Characterization of systemic immunosuppression by IDH mutant glioma small extracellular vesicles

Abstract

Background

Gliomas are the most common primary brain tumors and are universally fatal. Mutations in the isocitrate dehydrogenase genes (IDH1 and IDH2) define a distinct glioma subtype associated with an immunosuppressive tumor microenvironment. Mechanisms underlying systemic immunosuppression in IDH mutant (mutIDH) gliomas are largely unknown. Here, we define genotype-specific local and systemic tumor immunomodulatory functions of tumor-derived glioma small extracellular vesicles (TEX).

Methods

TEX produced by human and murine wildtype and mutant IDH glioma cells (wtIDH and mutIDH, respectively) were isolated by size exclusion chromatography (SEC). TEX morphology, size, quantity, molecular profiles and biodistribution were characterized. TEX were injected into naive and tumor-bearing mice, and the local and systemic immune microenvironment composition was characterized.

Results

Using in vitro and in vivo glioma models, we show that mutIDH TEX are more numerous, possess distinct morphological features and are more immunosuppressive than wtIDH TEX. mutIDH TEX cargo mimics their parental cells, and induces systemic immune suppression in naive and tumor-bearing mice. TEX derived from mutIDH gliomas and injected into wtIDH tumor-bearing mice reduce tumor-infiltrating effector lymphocytes, dendritic cells and macrophages, and increase circulating monocytes. Astonishingly, mutIDH TEX injected into brain tumor-bearing syngeneic mice accelerate tumor growth and increase mortality compared with wtIDH TEX.

Conclusions

Targeting of mutIDH TEX represents a novel therapeutic approach in gliomas.

Isocitrate Dehydrogenase gene (IDH) mutations drive multiple solid malignancies including a subset of gliomas.¹ Oncogenic mutations in IDH1 or IDH2 produce single nucleotide substitutions that convert arginine to histidine at residues 132 and 172 of IDH1 and 140 of IDH2 tumors. Epigenomic reprogramming is responsible for the induction of the glioma-CpG island methylator phenotype (g-CIMP) in IDH gliomas.² Wildtype IDH (wtIDH) and mutant IDH (mutIDH) gliomas have emerged as distinct clinic-pathologic entities.³ Tumors with the IDH mutation exhibit unique genetic landscapes and better survival outcomes than wtIDH gliomas.^3–5 Nevertheless, mutIDH gliomas remain universally fatal despite maximum medical and surgical therapy.⁵ MutIDH gliomas are typically discovered as lower-grade neoplasms and are less aggressive than their high-grade counterparts. They may lie dormant for several years before progressing to high-grade (WHO Gr. III/IV) gliomas. The persistence of mutIDH glioma dormancy suggests ineffective tumor immune surveillance. We and others have reported reduced anti-tumor cytotoxic effector lymphocyte immunity within the tumor-infiltrating immune compartment of mutIDH compared to high-grade gliomas, revealing that mutIDH glioma are inherently equipped to escape immune surveillance.^6–11

Mechanisms underlying immune escape in mutIDH glioma are both tumor cell-intrinsic or tumor cell-extrinsic mechanisms.^6–8 While tumor cell-intrinsic immune escape is mediated by epigenetic dysregulation of anti-tumor immune pathways, mechanisms of cell-extrinsic immune escape (including reduced immune cell infiltration) remain obscure in mutIDH glioma. The physiologic origins and biodistribution properties of tumor-derived small extracellular vesicles (TEX) make them an attractive candidate for modulating systemic tumor immune responses. TEX produced by tumor cells are 30 to 150 nm and mediate intercellular communication.¹² Studies in other malignancies reveal that TEX-mediated immunosuppressive effects depend on their immunosuppressive cargo and the intercellular delivery of these cargo components induces pro-tumorigenic reprogramming of immune cells, ultimately stimulating tumor progression.^13–15 TEX contain a spectrum of proteins, nucleic acids, and lipids that mimic their tumor cells of origin.¹⁶ In essence, TEX are a transport system that enables cell-cell communication by shuttling messages from the tumor to cells residing tumor-adjacent or at distant locations, including bone marrow and the lymphatic system.¹⁷ TEX can mediate juxtacrine, paracrine, and endocrine-like signals to drive tumor progression.^18–21 Beyond immunomodulation, TEX also modulate oncogenic functions of autologous tumor cells via autocrine signaling.²² To date, immunomodulation of mutIDH gliomas by TEX has not been studied.

