Membrane-bound IL-7 immobilized by the CD8 transmembrane region improves efficacy of CD19 CAR-T cell therapy

Abstract

Enhancing the efficacy of CD19 CAR-T cell therapy can significantly improve patient outcomes by reducing relapse rates in CD19 + B cell malignancies. Exogenous or transgenic cytokines are often used to boost the expansion and durability of CAR-T cells but pose risks of severe toxicities. A promising approach to address these limitations is to immobilize cytokines on the surface of CAR-T cells using transmembrane (TM) anchor domains. Given IL-7 can enhance T-cell proliferation and antitumor activity, our study developed membrane-bound IL-7 constructs using different TM anchor domains (CD8, CD28 and B7-1). We primarily found that the CD8 TM provided superior anchoring for IL-7 compared to CD28 and B7-1. Moreover, the IL-7 construct with a CD8 TM (IL7/CD8) enhanced naïve T cell proliferation and effector functions, and improved the in vitro and in vivo antitumor activity of CD19 CAR-T cells. Importantly, although IL7/CD8 could promote T-cell proliferation, it did not sustain long-term autonomous expansion, which could ensure the safety of CD19 CAR-T cells expressing IL7/CD8 in clinical applications. Collectively, the IL7/CD8 construct represents a promising strategy for enhancing the therapeutic potential of CD19 CAR-T cell therapy.

To the editor,

Cytokines have the ability to improve the proliferative and antitumor functions of CD19 CAR-T cells [1, 2]. However, there have been cases of toxicities resulting from uncontrolled systemic administration and release of cytokines from genetically modified T cells [3,4,5]. A previous conference abstract reported that membrane-bound IL-7 could enhance CAR-T proliferation without the need for additional cytokine support [6]. However, previous research reported that different transmembrane domains can vary in their ability to anchor cytokines to the cell membrane [7]. Our study attempted to develop membrane-bound IL-7 using different transmembrane (TM) anchor domains to enhance proliferation and anti-tumor activity of CD19 CAR-T cells and meanwhile minimize unexpected systemic toxicity.

To construct membrane-bound IL-7, we tested three TM anchors: CD8, B7-1, and CD28 (Fig. 1A and Table S1). The cell surface expression levels of IL-7 constructs with the CD8 TM and B7-1 TM were similar and significantly higher than those of the IL-7 construct with the CD28 TM (Fig. 1B and C). The culture supernatant of T cells expressing the CD8 TM construct had significantly lower soluble IL-7 compared to the culture supernatant of T cells expressing the B7-1 or CD28 TM constructs (Fig. 1D). To further evaluate the cell surface expression levels of different membrane-bound IL-7 constructs in CD19 CAR-T cells, we generated multi-cistronic lentiviral vectors to express either a conventional anti-CD19 targeting CAR with 4-1BB and CD3ζ signaling domains (CD19 CAR) or CD19 CAR also expressing membrane-bound IL-7 following a ribosome skipping 2 A element (Fig. 1E). As expected, we found that the cell surface expression levels of IL-7 constructs with the CD8 TM and B7-1 TM were comparable and markedly exceeded those of the IL-7 construct with the CD28 TM and yet expression levels of CD19 CAR were similar in different constructs (Fig. 1F-H). Meanwhile, the culture supernatant of CD19 CAR-T cells expressing the CD8 TM construct had significantly lower soluble IL-7 compared to the culture supernatant of T cells expressing the B7-1 or CD28 TM constructs (Fig. 1I). In summary, due to its relatively high membrane expression and relatively low shedding, the IL-7 construct with a CD8 TM (IL7/CD8) was selected for further study.

