline (NS), as well as co-cultured with the tumor cells without CD20 expression (Raji-CD20-/- groups and K562-wt groups), no significant cytotoxicity of CFR T cells was observed (Fig. 2b–h). These results indicate the cytotoxicity of the CFR T cells against CD20+ tumor cell lines were mediated by target-specific RTX.
The CFR T cells showed RTX-mediated specific cytotoxicity against CD20+ human primary cells
The CD20+ primary mononuclear cells from peripheral blood (PBMCs) or bone marrow (BMMCs) from 3 patients diagnosed respectively with mantle cell lymphoma (MCL), diffuse large B cell lymphoma (DLBCL) or chronic lymphocytic leukemia (CLL) were collected. The CD20 expression in these primary cells was confirmed in Fig. S4c. Compared to the mock T cells, the CFR T cells in the presence of RTX had higher levels of CD107a expression (Fig. 3a, b), increased primary cell lysis (Fig. 3c, d), and increased release of cytokines (Fig. 3e). When herceptin or normal saline was added, no such increases were observed. These findings further corroborate that the CFR T cells can also confer significant cytotoxicity against CD20+ human primary cells mediated by target-specific RTX.
The CFR T cells showed herceptin-mediated specific cytotoxicity against HER2+ solid tumor cell lines
HER2 expressing solid tumor cell lines (SK-OV-3 and SK-BR-3) were used as target cells (Fig. S5a).When co-cultured with SK-OV-3 cells, the CFR T cells showed significantly increased CD107a production in the presence of herceptin (Fig. 4a, b). The specific lysis of SK-OV-3 and SK-BR-3 by the CFR T cells in the presence of herceptin was substantially elevated in comparison to that by the mock T cells (Fig. 4c, d).
The RTCA system was utilized to estimate the specific lysis of adherent target cells, as indicated by CI values [42, 43]. A lower CI value indicates higher cytotoxicity. Remarkably reduced CI values were observed for SK-OV-3 co-cultured with 16s3 and 64-8a CFR T cells (E:T = 2:1) in the presence of herceptin (Fig. 4e, g). However, no significant reduction in CI values was observed for SK-OV-3 co-cultured with 32-8a CFR T cells mediated by herceptin (Fig. 4e, g).
To confirm this target-specific lysis, we examined another cell line SK-BR-3. Similarly, significant reductions in CI values of SK-BR-3 were found when co-cultured with 16s3 and 64-8a CFR T cells in the presence of herceptin (Fig. 4f, h). Additionally, the CI values of SK-BR-3 cells also decreased significantly when co-cultured with 32-8a CFR T cells in the presence of herceptin at 48h (Fig. 4f, h).
Cytokine levels of IFN-g, TNF-α, IL-2, and IL-6 were determined after co-culture of CFR T cells with SK-OV-3 or SK-BR-3 cells. When co-cultured with SK-OV-3 in the presence of herceptin, the levels of TNF-α and IL-6 significantly increased with 16s3, 32-8a and 64-8a CFR T cells; the increase of IFN-g levels was observed with 16s3 and 64-8a CFR T cells. However, the IFN-g released by the 32-8a CFR T cells and the IL-2 released by 16s3, 32-8a and 64-8a CFR T cells were not significantly increased (Fig. 4i).
When co-cultured with SK-BR-3 in the presence of herceptin, elevated levels of IFN-g and TNF-α were observed with 16s3 and 64-8a CFR T cells; a significant increase in IL-2 levels was observed only with 64-8a CFR T cells; the levels of the above four cytokines were not significantly increased in the 32-8a CFR T cell group (Fig. 4j).
For these solid tumor cell lines, no significant cytotoxicity of the CFR T cells was observed when RTX or normal saline was added (Fig.4a–e, i, j), proving again that the specificity of the CFR T cytotoxicity was directed by the specific MoAb.
The CFR T cells confer cytotoxicity against the herceptin-resistant tumor cell line HCC1954
Due to a PI3K mutation, HER2+ HCC1954 cell line exhibits inherent resistance to herceptin [47, 48].This resistance to herceptin was independently confirmed (Fig. S5b, c). Compared with HER2+ HCC1954, breast cancer cell line MCF-7 has no HER2 expression (Fig. S5a), and was used as a negative target cell control. After CFR T cells co-cultured with HCC1954 cells, the CD107a expression in the presence of herceptin increased significantly (Fig. 5a–b). The specific killing of HCC1954 by the CFR T cells in the presence of herceptin was also significantly increased in comparison to that by the mock T cells (Fig. 5c).
The cytokines (IFN-γ, TNF-α, and IL-6) secreted by both 16s3 and 64-8a CFR T cells in the presence of herceptin were significantly elevated (Fig. 5d). However, the IL-2 release by 16s3 and 64-8a CFR T cells in the presence of herceptin was not significantly increased. None of the four cytokines (IFN-γ, TNF-α, IL-2 and IL-6) were significantly increased with 32-8a CFR T cells in the presence of herceptin.
When RTX or normal saline was added to the co-culture, no significant cytotoxicity of the CFR T cells against the HER2+ HCC1954 was observed (Fig. 5a–d). When CFR T cells were co-cultured with the negative control MCF-7 cells in the presence of herceptin, no significant cytotoxicity was observed (Fig. 5a–d).
These results suggest that the 16s3 and 64-8a CFR T cells can confer specific cytotoxicity against HER2+ target cells which were resistant to herceptin alone.