Using patient glioma stem cell (GSC) cultures from our annotated glioma bank, we cultured mutIDH and wtIDH GSCs for TEX extraction. TEX derived from supernatants of early passage GSCs and other literature-validated mutIDH/wtIDH GSC cell lines²³ were used to assess immunomodulatory characteristics in vitro and in vivo. Here, we report that TEX from mutIDH GSCs carry an immunosuppressive cargo and negatively reprogram local and circulating immune cells in tumor-naïve and glioma-bearing mice. We show that mutIDH TEX are particularly effective at inducing suppressive myeloid cell types. Our work describes a hitherto unknown mechanism of systemic immunosuppression in mutIDH gliomas.

Materials and Methods

Cell Lines and Cell Culture

Previously characterized glioma stem cells TS603 (mutIDH) and TS667 (wtIDH) were obtained from Dr. Timothy A. Chan (Cleveland Clinic, Cleveland OH).²³ Cells were grown in Neurocult™ (Stemcell Technologies, Vancouver) as previously described.⁸

C57BL/6 mice were implanted in the frontal cortex with 10⁵ mutIDH or wtIDH glioma cells generated as described previously.⁶ Brain tumors were harvested, dissociated into single-cell suspensions, and grown in DMEM (Lonza Inc.) supplemented with 1% (v/v) penicillin/streptomycin and 10% (v/v) heat-inactivated FBS (Gibco, Thermo Fisher Scientific). FBS was depleted of small extracellular vesicles (sEVs) by ultracentrifugation (100 000×g, 3h). Study was approved by our institutional IRB and IACUC.

TEX Isolation

TEX were isolated from cell supernatants using SEC as we previously described.^24,25 TS603/TS667 supernatants were decanted after 72 h. For mouse glioma cultures, 4 × 10⁶ mutIDH or wtIDH cells were cultured (72 h, 150 cm² cell culture flasks in 25 mL of culture medium).²⁵ Supernatants were centrifuged at room temperature (RT) at 2000×g and again at 10 000×g at 4°C for 30 min. Final suspensions were filtered using 0.22 µm filters. 25 mL aliquots were concentrated using Vivacell 100. Concentrated supernatant (1 mL) was eluted with a 10 cm-long Sepharose 2-B column using PBS. Fraction #4, containing non-aggregated morphologically intact TEX, was used in all subsequent experiments.

TEX Characterization

TEX were characterized using MISEV2018 guidelines.²⁶ Protein concentrations were determined with BCA assay (Pierce Biotechnology). TEX were visualized by TEM.²⁴ Size distributions/particle concentrations were analyzed with tunable-resistive pulse sensing (TRPS, qNano, Izon).²⁰ Samples were measured without dilution (NP100, stretch 45.64 mm, voltage 0.7 V; two pressure steps of 3 to 7 mbar). Particle calibration (Part#: CRC100b, mean diameter: 114 nm, dilution: 1:1000) was assessed under identical conditions. Fraction #4 samples were concentrated for western blotting (0.5 mL 100K Amicon Ultra centrifugal filters, EMD Millipore, 4000×g) at 5 μg protein/lane. PVDF membranes were incubated o/n at 4 °C with primary antibody as described²⁰ (see comprehensive antibody list, Supplementary Table 1). Band intensities (BI), band area (BA), and background values (BV) were quantified with ImageJ (http://rsbweb.nih.gov/ij/). Integrated pixel value = (BI × BV)–BV.

Quantitative Real-Time PCR

TEX RNA was extracted with Total Exosome RNA & Protein Isolation Kit (Invitrogen, No. 4478545) using the manufacturer’s instructions. Purity was evaluated with Nanodrop. cDNA was reverse-transcribed (iScript^TM, Bio-rad, No. 178890). Real-time PCR was performed on a CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad) in 96-well reaction plates. Reactions were run in triplicates. The cycle number at which the reaction crossed an arbitrarily placed threshold (C_T) was determined for each and compared to 18S rRNA (control) using equation 2^−ΔC_T where ΔC_T = (C_TtargetRNA−C_{T18S rRNA}).

