Year: 2020
Type: article
Abstract: Diffuse intrinsic pontine gliomas (DIPGs) are high-grade tumors of the brain stem that occur in young children and that are characterized by the presence of histone H3.3/H3.1 missense mutations K27M in more than 80% of cases.1, 2 The H3K27M mutation is located in the histone tail region that is subject to extensive post-translational modifications (PTMs). The K27M substitution has been shown to disrupt PTM patterns resulting in global epigenetic deregulation. In fact, H3K27M mutant tumors and cell lines have reduced levels of H3K27me3 and H3K27me2, changes in histone acetylation patterns, and reduced DNA 5mC-methylation as well as disruptions in other methyl marks such as H3K4me3 and H3K36me2. Combined, these observations suggest that disruption of H3K27-centric PTM biology has pan-genomic consequences. Functionally, the H3K27M substitution drives or accelerates DIPG growth both in vitro and in animal models, suggesting that epigenetic misregulation promotes DIPG tumorigenesis; however, mechanistic details on how aberrations in chromatin and DNA modifications affect DIPG tumorigenesis are not fully understood. In a recent issue of Cancer Cell, Anastas and colleagues3 hypothesized that the aberrant chromatin induced by H3K27M mutation may offer vulnerabilities to epigenetic-based therapies. Building on prior examples of successful inhibition of DIPG growth in vitro and in vivo using single-agent small-molecule inhibitors of bromodomain-containing proteins (eg, BRD4), EZH2, and histone deacetylases (HDACs), the authors surmised that combinatorial inhibition of additional chromatin regulators along with HDAC inhibition may offer additional benefit against DIPG. To achieve this, they performed a chromatin gene-focused CRISPR screen in H3K27M mutant DIPG and identified the histone H3K4me1/2-specific demethylase lysine-specific demethylase 1 (LSD1) as a candidate that can be targeted to enhance the growth inhibitory effects of HDAC inhibition in H3K27M mutant DIPG cells. The authors validated these results pharmacologically using a combination of single-agent HDAC and LSD1 inhibitors and a bifunctional HDAC/LSD1 inhibitor named Corin. Dual HDAC and LSD1 inhibition treatments resulted in a synergistic effect on DIPG cell growth reduction and cell death. In addition, the authors demonstrate that dual HDAC/LSD1 inhibition using Corin treatment reduced DIPG cell proliferation and promoted differentiation both in vitro and in animal models of DIPG. Mechanistically, inhibition of HDAC and LSD1 leads to changes in global histone methylation and acetylation patterns at promoters and enhancers of cell cycle genes and cell differentiation genes. Finally, the clinical relevance of the work was demonstrated in a series of studies that shows that Corin-induced changes in gene signatures observed in vitro compares with gene expression changes that are associated with better DIPG patient outcomes. Current standard-of-care treatments for children with DIPG have failed, and less than 10% of patients survive longer than 2 years after diagnosis. This study reveals that HDACs and LSD1 co-inhibition with Corin reduces DIPG growth by modifying chromatin in a manner that induces differentiation and cell death while suppressing the cell cycle. Overall, the results reported by Anastas et al. suggest that combinatorial targeting of several chromatin regulators might represent a path forward to treat DIPG. Increasing evidence indicates that the tumor microenvironment plays an important role in the pathogenesis of primary and secondary brain tumors, including malignant glioma. The influence of neuronal constituents and their interconnected networks on glioma growth has recently been examined more intensively. In particular, multiple reports have shown that neuronal secretion of neuroligin-3 (NLGN3) facilitates gliomagenesis. These findings build on older literature implicating neuronal activity in glioma growth. In two recent Nature articles, Venkataramani et al. and Venkatesh et al. reported the results of similar studies designed to explore the mechanisms by which neuronal activity promotes gliomagenesis.1, 2 Using a combination of in vitro co-culture systems, patient-derived glioma cells in the context of murine xenografts, freshly resected glioma tissue, and genetically engineered glioma mouse models, the authors documented the existence of neuron-to-glioma (neuroglioma) synapses involving a significant subset of infiltrating tumor cells. These synapses were reminiscent of so-called tumor microtubes, which had previously been shown to connect glioma cells in a communicating network, facilitating both tumor proliferation and invasion. The authors of both studies then demonstrated that neuroglioma synapses were enriched in glutamatergic AMPA receptors and transmitted excitatory postsynaptic currents in associated tumor cells. Genetic and pharmacologic studies confirmed the dependence of these electrical