NEW YORK – A team from Israel, the US, Germany, and Switzerland mapped out the spatial features in dozens of glioma cases, including high-grade glioblastoma (GBM) tumors, identifying regions with varying cell state structure that appeared to reflect low oxygen, or hypoxia, conditions with the tumor tissue.
The study, published in Cell on Monday, "adds a spatial dimension to our growing understanding of the glioma ecosystem and may aid in the development of future treatments," co-senior and co-corresponding authors Itay Tirosh, a molecular cell biology researcher at the Weizmann institute of Science, and Mario Suvà, a pathology researcher affiliated with Massachusetts General Hospital and the Broad Institute, and their colleagues wrote.
For their analyses, the researchers considered 10x Genomics Visium-based spatial transcriptomics on 19 glioma samples and spatial proteomic profiles done on a dozen more samples using single-cell codetection by indexing (CODEX) technology, together with published Visium spatial transcriptomic data for 13 high-grade glioma, or GBM, cases.
By integrating these data, the team distinguished between structured and disorganized forms of glioma, which fell along a continuum that coincided with enhanced or reduced levels of hypoxia. Along with local cell clusters that shared a given cell state, for example, the spatial maps highlighted cell states that tended to neighbor one another, while a broader view of structured tumor regions revealed consistent layers of cell states.
"Although tissue organization may be intuitively linked to normal physiology while chaos is reminiscent of aggressive tumor phenotypes, we note that hypoxia and necrosis, and hence cancer state organization, are hallmarks of high-grade glioma," the authors explained.
"Such organization may imply that certain layers are less accessible to drugs or to immune cells," they added, "and thereby more resistant to particular therapies."
In samples that did contain structure, the investigators highlighted five recurrent cell state layers falling along a hypoxic, hypoxia response, and infiltrating continuum: a so-called hypoxic niche; a hypoxia-adjacent layer; a layer marked by angiogenesis immune-hub features; a neurodevelopment/GBM state layer; and a normal brain layer infiltrated by tumor tendrils.
In tumor sections with little to no sign of hypoxia, on the other hand, the team tended to see a disorganized amalgamation of cells with distinct cell state expression features, suggesting that hypoxia and the presence of hypoxic cells has a role in helping to organize a tumor's cell states.
"[W]e provide an extensive spatial description of glioma that demonstrates stereotypical spatial organization at multiple scales, with a prominent role of hypoxia as an organizer and with relative disorganization of regions that lack hypoxia," the authors reported.
Along with analyses focused on "metaprograms" consisting of gene clusters with shared expression features in the tumors, the team also went on to tease out features found in glioma and GBM tumors containing wild-type or mutant versions of the IDH gene.
For example, they noted that the lower-grade glioma cases containing IDH mutations tended to be disorganized and were missing marks of hypoxia, whereas atypical versions of the IDH-mutant gliomas had enhanced levels of hypoxia, along with more pronounced spatial organization features.
Together, the team suggested, the combined spatial transcriptomic and proteomic approach appeared to provide a more detailed look at processes at play in the tumors compared to prior studies focused on spatial transcriptomic data alone.
"We combined Visium with CODEX and computational approaches to define the organization of gliomas," the authors explained, adding that "we identified several [metaprograms] in the spatial transcriptomics data that were not previously defined from single-cell or single-nuclei data."