Bulk RNA Sequencing

TEX RNA was extracted as above. 3 technical replicates were included per group. RNA-Seq processing (including library prep/q.c.) was performed at the Pitt Health Sciences Genomic Core according to core standard operating procedures using Illumina NextSeq 550 system (800 000 reads/sample). RNA q.c. was performed on Agilent tapestation. RNA library prep utilized SMARTer Seq Total RNA pico mammalian v2 LT (<48) kit (Takara Bio) according to standard protocols (no PolyA selection). Differential gene expression (DGE) was determined using the CLC Genomics Workbench 20 (Qiagen). All samples were subjected to Trimmed Mean of M values (TMM) normalization. Generalized linear model (GLM) was used to calculate differential expression. FDR P-value <.05 was used as a DGE cutoff based on sample size (Pitt Genomics Research Core protocol). Pathway analysis of DGEs was performed with Ingenuity Pathway Analysis software (Qiagen). For pathway/network analysis, FDR p-value cutoff of 0.03 was used.

TEX Injections in Tumor Naive Mice

In vivo injections: 20 μg of concentrated mouse mutIDH (n = 4) or wtIDH (n = 3) glioma cell-derived sEVs or PBS (n = 3) was injected (retro-orbital) into six-week-old immunocompetent C57BL/6 mice every 72 h, for two weeks as previously described.¹⁴ PBMCs were extracted via tail-vein before treatment, at one week and two weeks post-treatment.

Xenogeneic Glioma Mouse Model (Flank)

Six-week-old Balb/c/Nu mice were injected with 10⁶ human mutIDH or wtIDH. In cohort #1, mice were sacrificed 30 days after tumor cell inoculation. PBMCs/tumor-infiltrating lymphocytes (TILs) were obtained from mice bearing mutIDH (n = 3) and wtIDH (n = 3) tumors. In cohort #2, tumor-bearing mice were injected with 20 μg of human TEX derived from mutIDH (n = 3) or wtIDH (n = 3) cells or PBS (n = 3) as above. Mice were randomized and treatments began at ~25 mm² tumor size. PBMCs were isolated as above. Tumor size was measured with calipers every 3 days. Mice were sacrificed two weeks after experiment inception and TILs isolated.

Syngeneic Orthotopic Mouse Glioma Model

Sleeping Beauty (SB) Transposase System was used to generate genetically engineered immunocompetent wtIDH and mutIDH mouse models of glioma harboring ATRX and TP53 loss as recently described (gift from Maria Castro, University of Michigan).²⁷ See Supplementary Methods.

Evaluation of the Biodistribution of Fluorescently- and Radiolabeled TEX in Mice

Detailed protocol for TEX labeling and biodistribution studies¹⁷ described in Supplementary Methods.

Flow Cytometric Analysis of Tumor-Infiltrating Lymphocytes (TILs) and PBMCs

TILs and PBMCs were isolated from tumor naïve and tumor-bearing mice and analyzed by flow cytometry as described in Supplementary Methods.

Statistical Analysis

Data were analyzed using GraphPad Prism software (v7.0). Values are expressed as means ± SEM (see figure legends). Differences between groups were assessed by Student t test or one‐way analysis of variance (ANOVA). Student–Newman–Keuls post hoc tests were used to calculate differences between pairs of groups. Statistical significance set at P < .05.

Results

MutIDH Glioma Cells Produce More TEX Than wtIDH Gliomas

Analysis of the SEC #4 fractions of wtIDH or mutIDH glioma cell supernatants by TEM showed that TEX featured vesicular morphology and were 30 to 150 nm in diameter (Figure 1A and B). TEX carried sEV markers TSG101, CD9, and Alix irrespective of genetic alterations (Figure 1C). Grp94, a negative sEV marker, was absent in SEC fraction #4 (Figure 1C). Glioma TEX carried a variety of immunosuppressive proteins including PD-L1, CD39, CD73, and CTLA-4, independent of the mutation status of the primary cells (Figure 1C). mutIDH TEX were enriched in the adenosinergic pathway components CD39/CD73 as well as immunosuppressive proteins TGF-β and TRAIL (Figure 1D). Mean sizes of wtIDH and mutIDH TEX were not significantly different (mutIDH = 86 nm; wtIDH = 82 nm; range 50 to 150 nm) (Figure 1E). TEX particles were more numerous in supernatants of mutIDH vs. wtIDH cells (1.60E+12 particles/mL/3.08E+11 particles/mL, respectively; P < .05; Figure 1E and G). mutIDH TEX contained higher overall protein content (P < .05; Figure 1F).

