JITC Digest March 2021

Other Recent JITC Articles

view other articles from this issue

JITC Editor Picks

In silico trials predict that combination strategies for enhancing vesicular stomatitis oncolytic virus are determined by tumor aggressivity

Adrianne L Jenner, Tyler Cassidy, Katia Belaid, Marie-Claude Bourgeois-Daigneault, Morgan Craig
Journal for ImmunoTherapy of Cancer 2021;9:e001387 (19 February 2021)
Research

Summary:

A major obstacle to translating combination immunotherapies from pre-clinical models into clinically efficacious regimens is tailoring dosing and scheduling to maximize patient benefit, especially for newer modalities such as oncolytic viruses where anti-tumor immunity is mediated by several sometimes overlapping, possibly oppositional mechanisms. Using a computational biology model describing tumor-immune interactions and systemic cytokine concentrations over time, Adrienne L Jenner et al evaluated different dosing regimens for a dual oncolytic virus therapy in cohorts of heterogeneous virtual individuals. Previously, in vivo data has demonstrated that vaccinia virus (VV) enhances the ability of vesicular stomatitis virus (VSV) to replicate in tumor cells through the natural expression of an interferon receptor decoy, while, simultaneously, VSV may augment VV spread via the expression of a recombinant expressing fusion-associated small-transmembrane protein, leading to a ‘ping pong’ oncolytic effect. By simulating all possible treatment protocols, the optimal number of VV doses prior to VSV administration as well as the ideal lag time between delivery of the two viruses was identified for each virtual individual. Importantly, the number of VV doses for optimal outcomes was tightly linked with the intrinsic tumor growth rate for each in silico patient. For example, virtual patients with aggressively growing tumors required a single VV administration followed by VSV injection 1 day later, whereas patients with the slowest growing tumors needed multiple courses of VV, followed by VSV after a delay of 15 days, results that were validated in independent cohorts of entirely new virtual individuals. The study suggests that tumor-intrinsic growth rates may be important for patient segregation in clinical trials, and highlight how computational biology may help bridge the gap between preclinical and human studies for future drug development.

Integrating genome-wide CRISPR immune screen with multi-omic clinical data reveals distinct classes of tumor intrinsic immune regulators

Research

Summary:

Using a genome-wide CRISPR screen integrated with multi-omics clinical data, Jiakai Hou et al provide a rich resource of potential biomarkers and novel targets for tumor-intrinsic immune escape. The drop-out screen leveraged murine colon adenocarcinoma cell line MC38 stably expressing the melanoma antigen gp100 as well as a library of guide RNAs and Cas9 co-cultured with tumor-reactive gp100-specific murine Pmel T cells in effector-to-target ratios of 0.3:1 and 1:1 to represent conditions of weak and moderate T cell-selection pressure. A 7-day in vitro expansion step before co-culture removed clones expressing guide RNAs targeting genes broadly involved in tumor proliferation and survival—essential genes were also identified using data from an earlier genetic vulnerability screen. A total of 139 genes and 199 genes were significantly depleted under conditions of weak and moderate selection pressure, respectively, 55 of which overlapped. Gene ontogeny analysis identified several validated immunosuppressive pathways in the depleted genes, including PI3K signaling and oxidative phosphorylation. Alignment against human genome sequences identified 189 orthologues. Comparison of expression levels of depleted genes in the screen against published lymphocyte infiltration score datasets from The Cancer Genome Atlas (TCGA) allowed for the identification of two broad categories of resistance regulators: dual regulators (expression correlated with infiltration score) and cytotoxicity regulators (expression did not correlate with infiltration score). Knockout cell lines lacking a representative dual regulator, PRMT1, as well as a representative cytotoxicity regulator, RIPK1, both were more sensitive to T cell killing in vitro, although the kinetics were different as visualized via a recently optimized live cell imaging method called time-lapse imaging microscopy in nanowell grids that monitors T cell-induced killing of genetically modified individual tumor cells in real time. In vivo, tumors lacking PRMT1 or RIPK1 were sensitized to anti-PD-1, but not anti-OX40. Intriguingly, pharmacologic inhibition of RIPK1 did not enhance anti-PD-1 anti-tumor effects, but moderate synergy was observed with anti-OX-40.

