SNO Basic and Translational Science Conference: Day 1 Summary
Radhika Mathur, PhD
The SNO Basic and Translational Science Conference, held virtually on July 15-16th, 2021, brought together world experts and trainees in neuro-oncology for an enlightening series of talks and panel discussions. The conference was kicked off by SNO President Gelareh Zadeh and meeting Co-Chairs Joe Costello, Frank Furnari, and Annette Molinaro, who introduced the theme for the first day in the context of the evolution of our understanding of neuro-oncology. From previous eras of brain tumor histological classification and mapping of individual oncogenes and tumor suppressor genes, we have found ourselves in a period of high-resolution ‘omics’ where whole genomes, transcriptomes, epigenomes, etc. are analyzed for tumor cells as well as for other cells present in the tumor micro-environment. Research talks for the day are thus organized into two sessions on ‘Tumor Omics’ and ‘Microenvironmental Omics,’ followed by workshop sessions in a new format featuring experts on single-cell and multi-omic data analysis. While the Co-Chairs applaud tremendous successes by neuro-oncology researchers in the omics era, they highlight remaining challenges of intratumoral heterogeneity and ongoing tumor evolution. Celebrations of omics era success are also tempered by the knowledge that data generation and analysis cannot alone improve patient care. The omics era must be followed by new periods of integrated diagnosis and functional omics where the insights we have gleaned are applied to the design of new therapeutics that substantially improve clinical outcomes.
Some of the most compelling and complete stories came from invited speakers Nada Jabado (McGill University) and Mariella Filbin (Dana-Farber Cancer Institute) on pediatric brain cancers. These cancers often occur in specific brain regions at particular ages, indicating developmental origins, and are characterized by specific mutations in histone proteins, implicating epigenetic dysregulation as a mechanism of oncogenesis. Collectively, pediatric brain cancers seem to be explained by a ‘Peter Pan syndrome’ of stalled development where cells become stuck in an early progenitor cell from which they cannot differentiate. Research is now elucidating the developmental lineage-of-origin and detailed molecular mechanisms underlying each of these spatiotemporally and genomically distinct tumors. For example, tumors with the H3G34R/V mutation are found to originate from GABAergic interneuron progenitors and show an aberrant chromatin looping interaction between the PDGFRA oncogene and an enhancer for the neighboring gene GSX2. In the case of H3K27M glioma, research findings have led to the initiation of new multi-institutional clinical trials for PDGFRA and BMI inhibitors.
In adult brain cancers such as glioblastoma, treatment failure is often attributed to intratumoral heterogeneity, which fuels tumor evolution and the selection of resistant cells. The publication of single-cell RNA-sequencing studies has generated excitement in the field by describing distinct ‘cellular states’ that co-exist amongst malignant cells from an individual tumor. One of the lead authors of this work, Itay Tirosh (Weizmann Institute of Science), provided a useful technical overview of the single-cell RNA-Seq method and analysis workflow in a workshop session followed by a lively Q&A. He noted limitations of the RNA-Seq method, such as the ability to capture only ~10% of RNA molecules and the high degree of random noise per gene. He also cautioned that divisions between cellular states are somewhat arbitrary and that cellular states are not the same as the normal developmental cell types they are named after. For example, the glioblastoma ‘MES-like’ state does not exist in the normal brain and should not be confused with differentiated mesenchymal cells. Importantly, all four glioblastoma cell states contain cycling cells and are characterized by plasticity, questioning the relevance of the glioma stem cell model. This point was reinforced in an invited talk by Dan Landau (Weill Cornell Medicine), who conducted multi-omic single-cell analysis using both transcriptomic and DNA methylation data. Methylation-based lineage tracing and mathematical modeling reveal that glioblastomas show a bidirectional hierarchy with an order of magnitude greater de-differentiation than slower-growing IDH-mutant gliomas. Many of the stem-like cells in glioblastomas are inferred to have undergone de-differentiation from more mature cell types, revealing the need for therapeutic strategies to target the totality of tumor cells and/or their ability to de-differentiate.
Despite the attention single-cell methods have garnered, Mike Oldham (University of California, San Francisco) and Fred Varn (The Jackson Laboratory) reminded us in another workshop session of many advantages provided by bulk analysis, including the ready availability of matched multi-omic datasets with large sample sizes and diverse study designs (spatial, longitudinal, etc.) The best of both worlds can be achieved by integrative analysis of both bulk and single-cell datasets. Mike Oldham further described the insights we can generate from the massive deconvolution of bulk tissue datasets with appropriate sample size and statistical analyses. Although microenvironmental cell types represent only small proportions of tumor tissues, analysis of gene expression datasets reveals the relative abundance of each cell type across samples. This method can further identify new marker genes and show expression differences between tumor and normal brain microenvironmental cells. For example, the analysis identified ENPEP as a marker specific to tumor vascular cells, which was then validated in tumor tissue.
The immune and neural components of the tumor microenvironment have emerged as exciting areas of research, as highlighted by Johanna Joyce (University of Lausanne) and Humsa Venkatesh (Harvard Medical School), respectively. Brain tumors are associated with diverse immune landscapes, with low-grade gliomas showing enrichment of resident microglia and higher-grade tumors showing periphery contributions such as monocyte-derived macrophages and neutrophils. Efforts to therapeutically exploit the tumor microenvironment are underway, such as by inhibiting CSF-1R on the surface of tumor-associated macrophages, but intrinsic and adaptive resistance challenges must be overcome. The intriguing discovery that glioma is an electrically active tissue has initiated a new field of study of interactions between tumor cells and neurons termed ‘cancer neuroscience.’ Glioma-axon synapse formation has been visualized, and neuronal activity has been shown to promote glioma progression, raising hopes for a new avenue of cancer therapeutics informed by the long history of pharmacology in neuroscience.
The theme of basic and translational omics of brain tumors and the microenvironment was thoroughly explored by invited speakers and selected live oral abstract presenters, many of whom were trainees. The day exemplified scientific innovation and discovery, fostered productive discussions and a collaborative environment with an eye towards the ultimate goal of improving patient care.