FAP-targeted CAR-T suppresses MDSCs recruitment to improve the antitumor efficacy of claudin18.2-targeted CAR-T against pancreatic cancer

Abstract

Purpose: The claudin 18.2 (CLDN18.2) antigen is frequently expressed in malignant tumors, including pancreatic ductal adenocarcinoma (PDAC). Although CLDN18.2-targeted CAR-T cells demonstrated some therapeutic efficacy in PDAC patients, further improvement is needed. One of the major obstacles might be the abundant cancer-associated fibroblasts (CAFs) in the PDAC tumor microenvironment (TME). Targeting fibroblast activation protein (FAP), a vital characteristic of CAFs provides a potential way to overcome this obstacle. In this study, we explored the combined antitumor activity of FAP-targeted and CLDN18.2-targeted CAR-T cells against PDAC.

Methods: Novel FAP-targeted CAR-T cells were developed. Sequential treatment of FAP-targeted and CLDN18.2-targeted CAR-T cells as well as the corresponding mechanism were explored in immunocompetent mouse models of PDAC.

Results: The results indicated that the priorly FAP-targeted CAR-T cells infusion could significantly eliminate CAFs and enhance the anti-PDAC efficacy of subsequently CLDN18.2-targeted CAR-T cells in vivo. Interestingly, we observed that FAP-targeted CAR-T cells could suppress the recruitment of myeloid-derived suppressor cells (MDSCs) and promote the survival of CD8⁺ T cells and CAR-T cells in tumor tissue.

Conclusion: In summary, our finding demonstrated that FAP-targeted CAR-T cells could increase the antitumor activities of sequential CAR-T therapy via remodeling TME, at least partially through inhibiting MDSCs recruitment. Sequential infusion of FAP-targeted and CLDN18.2-targeted CAR-T cells might be a feasible approach to enhance the clinical outcome of PDAC.

Fig. 1

Bioinformatics analysis of FAP biological and tumor microenvironment characteristics in pancreatic cancer. a Pan-cancer analysis of FAP expression. b Identification of optimal FAP expression cutoff dividing PDAC cohort. The upper scatter plot shows each cutoff point’s standardized log-rank statistic value. The lower Kaplan-Meier plot shows overall survival for patients divided by the optimal cutoff. c–f Biology and TME characteristics in high-FAP vs. low-FAP groups of PDAC cohort. c Density distribution of tumor purity. d Enrichment score of HALLMARK gene-sets related to high FAP expression. e Volcano map of the chemokines. f Immune suppressive characteristics estimated by IOBR. g IHC staining of FAP protein expression in 9 pancreatic cancer samples

Fig. 2

CAFs impair the antitumor activity of CAR-T cells in vivo. a IHC staining of CAFs in 4T1 and E0771 breast and PANC02 and KPC1199 pancreatic tumor-bearing mice. b-f In vivo experimental of Antigen-Positive E0771 allografts. b Experimental scheme. Antigen-Positive E0771 tumor cells alone or combined with NIH 3T3 fibroblast were in situ inoculated into C57BL/6 mice, treated i.p. injection with CPA, and then given CAR-T cells (i.v.; n = 5 mice per group). c The tumor volume of tumors of each treatment group. d The tumor growth inhibition of each treatment group. e The tumor weight at the end point of the animal experiment. f CAR copy numbers in genomic DNA of residual tumors after therapy were measured by qRT-PCR (TaqMan probe). The images were obtained under original magnification 200×. Scale bars, 100 μm. All data are presented as the mean ± SEM of triplicate experiments. *p < 0.05, **p < 0.01, ***p < 0.001

Fig. 3

Generation of FAP-targeted CAR-T cells and CLDN18.2-targeted CAR-T cells. a Binding specificity of the anti-FAP antibody to mFAP-transfected 3T3 or huFAP-transfected HT-1080. b Affinity measurements of anti-FAP antibody binding to FAP through BiacoreSurface Plasmon resonance (GE Healthcare). The upper shows its binding to mu-FAP and, the lower shows its binding to hu-FAP. c The construction of 8E3-mBBZ and 8E5-mBBZ CAR-T cells. This construct includes an extracellular antigen recognition region (8E5 targeting CLDN18.2 and 8E3 targeting FAP), a hinge, a transmembrane domain, an intracellular region of mouse 41-BB costimulatory molecules, and a mouse CD3-ζ chain. d The transduction efficiency of FAP-mBBZ and CLDN18.2-BBZ CAR-T CAR on splenic T cells derived from C57BL/6 was determined by flow cytometry. UTD cells served as negative controls. e Western blot of CD3-ζ in CAR-T. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as a loading control. f The expression of CLDN18.2 on KPC1199, PANC02, and PANC02-A2. Cells incubated with a mouse anti-mouse IgG antibody as a negative control. g The establishment of mFAP NIH 3T3 mouse fibroblast cells. NIH 3T3 was transduced by pWPT-GFP-mFAP lentivirus and further sorted by flow cytometry. h-i CAR-T cells were incubated with tumor or fibroblast cells at three effector: target (E: T) ratios for 18 h. Cell lysis was tested using a standard nonradioactive cytotoxicity assay. h Cytotoxic activities of CLDN18.2-mBBZ CAR-T cells on CLDN18.2-positive or CLDN18.2-negative tumor cells. i Cytotoxic activities of FAP-mBBZ CAR-T cells on FAP-positive or FAP-negative fibroblast cells