signals on the AMPA receptor apparatus. Subsequent functional analyses revealed that active neuroglioma synapses drive tumor proliferation in an AMPA-dependent manner, with Venkataramani et al. also demonstrating effects on glioma cell invasion.1 Finally, Venkatesh et al. documented increased neuronal excitability in association with infiltrating high-grade glioma, implying the existence of vicious cycles combining neuronal excitation with tumor growth/invasion.2 The findings of these two studies are highly significant in that they describe a novel and targetable mechanism by which the unique neuronal microenvironment directly impacts gliomagenesis. Aberrant DNA methylation in cancer cells is commonly associated with transcriptional alterations, such as silencing of tumor suppressor genes. Epigenetic mechanisms contribute to gene deregulation. In a recent paper, Court et al studied 70 samples of adult glioma to explore the relative contribution of DNA methylation-dependent vs. independent mechanisms to transcriptional alteration at CpG Island (CGI)/promoter-associated genes.1 The authors report that altered histone H3 trimethylation at lysine 27 (H3K27me3) was the predominant molecular defect at both aberrantly repressed and expressed genes. In many genes, ectopic expression was associated with CGI/promoters that gained methylation at their borders, whereas the H3K4me3-marked transcriptional start site was methylation-free. Highly expressed genes were more likely to be repressed by H3K27me3 than by DNA methylation. The main causes of transcriptional alteration in glioma cells were defects in the interplay between polycomb protein complexes and the brain-specific transcriptional machinery, mainly in the form of H3K27me3 dynamics deregulation, particularly when present in a bivalent context. The authors provided integrative analyses that quantified the main types of transcriptional defects in glioma, recapitulating the DNA methylation- and H3K27me3-based molecular signatures at aberrantly repressed and expressed CGI/promoter-associated genes. They described these defects in IDHwt and also in IDHmut glioma samples, indicating that defects occur regardless of the G-CIMP status and tumor aggressiveness. While the number of deregulated genes with CGI/promoters associated with DNA methylation defects was similar between glioma subtypes, there was a higher frequency (about two times) of DNA methylation-independent transcriptional alterations in IDHwt than in IDHmut samples. This could contribute to the difference in outcome between these glioma subtypes. The CIMP-positive status, by promoting stable gene repression, could also act as a protective mechanism against cancer progression by restricting epigenetic plasticity as well as the ability of glioma cells to adapt to stresses. Court et al found that healthy tissue may play an important role in the (epi)genetic landscape of glioma. Gene expression strength in healthy brain cells appeared to influence the choice between DNA methylation- and H3K27me3-associated silencing. CIMP was observed mainly at genes that are already repressed in healthy brain. Highly expressed genes were more likely to be repressed by H3K27me3 rather than DNA methylation. The authors identified a subset of genes with an apparent constitutive H3K4me3-only signature in healthy embryonal stem cells from brain and glioma samples. This group of genes could contribute to glioma aggressiveness by maintaining tumor cells in a stem cell-like state. Several groups2 have shown that DNA methylation patterns correlate with tumor types, whereas Court et al suggest that the transcriptome may be dominantly regulated by epigenetic events other than methylation. It is unclear how to reconcile this disparity. Nonetheless, Court at al provide a framework to explain the counterintuitive observation that patients with CIMP-positive IDHmut have a better clinical outcome than patients with CIMP-negative IDHwt. The study may also inform development of therapies targeting epigenetic defects in cancer. Immunotherapies against glioblastoma (GBM) have steadily gained momentum in the past few years. One of the approaches that continues to improve and show considerable promise against solid tumors is the use of chimeric antigen receptor-expressing T cells (CAR-Ts), which recognize specific antigens on the tumor cell surface and deliver a strong cell-to-cell cytotoxic response.1 The exquisite specificity of CAR-Ts is at the same time a limitation for their efficacy against highly heterogeneous cancers like GBM because malignant clones lacking the target antigen can escape cytotoxicity and recur as resistant tumors. To address this issue, Choi and colleagues have recently engineered T cells that not only express a chimeric antigen receptor but also secrete a protein that engages adjacent, non-CAR T cells against the tumor cells.2 Bi-specific T cell engagers (BiTEs) are recombinant “mini-antibodies” generated by the