The impact of free human IgG on the cytotoxicity of the CFR T cells
To investigate the potential impact of free human IgG (hIgG) in the plasma on the CFR T cells, we evaluated the effect of hIgG on the cytotoxicity of the CFR T cells at both a physiological concentration (10 g/L) and a supraphysiological concentration (30 g/L).
In the presence of 10 g/L hIgG, there was no significant effect on the cytotoxicity of 16s3 and 32-8a CFR T cells (Fig. 6a, c). However, the cytotoxicity of 64-8a CFR T cells was partially but significantly inhibited compared to no hIgG addition (Fig. 6a, c). When exposed to 30 g/L hIgG, both 16s3 and 32-8a CFR T cells still exhibited significant tumor cell lysis (Fig. 6b, d). However, target cell lysis by 64-8a CFR T cells was completely inhibited when compared to the mock T cells (Fig. 6b, d).
Overall, the physiological dose (10 g/L) and supraphysiological dose (30 g/L) of free hIgG have no significant impact on 16s3 and 32-8a CFR T cells but significant impact on the cytotoxicity of 64-8a CFR T cells.
16s3 CFR T cells exhibited superior long-term cytotoxicity to 32-8a CFR T cells in vitro
In the presence of herceptin, 16s3 CFR T cells exhibited significantly higher levels of CD107a, more target cell lysis and cytokine secretion than those of 32-8a CFR T cells (Fig. S6). No significant differences in CD20+ target cell lysis was observed between 16s3 CFR T cells and 32-8a CFR T cells in the presence of RTX (data not shown).
To further investigate the long-term cytotoxicity of 16s3 and 32-8a CFR T cells mediated by RTX, we implemented a repetitive antigen stimulation experiment (Fig. S7). To exclude the confounding effects of cytokine alterations, groups with or without cytokines were included. Half of the supernatant (500 μl) from the preceding round of co-culture was used to provide cytokines.
The long-term target lysis by 16s3 CFR T cells was significantly higher when compared to that of 32-8a CFR T cells at Round (R)3-Day 5, irrespective of the addition of cytokines (Fig. 7a, d). The peak value of the median fluorescence intensity (MFI) of 32-8a CFR T cells at R1-72 hours was significantly higher than that of 16s3 CFR T cells in the groups without cytokines, potentially contributing to the enhanced target cell lysis observed at R2-72 hours by 32-8a CFR T cells (Fig.7a). However, the difference of target lysis between 16s3 and 32-8a CFR T cells was not found in the cytokine addition groups at R2-72 hours (Fig.7d). Notably, the CFR-MFI of 16s3 CFR T cells exhibited a significant increase from R2-72 hours to R3-72 hours, whereas no such trend was observed in 32-8a CFR T cells (Fig. 7c, f). Furthermore, the CFR-MFI of 16s3 CFR T cells was significantly higher than that of 32-8a CFR T cells at R3-Day 5 (Fig. 7b, e), which may partially account for the enhanced long-term cytotoxicity observed in the former.
Our findings suggest that 16s3 CFR T cells exhibited superior long-term anti-tumor potential in the presence of RTX in vitro. Therefore, 16s3 CFR T cells were selected to investigate the tumor clearance in mouse tumor transplantation models.
Specific anti-tumor effect of 16s3 CFR T cells mediated by RTX in vivo
The schematic diagram of the experiments in the human lymphoma mouse model was presented in Fig. 8a. The median survival time of the 16s3 CFR T-RTX group (n = 7) was 242 days, which was significantly longer than those of the control groups [P < 0.001; mock T-RTX, 39 days (n = 7); PBS-RTX, 39 days (n = 7); Raji-CD20-/-, 30 days (n = 7); 16s3 CFR T-herceptin, 33 days (n = 6); and mock T-herceptin, 32.5 days (n = 6)] (Fig. 8b). Among the five control groups, the survival time showed no significant difference (P > 0.05). 16s3 CFR T cells demonstrated a significant RTX-mediated anti-tumor effect in the CD20+ human lymphoma xenograft model.
Specific anti-tumor effect of 16s3 CFR T cells mediated by herceptin in vivo
The schematic diagram of the experiments in the human ovarian cancer mouse model was presented in Fig. 8c. On Day 25, euthanasia was performed, and tumor masses were resected and examined (Fig. 8d). The average tumor weight of the 16s3 CFR T-herceptin group (n = 5) was 0.01 ± 0.03 g (95%CI: 0–0.05), which was significantly decreased when compared to those of the control groups [P < 0.01; mock T-herceptin, 0.17 ± 0.05 g (95%CI: 0.11–0.23) (n = 5); 16s3 CFR T-RTX, 0.33 ± 0.10 g (95%CI: 0.21–0.45) (n = 5); 16s3 CFR T-NS, 0.30 ± 0.10 g (95%CI: 0.17–0.43) (n = 5)] (Fig. 8e). Similarly, a significant reduction of tumor volume was observed in the 16s3 CFR T-herceptin group (Fig. 8f) (P < 0.001). It was noticed that the tumor weight and volume in the mock T-herceptin group was significantly decreased than those of 16s3 CFR T-RTX and 16s3 CFR T-NS groups (P < 0.05). Whereas no significant differences in tumor weight and volume were noted between the 16s3 CFR T-RTX and the 16s3 CFR T-NS groups (P > 0.05). 16s3 CFR T cells in the presence of herceptin demonstrated specific in vivo anti-tumor effect.