Tumor cell networks formed by tumor microtubes (TMs) may play a key role in the development of therapy resistance in glioblastoma (GB). TM-mediated detoxification from radiation-induced reactive oxygen species (ROS) may infer radioresistance. We hypothesize that high linear energy transfer (LET) radiation, which describes the amount of energy deposited by radiation per unit length, interacts directly with the DNA backbone to induce complex lesions and thus might be less dependent on TM-mediated resistance mechanisms. Therefore, we sought to systematically investigate the impact of LET-induced complex DNA damage on TM formation and GB survival. To this end, the formation of TMs, radiation-induced nuclear DNA damage repair foci (RIF), and GB survival were correlated with a gradual increase in LET using a dose series of clinical protons (low), helium (intermediate), and carbon (high) ion beams. Consistent with conventional photon/X-rays, low-LET proton irradiation promoted TM formation in a dose-dependent manner. In contrast, an anti-correlation between LET and TM induction was found, i.e., a decreased network connectivity with gradual increase of LET and formation of complex DNA damage. Consequently, LET increase correlated with reduced cell survival, with the most pronounced cell killing observed after high-LET carbon irradiation. Moreover, the inverse correlation between LET and TM density was further confirmed for a broad range of LET modulated within the carbon ion irradiation spectrum. This is the first report on the relevance of LET as a novel mean to overcome TM network-mediated radioresistance in GB, with ramifications for the clinical translation of high-LET particle radiotherapy to further improve outcome in this still devastating disease. ### Competing Interest Statement JD reports grants from CRI The Clinical Research Institue GmbH grants from View Ray Inc., grants from Accuray International Sarl, grants from Accuray Incorposrated, grants from RaySearch Laboratories AB, grants from Vision RT limited, grants from Merck Serono GmbH, grants from Astellas Pharma GmbH, grants from Astra Zeneca GmbH, grants from Siemens Healthcare GmbH, grants from Merck KGaA Accounts Payable, grants from Solution Akademie GmbH, grants from Ergomed PLC Surrey Research Park, grants from Siemens Healthcare GmbH, grants from Quintiles GmbH, grants from Pharmaceutecal Research Associates GmbH, grants from Boehringer Ingelheim Pharma GmbH Co, grants from PTW-Freiburg Dr. Pychlau GmbH. AA report grants and other from Merck and EMD, grants and other from Fibrogen, other from BMS, other from BioMedX, other from Roche, outside the submitted work. All other authors declare no competing interests.

Protein kinases control most cellular processes and aberrant kinase activity is involved in numerous diseases. Here we introduce molecular recorders of kinase activities for later analysis to investigate the link between specific kinase activities and cellular phenotypes in heterogeneous cell populations and in vivo. Based on split-HaloTag and a phosphorylation-dependent molecular switch, our recorders become rapidly labeled in the presence of a specific kinase activity and a fluorescent HaloTag substrate. The kinase activity in a given cell controls the degree of fluorescent labeling, whereas the recording window is set by the presence of the fluorescent substrate. We designed specific recorders for four protein kinases, including protein kinase A. We apply our protein kinase A recorder to sort heterogeneous cell populations for subsequent transcriptome analysis, in genome-wide CRISPR screens to discover regulators of PKA activity and to track neuromodulation in freely moving mice.

BACKGROUND:Epstein-Barr virus (EBV)+ and EBV- primary CNS lymphomas (PCNSL) carry distinct mutational landscapes, but their transcriptional and epigenetic profiles have not been integrated and compared. This precludes further insights into pathobiology and molecular differences, relevant for classification and targeted therapy. METHODS:Twenty-three EBV- and 15 EBV+ PCNSL, histologically classified as diffuse large B-cell lymphomas, were subjected to RNA-sequencing and EPIC methylation arrays. Unsupervised clustering analyses were performed. Differentially expressed and differentially methylated genes were identified and integrated. RESULTS:Two distinct transcriptional clusters were found, which separated EBV- and EBV+ PCNSL (P < .0001). The EBV+ transcriptional signature contained genes (GPR15, FCER2/CD23, SLAMF1/CD150) closely regulated by EBV oncogenes in B cells. Pathway enrichment analysis uncovered enhanced B-cell receptor (BCR) and WNT/beta-catenin signaling in EBV- lymphomas, whereas Interleukin-10, NOTCH, and viral life cycle pathways were upregulated in EBV+ PCNSL. Correspondingly, BCR-associated SYK kinase activity was enriched in EBV- tumors while JAK2 was overrepresented in EBV+ PCNSL. Epigenetic profiling revealed reduced global promoter methylation in EBV+ PCNSL. Two methylation clusters were recognized, which separated EBV- and EBV+ PCNSL (P < .0001). Epigenetic profiles were distinct from 2,788 other brain tumor and nonmalignant reference samples. Promoter region hypermethylation of CD79B, a BCR subunit critical for sustained proliferation in EBV- disease, highly correlated (R = -0.7) with its transcriptional downregulation in EBV+ PCNSL. CONCLUSIONS:EBV+ and EBV- PCNSL harbor distinct transcriptional and epigenetic profiles, corroborating them as distinctive biological subtypes. Uncovered differences provide novel insights into their pathobiology, may guide molecular diagnostics and targeted therapies.