Fig. 4

Sequential therapy of FAP-targeted and CLDN18.2-targeted cells dramatically improved antitumor efficacy against CLDN18.2-positive tumor-bearing mice. a-d In vivo experimental design. C57BL/6 mice were injected s.c. with PANC02-A2 cells and allowed to establish for 10 days. Mice were assigned to four experimental groups. The first and second CAR-T cells were injected on day 10 and day 18, respectively (i.v.; n = 5 mice per group). b The tumor growth inhibition of each treatment group. c The volume of tumors of each treatment group. d The body weight of mice of each treatment group. e-h In vivo experimental design. C57BL/6 mice were injected s.c. with KPC1199 cells and allowed to establish for 12 days. Mice were assigned to four experimental groups. The first and second CAR-T cells were injected on day 12 and day 19, respectively (i.v.; n = 5 mice per group). f The tumor growth inhibition of each treatment group. g The volume of tumors of each treatment group. h The body weight of mice of each treatment group. All data are presented as the mean ± SEM of triplicate experiments. *p < 0.05, **p < 0.01, ***p < 0.001

Fig. 5

Sequential therapy of FAP-targeted and CLDN18.2-targeted CAR-T cells dramatically increase the accumulation of CD8⁺ T and CAR-T cells. a–c Tumor tissues were harvested from the mice at the end of therapy in the PANC02 model. a The representative images of tumor-infiltrating T cells immunostained with anti-CD4 and anti-CD8 from each treatment group. b Histograms show the quantification of T cells in tumor tissues. c CAR copy numbers in genomic DNA of residual tumors from each treatment group were measured by qRT-PCR (TaqMan probe). d–f Tumor tissues were harvested from the mice at the end of therapy in the KPC1199 model. d The representative images of tumor-infiltrating T cells immunostained with anti-CD4 and anti-CD8 from each treatment group. e Histograms show the quantification of T cells in tumor tissues. f CAR copy numbers in genomic DNA of residual tumors from each treatment group were measured by qRT-PCR (TaqMan probe). The images were obtained under original magnification 200×. Scale bars, 100 μm. All data are presented as the mean ± SEM of triplicate experiments. *p < 0.05, **p < 0.01

Fig. 6

FAP-targeted CAR-T cells eliminate CAFs and change tumor microenvironment in vivo. C57BL/6 mice were injected s.c. with PANC02-A2 or KPC1199 cells and treated with CAR-T cells after the tumor was established. To exclude the influence of tumor volume difference by distinct CAR-T cells, tumor tissue was harvested 7 days after treatment. a–c Analysis of the CAFs in tumor tissues of PANC02-A2 and KPC1199 allografts. a The representative images of CAFs immunostained with α-SMA in tumor tissues from each treatment group. b q-PCR measured the mRNA expression of CXCL12 in tumor tissues from each treatment group. c PET image labeling FAP by FAPI in PANC02-A2 allografts 7 days after the treatment of FAP-mBBZ or UTD T cells.  d–f  Analysis of the immune cell in tumor tissues of PANC02-A2 allografts. d The representative flow cytometry plots showing the frequencies of tumor-infiltrating CD8⁺ and CD4⁺ cells in CD3⁺ T of each treatment group. e The representative flow cytometry plots showing the frequencies of tumor-infiltrating CD45⁺ immune cells and MDSCs of each treatment group. f Flow cytometry quantitation of tumor-infiltrating immune cells of each treatment group. g–i Analysis of the immune cell in tumor tissues of KPC1199 allografts. g The representative flow cytometry plots showing the frequencies of tumor-infiltrating CD8⁺ and CD4⁺ cells in CD3⁺ T of each treatment group. h The representative flow cytometry plots showing the frequencies of tumor-infiltrating CD45⁺ immune cells and MDSCs of each treatment group. i Flow cytometry quantitation of tumor-infiltrating immune cells of each treatment group. The images were obtained under original magnification 200×. Scale bars, 100 μm. All data are presented as the mean ± SEM of triplicate experiments. *p < 0.05, **p < 0.01, *** p < 0.001

Fig. 7

Schema: Sequential Therapy enhances the antitumor ability of CAR-T by suppressing the infiltration of MDSCs. Left: CAFs are crucial in TME, forming an immune surveillance barrier and perpetuating tumor-promoting. The CAFs could also promote recruit monocytes from the bone marrow, such as MDSCs, to form a tumor-suppressing microenvironment, suppressing CAR-T cell function. Right: FAP-targeted CAR-T cells eliminate CAFs via specific recognition of FAP on CAFs in the tumor microenvironment. Further, CLDN18.2-targeted CAR-T cells could also increase cytotoxic T cells and inhibit the recruiting of MDSCs. With an improved immune suppressive microenvironment, the antitumor effect of the sequential infusion of CLDN18.2-targeted CAR-T cells were enhanced