combination of two antibody fragments, one binding the desired antigen and the other engaging T cells by binding their cell-surface CD3 receptor.3 BiTEs can redirect naïve T cells against the target antigen in MHC-independent fashion, achieving an effect comparable to CAR-Ts without cell transduction, although a continuous delivery of BiTEs is required. Choi et al. modified T cells to express a CAR against the tumor-specific truncated variant of the epidermal growth factor receptor (EGFRvIII, expressed in 30% of GBM cells), together with a secreted BiTE against wild-type EGFR (EGFRwt, present in 80% or more of the tumor cells). Although targeting EGFRwt could potentially damage normal cells the investigators chose this receptor because it is weakly expressed in the normal CNS but it is highly upregulated in GBM and remains expressed even if EGFRvIII is lost. When CAR-Ts targeting only EGFRvIII were tested against GBM cells with heterogeneous EGFRvIII expression, the team found that, despite a good initial response, the immune cells failed to eradicate the tumors, which expectedly progressed as EGFRvIII-negative (but EGFRwt-positive) malignancies. In contrast, CAR-Ts targeting EGFRvIII and secreting EGFRwt-specific BiTEs not only eliminated the EGFRvIII-positive GBM cells but also triggered a larger T cell response against the remaining GBM cells, preventing the escape of CAR-resistant clones. Interestingly, CAR-Ts secreting EGFRwt-specific BiTEs did not attack normal skin cells (which express EGFRwt), whereas CAR-Ts expressing an EGFRwt-specific CAR did cause skin damage; these results underscore important differences in CAR- versus BiTE-mediated T cell activation that should be further studied to increase the safety of engineered T cells. Overall, this study suggests promising avenues to improve T cell therapies that can overcome, in part, the heterogeneity of malignant gliomas. T cell-based immunotherapies have become increasingly promising treatments for glioblastoma (GBM) due to the potent tumor-killing ability of these immune cells. Nevertheless, the efficacy of T cells continues to be hampered by the heterogeneity of GBM, the immunosuppressive tumor microenvironment, and the limited duration of anti-tumor effects before T cell exhaustion. As a result, immune-boosting strategies such as immune checkpoint blockade, which has remarkable efficacy against other solid tumors, have not yet shown survival benefits in GBM patients.1 Seeking to improve the anti-tumor effects of T cells, Ye and colleagues have developed a CRISPR-based screening to identify intrinsic factors that limit the efficacy of these cells in the tumor.2 Briefly, the investigators developed a library of RNA guide sequences (sgRNAs) for CRISPR-based gene disruption, focusing on 1658 cell-membrane receptors and cell surface-associated target genes. The library constructs were virally introduced in mouse T cells, which were injected intravenously in mice carrying intracranial gliomas. The investigators reasoned that T cells enriched in the tumors were likely the “winners” that had lost immune inhibition thanks to CRISPR disruption. Accordingly, quantification of enriched sgRNA sequences in T cells recovered from tumors compared with sequences of the pre-injected T cell pool allowed the team to identify putative genes that had been disrupted. Those genes were subsequently validated to confirm that their disruption effectively enhanced the infiltration and response of T cells against the tumor. Surprisingly, only one immune checkpoint gene (the cell-membrane receptor LAG3, which competes with the CD3 receptor and inhibits antigen binding3) was preferentially disrupted in tumor-enriched T cells, with no significant changes in other checkpoint targets such as PD-1. Even more remarkable was the absence of other cell membrane receptors that were significantly disrupted in the tumor-infiltrated T cells. Instead, two genes consistently disrupted were the glycosyltransferase MGAT5 and the chaperone PDIA3, both of which are involved in the normal processing of cell surface proteins as they travel through the endoplasmic reticulum. Both genes were further confirmed to limit T cell activation, and their disruption enhanced anti-tumor T cell effects and extended animal survival. As a separate demonstration the investigators disrupted PDIA3 in human CAR-T cells targeting a tumor-specific variant of the epidermal growth factor receptor (EGFRvIII), which increased the efficacy of those CAR-Ts against GBM cells. Although the study did not address the mechanisms by which GBM cells can still circumvent activated T cells, it provides a relatively unbiased identification of novel targets worth considering for T cell boosting. Disruption of these targets could help develop combinatorial strategies to potentiate T cell therapy of glioblastoma.
Source: Neuro-Oncology
Language: English
Subfield: Genetics
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W4206840596