Tailored chemokine receptor modification improves homing of adoptive therapy T cells in a spontaneous tumor model

Abstract

In recent years, tumor Adoptive Cell Therapy (ACT), using administration of ex vivo-enhanced T cells from the cancer patient, has become a promising therapeutic strategy. However, efficient homing of the anti-tumoral T cells to the tumor or metastatic site still remains a substantial hurdle. Yet the tumor site itself attracts both tumor-promoting and anti-tumoral immune cell populations through the secretion of chemokines. We attempted to identify these chemokines in a model of spontaneous metastasis, in order to "hijack" their function by expressing matching chemokine receptors on the cytotoxic T cells used in ACT, thus allowing us to enhance the recruitment of these therapeutic cells. Here we show that this enabled the modified T cells to preferentially home into spontaneous lymph node metastases in the TRAMP model, as well as in an inducible tumor model, E.G7-OVA. Due to the improved homing, the modified CD8+ T cells displayed an enhanced in vivo protective effect, as seen by a significant delay in E.G7-OVA tumor growth. These results offer a proof of principle for the tailored application of chemokine receptor modification as a means of improving T cell homing to the target tumor, thus enhancing ACT efficacy. Surprisingly, we also uncover that the formation of the peri-tumoral fibrotic capsule, which has been shown to impede T cell access to tumor, is partially dependent on host T cell presence. This finding, which would be impossible to observe in immunodeficient model studies, highlights possible conflicting roles that T cells may play in a therapeutic context.

Keywords: Adoptive Cell Therapy; T cells; chemokines; fibrosis; tumor.

Figure 1. Ccl2 mRNA is differentially upregulated in TRAMP metastatic lymph nodes

A: Ccl2 mRNA relative expression in TRAMP SV40⁺ or healthy lymph nodes; Mann Whitney test: (***) P<0.001. B: Ccl5 mRNA relative expression in TRAMP SV40⁺ or healthy lymph nodes; Mann Whitney test: (ns) P>0.05. C: Cxcl12 mRNA relative expression in TRAMP SV40⁺ or healthy lymph nodes; Mann Whitney test: (*) P< 0.05. D: Cxcl9 mRNA relative expression in TRAMP SV40⁺ or healthy lymph nodes; Mann Whitney test: (**) P< 0.01. E: Cxcl10 mRNA relative expression in TRAMP SV40⁺ or healthy lymph nodes; Mann Whitney test: (***) P< 0.001. F: Cxcl11 mRNA relative expression in TRAMP SV40⁺ or healthy lymph nodes; Mann Whitney test: (**) P< 0.01.

Figure 2. CCR2 transduction significantly improves homing to the metastatic lymph nodes

A: percentage of GFP-CCR2 (n=13) or GFP-control (n=9) injected cells recovered from TRAMP SV40 TAg⁺ lymph nodes, 24 hours after adoptive transfer, analyzed by FACS; Mann Whitney test: (*) P<0.05. B: percentage of GFP-CCR2 (n=9) or GFP-control (n=10) injected cells, recovered from TRAMP spleens, 24 hours after adoptive transfer, analyzed by FACS; Mann Whitney test: (ns) P>0.05. C: density of GFP-CCR2 (n=4) versus GFP-control (n=6) in TRAMP SV40 TAg⁺ lymph nodes (cells per mm³), 24 hours after adoptive transfer, analyzed by 2-photon microscopy; unpaired t test: (*) P<0.05 (top). D: Representative images of SV40 TAg⁺ lymph nodes acquired by 2-photon microscopy. Lymph node from a recipient of GFP-control CD8⁺ T cells on the left; from a recipient of GFP-CCR2 CD8⁺ T cells on the right. Green signal indicates GFP⁺ cells; collagen stain by SHG shown in white.

Figure 3. Chemokine receptor transduction does not affect killing capacity or activation state of CD8⁺ T cells

A: Lack of effect of chemokine receptor transduction in in vitro T cell-mediated cytotoxicity. Target (5×10⁴ E.G7-OVA) cells were co-incubated with increasing concentrations of OVA-specific TCR/GFP-control or OVA-specific TCR/GFP-CCR2 transduced CD8⁺ T cells as effector cells. The target cells were labeled with Orange CMTMR Dye and cell death was assessed by 7-AAD. The percentage of dead target cells was evaluated by flow cytometry; 2-way repeated measures ANOVA: (ns) (P>0.05). B: Lack of costimulatory effects of chemokine ligand on chemokine receptor-transduced CD8⁺ T cells. CCL2 costimulation assays were performed with GFP-CCR2 or GFP-control-transduced CD8⁺ T cells. Activation markers CD25 and CD69 were measured by flow cytometry 18h after stimulation. 2-way ANOVA: (***) P<0.001; (ns) P>0.05.

