Tumor Regression Following Engineered Polyomavirus-Specific T Cell Therapy in Immune Checkpoint Inhibitor-Refractory Merkel Cell Carcinoma

Abstract

Although immune check-point inhibitors (CPIs) revolutionized treatment of Merkel cell carcinoma (MCC), patients with CPI-refractory MCC lack effective therapy. More than 80% of MCC express T-antigens encoded by Merkel cell polyomavirus, which is an ideal target for T-cell receptor (TCR)-based immunotherapy. However, MCC often repress HLA expression, requiring additional strategies to reverse the downregulation for allowing T cells to recognize their targets. We identified TCR_MCC1 that recognizes a T-antigen epitope restricted to human leukocyte antigen (HLA)-A*02:01. Seven CPI-refractory metastatic MCC patients received CD4 and CD8 T cells transduced with TCR_MCC1 (T_TCR-MCC1) preceded either by lymphodepleting chemotherapy or an HLA-upregulating regimen (single-fraction radiation therapy (SFRT) or systemic interferon gamma (IFNγ)) with concurrent avelumab. Two patients who received preceding SFRT and IFNγ respectively experienced tumor regression. One experienced regression of 13/14 subcutaneous lesions with 1 'escape' lesion and the other had delayed tumor regression in all lesions after initial progression. Although T_TCR-MCC1 cells with an activated phenotype infiltrated tumors including the 'escape' lesion, all progressing lesions transcriptionally lacked HLA expression. While SFRT/IFNγ did not immediately upregulate tumor HLA expression, a secondary endogenous antigen-specific T cell infiltrate was detected in one of the regressing tumors and associated with HLA upregulation, indicating in situ immune responses have the potential to reverse HLA downregulation. Indeed, supplying a strong co-stimulatory signal via a CD200R-CD28 switch receptor allows T_TCR-MCC1 cells to control HLA-downregulated MCC cells in a xenograft mouse model, upregulating HLA expression. Our results demonstrate the potential of TCR gene therapy for metastatic MCC and propose a next strategy for overcoming epigenetic downregulation of HLA in MCC.

Figure 1.. T cells transduced with the anti-MCPyV TCR identified from healthy donor PBMCs showed similar functional avidity but broader recognition of LTAg_15–23 variants compared to those identified from responding patient-derived tumor-infiltrating T cells.

a Schema of the expanded autologous T cell therapy with HLA-A*02:01-restricted LTAg_15–23-specific CD8+ T cells resulting in MCC regression. b distribution in the infusion product (left) and frequency of the infusion-enriched clonotypes before and after T cell infusion in a regressing tumor biopsy (right). c CD3 expression (left) indicating surface expression of a functional TCR and LTAg_15–23 p/HLA binding (right) by endogenous TCR knock-out Jurkat cells. d IFNγ production by CD8 T cells transduced with indicated TCRs exposed to TAP-deficient T2 B lymphoblastoid cell lines (T2-BLCL) exogenously loaded with indicated concentrations of LTAg_15–23 (left) and mean half-maximal concentrations (EC50) (right) of three biological replicates for each TCR. 1 ng/mL LTAg_15–23 equals 1 nM. e, f Growth kinetics of HLA-A2+ LTAg+ dermal fibroblasts (e) or WaGa cells (f) in the presence of CD8 T cells transduced with indicated TCRs. g IFNγ production by CD8 T cells transduced with indicated TCRs upon coculture with exogenously loaded LTAg_15–23 variants (10 μg/mL). Horizontal line indicates 50% threshold below which variant recognition was impaired. Data from two biological replicates are shown. 50% h, i Same as (d, e) comparing CD8 vs. CD4 T cells transduced with TCR_MCC1.

Figure 2.. forced HLA expression, TCR_MCC1, and CPIs induced tumor regression in 2/7 previously CPI-refractory metastatic MCC patients.

a A swimmer’s plot showing clinical courses of the five enrolled patients. b Images and map of Pt 2 tumors at enrollment. c CT and PET scan images of Pt. 2’s remaining ‘escape’ lesion 105 days after TCR_MCC1 infusion. d-f Pt 6 circulating tumor DNA (ctDNA) measurements (d), CT scan (e) and PET images (f) at indicated timepoints after TCR_MCC1 infusion. Target lesions are indicated with red arrows and a circle.

Figure 3.. Functional T_TCR-MCC1 cells persist in blood and preferentially infiltrate tumors.

a Percent of CD4 and CD8 T_TCR-MCC1 within the infusion products that produced IFNg, TNFa or IL-2 in response to LTAg_15–23. b Circulating T_TCR-MCC1 (WPRE copies/10⁶ cells) (y axis) at indicated timepoints (x axis) for all pts. c Percent of CD4 and CD8 T_TCR-MCC1 that produced IFNg in response to 1μM LTAg_15–23 at indicated timepoints. d CD3/CD8/WPRE (FISH) staining of Pt 2 ‘escape’ lesion. The right panel shows a zoomed-in image of the selected region in the left panel. e Quantification of CD8+ or CD8- CD3+ WPRE+ cells in (d). f Density of TCR_MCC1 cells projected on the UMAP coordinate. g A UMAP plot of T cell populations with cluster annotations. h Frequency of T_TCR-_MCC1 cells in each cluster (y-axis) and frequency of each cluster in T_TCR-_MCC1 cells. i, j Density of cells that express indicated genes projected on the UMAP coordinate. k Cytokine production from expanded tumor-infiltrating T cells in Pt 6 biopsies.

Figure 4.. HLA expression/downregulation is associated with tumor regression/progression.

a IHC of class-I HLA in tumor tissues. b Serum IFNγ level detected by Luminex in Pts 6 and 7 during the systemic IFNγ regimen. c IFNγ downstream gene scores calculated in Pts 6 and PBMC scRNA-seq data up to Day 28 post-T cell infusion. Each data point represents a single time point where scores of all cells were averaged. Dotted lines indicate basal levels at pre-IFNγ time points.

Figure 5.. Activated endogenous T and NK cells expanded in a tumor showing delayed regression.

a, b scRNA-seq cluster annotation of major cell types (a) and T/NK cell sub-clusters (b). c, d cluster frequency change from pre-treatment to Day 118, with (c, d) corresponding to (a, b) respectively. e Effector gene expression in T cells of sub-cluster 1–3 combined at the indicated time points. f TCR clonotypes in the infusion products for a previous endogenous T cell trial that were re-discovered in Pt 6 PBMC and tumors. “Max in PBMC” indicates the maximum frequency across tested PBMC time points. g Frequency of old HLA-A2-targeting infusion clones in T cell sub-clusters. h Distribution of the A2-targeting clones on the UMAP plot. i Density of cells that express indicated genes projected on the UMAP coordinate. j, k Effector (j) and phenotypic marker (k) gene expression in NK cells at the indicated time points.

Figure 6.. A CD200-targeting switch receptor allows TTCR-MCC1 cells control HLA-low MCC cells in mice.

a CD200 expression in MCC. A representative image from Pt 1 is shown. b Design of TCR_MCC1/CD200R-CD28/CD8αβ construct. c In vivo tumor control by T cells with the indicated transgene. 4 mice per arm. d Frequency of human CD4 and CD8 T cells in digested tumor single cell suspension. e HLA expression on tumor (huCD45- CD56+) cells measured by flow cytometry.