Mining exomic sequencing data to identify mutated antigens recognized by adoptively transferred tumor-reactive T cells

Abstract

Substantial regressions of metastatic lesions have been observed in up to 70% of patients with melanoma who received adoptively transferred autologous tumor-infiltrating lymphocytes (TILs) in phase 2 clinical trials. In addition, 40% of patients treated in a recent trial experienced complete regressions of all measurable lesions for at least 5 years following TIL treatment. To evaluate the potential association between the ability of TILs to mediate durable regressions and their ability to recognize potent antigens that presumably include mutated gene products, we developed a new screening approach involving mining whole-exome sequence data to identify mutated proteins expressed in patient tumors. We then synthesized and evaluated candidate mutated T cell epitopes that were identified using a major histocompatibility complex-binding algorithm for recognition by TILs. Using this approach, we identified mutated antigens expressed on autologous tumor cells that were recognized by three bulk TIL lines from three individuals with melanoma that were associated with objective tumor regressions following adoptive transfer. This simplified approach for identifying mutated antigens recognized by T cells avoids the need to generate and laboriously screen cDNA libraries from tumors and may represent a generally applicable method for identifying mutated antigens expressed in a variety of tumor types.

Figure 1. Response of TIL 2098 to candidate epitopes identified from autologous tumors

(a) A screening assay was carried out to evaluate the release of IFN-γ from TIL 2098 in an overnight co-culture with peptide-pulsed T2 cells that were pulsed individually with the top 61 candidate HLA-A*0201 binding peptides (Supplementary Table 1) identified from 2098 mel, with the exception of peptides 9,10,14,23,24,31,33 and 37, which overlapped with peptides 1,4,4,25,12,15,16 and 4, respectively (NT). Peptides 39-62 (Supplementary Table 1) stimulated the release of 100 pg/ml or less of IFN-γ from TIL 2098 and are not depicted in this graph. The autologous 2098 mel stimulate the release of 10,000 pg/ml of IFN-γ from 2098 TIL in this assay. (b-d) T2 cells were pulsed with tittered doses of the indicated mutated or wild type (WT) peptides, and IFN-γ release measured in an overnight co-culture with TIL 2098. (e) Stable transfected of COS7 cells (COS-A2) expressing HLA-A*0201 were transiently transfected with the indicated transcripts and evaluated for their ability to stimulate IFN-γ release from TIL 2098 in an overnight co-culture. (f) 293 cells were transiently transfected with a construct encoding HLA-*0201 as well as the indicated transcripts and evaluated for their ability to stimulate IFN-γ release from TIL 2098 in an overnight co-culture.

Figure 2. Response of TIL 2369 to candidate epitopes identified from autologous tumors

(a) A screening assay was carried out to evaluate the release of IFN-γ from TIL 2369 in an overnight co-culture with peptide-pulsed 293-A1 cells that were pulsed individually with the top 56 candidate HLA-A*0101-binding peptides identified from 2369 mel (Supplementary Table 2), with the exception of peptides 35, 39 and 44, which overlapped with peptides 20, 33 and 43, respectively, and were not tested (NT). Peptides 24-56 (Supplementary Table 2) stimulated the release of less than 100 pg/ml of IFN-γ from TIL 2369 and are not depicted in this graph. The autologous 2369 mel stimulated the release of 20,838 pg/ml of IFN-γ from TIL 2369 in this assay. (b,c) 293-A1 cells were pulsed with the indicated concentrations of peptides in an overnight co-culture with TIL 2098, and IFN-γ release measured. d. 293-A1 cells were transiently transfected with the indicated constructs and evaluated for their ability to stimulate IFN-γ release from TIL 2369 in an overnight co-culture.

Figure 3. Response of TIL 3309 to candidate epitopes identified from autologous tumors

(a) A screening assay was carried out to evaluate the release of IFN-γ from TIL 3309 in an overnight co-culture with COS-A11 cells that were pulsed individually with the top 46 candidate HLA-A*1101-binding peptides identified from 3309 mel (Supplementary Table 5). Peptides 37-46 (Supplementary Table 5) stimulated the release of less than 100 pg/ml of IFN-γ from TIL 3309 and are not depicted in this graph. The autologous 3309 mel stimulated the release of 4,600 pg/ml of IFN-γ from TIL 3309 in this assay. (b,c) COS7 cells stably transduced with a retroviral vector expressing HLA-A*1101 (COS-A11) were pulsed with the indicated concentrations of peptides in an overnight co-culture with TIL 3309, and IFN-γ release measured. d. COS-A11 cells were transiently transfected with the indicated constructs and evaluated for their ability to stimulate IFN-γ release from TIL 3309 in an overnight co-culture.

Figure 4. IFN-γ ELISPOT responses of TIL and PBMC obtained prior to and following autologous TIL transfer

Samples of PBMC prior to adoptive TIL transfer as well as samples of PBMC obtained approximately one month following transfer were evaluated along with samples of infused TIL. A pre-treatment sample was not available for subject 1, and the pre-treatment sample analyzed from subject 2 was obtained approximately one month following an adoptive TIL transfer that was administered in the absence of non-myeloablative chemotherapy one month prior to infusion of the TIL analyzed in (b). TIL 2098 from subject 1 was cultured with HLA-A*0201⁺ COS-7 cells (a), TIL 2369 from subject 2 was cultured with HLA*0101⁺ COS7 cells (b), and TIL 3309 from subject 3 were cultured with HLA*1101⁺ COS-7 cells (c) that were pulsed with candidate nonamer peptides for two hours at 37⁰C. T cells were also cultured with autologous as well as allogeneic tumor cells, and in addition, each of the individual T cell populations were stimulated in parallel cultures with PMA plus ionomycin (PMA/I). The numbers of IFN-γ ELISPOTS per 10⁵ T cells generated in responses to PMA/I were as follows: Subject 1 TIL 2098:7,300, subject 1 post transfer PBMC:4,700, Subject 2 TIL 2369:21,800, Subject 2 pre-transfer PBMC:21,800, Subject 2 post-transfer PBMC:11,000, Subject 3 TIL 3309:7,700, Subject 3 pre-transfer PBMC: 12,500, Subject 3 post-transfer PBMC:26,300. TIL 2098 was plated at 2,000 cells per well, TIL 2369 and 3309 were plated at 1,000 cells per well, and PBMC were plated at 100,000 cells per well, whereas all of the groups stimulated with PMA/I were plated at 1,000 cells per well. The error bars represent the mean + and − the standard error of the mean of three replicate wells per group.