Intratumoral CXCL13⁺CD8⁺T cell infiltration determines poor clinical outcomes and immunoevasive contexture in patients with clear cell renal cell carcinoma

Siyuan Dai, Han Zeng, Zhaopei Liu, Kaifeng Jin, Wenbin Jiang, Zewei Wang, Zhiyuan Lin, Ying Xiong, Jiajun Wang, Yuan Chang, Qi Bai, Yu Xia, Li Liu, Yu Zhu, Le Xu, Yang Qu, Jianming Guo, Jiejie Xu
Journal for ImmunoTherapy of Cancer 2020;9:e001823 (15 February 2021)
Research

Summary:

CXCL13, a B cell chemoattractant that contributes to tertiary lymphoid structure formation, has been well-characterized as a product of follicular helper T cells, yet the effects of the chemokine on anti-tumor immunity when secreted by CD8⁺ T cells remain unknown. Through a combination of in silico analyses and validation in vitro, Siyuan Dai et al demonstrate an exhausted phenotype in intratumoral CXCL13⁺CD8⁺ T cells, causing impaired effector functions in the total CD8⁺ T cell populations in clear cell renal cell carcinoma (ccRCC). In datasets from 223 patients from Zhongshan Hospital cohort and 535 patients from The Cancer Genome Atlas (TCGA), the levels of infiltration of CXCL13⁺CD8⁺ T cells positively correlated with tumor stage, UCLA Integrated Staging System (UISS) stage, and stage size grade and necrosis score (SSIGN) grade. In three independent cohorts (discovery cohort, internal validation, and TCGA external validation), high levels of tumor-infiltrating CXCL13⁺CD8⁺ T cells were associated with significantly shorter overall survival and recurrence-free survival. Flow cytometry analyses on 42 fresh ccRCC tumor samples showed that CXCL13⁺CD8⁺ T cells expressed high levels of checkpoint proteins, including PD-1, TIM-3, and TIGIT, high proliferative capacity as measured by Ki67, and decreased production of effectors, such as interferon gamma and tumor necrosis factor alpha. Notably, in the larger datasets, high levels of CXCL13⁺CD8⁺ T cells were also associated with an exhausted phenotype in the CD8⁺ population as a whole, similarly characterized by increased checkpoints and decreased effectors. High levels of intratumoral CXCL13⁺CD8⁺ T cells led to a microenvironment with more CD4⁺ T cells, Th2 cells, regulatory T cells, and tumor-associated macrophages, along with fewer natural killer cells and granzyme B positive cells. Merged survival curves showed that CD8⁺ T cell abundance independently predicted prognosis in ccRCC patients with high levels of CXCL13⁺CD8⁺ T cells, but the predictive power was lost for the group with low CXCL13⁺CD8⁺ T cell infiltration. This is the first demonstration of the clinical implications and immune contexture of CXCL13⁺CD8⁺ T cells, opening the door to future prospective studies of the utility of CXCL13 as a biomarker or immunotherapeutic target.

Therapeutic depletion of CCR8+ tumor-infiltrating regulatory T cells elicits antitumor immunity and synergizes with anti-PD-1 therapy

Summary:

Strategies for systemic depletion of regulatory T cells (Tregs) to enhance anti-tumor responses have been limited because all approaches tested to date also unleash severe autoimmunity. Helena Van Damme and colleagues identify the CC chemokine receptor CCR8 as a tumor-infiltrating Treg-specific marker and a potential target to enhance tumor immunity in murine models. Single-cell RNA sequencing revealed two distinct clusters of Tregs infiltrating Lewis lung carcinomas: a subset with prominent expression of activation markers, immune checkpoints, and costimulatory molecules indicative of a highly suppressive phenotype that mainly expressed Cccl8, along with a low-Ccl8-expressing group with expression profiles akin to Th17 cells. Gene expression data were largely confirmed by multi-color flow cytometry across multiple tumor models, including MC38 and B16—namely, no CCR8 was detected on non-hematopoietic or myeloid populations, whereas within the lymphoid lineage, CCR8 was present on approximately 25% and 60% of CD4⁺Foxp3⁺LAG-3^Low and CD4⁺Foxp3⁺LAG-3^High Tregs, respectively. In tumor-naïve mice, CCR8 was largely absent from Tregs isolated from all organs except the thymus, where expression levels were roughly 8%. Lewis lung carcinoma tumor-bearing mice had CCR8-expressing Tregs in the thymus and spleen, yet the highest proportion of CCR8-positive Tregs by far was found in the tumor microenvironment, at 57%. Antigen-specific stimulation was required for splenic and lymph-node Tregs to upregulate CCR8. Reclustering of the RNA sequencing data revealed NF-kappa B as a potential regulator of CCR8 upregulation, and splenocyte stimulation in the presence of the NF-kappa B inhibitor CAPE led to a dose-dependent reduction of CCR8 upregulation on Tregs with minimal impacts on viability. Notably, tumor-bearing mice lacking Ccr8 did not display any notable phenotypic differences from wild type controls, including Treg tumor homing, immune infiltrate activation status, and tumor growth kinetics. Blocking CCR8 with tetravalent nanobodies lacking Fc domains (generated from fusions of four camelid single domain antibody fragments) similarly did not alter tumor growth kinetics. Nanobodies engineered with an autofucosylated Fc domain to enhance antibody-dependent cellular cytotoxicity, by contrast, led to tumor control without any evidence of immune-mediated colitis in lung tumor-bearing mice, and complete rejection of tumors when combined with anti-PD-1. Nanobody-mediated Treg depletion required natural killer cells and, importantly, did not affect peripheral Tregs.

TGF beta drives NK cell metabolic dysfunction in human metastatic breast cancer

Summary:

Emerging evidence suggests that natural killer (NK) cell dysfunction during cancer inhibits tumor immune control. Using single-cell analysis, metabolic flux measurements, and confocal microscopy, Karen Slattery and colleagues identify a novel TGF beta-dependent autocrine regulatory mechanism that causes metabolic dysregulation and impaired effector function in peripheral NK cells from patients with breast cancer. Both responses to stimulation with IL-2 or IL-12/15 as well as direct killing of target cells were markedly impaired in peripheral NK cells from patients with breast cancer (mostly with hormone receptor positive tumors, though other subtypes were included) compared to healthy controls. Ex vivo analysis (without manipulation) revealed evidence of mitochondrial fragmentation in patient peripheral NK cells, including increased mitochondrial mass, elevated levels of reactive oxygen species, and punctate morphology. Additionally, patient peripheral NK cells showed defects in cytokine-induced metabolic reprogramming compared to healthy controls, including significantly reduced glycolytic capacity and reduced flux through the oxidative phosphorylation pathway as well as lower levels of mammalian target of rapamycin complex 1 (mTORC1) after IL-2 stimulation. Despite no differences in systemic TGF beta levels, peripheral NK cells from patients trended toward elevated amounts of phosphorylated SMAD2/3. Notably, compared to controls, NK cells from patients with metastatic breast cancer expressed significantly higher levels of the transmembrane protein GARP, which has been described to play an autocrine function in regulatory T cells by sequestering TGF beta at the cell surface. High GARP expression was associated with higher TGF beta signaling across NK cell subsets. Neutralization of TGF beta in vitro restored CD25 expression and mTORC1 levels in response to cytokine stimulation to levels seen in healthy controls in patient-derived peripheral NK cells. Patient-derived NK cells treated with anti-GARP antibodies phenocopied TGF beta neutralization, including rescued expression of activation and cytotoxicity markers, as well as upregulated metabolic flux through oxidative phosphorylation. The findings not only demonstrate that cancer causes systemic NK cell dysregulation, but also support the notion of GARP as a target for immunotherapy.