Mutant p53-specific CD8TCR-therapy combined with a CD4TCR prevents relapse of cancer and outgrowth of micrometastases

Abstract

Relapse remains challenging in the treatment of metastatic cancers. More than 50% of human cancers harbor mutant p53 (mp53) as a cancer-specific target. We present the spontaneously metastasizing tumor model Ag104A to advance mp53-specific T cell receptor engineered T cell therapy (TCR-therapy). We identified in Ag104A an autochthonous p53^D256E mutation as neoantigen recognized by a TCR isolated from CD8⁺ T cells (CD8TCR). Cloning of the Ag104A cancer revealed mp53 expression in >99% of cancer cells. Targeting mp53 by CD8TCR-therapy was initially therapeutic, but tumors escaped as cancer cells with reduced or lack of antigen expression. Therefore, we determined whether escape could be prevented by combining the mp53-specific CD8TCR with a CD4⁺ T cell-derived TCR (CD4TCR) recognizing a mutant antigen presented on the stroma of the cancer. No relapse occurred when the mp53-specific CD8TCR was combined with the stroma-recognizing CD4TCR. The combination therapy also prevented the development of macrometastases from cancer cells that had already spread to the lung at the time of TCR-therapy. Macrometastases were only observed after monotherapy. Thus, in a spontaneously metastatic model, tumor relapse and development of macrometastases can be prevented by combining a CD8TCR targeting an autochthonous p53-mutation with a mutation-specific CD4TCR recognizing tumor stroma.

Keywords: Adoptive cell transfer; TCR-therapy; metastatic tumor model; mutant p53; neoantigen.

Figure 1.

Selection of a CD8TCR from an Ag104A-specific CTL line. (a) overview of the generation of an Ag104A-specific CTL line done by Li et. al. (b) specificity of the CTL line YL1 and its derived clones A1 and A2 was determined by a 4 h Cr⁵¹-release assay. (c) the Ag104A-specific clones A1 and A2 express the most frequent TCR α- and β-chains of the CTL line YL1 as detected by 5’-RACE PCR. (d) T cells from the spleen of C3H/HeN mice were retrovirally transduced to express the M2/3-CD8TCR. (e) TCR-engineered T cells were cocultured for 24 h with indicated cancer cells and supernatants were analyzed for IFN-γ concentrations by ELISA. TCR-engineered T cells specifically recognized Ag104A cancer cells. Shown is one out of three independent biological repeats. Data are mean ± standard deviation from technical duplicates. Stimulation with αCD3/28 (MAX) was used as nonspecific T cell activation.

Figure 2.

The CD8TCR M2/3 is specific for a D256E mutation in the tumor suppressor gene Trp53. (a–b) every position on a tandem minigene supports antigen processing for recognition by TCR-engineered T cells. (a) overview of minigenes encoding known antigens. Minigenes depicted as black box represent unrelated antigen sequences. Four different minigenes were designed to cover all ten positions in a decamer tandem minigene construct (CG-1, −2, −3 and − 4). (b) EL4 cells were electroporated with the different CG constructs and a minigene encoding unrelated sequences (mock). EL4 cells were then cultured for 24 h with TCR-engineered T cells. Supernatants were analyzed for IFN-γ concentrations by ELISA. Spleen of C57BL/6 mice were used as T cell source for TCR-engineering. Shown is one out of two independent biological repeats. Data are mean ± standard deviation from technical triplicates. (c) results of whole-exome-sequencing and RNA-sequencing were compared between Ag104A cancer cells and Ag104 heart-lung fibroblasts (HLF) as syngeneic normal control to determine expression of 77 Ag104A-specific mutations. (d) schematic representation of the tandem decamer minigene (TMG) design encoding for 25mer neoepitope sequences with the mutated amino acid at the center position, concatenated with an AAY proteasomal cleavage site and linked via 2A to GFP. Eight tandem minigenes were constructed to cover 77 mutations. (e – g) spleen of C3H/HeN mice were used as T cell source for TCR-engineering. EL4 cells were electroporated with different TMG constructs and cultured for 24 h with M2/3-engineered T cells. Supernatants were analyzed for IFN-γ concentrations by ELISA. Shown is always one out of two independent biological repeats. Data are mean ± standard deviation from technical (e) quadruplicates or (f and g) triplicates. (e) eight different TMG constructs were introduced into EL4 cells either positive for H-2K^k or H-2D^k MHC class I alleles. Only TMG7 expressed by H-2K^k positive EL4 cells was recognized by M2/3-engineered CD8⁺ T cells. (f) TMG7 was divided into sub-TMGs 7–1, 7–2 and 7–3 containing only three or four minigenes. EL4 H-2K^k cells expressing TMG7–3 were recognized. (G) TMG7–3 was truncated to contain only one (7-3a), two (7-3b), three (7-3c) or four (7-3d) neoantigens. TMG7-3b, -3c and 3-d were all recognized indicating position eight of the original TMG7 as target for M2/3-engineered CD8⁺ T cells. (h) NetMHC 4.0 predicted 21 out of 77 nsSNVs to bind to H-2K^k. The affinities of these 21 neoantigens were plotted against their total number of reads obtained by RNA-sequencing. (i) position eight of TMG7 contained the D256E mutation in Trp53. The 9mer sequence LEESSGNLL was predicted to be one of the highest H-2K^k binders. Splenocytes from C3H/HeN were cultured either with graded concentrations of mutant or wild type 9mer peptide and M2/3-engineered T cells. Spleen of C3H/HeN mice were used as T cell source. Supernatants were analyzed for IFN-γ concentrations by ELISA. Shown is one out of two independent biological repeats. Data are mean ± standard deviation from technical triplicates.

