Targeting CD276 with Adapter-CAR T-cells provides a novel therapeutic strategy in small cell lung cancer and prevents CD276-dependent fratricide

Abstract

Background: Survival rates in Small Cell Lung Cancer (SCLC) remain dismal, posing a huge medical need for novel therapies. T-cells, engineered to express chimeric antigen receptors (CAR-T) have demonstrated clinical activity against a variety of haematological malignancies. Yet, efficacy against solid tumour entities remains limited.

Methods: In this study, we investigated the expression of CD276 (B7-H3), an immune checkpoint molecule and promising target antigen for CAR-T therapy in SCLC, at the RNA and protein level. We further developed novel Fab-based adapter molecules (AM) targeting CD276 and optimized our previously established modular Adapter CAR-T (AdCAR-T) platform as well as AM dosing schemes.

Results: CD276 is broadly expressed across SCLC subtypes, representing a promising target for CAR-T therapy. We describe that T-cell activation and CAR-signalling induces CD276-expression on CAR-T, resulting in CD276-dependent fratricide, limiting anti-CD276-CAR-T expansion and activity. The AdCAR-T platform allows CAR-T expansion in absence of CD276 targeting. Novel CD276 targeted AMs demonstrate potent in vitro and in vivo activity against SCLC. Intermittent AM-dosing allows functional persistence of AdCAR-T in vivo in contrast to CD276-targeted conventional CAR-T. AdCAR-T in vivo expansion and activity is further promoted by introducing activation-induced, AM remote controlled, IL-18 secretion into the AdCAR-T design.

Conclusion: We identified CD276 as a promising target antigen, uniformly expressed in SCLC and demonstrate the therapeutic potential of novel anti-CD276 Fab-based AM in combination with optimized, IL-18 armoured AdCAR-T.

Keywords: AdCAR-T; CD276; Fab-based adapter molecule; Fratricide; Immunotherapy; In vivo solid tumor targeting; SCLC.

Fig. 1

CD276-expression is high on SCLC primary patient samples. (A) Analysis of 79 primary patient samples showed a higher gene expression of CD276 compared to DLL3. n = 79. ****p < 0.0001 (unpaired t-test). (B) Based on the mRNA analysis of cell lines derived from tumor patient samples, CD276 is broadly expressed on all four subtypes of SCLC. CD276-expression is significantly higher than DLL3-expression in the subtypes ASCL1 and YAP1. n as shown. ***p < 0.001. ****p < 0.0001 (Two-way ANOVA). (C) CD276-expression on four different SCLC cell lines was analysed via flow cytometry and showed high expression on all. (D) IHC staining of CD276 on a TMA of 39 primary patient samples was evaluated. Exemplary pictures show low expression (left; score 1), medium expression (middle; score 2) and high expression (right; score 3). Scale bar represents 50 μm. Positive cell to negative cell ratio was calculated from tumor regions (exemplarily highlighted by square) and showed (E) high expression in 71.8% of the samples. (F) H-score was calculated by multiplying the percentage of positive cells with the score

Fig. 2

Upregulation of CD276 on activated CAR-T leads to decrease in cell number and fratricide in vitro. (A) Schematic overview of the AM and AdCAR system, as well as the d CAR. (B) dCAR-T culture at day six results in lower cell numbers compared to untransduced activated T-cells and AdCAR-T. n=6. ****p<0.0001 (One-way ANOVA). Error bars indicate SD between donors. (C) After six days of co-culture with SHP-77, dCAR-T show significantly less CD276-positive cells compared to AdCAR-T. n=3. *p<0.05 (paired t-test). Error bars denote SD between donors. (D) CD276 is mainly expressed on the activated CD4-positive fraction of AdCAR-T during in vitro culture and the fractions of CD4/CD8 show comparable CD276-expression to untransduced control T-cells. n=3. ***p<0.0005 (Two-way ANOVA). (E) During expansion of CD19-CAR-T and AdCAR-T ex vivo, the initial high expression of CD276 decreases over time. n=3. Error bars denote SD between donors. (F) CD276-expression on CD19CAR-T and AdCAR-T can be re-induced by stimulation with CD19-expressing target cells every 24 h. n=4. Error bars denote SD between donors. (G) In comparison to AdCAR-T only, AdCAR-T incubated with CD276-directed Fab or full mAb show and increased fraction of dead cells (7AAD+) as well as (H) a higher frequency of activated (CD69+/CD25+) cells. n=3. Error bars denote SD between donors. *p<0.05. **p<0.01. ***p<0.0005. ****p<0.0001 (Two-way ANOVA).

