Neoantigens arise from somatic mutation caused by carcinogen exposure or other genomic alterations and therefore are truly cancer-specific. CD8+ T cells harbour specific T cell receptors (TCRs) that can recognize neoantigens expressed by cancer cells as small peptides bound to major histocompatibility complex (MHC) class I molecules and they have the potential to drive anti-tumor T cell responses. Several studies have demonstrated a correlation between the benefits of checkpoint blockade immunotherapy and the tumor mutational burden in patients with melanoma and non-small cell lung cancer (NSCLC), and showed that patients with tumors enriched for clonal neoantigens have increased sensitivity to anti-PD-1/anti-CTLA-4 immunotherapy (1-4). As a result, neoantigens are currently considered promising targets for personalized cancer immunotherapy. Methylcholanthrene (MCA)-induced sarcoma is a type of tumor which is susceptible to treatment with anti-CTLA-4 and/or anti-PD-1. In order to identify therapy-relevant neoantigens and to measure the effects of checkpoint blockade on neoantigen-specific T cells, high-dimensional mass cytometry in combination with combinatorial MHC-tetramer multiplexing was employed. CD8+ T cell responses to a broad range of predicted epitope candidates were screened across multiple tissues from tumor-bearing mice.
Figure 1. In MCA-induced sarcoma, two immunodominant neoantigens, mLama4 and mAlg8, were identified following treatment with anti-CTLA-4 and/or anti-PD-1, which led to tumor rejection.
The detection of neoantigen-specific CD8+ T cells is challenging due to the small pool of tumor antigen-specific T cells. In this study, among 81 candidate antigens tested, T cells restricted to two major neoantigens, mLama4 and mAlg8, were identified in tumors, spleens and lymph nodes of tumor-bearing mice. Anti-CTLA-4 treatment induced increased proliferation of neoantigen-specific T cells selectively in the tumor but not in peripheral lymphoid organs. Neoantigen-specific TILs, which had stronger hallmarks of exhaustion compared with their lymphoid organs counterparts before treatment, consistently shower marked alteration in expression of markers associated with dysfunction (PD-1, Lag-3 and CD160), activation (CD25, GITR and CD38), co-stimulation and development (CD27 and CD127) upon treatment. These results demonstrate the utility of high-dimensional mass cytometry for identifying tumor-specific T cells and measuring the biological activity of checkpoint blockade in a preclinical setting.
Using whole exome and RNA sequencing, tumor neoantigens predicted to bind to MHC-I were identified. Mass cytometry coupled with a 28-marker staining panel was used to phenotypically profile immune cells across tumors, spleens and lymph nodes from tumor-bearing mice. To screen for antigen-specific CD8+ T cells, a highly-multiplexed combinatorial MHC-tetramer staining using nine different metal-labelled streptavidins was employed (9) and 84 possible combinations were generated to screen for 81 H-2Kb putative neoantigens. Single-cell suspensions from tumors, spleens, draining and non-draining lymph nodes were obtained on day 12 and were probed simultaneously for all 81 potential T cell specificities (fig. 2a).
Consistent with previously published data (6), substantial numbers of CD8+ T cells restricted to two mutant tumor epitopes, mLama4 and mAlg8, were identified in tumor-bearing mice. In addition, CD8+ T cells reacting with tetramers specific for both epitopes were detected in spleens, draining and non-draining lymph nodes from the same group of animals (fig. 2b).
Figure 2. Analysis of neoantigen-specific T cells. (a) Screening for CD8+ T cells targeting 81 potential mutant peptide–MHC complexes by a combinatorial peptide–MHC-tetramer staining approach. (b) Representative example for a triple-coded tetramer staining from dLN to identify antigen-specific CD8+ T cells.
Large numbers of tetramer-positive cells infiltrating the tumors expressed PD-1 and Tim-3, and these markers could only be identified on a very low percentage of their antigen-specific T cell counterparts in the periphery of tumor-bearing animals (fig. 3).
Figure 3. Percentages of mLama-4- and mAlg8-specific CD8+ T cells expressing PD-1 and Tim-3 in tumors,
spleen, dLN and ndLN.
Furthermore, a significant increase in the magnitude of mLama4 and mAlg8 specific T-cells was detected only in tumors but not in the periphery from mice that received anti-CTLA-4 therapy. These results suggest that anti-CTLA-4 treatment induces increased proliferation of neoantigen-specific T cells selectively in the tumor (fig. 4).
