CAP TODAY Pathology/Laboratory Medicine/Laboratory Management
Editors: Donna E. Hansel, MD, PhD, chief, Division of Anatomic Pathology, and professor, Department of Pathology, University of California, San Diego; James Solomon, MD, PhD, resident, Department of Pathology, UCSD; Richard Wong, MD, PhD, molecular pathology fellow, Department of Pathology, UCSD; and Sounak Gupta, MBBS, PhD, molecular pathology fellow, Memorial Sloan Kettering Cancer Center, New York.
Link between immunogenic neoantigens derived from gene fusions and T-cell responses
June 2019—Immunotherapy is quickly emerging as an important therapeutic strategy for hematologic and solid tumors. In principle, immune cells recognize unique non-self antigens expressed by cancer cells and this serves as the initial step in eliminating such cells. Among the metrics used to assess the likelihood of response to immunotherapy is the presence of a high level of somatic mutations, referred to as tumor mutation burden (TMB). This serves as a surrogate for extrapolating the level of neoantigens ex- pressed by cancer cells. Tumors that are driven by fusion events—that is, two genes combine to produce a new protein that includes products of both genes, referred to as chimeric proteins—do not fit this paradigm, especially as many such tumors have extremely low levels of background somatic mutations, or low TMB. The authors conducted a study in which they demonstrated that high levels of expression of neoantigens secondary to fusion events can drive anticancer immune responses independent of the level of somatic mutations (TMB) in such tumors. Tumors that were profiled and that supported this finding included a head and neck squamous cell carcinoma with a poorly understood DEK-AFF2 fusion and an adenoid cystic carcinoma (a rarer subtype of head and neck cancer) that harbored a better-characterized MYB-NFIB fusion. The head and neck squamous cell carcinoma was considered unresectable, had metastasized to the lung, and failed conventional chemotherapy. It showed complete regression following immunotherapy. The authors excluded variables that could explain this response, such as a high neoantigen load (high TMB); viral neoantigen expression, including that related to human papillomavirus, commonly seen in some head and neck squamous cell carcinomas; high levels of immune infiltration; and high expression of immune cell biomarkers of response (PD-L1). In a series of in vitro experiments, they demonstrated that specific antigenic peptides derived from the fusion-driven chimeric protein were capable of eliciting strong T-cell–mediated cytotoxic responses. They also demonstrated that circulating immune cells in the patient’s blood were enriched for those that were capable of recognizing such immunogenic peptides. These immune cells were undetectable in pre-immunotherapy specimens and were detectable as late as 21 months on immunotherapy. These results were extrapolated to 5,825 fusion-positive cancer samples from The Cancer Genome Atlas data sets. At least a quarter of these tumors were predicted to produce neoantigens that were highly likely to be expressed by the tumor cells based on each patient’s genetic makeup, using human leukocyte antigen (HLA) subtyping. In a comparison of fusion-positive tumors in matched pre-immunotherapy patients and patients on immunotherapy, the tumors that responded were found to be more likely to generate neoantigens that were matched to a patient’s HLA type. In summary, this study provides evidence that peptides derived from gene fusions are an important source of cancer cell-specific neoantigens. This has implications for the future use of immunotherapy to treat such tumors, especially those that do not have effective targetable therapies.
Yang W, Lee KW, Srivastava RM, et al. Immunogenic neoantigens derived from gene fusions stimulate T cell responses. Nat Med. 2019. doi:10.1038/s41591-019-0434-2.
Correspondence: Dr. Timothy A. Chan at chant@mskcc.org or Dr. Luc G. T. Morris at morrisl @mskcc.org
Findings from evolutionary trajectories of IDHWT glioblastomas
Glioblastomas are aggressive primary brain tumors with a dismal five-year survival rate of approximately six percent. These tumors are thought to progress from lower grade tumors or arise de novo. The latter category of tumors are typically characterized by isocitrate dehydrogenase (IDH)-unmutated status. Clinical management of these tumors involves neurosurgical resection followed by other forms of treatment, such as chemotherapy or radiation. Even with such interventions, the vast majority of these tumors recur. The authors conducted a study in which they evaluated paired samples of primary or de novo glioblastomas and recurrences that occurred post-surgery and post-chemoradiation and that regrew at the same location as the primary glioblastoma in most cases. These primary and recurrent tumors were profiled using multiple molecular profiling tools to better understand their molecular evolution over time. Overall, most primary and recurrent tumors showed similar mutational profiles, except for a small number of cases that acquired a hypermutated molecular profile linked to chemotherapy with temozolomide. Some cases in the latter category were enriched for a pattern of mutations referred to as signature 11, commonly seen in response to alkylating chemotherapeutic agents. This suggests that standard therapy for primary glioblastomas exerts little selective pressure on recurrences. The molecular evolution of these tumors was deduced by looking at the frequency of shared alterations between primary and recurrent tumors. For instance, mutations that were rare in the primary tumor but present at high levels in the recurrence were considered to be enriched by clonal evolution in the latter, while mutations present at high levels in both primaries and recurrences were considered to be shared founder cell alterations. Knowing this, the authors were able to evaluate tumor evolution based on the acquisition of alterations in the cancer genome over time. In addition, the authors modeled rates of cancer cell proliferation and cancer cell death and the acquisition of novel mutations by daughter cells to illustrate tumor evolution. Three common alterations related to deviations from the normal two copies of genes present in all cells (diploid status) were found to be shared at high levels by the primary tumors and recurrences, suggesting that these represented early oncogenic events. These included losses on the short arm of chromosome 9, including the locus for CDKN2A/B genes; losses on the long arm of chromosome 10, including the locus for the PTEN tumor-suppressor gene; and gains of chromosome 7, including the EGFR oncogene. Recurring hotspot mutations of the promoter region of the TERT gene, on the other hand, while present in almost all tumors, were found at lower or subclonal levels in these tumors, suggesting that this is an event that occurs later in disease progression. Among the more interesting insights from the mathematical and molecular modeling of tumor evolution is the prediction that an undetectable cell of tumor origin or founder cell will evolve over a period of two to seven years before a clinically detectable tumor is identified. Overall, the findings provide interesting insights into the tumorigenesis of glioblastoma and the effects of therapy.