wtIDH or mutIDH Balb/c/Nu subcutaneous xenograft models were used for in vivo validation of TEX morphology and content. Circulating total sEVs in the plasma were extracted when gross tumor size reached 200 mm². mutIDH glioma-bearing mice showed increased circulating sEVs (P < .05; Figure 1H). These results were further validated with additional patient-derived mutIDH (n = 2) and wtIDH (n = 1) xenograft models (P < .01; Figure 1I).

TEX Cargo Reflect Parental Cell Gene Expression

We next investigated gene expression differences in TEX cargo from each glioma genotype using RNA-Seq. Statistically significant global differences between wtIDH and mutIDH TEX gene expression were observed (volcano plot, Figure 2A; P < .05). We performed pathway analysis of genes with a particular focus on genes downregulated in mutIDH relative to wtIDH TEX (consistent with epigenetic transcriptional repression). Intriguingly, gene expression associated with immune cell trafficking was robustly and significantly downregulated in mutIDH TEX, suggesting that mutIDH TEX may modulate immune cell chemotaxis from the peripheral circulation into the local TME (Figure 2B). Molecular interactions and the list of differentially regulated genes associated with immune cell trafficking in mutIDH TEX are shown in Figure 2C and Supplementary Table 2, respectively.

Next, we interrogated total DGEs gleaned from the RNA-Seq analysis. We focused on interrogating expression levels of immune-related genes (NKG2DLs, retinoic acid pathway regulators, and CCL2 chemokines) corresponding to the CIMP signature, to investigate recapitulation of parental immunosuppressive pathways.² These genes have been validated in vivo and ex vivo in both mouse and human mutIDH glioma.^6,8,28 In general, all genes analyzed in both TEX genotypes mimicked the gene expression levels observed in the glioma parent cells (Figure 2D). Notably, downregulated genes in mutIDH TEX included activating NK ligands ULBP1 and ULBP3; immune homeostasis regulator RBP1; and the CCL2 chemokine and was consistent with previously defined mutIDH-induced immunosuppressive genes (Figure 2D, P < .05). Interestingly, transcriptional repression of the CCL2 chemokine and RBP1 (both significant modulators of immune suppression in mutIDH glioma) was more pronounced in TEX compared to parental cells (Figure 2D). Although the pathophysiologic consequence of this phenomenon is unclear, it appears that TEX cargo exaggerates specific immunosuppressive messages generated by mutIDH cells. Our data suggest that mutIDH TEX may be couriers for the delivery of immune-related genetic signals from glioma cells to local and distant components of the TME.

TEX Derived from mutIDH Cells Cross the Blood-Brain Barrier and Show a Distinct Biodistribution Profile

To test whether glioma-derived TEX cross the blood-brain barrier (BBB) and interact with cells in distant organs, genotype-specific TEX were fluorescently labeled and injected intracerebrally into 4-day old mice. Fluorescent signals were analyzed with bioluminescence (IVIS) after 20h. We discovered that the distribution of TEX was ubiquitous regardless of genotype (Figure 3A). Spleen and hindlimbs/sternal marrow, major production sites of immune cells, displayed markedly pronounced uptake of TEX (Figure 3A).

For quantitative analysis of TEX biodistribution, radiolabeled TEX were injected into the jugular vein of mice and unidirectional blood-tissue uptake was determined (See Supplementary Materials). Capillary depletion was utilized to assess whether radioactive material was sequestered in capillaries or crossed the BBB into brain parenchyma. All TEX showed a good brain penetration, with ~80% brain parenchymal uptake and ~20% capillary retention. There were no genotype-dependent differences in BBB penetration (Figure 3B). Analysis of the TEX blood clearance revealed an early clearance phase followed by a steady-state phase for both TEX genotypes. Notably, however, mutIDH TEX showed larger volumes of distribution as evidenced by a lower value for %Inj/ml at t = 0, whereas the steady-state was similar compared to the wtIDH TEX (Figure 3C). Quantification of TEX uptake rate showed significantly more rapid penetration into brain, liver, lung, and spleen (p < .05) in mutIDH TEX compared to wtIDH counterparts (Figure 3D-H).