Figure 4. Chemokine receptor-modified CD8⁺ T cells lead to significant delay in tumor growth

Top: experimental design scheme. Images and plots: (DAY 16) in vivo Integrisense radiance emission detection of implanted E.G7-OVA tumor, 8 days after the first adoptive transfer of CCR2 (n=8) or GFP-control (n=8) transduced cells or no transfer (n=9); 1-way ANOVA and Tukey's multiple comparison test: (*) P<0.05. (DAY 24) Integrisense radiance emission of harvested E.G7-OVA tumor, 8 days after the second adoptive transfer of CCR2 (n=8) or GFP-control (n=7) transduced cells or no transfer (n=9); 1-way ANOVA and Tukey's multiple comparison test: (*) P<0.05. Data are displayed as mean + SEM of radiance for each group of animals at a given time point.

Figure 5. Enhancement in ACT efficacy is associated with improved homing of chemokine receptor-modified CD8⁺ T cells

A: Density of injected cells in E.G7-OVA implanted tumors (cells per μm²), at the end of therapy; Mann Whitney test, recipients of GFP-CCR2 (n=8) versus GFP-control (n=7) transduced T cells: (**) P<0.01 B: Representative images of immunohistochemical analysis: Brown staining/red arrows: GFP⁺ cells (GFP-ctrl⁺ or GFP-CCR2⁺ CD8⁺ T cells).

Figure 6. Host T cell presence inhibits tumor growth but simultaneously favors fibrosis formation

A: PANC02 tumor size, shown as tumor area after 3 weeks of in vivo growth, in the presence (C57BL/6J recipients) or absence (RAG2^−/− recipients) of host T cells; Mann Whitney test: (*) P<0.05. Each dot represents one animal. B: Representative images of Masson's trichrome staining of harvested PANC02 tumor sections. Collagen stain visible in green/cyan color. Boxes show the areas from which regions of interest were obtained to calculate mean collagen density for every sample. Scale bars represent 50 μm. C: collagen density in the presence (C57BL/6J recipients) or absence (RAG2^−/− recipients) of host T cells; Mann Whitney test (**) P<0.01. Each dot represents one animal. D: T cell density distribution across PANC02 tumors grown for 3 weeks in C57BL/6 recipients that are T cell-sufficient (n=5) or in RAG2^−/− recipients in which CD3⁺ cells were injected 3 days prior to analysis (n=8). Y-axis values represent the mean T cell density ± SEM in concentric regions starting from the outer tumor capsule border and moving towards the center of the tumor (examples of the regions shown in Supplemental Figure 7). X-axis values show the distance of the region from the outer tumor capsule border. The vertical dotted line marks the estimated average inner edge of the tumor capsule, derived from 2PM SHG imaging. 95% confidence intervals are also shown. The slope of the RAG2^−/− recipients is not different from a flat line (P>0.05, linear regression). E: Natural logarithmic transform of the T cell distributions, for the region covering the average capsule width (0-250 μm from the outer tumor border). Data points shown as mean ± SEM; 95% confidence intervals are also shown. The two slopes are significantly different (*: P<0.05, linear regression).

Figure 7. Host T cell but not B cell depletion significantly affects peritumoral fibrosis formation

PANC02 tumors were implanted into C57BL/6J recipients, after depletion of B cells (via anti-CD20 administration) or T cells (via anti-CD3 administration). A: Representative images of Masson's trichrome staining of harvested PANC02 tumor sections, after 3 weeks of in vivo growth. Collagen stain visible in green/cyan color. Boxes show the areas from which regions of interest were obtained to calculate mean collagen density for every sample. Scale bars represent 50μm. “T” marks the tumor mass, “P” marks the peritumoral capsule, “C” marks the irrelevant fibrotic area associated with cutaneous structures (skin). B: collagen density in B cell-depleted recipients (and controls) or T cell-depleted recipients (and controls); 1-way ANOVA and Tukey's multiple comparison test: (*) P<0.05. Each dot represents one animal (n=5 per group). C: The efficiency of B cell depletion via anti-CD20 treatment, as measured by FACS analysis of blood samples for B cell marker CD19. D: The efficiency of T cell depletion via anti-CD3 treatment, as measured by FACS analysis of blood samples for T cell markers CD4 and CD8. 1-way ANOVA and Tukey's multiple comparison test (between each experimental group and the matching control): (***) P<0.001.