Highlights from the Literature

The temperature of silence: A spectrum of immunogenicity revealed in rhabdoid tumors

Immunotherapeutic treatment strategies provide novel avenues to improve clinical outcomes in patients. Significant anti-tumor responses have been observed in several subsets of adult tumors through targeting immune checkpoint blockade.¹ The vast majority of pediatric tumors have previously been considered as poor candidates for immunotherapeutic treatment due to their stable genomes. Immune checkpoint inhibitors (ICI) have remained confined to specific pediatric tumors that often demonstrate high mutational burden and possess elevated tumor antigen, canonical hallmarks of effective ICI response.² Nevertheless, certain cancer subtypes, such as renal cell carcinoma, have demonstrated sensitivity to ICIs despite low mutational burden, suggesting additional factors are responsible for ICI efficacy. This in turn has led to renewed interest in evaluating the immunogenic potential within pediatric tumors and evaluate their susceptibility to ICI.

A study by Leruste and colleagues in Cancer Cell³ is a landmark paper in characterizing the immunogenic potential of genomically stable pediatric tumors. The authors investigated the immune profile of rhabdoid tumors (RT), a rare aggressive pediatric tumor that contains one of the lowest base rate changes amongst sequenced tumor subtypes. The primary mutation observed in RT occurs at the SMARCB1 locus, a core subunit of the SWI/SNF complex. Despite the quite coding genome, RTs present heterogeneously in the clinic and have been stratified into subgroups with distinct clinical phenotypes, epigenomic landscapes, and drug susceptibility.⁴ Initial deconvolution studies performed by the authors using RNA-seq transcriptomic and DNA methylation data, which was subsequently validated via immunohistochemistry, revealed significant intra-tumoral heterogeneity and enrichment for immune infiltrates in a subgroup-specific manner. Macrophages, T cells, and NK cells were prominent features of the tumor microenvironment, suggestive of an active immune infiltrate phenotype in certain RT subgroups.

To further interrogate the immunological profile of RTs and to improve the understanding of the T cell composition of RTs, the authors performed single-cell RNA-seq (scRNA) on matched tumor and peripheral blood mononuclear cells to reveal T cell heterogeneity. Across all four groups of RTs, there was a highly heterogenous T cell population demonstrating varying states of exhaustion and activation. Further investigation utilizing coupled T cell receptor sequencing (TCR-seq) elucidated the granularity of the T cell infiltrates, revealing distinct differences in T cell composition of immunosuppressive and immunogenic subgroups, with immunogenic tumors phenocopying adult immunosensitive tumors. Furthermore, these findings were corroborated in a genetically engineered mouse model of RT,⁵ whereby PD-1-positive T cell infiltrates were identified and shown to respond to ICI treatments that proved durable once challenged again with tumor cells. The authors then proceeded to investigate endogenous retrovirus (ERVs) as an immunogenic determinant in RT and provided evidence that de-repression of ERVs in RTs is at the heart of evoking interferon pathways and abundance of ERV expression correlated with immunogenicity of the tumor.

Leruste et al have provided a comprehensive analysis of the immune profile of RTs and have identified a therapeutically relevant avenue within RTs. The authors provide evidence of immunogenic responses in a low mutational burden pediatric tumor background and demonstrate heterogeneous immune infiltration mechanisms that they suggest is dependent on ERV-induced interferon signaling.

Modulation of the CNS lymphatic vasculature may influence efficacy of immunotherapy

In a recent article in Nature, Song et al observed that mice treated with the vascular endothelial growth factor-C (VEGF-C; a growth factor for lymphatic vessels) rejected intracranial tumor implants.¹ These mice also rejected a later re-challenge with flank tumor implants, demonstrating memory responses. Depletion of CD4 or CD8 T cells (not B cells) removed this protection.

Immune surveillance and antigen presentation require lymphatic drainage to lymph nodes. The brain, unlike extracranial organs, contains no lymphatic network. The cerebral spinal fluid (CSF) is drained through meningeal lymphatics into the deep cervical lymph nodes (dCLNs), where brain-derived antigens and peripheral T cells interact.

Song et al ligated the afferent lymphatic vessels draining into the dCLNs in VEGF-C treated mice. Ligation of the dCLNs removed the survival benefit seen with VEGF-C. VEGF-C treatment resulted in increases in the percentage and absolute number of tumor-specific T cells in the dCLNs, suggesting that VEGF-C-mediated protection against glioblastoma (GBM) requires lymph drainage to dCLNs and that the VEGF-C-generated protection is immune mediated. To be effective, peripheral T cells may not require access to the brain, because they can experience the parenchymal antigenic repertoire within the dCLNs.

Song et al then demonstrated that VEGF-C given with an anti-PD-1 antibody leads to tumor regression and longer survival. Treated mice were also resistant to re-challenge with tumors in the contralateral hemisphere. Furthermore, T cells transferred from the draining lymph nodes and spleen of mice rejecting tumors provided protection against tumor implants in naïve mice. Even when tumors were allowed to grow for long periods (20 days), the combination therapy still yielded a survival benefit.

To examine whether VEGF-C is effective in treating “non-GBM” types of cancer in the CNS, Song et al also studied melanoma cell lines. In contrast to GBM, combined anti-PD1 and anti-CTLA4 have intra-and extracranial activity in melanoma. Consistent with clinical observations, mice with both a flank and an intracranial tumor responded better to immunotherapy than did those with just intracranial melanoma.

Mice with both intracranial and flank tumors benefited from checkpoint inhibitor therapy regardless of VEGF-C treatment. Survival benefit in mice with only intracranial tumors treated with VEGF-C/anti-PD1 combination therapy was similar to that in mice with both intracranial and flank tumors treated with checkpoint inhibitor therapy alone. Ligation of the dCLNs removed the VEGF-C benefit in mice with intracranial tumors but not in mice with both intracranial and flank implants.