To assess the effect of global innate immune activation on TEX uptake, radiolabeling experiments were repeated after systemic LPS infection. Rates of mutIDH TEX entry into the brain and lung were significantly increased under these conditions, while spleen uptake was reduced. For wtIDH TEX, LPS induced increased uptake in the liver and spleen (Supplementary Figure S1A). The data suggest that TEX produced by glioma cells display genotype-specific tissue uptake in a manner modulable by systemic inflammatory states.

Recent studies reveal that capillary penetration of TEX occurs via adsorptive transcytosis utilizing wheat germ agglutinin (WGA) glycoproteins or the mannose-6 phosphate (M6) receptor.¹⁷ To determine which glycoprotein receptor mechanism predominates glioma-derived TEX capillary penetration, co-injection of WGA or M6P with TEX was performed. WGA significantly decreased liver uptake and increased brain, kidney, and lung uptake of mutIDH TEX while increasing uptake by brain and lung in wtIDH TEX (Supplementary Figure S1B). Notably, M6P did not affect the uptake of either TEX by any tested tissue and no significant differences were observed (Supplementary Figure S1B). Thus, penetration of the BBB by glioma TEX appears at least partially dependent on WGA.

TEX Generated by mutIDH Cells Increase Circulating Monocytes and Decrease Regulatory T Cells in Tumor Naive Mice

We next examined genotype-specific TEX effects on circulating immune cells in naive animals by injecting wtIDH or mutIDH TEX intravenously (every 72h over 2 weeks). PBMCs were harvested before, during, and after TEX delivery. Circulating monocyte proportions were significantly increased in mutIDH TEX-treated mice relative to wtIDH (P < .01; Figure 4F). Interestingly, no differences in immune cell proportions were observed after one week of TEX delivery, suggesting that reprogramming of circulating immune cells might involve non-instantaneous, complex biological mechanisms such as epigenetics. Decreased circulating regulatory T cell proportions were also observed in mutIDH TEX-treated mice (P < .01; Figure 4B). There was a trend towards decreased circulating dendritic cells in mice receiving mutIDH TEX (P = .0571; Figure 4D). No changes were observed in other circulating immune cells injected with either wtIDH or mutIDH TEX (Figure 4). These results suggest that mutIDH TEX contain the machinery necessary to reprogram certain subsets of immune cells (particularly monocytes) in the absence of malignancy and that prolonged exposure to mutIDH TEX may bias hematopoiesis towards global monocytic but not lymphocytic lineages.

It is unknown whether mutIDH TEX-induced immune cell reprogramming occurs at the level of the bone marrow or within existing circulating immune cells (Figure 4G and H). We explored the impact of glioma TEX on immune cell differentiation as an ex vivo substitute for a circulating immune cell model. We utilized naive PBMCs and PKH67-labeled wtIDH and mutIDH TEX to determine differential uptake of glioma TEX by disparate immune cell subsets ex vivo (see methods). Flow cytometric analysis 16h post-labeling showed equal uptake of both PKH67-labeled wtIDH and mutIDH TEX by total CD45(+) cells (Supplementary Figure S5A). We demonstrate that mutIDH TEX uptake by PBMCs is remarkably suppressive myeloid biased ex vivo. Indeed, proportions of phenotypically suppressive myeloid cells (monocytes, M2 macrophages, and mono-MDSCs) were markedly increased after mutIDH TEX internalization, whereas M1 macrophages were significantly decreased (Supplementary Figure S5B). These data, which require cautious interpretation, given the synthetic conditions, suggest that mutIDH TEX may impact monocyte survival and/or differentiation into mono-MDSCs.