Figure 3.

Escape of mutant-p53 specific TCR-therapy by antigen-loss variants. (a) cancer cell clones (n = 128) generated from the Ag104A cell line were cocultured for 24 h with M2/3-engineered T cells. Supernatants were analyzed for IFN-γ concentrations by ELISA. Spleen from C3H/HeN was used as source for T cells. Two independent biological repeats were done. One is shown. (b) overview of TCR-therapy. C3H Rag^−/− mice were injected s.c. with Ag104A. 14 days later, mice were treated with M2/3-engineered T cells. Spleen from C3H/HeN mice were used as source for T cells. (c) treatment of Ag104A tumor-bearing mice is indicated by the red arrowhead at day 0 (total numbers of mice, n = 8). Average tumor size at day of treatment: 0.098 ± 0.048 cm³. Mock-transduced CD8⁺ T cells were used for treatment of control mice (n = 4). Data are summarized from three independent biological repeats. (d) mice treated with either M2/3- (n = 8) or mock-engineered (n = 4) T cells were compared in a Kaplan-Meier survival analysis (***p < 0.001). Log-rank test was used to determine significance. (e) relapse variants M1, M2, M3 and M4 from (c) were readapted in vitro and stained for expression of H-2K^k and H-2D^k. (f) relapse variants (M1 – M8) and the parental Ag104A cells were cultured for 24 h with M2/3-engineered CD8⁺ T cells. Supernatants were analyzed for IFN-γ concentrations by ELISA. Spleen from C3H/HeN mice were used as source for T cells. T cells were cultured alone (none) as negative control. Ionomycin and PMA (MAX) was used as non-specific positive control for T cell stimulation. Shown is one out of three independent biological repeats. Data are mean ± standard deviation from technical triplicates. (g) PCR was used to amplify the genomic Trp53 region harboring the D256E mutation and analyzed by Sanger sequencing. Electropherograms of relapse variants are shown. Arrows indicate the position where the C to G mutation is located. Nucleotide codon GAC, indicated in blue, encodes aspartic acid (D, wild type p53), while GAG, indicated in black, encodes glutamic acid (E, mutant p53). (H–I) Trp53-specific sequencing of either (H) the Ag104A bulk cell line showing an equal ratio of wild type and mutant Trp53 reads or (I) Ag104A clones (n = 1,828). 99.5% of all clones (n = 1,820) harbor the heterogeneous Trp53^wt/mt genotype while 0.2% harbor only Trp53^mt (n = 3) and 0.3 % harbor only Trp53^wt (n = 5).

Figure 4.

Prevention of relapse and outgrowth of macrometastases of cancer that had spread to the lung at time of combination TCR-therapy. (left) established Ag104A-mL9-GFP tumors were treated by TCR-therapy 25 days after s.c. injection in C3H Rag^−/− mice as indicated by the arrow head. Average tumor size at day of treatment was 0.055 ± 0.031 cm³ standard deviation. Mice were treated either with the M2/3-CD8TCR (top, n = 5), or the H6-CD4TCR (middle, n = 7) or with a combination of both (bottom, n = 7). Untreated outgrowth controls (n = 2) are indicated. Spleen from C3H CD8^−/− or C3H CD4^−/− mice were used as source for T cells. Data are summarized from three independent biological repeats. (right) shown are representative pictures of lungs from each treatment group under natural or fluorescent light. Under fluorescent light, GFP-expressing cancer cells become visible. Top and bottom view of the same lung is indicated. White sizing bar indicates a length of 500 µm. (top) lung from a mouse treated with the M2/3-CD8TCR. (middle) lung from a mouse treated with the H6-CD4TCR and inflated using India ink. (bottom) lung from a mouse treated with the combination of the M2/3-CD8TCR and the H6-CD4TCR.