Fig. 3

Novel biotinylated anti-CD276 AM mediates effective tumor cell elimination through AdCAR-T in vitro. (A) AM candidates were generated of parental clones through a naïve human phage display library screening. Through affinity maturation and exclusion of candidates with potential glycosylation sites and unpaired cysteines, unique clones were identified. Lead candidates were chosen based on manufacturability and specificity. (B) EC₅₀ of the AM candidates in the AdCAR system calculated by linear regression and (C) the respective 95% confidence limits and R². (D) Five novel AM candidates mediate specific cell lysis of the SCLC cell line SHP-77 by AdCAR-T at different E: T ratios after 48 h of co-incubation (5 nM AM). Negative lysis is a result of the calculation and equals target cell growth and is therefore excluded. n = 3. *p < 0.05 (Two-way ANOVA). Error bars indicate SD between donors. (E) Re-challenge cycles in which AdCAR-T were repeatedly exposed to SHP-77. AdCAR-T showed target cell lysis up to three re-challenge cycles and 192 h of co-culture (2 nM AM). Error bars denote SD between donors. n = 3. *p < 0.05. **p < 0.01 (Two-way ANOVA)

Fig. 4

AdCAR-T can actively control disease progression of SCLC in vivo. (A) Schema of in vivo experiment using 0.5*10^6 SHP-77 tumor cells, engrafted into the lungs, and challenged by 30*10^6 AdCAR-T in combination with the AM46 (100 µg). (B) Representative BLI images of the SHP-77 mouse model on day -3, 0, 7, 21, and 35. Color scale represents radiance from 7x103 to 9x104 photons/sec. Mice were sacrificed when reaching termination criteria approx. at a total flux [p/s] of 1010. (C) Representative BLI kinetics of tumor signal of the mice receiving the different treatment options (tumor only, AdCAR-T only, dCAR-T and AdCAR-T+AM) of the respective imaging days. **p<0.01 (Mixed-effects Model with Geisser-Greenhous correction). (D) Whole mouse blood of AdCAR-T only, dCAR-T and AdCAR-T without AM46, analysed on day 14 and 35 for the frequency of human CD45-positive cells. Dead and murine cells were excluded prior. n=5. SD between mice is indicated by error bars (paired t-test). CAR-T frequency of mice treated with (E) AdCAR-T+AM46 or (F) dCAR-T in whole mouse blood on day 35, on day 59 and at the End. Re-challenge was done on day 46 with 0.5x106 SHP-77. n=5. SD between mice is indicated by error bars. *p<0.05 (Wilcoxon test).

Fig. 5

Dosing scheme enhances T-cell expansion after re-challenge, but daily dosing shows efficient anti-tumor effect. (A) Mice were ingrafted with SHP-77 cells and three days later infused with AdCAR-T. Re-challenge with 1 × 10⁶ SHP-77 at day 38. AM46 dosing either daily (Cond I), four days on / three days off (Cond II), or Two days on, two days off (Cond III). (B) BLI kinetics of the mouse groups receiving AdCAR-T treatment plus AM46 in different dosing schemes, and a re-challenge with SHP-77 on day 38. Whole mouse blood was analysed for the frequency of human CD45-positive cells on day 3, 17, 29, and at the End of the treatment groups: (C) dosing regimen I: AdCAR-T with AM46 daily administration; (D) dosing regimen II: 4 days AM46 administration, followed by 3 day break; (E) dosing regimen III: 2 day AM46 administration, followed by 2 day break and (F) condition IV: dCAR-T. n = 5. ***p < 0.0005. ****p < 0.0001 (Mixed-effects model). T-cell phenotypes of differentiation subsets in (G) dosing regimen I, (H) dosing regimen II, (I) dosing regimen III or (J) condition IV: dCAR-T were analysed. n = 5. **p < 0.01 (Two-way ANOVA)

Fig. 6

AdCAR-T loaded with IL-18 enhance T_CM differentiation, cell expansion, and anti-tumor effect. (A) Schematic overview of the TRUCK AdCAR-T construct. Binding of the AdCAR-T to its target mediated by AM46 leads to activation of NFAT. NFAT induces the secretion of IL-18 by the cell. (B) IL-18 secretion of AdCAR_IL-18-T upon co-incubation with H-446. n=3. SD is indicated by Error bars. ****p<0.0001 (Two-way ANOVA). (C) Specific lysis of AdCAR_IL-18-T co-incubated with SHP-77 in vitro and re-challenged every 48 h, as indicated by arrows. n=3. SD is denoted by error bars (Two-way ANOVA). (D) Representative BLI mouse images of the groups receiving no CAR-T, AdCAR-T or AdCAR_IL-18-T plus daily AM46. Images taken on day 0, 6, 32, and on day 42 after re-challenge with 0.5x106 SHP-77. (E) BLI kinetics of mice treated with either AdCAR-T or AdCAR_IL-18-T with daily AM46. Re-challenge was performed at day 38. (F) BLI analysis of the lungs of mice receiving either AdCAR-T or AdCAR_IL-18-T with daily dosing of AM46 before and after re-challenge with SHP-77. n=5. *p<0.05. **p<0.01 (Two-way ANOVA). (G) Proportion of CD4-positive T-cells and CD8-positive T-cells of all CD45+ T-cells of the mice treated with AdCAR-T and AdCAR_IL-18-T. Depicted on day 3, day 29 and the End. n=5. SD between mice is indicated by error bars. *p<0.05 (Two-way ANOVA). Error bars indicate SD between mice. (H) Whole mouse blood of the AdCAR-T and AdCAR_IL-18-T groups were analysed for the frequency of human CD45-positive cells on day 3, 29, and at the End, showing higher frequencies for AdCAR_IL-18-t at d29 and at the End. n=5. **p<0.01 (Two-way ANOVA). (I) Flow cytometry analysis of T-cell subsets of the AdCAR_IL-18-T treatment group and frequency of T-cells on day 3, 17, 29, and at the End. n=5. *p<0.05 (Two-way-ANOVA).