T cell priming through expression of VEGF-C in the CSF, or through a flank tumor, enables checkpoint inhibition in the CNS. However, in the case of a tumor that is confined to the CNS at steady state (e.g., GBM), immune checkpoint inhibitors alone do not confer notable benefits.

If these experiments are able to be reproduced in other, more resistant models, then the dCLNs (and VEGF-C) may hold the key to a new strategy to increase lymphatic drainage, thereby enhancing immunosurveillance and overcoming the immune ignorance of GBM.

Glioma cells form coordinated networks to invade brain tissue

One of the hallmarks of malignant gliomas is their diffuse infiltration through neural tissue, which ultimately makes these tumors impossible to contain by surgical resection alone. Multiple mechanisms of cell adhesion, force generation, and remodeling of the unique neural extracellular matrix have been shown to underlie the ability of individual glioma cells to invade the brain.¹ However, glioma cells have also been shown to form extensive multicellular networks connected by gap junctions and microtubes bridging neighboring cells, which are important for the expansion and apoptotic resistance of the tumor’s leading edge.²

Because the relevance of these “glioma networks” for tumor invasion has not been determined, a recent study by Gritsenko and colleagues³ analyzed whether the cell-to-cell bridges and adhesions actually promote or self-limit tumor invasion, hoping to identify novel targets to contain the tumor infiltration that continues after surgery. The investigators reproduced network-like invasion out of glioma spheroids cultured in three-dimensional scaffolds and combined this with imaging of glioma networks in vivo after pharmacological and genetic manipulations. Histological data from intracranial xenografts confirmed that glioblastomas form extensive networks as they invade through the brain parenchyma or along brain vessels. When tested ex vivo, these networks were necessary to synchronize calcium oscillations across multiple cells; accordingly, disruption of gap junctions or cell adhesion molecules disturbed the propagation of calcium waves and slowed down the invasion at the tumor front. Remarkably, inhibition of RhoA/ROCK signaling, which is an important approach to block the invasion of dissociated glioma cells,⁴ promoted instead the formation of glioma networks, suggesting that disrupting the motility of individual cells may not be sufficient to limit the infiltration of the tumor border. A target of particular relevance was p120-catenin, which is a key regulator of cell-cell adhesion and has been associated with tumor progression and poor survival in glioblastoma and other solid tumors.⁵ Knockdown of p120-catenin was sufficient to disassemble glioma networks, diminishing tumor invasion in vivo and re-sensitizing the dissociated glioma cells to Rho/ROCK inhibition.

This study underscores the importance of carefully characterizing the process of glioma invasion either in vivo or in conditions reproducing the structure and topology of brain tissue. When studied in these conditions, tumor invasion appears as a cooperative process that is essentially different from the behavior of dissociated cells in conventional in vitro assays. Targeting of p120-catenin and other molecules required for the formation of multicellular glioma networks may open new avenues to disrupt collective tumor invasion and potentiate the long-term effect of conventional therapies.

Safety and feasibility of multiple blood-brain barrier disruptions for the treatment of glioblastoma in patients undergoing standard adjuvant chemotherapy

The major cause of death due to glioblastoma (GBM) after treatment is tumor recurrence, which usually develops within 2 cm of the original lesion. This area is commonly infiltrated with tumor cells, but the blood-brain barrier (BBB) in this region is mostly intact. Various strategies to overcome the BBB have been investigated; however, all approaches have been limited by the lack of specificity, safety concerns, and failure to deliver adequate concentrations of therapeutic agents to brain tissue.¹ Recently, MR-guided focused ultrasound (MRgFUS) has been attracting attention as a noninvasive means of temporarily disrupting the BBB. MRgFUS delivers ultrasonic energy to the target with intraoperative imaging guidance and real-time feedback for monitoring without opening the scalp or skull.²

The purpose of a recent study by Park and colleagues, published online in January 2020 in the Journal of Neurosurgery, was to evaluate the safety and feasibility of repeated disruption of the BBB (BBBD) with MR-guided focused ultrasound (MRgFUS) in patients with GBM during standard adjuvant temozolomide (TMZ) chemotherapy.

This study was a prospective, single-center, single-arm study. BBBD with MRgFUS was performed adjacent to the tumor resection margin on the first or second day of the adjuvant TMZ chemotherapy at the same targets for 6 cycles. T2*-weighted/gradient echo (GRE) MRI was performed immediately after every sonication trial, and comprehensive MRI was performed at the completion of all sonication sessions. Radiological, laboratory, and clinical evaluations were performed two days before each planned BBBD.

From September 2018, 6 patients underwent 145 BBBD trials at various locations in the brain. The authors observed gadolinium-enhancing spots at the site of BBBD on T1-weighted MRI in 131 trials (90.3%), and 93 trials (64.1%) showed similar spots on T2*-weighted/GRE MRI. When the two sequences were combined, BBBD was observed in 134 targets (92.4%). The spots disappeared on follow-up MRI. There were no imaging changes related to BBBD and no clinical adverse effects during the 6 cycles.

This is the first study in which repetitive MRgFUS was performed at the same targets with a standard chemotherapy protocol for a malignant brain tumor. BBBD with MRgFUS was performed accurately, repeatedly, and safely, but a longer follow-up period is needed. A limitation of this study is that it did not quantitate the increase in TMZ concentrations following BBBD. Future investigations should be focused on which MRI method is more appropriate to quantify BBBD. Overall, this study allows for the possibility of employing other therapeutic agents that previously could not be used due to the restriction of the BBB.