TEX Derived From mutIDH Cells Reprogram the Immune Landscape in Tumor-Bearing Mice

wtIDH and mutIDH glioma harbor distinct immune TMEs (Supplementary Figures S2 and S3),^6,8 but the contribution of TEX to the immune TME is unknown. We first used a xenogeneic (human mutIDH/wtIDH cell), murine model, to elucidate whether the effect of exogenous TEX administration modulates the immune TME in a genotype-specific manner. wtIDH or mutIDH tumor-bearing mice were treated with either PBS or mutIDH and wtIDH TEX (see Methods). Although no significant differences in the tumor growth kinetics were observed in this model (Figure 5A), we observed substantial TEX genotype-dependent changes in the composition of the immune landscape (Figure 5B and C). wtIDH glioma-bearing mice treated with mutIDH TEX demonstrated a significant reduction of both circulating and infiltrating NK, DC, and macrophage proportions (P < .05; Figure 5B and C). mutIDH TEX also significantly increased circulating monocyte and mono-MDSC proportions under these conditions (Figure 5C, P < .05). In contrast, mutIDH glioma-bearing mice exhibited reduced circulating monocyte and mono-MDSC proportions. wtIDH TEX had no effect on circulating NK cell, DC, macrophage, and PMN-MDSC proportions in mutIDH tumor-bearing mice (Figure 5C). These data imply that exogenous TEX overwhelms the biological effects of endogenous TEX produced by the existing tumors.

Genotype-Specific TEX Alter the Local and Circulating Immune Microenvironment and Impact Animal Survival in Syngeneic Murine Glioma

To examine effects of genotype-specific TEX in a fully immunocompetent intracranial model, we turned to the SB transposase (C57BL/6) orthotopic glioma (see Methods for details).²⁷ 14 days post-implantation wtIDH and mutIDH tumor-bearing mice were either treated with PBS (n = 5), mutIDH, or wtIDH TEX (n = 7). We observed more robust effects of glioma TEX in this syngeneic model compared with our Balb-C/Nu model, emphasizing a role for TEX-mediated adaptive immune reprogramming (Figure 6). mutIDH and wtIDH TEX adversely affected the survival of wtIDH and mutIDH tumor-bearing mice relative to the PBS treated mice (P = .0009 and P = .0016, respectively; Figure 6A). Surprisingly, the worst survival outcomes were observed in wtIDH glioma with mutIDH TEX (Figure 6B and C). Our analysis of TILs mirrored our xenogeneic model wherein mutIDH TEX increased tumor-promoting immune cell populations, including Tregs and immunosuppressive myeloid populations, including mono-MDSCs (Figure 6B) while decreasing lymphocyte effector immune cell populations, such as NK cells (relative to PBS cohort, Figure 6B). Conversely, mutIDH tumors treated with wtIDH TEX showed increased effector lymphocytes and M1 macrophages, while reducing suppressive myeloid cells (Figure 6B). TEX also reprogrammed the peripheral immune landscape in a genotype-specific manner. NK cells, DCs, and macrophages were reduced in wtIDH tumor-bearing mice after injection of mutIDH TEX and the inverse effect was observed in the mutIDH tumor-bearing mice treated with wtIDH TEX (Figure 6C). M2 macrophages, monocytes, and MDSCs increased with mutIDH TEX treatment, whereas the opposite effect was observed for the wtIDH TEX relative to the respective PBS cohorts (Figure 6C).

Our findings validate a powerful genotype-driven immune regulatory effect of TEX in local and circulating immune TMEs of multiple glioma models. Surprisingly, mutIDH TEX encourage a globally tumor-permissive TME.

Discussion

In this study, we set out to address the potential impact of genotype-restricted TEX on the immune pathogenesis of gliomas. This study provides a number of important insights into differential effects of mutIDH TEX vs. wtIDH TEX and identifies remarkably tumor-permissive functions of mutIDH TEX characterized by increased TEX production; support of suppressive myeloid cell persistence; recapitulation of parental immunosuppressive genetic cargo; robust transcytosis across capillary vasculature that constitute the BBB; and acceleration of tumor growth. The surprisingly immunosuppressive capacity of peripherally injected mutIDH TEX and their capacity to accelerate wtIDH gliomas in a syngeneic model indicates their important role in modulating the anti-tumor immune response. While mutIDH tumor cells are less proliferative, their systemic immune evasion machinery including TEX, are more robust in certain respects compared to wtIDH glioma. This paradox might explain why mutIDH glioma cells, which are less intrinsically oncogenic than wtIDH glioma cells, persist indefinitely. It may also explain why aggressive surgical resection of mutIDH tumors (which theoretically reduces TEX production) produces superior survival advantages compared with wtIDH gliomas.