Fig. 1

Construct, optimization and effect function of membrane-bound IL-7 A, Schematic representations of membrane-bound IL-7 using different TM anchor domains (CD8, CD28, and B7-1). B-C, Representative flow cytometric results showing membrane-bound IL-7 with different TM anchor domains on the cell surface (B) and corresponding quantitative analysis (C). D, IL-7 concentrations in the supernatant of T cells expressing membrane-bound IL-7 with different TM anchor domains, measured by ELISA. E, Schematic representations of CD19 CAR and membrane-bound IL-7 with different TM anchor domains. F-H, Representative flow cytometric results showing expression level of CD19 CAR and membrane-bound IL-7 with different TM anchor domains on the surface of T cells (F) and corresponding quantitative analysis of CD19 CAR expression level (G), and membrane-bound IL-7 expression level (H). I, IL-7 concentrations in the supernatant of CD19 CAR-T cells expressing membrane-bound IL-7 with different TM anchor domains, measured by ELISA J-L, IFN-γ (J), TNF-α (K), and IL-2 (L) secreted by T cells 24 h after coculture with Raji tumor cells, determined by ELISA. M-O, IFN-γ (M), TNF-α (N), and IL-2 (O) secreted by T cells 24 h after coculture with Nalm6 tumor cells, determined by ELISA. P-Q, Lysis of Raji cells (P) and Nalm6 cells (Q) by T cells. Data are representative of at least three independent experiments with more than three different donors. Mean values from each group are plotted. Error bars represent SEM (* p < 0.05, ** p < 0.01, *** p < 0.001, analyzed by one-way ANOVA with Bonferroni posttest)

Greater amounts of IFN-γ, IL-2, and TNF-α were produced by IL7/CD8&CD19 CAR-T cells compared to CD19 CAR-T cells when cocultured with Raji and Nalm6 cells (Fig. 1J-O). Compared with CD19 CAR-T cells, IL7/CD8&CD19 CAR-T cells exhibited more potent cytotoxicity activity against Raji and Nalm6 cells (Fig. 1P and Q). Collectively, IL7/CD8 enhances in vitro cytokine production and cytotoxicity of CD19 CAR-T cells.

After infusion into patients, CAR-T cells lack the support of anti-CD3 and anti-CD28 antibodies and IL-2, making it essential to evaluate the effect of IL7/CD8 on CD19 CAR-T cells following withdrawal of these antibodies and IL-2. Following a 4-day withdrawal, we found that the viability, absolute number, and proliferative capacity of CD19 CAR-T cells expressing IL7/CD8 were higher than those of CD19 CAR-T cells (Fig. 2A-E). Meanwhile, we found that IL7/CD8&CD19 CAR-T cells retained a significantly higher naïve phenotype and lower central memory phenotype compared with CD19 CAR-T cells (Fig. 2F and G). To evaluate whether IL7/CD8 could promote antigen-independent long-term expansion, we cultured CD19 CAR-T cells and IL7/CD8&CD19 CAR-T cells without the supplementation of anti-CD3 and anti-CD28 antibodies and IL-2. Although IL7/CD8&CD19 CAR-T cells persisted significantly longer than CD19 CAR-T cells in vitro, the IL7/CD8&CD19 CAR-T population began to contract by day 5 to 10, with all cells dying by day 25 (Fig. 2H). This confirmed that IL7/CD8 did not sustain long-term autonomous T-cell expansion, which is a crucial aspect of safety.

Fig. 2

Phenotypes and in vivo antitumor activity of CD19 CAR-T cells expressing IL7/CD8 A, Schematic representation of T cell culture following a 4-day withdrawal of anti-CD3 and anti-CD28 antibodies and IL-2. B-C, Percentages (B) and counts (C) of living T cells. D-E, Representative flow cytometric plots showing percentages of Ki67 + T cells (D) and corresponding quantitative analysis (E). F-G, Representative flow cytometric plots showing CD45RA and CCR7 expression (F) and corresponding quantitative analysis (G). H, Quantitative analysis of in vitro persistence of T cells following withdrawal of anti-CD3 and anti-CD28 antibodies and IL-2. Live cells were counted using trypan-blue exclusion. The X-axis denotes the number of days after anti-CD3 and anti-CD28 antibodies and IL-2 were withdrawn from the culture media. I, Schematic representation of the Raji xenograft model. NCG mice were intravenously inoculated via tail injection with 1 × 10⁶ Raji cells labeled with Firefly-luciferase and three days later received T cells intravenously. J-K, Representative bioluminescent images (J) and quantitated bioluminescent signals (K) showing tumor growth over time. L, Kaplan–Meier survival curve showing the survival of mice. Significance was determined by the log-rank test. ** p < 0.01, *** p < 0.001. M-N, Quantification of naïve T cells (M) and IL-7 (N) in the peripheral blood of treated mice at days 6 and 16 after T-cell infusions. Significance was determined by a two-tailed t test. * p < 0.05, ns, not significant. Data are representative of at least three independent experiments with more than three different donors. Mean values from each group are plotted. Error bars represent SEM (* p < 0.05, ** p < 0.01, *** p < 0.001, analyzed by two-tailed t test)