Introduction Glioblastoma (GBM) progression and therapeutic resistance are significantly influenced by complex interactions between tumor cells and the brain microenvironment, particularly neurons. However, studying these interactions in physiologically relevant conditions has remained challenging due to limitations in existing model systems. Objectives Here, we present hGliCS (human glioma-cortical spheroid), a novel fully human brain tumor model that overcomes key limitations of current approaches by combining patient-derived GBM cells with mature human cortical neurons derived from induced pluripotent stem cells. Results We demonstrate that GBM cells in hGliCS develop three critical hallmark features observed in patients: (i) formation of tumor microtubes enabling intercellular communication, (ii) establishment of neuron-glioma synapses, and (iii) development of an interconnected network with coordinated calcium signaling. Single-cell RNA sequencing reveals that tumor cells in hGliCS exhibit cellular heterogeneity and transcriptional profiles remarkably similar to those observed in mouse xenografts, including activation of key oncogenic pathways and neuronal-like features. Notably, while GBM cells showed substantial transcriptional adaptation to the neural environment, neurons maintained their core identity with only subtle alterations in glutamate signaling and structural gene expression. We validate hGliCS as a drug screening platform by demonstrating resistance patterns to standard chemotherapy and radiation similar to clinical observations. Furthermore, we show the model’s utility in testing standard and novel therapeutic compounds targeting cell proliferation and tumor-specific neurobiological features, respectively. Conclusion This physiologically relevant human model system provides new opportunities for studying GBM biology and tumor-neuron interactions in a controlled environment. By bridging the gap between simplified in vitro systems and complex in vivo models, hGliCS represents a promising platform for therapeutic development and personalized medicine approaches in GBM treatment.

BACKGROUND AND OBJECTIVES:Neurolymphomatosis (NL) refers to lymphomatous infiltration of the peripheral nervous system (PNS). NL diagnosis and treatment are challenging given the broad differential diagnosis of peripheral neuropathy, the lack of larger cohorts, and the subsequent unavailability of prognostic factors or consensus therapy. This study aimed to define characteristics and prognostic factors of NL. METHODS:A systematic review of the literature (2004-2023) was performed using PubMed and Scopus databases and reported following PRISMA guidelines. Studies reporting individual patient data on cases with definitive NL diagnosis were included. Clinical, radiologic, pathologic, and outcome information were extracted. Univariable and multivariable survival analyses were performed using log-rank tests and Cox proportional hazard models. RESULTS:A total of 459 NL cases from 264 studies were accumulated. NL was the first manifestation of malignancy (primary NL) in 197 patients. PNS relapse of known non-Hodgkin lymphoma (secondary NL) occurred in 262 cases after a median 12 months. NL predominantly presented with rapidly deteriorating, asymmetric painful polyneuropathy. Infiltrated structures included peripheral nerves (56%), nerve roots (52%), plexus (33%), and cranial nerves (32%). Diagnosis was established at a median of 3 months after symptom onset with substantial delays in primary NL. It mainly relied on PNS biopsy or FDG-PET, which carried high diagnostic yields (>90%). Postmortem diagnoses were rare (3%). Most cases were classified as B-cell (90%) lymphomas. Tumor-directed therapy was administered in 96% of patients and typically consisted of methotrexate or rituximab-based polychemotherapy. The median overall survival was 18 months. Primary NL without concurrent systemic disease outside the nervous system (hazard ratio [HR]: 0.44; 95% CI 0.25-0.78; p = 0.005), performance status (ECOG <2, HR: 0.30; 95% CI 0.18-0.52; p < 0.0001), and rituximab-based treatment (HR: 0.46; 95% CI 0.28-0.73; p = 0.001) were identified as favorable prognostic markers on multivariable analysis when adjusting for clinical and sociodemographic parameters. DISCUSSION:Advances in neuroimaging modalities, particularly FDG-PET, facilitate NL diagnosis and offer a high diagnostic yield. Yet, diagnostic delays in primary NL remain common. Rituximab-based therapy improves NL outcome. Findings may assist clinicians in early recognition, prognostic stratification, and treatment of NL.

