Presenters share latest insights from research into the tumor-host ecosystem
//
Estimated Read Time:
TNM staging is part of daily life in oncology. A growing body of evidence suggests that TNM staging is also myopic. It focuses on tumor characteristics and ignores the ecosystem in which tumors emerge, progress, metastasize, and kill—or in which they fail.
“Current models of cancer development and staging do not include the immune system,” said Jérôme Galon, PhD, INSERM. “T cells play an essential role in tumor progression. By determining their presence in the tumor, their type, their number, and the quality of the immune response, we have been able to more precisely predict relapse risk in cancer patients. It is a paradigm shift.”
Galon joined three other experts for the Plenary Session New Therapeutic Insights from the Tumor-Host Ecosystem on Monday, April 12. Registrants can watch a replay of the session through June 21, 2021.
Starting with the analysis of 122 pre-cancerous lesions from heavy smokers across nine developmental stages, Galon identified changes in the tumor microenvironment (TME) associated with progressing lesions that were increasingly immune privileged and progressed to active cancer.
Multi-omics analysis of metastasis yielded a similar interplay between tumor, host, TME, and immune response. Tracking the evolution of tumor clones that did or did not spread revealed that non-recurrent clones are edited out by the immune response. Progressing clones are immune privileged.
It is possible to score immune response to produce an immunoscore that correlates with long-term survival. Patients with a high score are more likely to survive. Immunoscore also correlates well with response to chemotherapy and CAR-T therapy in clinical trials.
Immunoscore has been incorporated into the WHO Classification of Tumours, 5th edition, and the latest ESMO Clinical Practice Guidelines for Gastrointestinal Cancers.
Mutational signatures in whole cancer genomes may have therapeutic applications
Every cancer genome carries the scars of mutagenic activity, DNA damage, and DNA repair that has occurred during the development of the tumor. These scars create characteristic imprints, or mutational signatures, that may be predictive of clinical response to specific anti-tumor agents.
“You can obtain a comprehensive picture from the whole cancer genome,” said Serena Nik-Zainal, MD, PhD, CRUK advanced clinician scientist in medical genetics, University of Cambridge School of Clinical Medicine. “When you use single genomic datapoints, you are ignoring the rest of the tumor context, which could be highly clinically relevant.”
Conventional genetic screening looks at a handful of variants, often just a single mutation, Nik-Zainal said. Mutational signatures, derived from the whole cancer genome, give a more comprehensive picture of tumor development, drivers, and potential therapeutic targets.
A weighted model called HRDetect, developed using whole breast cancer genomes from women with germline or inherited BRCA1/BRCA2 defects, can reliably identify triple-negative breast cancers (TNBC) that are missed by targeted sequencing. A proof-of-concept trial in women with sporadic TNBC show predictive therapeutic value.
Similar algorithms can be applied to whole genome sequencing from other tumor types to identify specific tumor drivers and predict sensitivity to specific therapeutic agents.
Targeting clonal evolution in myeloid malignancies
Acute myeloid leukemia (AML) is genetically simple and clinically complex. AML involves just 50-70 recurrently mutated genes with 3-8 coding mutations per AML genome. Recent technological developments are helping unravel the links between different mutational patterns and the complex heterogeneity in AML prognosis, biology, and outcome.
Single-cell profiling can help identify different sequences of somatic mutations that induce the progression from normal stem-cell progenitors to clonal hematopoiesis that can evolve to hematopoietic malignancies.
“Disease initiation is likely driven by single mutations or mutational combinations which modestly increase fitness of stem or progenitor cells,” said Ross L. Levine, MD, Laurence Joseph Dineen Chair in Leukemia Research and Chief of Molecular Cancer Medicine Service for the Human Oncology & Pathogenesis Program, Memorial Sloan Kettering Cancer Center, New York. “There is clear evidence of cooperation for some mutational combinations but not for others.”
Clonal evolution models have already produced a more faithful mouse model for NPM1/FLT3 human AML that shows an appropriate response to FLT3 inhibition with gilteritinib, approved by the FDA in 2018 for FLT3-mutated relapsed or refractory ALM. Similar models could be used to develop novel therapeutic approaches, including agents targeting the genetic or epigenetic events that drive clonal progression.
Taking pathology to the atomic scale
Pathology has become more clinically useful as techniques and technologies to analyze the tissue architecture move to ever-finer resolution and scale. Analyzing the tumor immune microenvironment at the cellular level can separate responders from non-responders based the 3D location of specific cell types and the distances between cell types.
“The atomic scale is where everybody needs to go to define the architecture of cells as defined by niches and inter-cellular relationships,” said Garry P. Nolan, PhD, Stanford University School of Medicine. “We are going to link visualizations at the atomic level to biochemistry and clinical outcomes to predict behaviors and functionality.”
Using CODEX, an instrument to perform multiplexing immunochemistry using up to 120 antibodies inside a conventional microscope platform, his lab analyzes the cellular neighborhoods by studying where different immune cell types cluster in different patterns within the tumor microenvironment.
Cell types in neighborhoods have prognostic value, Nolan explained. In colorectal cancer tumors, CD4+ and CD8+ enriched neighborhoods indicate better outcomes. The strongest signal for positive outcome is an increased density of PD-1+ CD4+ T cells in granulocyte-enriched neighborhoods.
In cutaneous T-cell lymphoma, the best prognostic marker is the spatial relationship between tumor cells, effector immune cells, and immunosuppressive cells. The closer the effector immune cells are to the tumor than the immunosuppressive cells, the better the likely outcome of pembrolizumab treatment.
His lab is developing a mass spectrometer-based atomic microscope with sub-angstrom resolution that can show the 3D spatial position and chemistry of every molecule within a cell. “You want to drive cell types and tumors toward that kind of structure because it might help these patients and identify potential therapies,” Nolan said.
[gap]