Figure 4. Frequencies of mLama4 and mAlg8-specific CD8+ T cells from tumors, spleen, dLN and ndLN of tumor-bearing mice treated with anti-CTLA-4 or control (isotype).
To measure whether anti-CTLA-4 treatment alters the phenotypic profile of mLama4- and mAlg8-specific T cells, samples were stained using a panel consisting of 28 surface markers and analysed by applying the t-distributed stochastic neighbour embedding (t-SNE) algorithm (7-9). t-SNE reduces the high dimensionality of the mass cytometry data into two dimensions and separates distinct cellular subsets while retaining the overall cellular relationships. Before the treatment mLama4- and mAlg8-specific T cells were unequally distributed across clusters, primarily C1-C6 (fig. 5a). Anti-CTLA-4 treatment induced substantial phenotypic alteration of mLama4- and mAlg8-specific T cells observed as a remarkable shift in their cluster positions in the two-dimensional t-SNE plot, from clusters C1–C5 towards clusters C7-C10 (fig. 5b). These data showed that neoantigen specific T cells constituted a heterogeneous cell population and anti-CTLA-4 treatment induced substantial phenotypic changes in these tumor-specific T cells.
Figure 5. Heterogeneity of tumor-specific TIL phenotypes changes following treatment. mLama4- and mAlg8-specific T cells from untreated tumor-bearing mice can be identified within six clusters (C1–C6) segregated by t-SNE (a). Anti-CTLA-4 treatment shifts mLama4- and mAlg8-specific TILs from clusters C1–C5 towards clusters C7-C10 on the t-SNE plot (b).
The phenotypic profiles of mLama4 and mAlg8 reactive T cells were then presented as heat plots to visualize the effect of treatment on each individual marker. CTLA-4 blockade significantly altered expression of markers associated with T cell dysfunction (PD-1, Lag-3 and CD160), activation (CD25, GITR and CD38), as well as co-stimulation and development (CD27 and CD127) on neoantigen-specific T cells infiltrating the tumors of treated animals. In the peripheral lymphoid organs, CD8+ T cells specific for either antigen, unlike neoantigen-specific TILs, were not affected by anti-CTLA-4 immunotherapy, as no differences were observed in the expression of any of the marker molecules assessed. Moreover, no significant differences in the phenotypes of tetramer-negative cells were found within the tumors in response to anti-CTLA-4 treatment (fig. 6). These results are a proof-of-concept for the mass screening of predicted neoantigens and the detection of tumor-specific T cells across multiple tissues. In addition, deep insights were obtained from the phenotypic profiles of neoantigen-specific T cells and their responsiveness to checkpoint blockade therapy.
This strategy may help to measure the biological activity of immunotherapy in preclinical or early clinical settings, to compare the potency of different combination treatments, and to define biomarkers predicting the clinical outcomes of cancer patients.
Figure 6. CTLA-4 blocking affects phenotypes of neoantigen-specific TILs. Heat plot showing the relative expression of each marker molecule assessed for mAlg8, mLama4-specific CD8+ T cells and bulk (tetramer negative) CD8+ T cells in the tumors from anti-CTLA-4 and control (isotype) treated tumor-bearing mice.
- Rizvi, N. A. et al. Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer. Science 348, 124–128 (2015).
- Snyder, A. et al. Genetic basis for clinical response to CTLA-4 blockade in melanoma. N. Engl. J. Med. 371, 2189–2199 (2014).
- Van Allen, E. M. et al. Genomic correlates of response to CTLA-4 blockade in metastatic melanoma. Science 350, 207–211 (2015).
- McGranahan, N. et al. Clonal neoantigens elicit T cell immunoreactivity and sensitivity to immune checkpoint blockade. Science 351, 1463–1469 (2016).
- Newell, E. W. et al. Combinatorial tetramer staining and mass cytometry analysis facilitate T-cell epitope mapping and characterization. Nat. Biotechnol. 31, 623–629 (2013).
- Gubin, M. M. et al. Checkpoint blockade cancer immunotherapy targets tumor-specific mutant antigens. Nature 515, 577 (2014).
- Amir, E. D. et al. viSNE enables visualization of high dimensional single-cell data and reveals phenotypic heterogeneity of leukemia. Nat. Biotechnol. 31, 545–552 (2013).
- Becher, B. et al. High-dimensional analysis of the murine myeloid cell system. Nat. Immunol. 15, 1181–1189 (2014).
- van der Maaten, L. & Hinton, G. Visualizing data using t-SNE. J. Machine Learn. Res 9, 2579–2605 (2008).