Sera from glioma patients contain IDH mRNA in sufficient quantities to differentiate wtIDH and mutIDH gliomas.^29,30 Ebrahimkhani et al. validated genotype-driven differences in glioma using deep sequencing of microRNA (miR) in plasma-derived sEVs. This effort identified 23 dysregulated miRs specific to mutIDH.³¹ However, the functional consequences of glioma sEV/TEX genotype specificity remained unknown.

mutIDH gliomas feature derangement of cellular metabolism²³ and high oxidative stress,^32,33 conditions known to stimulate TEX release.^34–36 Moreover, metabolic stress induced by therapeutic intervention in glioma may induce TEX release.^34,37,38 It is feasible that elevated production of TEX and the accelerated biodistribution of mutIDH TEX can be explained by intrinsically higher oxidative stress caused by IDH mutations.³²

mutIDH gliomas exhibit reduced infiltration with NK and CD8⁺ T cells compared to wtIDH gliomas and are resistant to innate cytotoxic immune mechanisms that define immune surveillance.^6,7 Such gliomas also exhibit cell-intrinsic transcriptional repression of NKG2D ligands.⁸ mutIDH TEX are well-positioned to explain the recapitulation of tumor cell-intrinsic regulatory immune signals in the local and circulating immune TME. Although immunosuppressive functions of TEX are well known, our investigation illustrates genotype-dependent differences in TEX that were previously unknown, and further, granular mechanistic studies may identify specific targetable mechanisms that disrupt the courier functions of mutIDH TEX.

Emerging data suggest that TEX mediate chemotactic effects in a broad swath of immune cells, including those circulating outside the BBB.³⁹ Immune regulatory roles of glioma TEX in the peripheral circulation is supported by the constant production of TEX by gliomas, the ease with which TEX traverse the BBB, and phenotypic/functionally modulation of circulating immune cells by TEX.¹⁷ Further, the cargo components of glioma TEX, which include chemokine/cytokine protein and mRNA, support its capacity to influence immune chemotaxis.^17,40 A remarkably constant feature of mutIDH gliomas is the transcriptional repression of CCL2, a chemokine with broad chemotactic targets and functions.⁶ We show that mutIDH TEX reflect genetic material of their parental cells and feature decreased expression of CIMP genes involved in immune cell trafficking, including CCL2. Further, our TEX biodistribution studies indicate that the target effects of TEX can be ubiquitous and can target anatomic sites with abundant immune cells (marrow, spleen). Interestingly, mutIDH TEX have larger volumes of distribution (Figure 3C) and are more rapidly incorporated into the brain, liver, kidney, lung, and spleen (Figure 3D-H). In conjunction with the fact that mutIDH gliomas produce more numerous TEX, we can feasibly hypothesize that mutIDH TEX are important modulators of immune suppression. This hypothesis is bolstered by studies shown that TEX educate bone marrow, reprogram the cancer progenitor cells, and contribute to tumor progression and metastasis.⁴¹

In summary, this exploratory study indicates that mutIDH TEX shape the immune cell composition of systemic and local immune landscapes. This effect might be orchestrated by multiple pathways, including direct TEX-immune cell interactions, attenuation of immune cell chemotaxis, and systemic reprogramming of the immune function.

Acknowledgments

N.L. was supported by the Leopoldina Fellowships LPDS 2017-12 and LPDR 2019-02, German National Academy of Sciences. T.L.W. was supported by the National Institutes of Health grant U01-DE029759. N.A. was supported by the Karp Foundation and United Initiative to Cure Brain Cancer.

Conflict of interest statement. None.

Authorship statement. N.L. performed TEX isolation/characterization. S.S.Y., A.D.S., K.M.B., K.M.H., and W.A.B. performed the biodistribution (fluorescent/radiolabeled TEX). A.R., P.S., and J.A. performed flow cytometry. A.R., J.A., and X.Z. performed RNA-Seq. C.S.H. performed western blotting. N.L., A.R., W.A.B., P.S., X.Z., K.G., T.L.W., and N.M.A. performed data analysis. N.L., X.Z., A.R., P.S., W.A.B., K.G., T.L.W., and N.M.A. performed manuscript writing. N.L., W.A.B., T.L.W., and N.M.A. provided funding.

References

Author notes