Acquired cancer resistance to combination immunotherapy from transcriptional loss of class I HLA

Abstract

Understanding mechanisms of late/acquired cancer immunotherapy resistance is critical to improve outcomes; cellular immunotherapy trials offer a means to probe complex tumor-immune interfaces through defined T cell/antigen interactions. We treated two patients with metastatic Merkel cell carcinoma with autologous Merkel cell polyomavirus specific CD8+ T cells and immune-checkpoint inhibitors. In both cases, dramatic remissions were associated with dense infiltration of activated CD8+s into the regressing tumors. However, late relapses developed at 22 and 18 months, respectively. Here we report single cell RNA sequencing identified dynamic transcriptional suppression of the specific HLA genes presenting the targeted viral epitope in the resistant tumor as a consequence of intense CD8-mediated immunologic pressure; this is distinguished from genetic HLA-loss by its reversibility with drugs. Transcriptional suppression of Class I loci may underlie resistance to other immunotherapies, including checkpoint inhibitors, and have implications for the design of improved immunotherapy treatments.

Fig. 1

Acquired resistance to combination immunotherapy. T cell infusions indicated by dashed lines. a Clinical course. b Immunotherapy treatments. c T cell persistence. CD8+ T cell persistence by MCPyV-sT_83-91 tetramer is on left axis (green line) and frequency of dominant infused clone by CDR3-beta sequencing on right axis (blue line; this clonotype was 85% of the infusion product). d Evaluation of tumor at acquired resistance/late relapse (post-treatment day + 832): Left: Immunohistochemistry demonstrating expression of MHC class I (pan-HLA–ABC) in tumor (T) and epidermis (E) and MCPyV T antigen in tumor. Scale bar indicates approximately 100 micrometers. Right: sequence of the targeted MCPyV-sT_83-91 epitope was unmutated. Discovery patient (2586-4) is shown

Fig. 2

scRNAseq of PBMC identifies an activated CD8+ T cell population at response. Four peripheral blood time points are shown, all from the discovery patient (2586-4): pre-treatment, early post-treatment (day + 27), treatment response (day + 376), and late/acquired resistance (day + 614). a t-Stochastic Neighbor Embedding (tSNE) visualization of clustering of peripheral blood. Peripheral blood mononuclear cells (PBMC; n = 11,021) clustered into populations as indicated. Representative marker genes shown in Supplementary Fig. 7. b, c Enrichment of cluster of activated CD8+ lymphocytes (red cluster, arrow) at response to immunotherapy. Clustering biostatistical analysis described in detail in the Methods, and proportion of CD8+s in each cluster at the various time points compared with Fisher’s exact test. d Heat map of selected significantly differentially expressed genes comparing CD8+ T cells in the red activated cluster (n = 170) to those in the blue effector/EM cluster (n = 429) at response (day + 376). For full list of all 45 differentially expressed genes, see Supplementary Table 3

Fig. 3

Activated CD8+ T cell infiltration into tumor at the time of treatment response. a Multiplex immunohistochemistry showing representative peritumoral and intratumoral CD8+ infiltrates. Arrows indicate CD3⁺CD8+ cells (not all indicated). Dark blue = DAPI (nuclei), Light blue = CD56 (MCC tumor), Red = CD3⁺CD8^–, Green = CD8+, Purple = HLA-DR. Three timepoints are shown from the discovery patient (2586-4): pre-treatment, post-treatment response (day + 349), and late/acquired resistance (day + 832). b Quantification of density of peritumoral and intratumoral CD8 cells. Three representative microscope scan areas were digitally scored for each patient, and significant differences in density determined with student’s T test. c Detection of infused T cell clonotypes in tumor at time of treatment response by sequencing of TCR CDR3-beta

Fig. 4

scRNAseq of tumor biopsies. a–e Discovery patient (2586-4). f, g Validation patient (9245-3). h Both patients. a tSNE of viably frozen cells (n = 7431) from tumor biopsies obtained pre-treatment (blue) and at late relapse/acquired resistance (day + 615; orange). Marked spatial separation of tumor cells from the two timepoints indicates substantial transcriptional change. b Heat map of selected significantly differentially expressed genes (DEGs) in tumor clusters. For full table of DEGs, see Supplementary Data 2. c tSNE of HLA-B expression. d Differential change in scRNAseq expression of HLA-A, -B by tumor cells. HLA (HLA-A, HLA-B) and time point (pre: pre-treatment, rel: acquired resistance/relapse) are indicated below. e qPCR validation and reversibility of HLA-B downregulation on a repeat tumor biopsy at time of acquired resistance, day + 832, graphed on a log2 scale. Tumor cells were cultured ex vivo for 48 h before RNA collection with either vehicle control (–), interferon gamma (1000 IU/mL; IFN) or the hypomethylating agent 5-azacytidine (1 μM; 5-aza) as indicated below. Bars represent mean and error bars represent range of two experimental runs, each performed with triplicate wells. f tSNE of specimens from late relapse in the validation patient (day + 565 post treatment). Viably frozen cells (n = 11267) from PBMC (pink) and tumor biopsy (blue) at late relapse/acquired resistance. Patient had received HLA-A-restricted CD8+ T cells with 18 months response followed by relapse (Supplemental Figures 10 and 11). g tSNE of HLA-A expression. h Differential expression of targeted and non-targeted HLA on relapsed tumor for both patients. Proportion MCC cells expressing gene by scRNAseq are indicated. Data for discovery patient (2586-4) are reproduced from panel 4d for clarity. Please see Methods section for details of scRNAseq biostatistical analysis and determination of DEGs