After CAR-T cells are infused into patients, they not only lack necessary support from anti-CD3 and anti-CD28 antibodies and IL-2, but could repeatedly encounter tumor cells. Therefore, it is essential to evaluate the impact of IL7/CD8 on CD19 CAR-T cells during consecutive in vitro tumor cell challenges following withdrawal of anti-CD3 and anti-CD28 antibodies and IL-2. IL7/CD8&CD19 CAR-T cells and CD19 CAR-T cells were then continuously stimulated with Raji cells twice at an E: T ratio of 1:2, with a three-day interval between each stimulation (Figure S1A). After each round of stimulation, CAR-T cells were recollected and cocultured with Raji cells at an E: T ratio of 1:1 for 24 h. During the initial round of coculture, both IL7/CD8&CD19 CAR-T cells and CD19 CAR-T cells effectively eliminated Raji cells (Figure S1B). However, after the second stimulation, IL7/CD8&CD19 CAR-T cells exhibited more potent cytotoxic activity against Raji compared to CD19 CAR-T cells (Figure S1B). Analysis of the phenotypes of CAR-T cells at the end of each round of stimulation was also conducted. Compared with CD19 CAR-T cells, IL7/CD8&CD19 CAR-T cells exhibited higher proliferative capacity and a more naïve phenotype after the second round of stimulation (Figure S1C-F). However, expression levels of exhausted makers, such as PD1, TIM3 and Lag3, were similar between CD19 CAR-T cells and IL7/CD8&CD19 CAR-T cells (Figure S1G-H). Overall, IL7/CD8 improves antitumor activity and expansion of CD19 CAR-T cells, but could not affect exhaustion levels during sequential encounters with tumor cells.

To evaluate the in vivo antitumor activity of IL7/CD8&CD19 CAR-T cells and CD19 CAR-T cells, we intravenously delivered a suboptimal dose (1 × 10⁶) of CAR-T cells into NCG mice engrafted with the Raji tumor cell line (Fig. 2I). Tumors grew rapidly in mice treated with mock cells, while both CD19 CAR-T cells and IL7/CD8&CD19 CAR-T cells controlled tumor growth (Fig. 2J and K). However, only IL7/CD8&CD19 CAR-T cells succeeded in controlling tumor growth and eliminating all tumor cells (Fig. 2J and K). Additionally, IL7/CD8&CD19 CAR-T cells induced long-term survival in all five mice, while all five mice receiving CD19 CAR-T cells died before day 54 (Fig. 2L). Consistent with the in vitro data, IL7/CD8&CD19 CAR-T cells were enriched in naïve T cells in the peripheral blood at days 6 and 16 after CAR-T cell infusion, compared to CD19 CAR-T cells (Fig. 2M). Additionally, the concentration of soluble IL7 were similar and relatively low in the serum of mice in both groups at days 6 and 16 after CAR-T cell infusion (Fig. 2N). To assess whether IL7/CD8&CD19 CAR-T cells remained present and active against tumor cells, we rechallenged three mice on day 54 that did not have detectable residual tumors after treatment with IL7/CD8&CD19 CAR-T cells (Figure S2A). These mice were not fully protected against tumor outgrowth, but tumor growth was delayed compared to that in naive age-matched control mice, suggesting residual presence and activity of tumor-reactive T cells (Figure S2B and S2C).

Having observed improved tumor control of CD19 CAR-T with the addition of IL7/CD8, we attempted to evaluate how IL7/CD8 acted on CAR-T cells: in cis, trans or even both? Since phosphorylation of STAT5, a signaling mediator downstream of IL-7R, is a marker of pathway activity, phosphorylation of STAT5 could be used to evaluate how IL7/CD8 acted on CAR-T cells. We found the IL7/CD8 positive T cells had more STAT5 phosphorylation than the IL7/CD8 negative T cells, implying that membrane-bound IL-7 not only acts on its own T cells but can also affect other T cells (Figure S3A and S3B).

In conclusion, we have successfully developed a technology that incorporates CD8 transmembrane region-anchored IL-7 into CD19 CAR-T cells. This technology represents a promising avenue for enhancing the therapeutic potential of CD19 CAR-T cell therapy and warrants further exploration in clinical trials.