In successful cancer immunotherapy, T cell responses appear to be directed toward neoantigens created by somatic mutations; however, direct evidence that neoantigen-specific T cells cause regression of established cancer is lacking. Here, we generated T cells expressing a mutation-specific transgenic T cell receptor (TCR) to target different immunogenic mutations in cyclin-dependent kinase 4 (CDK4) that naturally occur in human melanoma. Two mutant CDK4 isoforms (R24C, R24L) similarly stimulated T cell responses in vitro and were analyzed as therapeutic targets for TCR gene therapy. In a syngeneic HLA-A2–transgenic mouse model of large established tumors, we found that both mutations differed dramatically as targets for TCR-modified T cells in vivo. While T cells expanded efficiently and produced IFN-γ in response to R24L, R24C failed to induce an effective antitumor response. Such differences in neoantigen quality might explain why cancer immunotherapy induces tumor regression in some individuals, while others do not respond, despite similar mutational load. We confirmed the validity of the in vivo model by showing that the melan-A–specific (MART-1–specific) TCR DMF5 induces rejection of tumors expressing analog, but not native, MART-1 epitopes. The described model allows identification of those neoantigens in human cancer that serve as suitable T cell targets and may help to predict clinical efficacy.

Tumor-induced remodeling of the microenvironment relies on the formation of blood vessels, which go beyond the regulation of metabolism, shaping a maladapted survival niche for tumor cells. In high-grade B-cell lymphoma, angiogenesis correlates with poor prognosis, but attempts to target established proangiogenic pathways within the vascular niche have been inefficient. Here, we analyzed Myc-driven B-cell lymphoma–induced angiogenesis in mice. A few lymphoma cells were sufficient to activate the angiogenic switch in lymph nodes. A unique morphology of dense microvessels emerged without obvious tip cell guidance and reliance on blood endothelial cell (BEC) proliferation. The transcriptional response of BECs was inflammation independent. Conventional HIF1α or Notch signaling routes prevalent in solid tumors were not activated. Instead, a nonconventional hypersprouting morphology was orchestrated by lymphoma-provided VEGFC and lymphotoxin (LT). Interference with VEGF receptor-3 and LTβ receptor signaling pathways abrogated lymphoma angiogenesis, thus revealing targets to block lymphomagenesis. Significance: In lymphoma, transcriptomes and morphogenic patterns of the vasculature are distinct from processes in inflammation and solid tumors. Instead, LTβR and VEGFR3 signaling gain leading roles and are targets for lymphomagenesis blockade.

The success of chemotherapeutic intervention is limited because the necessary high local drug doses cannot be achieved without systemic toxicity. Application of suicide genes (SGs) and direct conversion of prodrugs (PDs) to toxic metabolites in situ by SGs may enhance the efficacy of chemotherapy. To evaluate this strategy in two murine breast cancer models, TS/A and GR, we injected cellulose sulfate capsules harboring cat kidney cells expressing the SGs cytosine deaminase and cytochrome P450 2B1 (CYP2B1) intratumorally. The PDs 5-fluorocytosine and ifosfamide were administered in 3-day intervals. The effect of in situ chemotherapy with each PD alone and the combination was analyzed over a period of 100 days. The results reveal that for TS/A tumors, the antitumoral effect mediated by CYP2B1 is more efficient than that of cytosine deaminase, whereas for GR tumors, both systems worked equally well. Furthermore, we find additive toxicity using both SG/PD systems for both TS/A and GR tumors.

Epstein-Barr virus (EBV) is associated with solid and hematopoietic malignancies. After allogeneic stem cell transplantation, EBV infection or reactivation represents a potentially life-threatening condition with no specific treatment available in clinical routine. In vitro expansion of naturally occurring EBV-specific T cells for adoptive transfer is time-consuming and influenced by the donor's T-cell receptor (TCR) repertoire and requires a specific memory compartment that is non-existent in seronegative individuals. The authors present highly efficient identification of EBV-specific TCRs that can be expressed on human T cells and recognize EBV-infected cells. Mononuclear cells from six stem cell grafts were expanded in vitro with three HLA-B*35:01- or four HLA-A*02:01-presented peptides derived from six EBV proteins expressed during latent and lytic infection. Epitope-specific T cells expanded on average 42-fold and were single-cell-sorted and TCRαβ-sequenced. To confirm specificity, 11 HLA-B*35:01- and six HLA-A*02:01-restricted dominant TCRs were expressed on reporter cell lines, and 16 of 17 TCRs recognized their presumed target peptides. To confirm recognition of virus-infected cells and assess their value for adoptive therapy, three selected HLA-B*35:01- and four HLA-A*02:01-restricted TCRs were expressed on human peripheral blood lymphocytes. All TCR-transduced cells recognized EBV-infected lymphoblastoid cell lines. The authors' approach provides sets of EBV epitope-specific TCRs in two different HLA contexts. Resulting cellular products do not require EBV-seropositive donors, can be adjusted to cell subsets of choice with exactly defined proportions of target-specific T cells, can be tracked in vivo and will help to overcome unmet clinical needs in the treatment and prophylaxis of EBV reactivation and associated malignancies.

Interleukin 4 (IL‐4) was shown to be tumor‐promoting in full carcinogenesis studies using 3‐methylcholanthrene (MCA). Because heretofore the role of IL‐4 in DMBA/TPA (9,10‐dimethyl‐1,2‐benz‐anthracene/12‐O‐tetradecanoylphorbol‐13‐acetate) two‐stage carcinogenesis was not studied, we performed such experiments using either IL‐4^−/− or IL‐4Rα^−/− mice. We found that IL‐4Rα^−/− but not IL‐4^−/− mice have enhanced papilloma formation, suggesting that IL‐13 may be involved. Indeed, IL‐13^−/− mice developed more papillomas after exposure to DMBA/TPA than their heterozygous IL‐13‐competent littermate controls. However, when tested in a full carcinogenesis experiment, exposure of mice to 25 μg of MCA, both IL‐13^−/− and IL‐13^+/− mice led to the same incidence of tumors. While IL‐4 enhances MCA carcinogenesis, it does not play a measurable role in our DMBA/TPA carcinogenesis experiments. Conversely, IL‐13 does not affect MCA carcinogenesis but protects mice from DMBA/TPA carcinogenesis. One possible explanation is that IL‐4 and IL‐13, although they share a common IL‐4Rα chain, regulate signaling in target cells differently by employing distinct JAK/STAT‐mediated signaling pathways downstream of IL‐13 or IL‐4 receptor complexes, resulting in different inflammatory transcriptional programs. Taken together, our results indicate that the course of DMBA/TPA‐ and MCA‐induced carcinogenesis is affected differently by IL‐4 versus IL‐13‐mediated inflammatory cascades.