Background Glioblastoma is the most frequent and a particularly malignant primary brain tumor with no efficacy-proven standard therapy for recurrence. It has recently been discovered that excitatory synapses of the AMPA-receptor subtype form between non-malignant brain neurons and tumor cells. This neuron-tumor network connectivity contributed to glioma progression and could be efficiently targeted with the EMA/FDA approved antiepileptic AMPA receptor inhibitor perampanel in preclinical studies. The PerSurge trial was designed to test the clinical potential of perampanel to reduce tumor cell network connectivity and tumor growth with an extended window-of-opportunity concept. Methods PerSurge is a phase IIa clinical and translational treatment study around surgical resection of progressive or recurrent glioblastoma. In this multicenter, 2-arm parallel-group, double-blind superiority trial, patients are 1:1 randomized to either receive placebo or perampanel ( n = 66 in total). It consists of a treatment and observation period of 60 days per patient, starting 30 days before a planned surgical resection, which itself is not part of the study interventions. Only patients with an expected safe waiting interval are included, and a safety MRI is performed. Tumor cell network connectivity from resected tumor tissue on single cell transcriptome level as well as AI-based assessment of tumor growth dynamics in T2/FLAIR MRI scans before resection will be analyzed as the co-primary endpoints. Secondary endpoints will include further imaging parameters such as pre- and postsurgical contrast enhanced MRI scans, postsurgical T2/FLAIR MRI scans, quality of life, cognitive testing, overall and progression-free survival as well as frequency of epileptic seizures. Further translational research will focus on additional biological aspects of neuron-tumor connectivity. Discussion This trial is set up to assess first indications of clinical efficacy and tolerability of perampanel in recurrent glioblastoma, a repurposed drug which inhibits neuron-glioma synapses and thereby glioblastoma growth in preclinical models. If perampanel proved to be successful in the clinical setting, it would provide the first evidence that interference with neuron-cancer interactions may indeed lead to a benefit for patients, which would lay the foundation for a larger confirmatory trial in the future. Trial registration EU-CT number: 2023-503938-52-00 30.11.2023.

Immunodeficiency-associated primary CNS lymphoma (PCNSL) represents a distinct clinicopathological entity, which is typically Epstein-Barr virus-positive (EBV + ) and carries an inferior prognosis. Genetic alterations that characterize EBV-related CNS lymphomagenesis remain unclear precluding molecular classification and targeted therapies. In this study, a comprehensive genetic analysis of 22 EBV + PCNSL, therefore, integrated clinical and pathological information with exome and RNA sequencing (RNASeq) data. EBV + PCNSL with germline controls carried a median of 55 protein-coding single nucleotide variants (SNVs; range 24–217) and 2 insertions/deletions (range 0–22). Genetic landscape was largely shaped by aberrant somatic hypermutation with a median of 41.01% (range 31.79–53.49%) of SNVs mapping to its target motifs. Tumors lacked established SNVs (MYD88, CD79B, PIM1) and copy number variants (CDKN2A, HLA loss) driving EBV − PCNSL. Instead, EBV + PCNSL were characterized by SOCS1 mutations (26%), predicted to disinhibit JAK/STAT signaling, and mutually exclusive gain-of-function NOTCH pathway SNVs (26%). Copy number gains were enriched on 11q23.3, a locus directly targeted for chromosomal aberrations by EBV, that includes SIK3 known to protect from cytotoxic T-cell responses. Losses covered 5q31.2 (STING), critical for sensing viral DNA, and 17q11 (NF1). Unsupervised clustering of RNASeq data revealed two distinct transcriptional groups, that shared strong expression of CD70 and IL1R2, previously linked to tolerogenic tumor microenvironments. Correspondingly, deconvolution of bulk RNASeq data revealed elevated M2-macrophage, T-regulatory cell, mast cell and monocyte fractions in EBV + PCNSL. In addition to novel insights into the pathobiology of EBV + PCNSL, the data provide the rationale for the exploration of targeted therapies including JAK-, NOTCH- and CD70-directed approaches.

BACKGROUND:Neuroligin 4 X-linked (NLGN4X) harbors a human leukocyte antigen (HLA)-A*02-restricted tumor-associated antigen, overexpressed in human gliomas, that was found to induce specific cytotoxic T cell responses following multi-peptide vaccination in patients with newly diagnosed glioblastoma.METHODS:T cell receptor (TCR) discovery was performed using droplet-based single-cell TCR sequencing of NLGN4X-tetramer-sorted T cells postvaccination. The identified TCR was delivered to Jurkat T cells and primary human T cells (NLGN4X-TCR-T). Functional profiling of NLGN4X-TCR-T was performed by flow cytometry and cytotoxicity assays. Therapeutic efficacy of intracerebroventricular NLGN4X-TCR-T was assessed in NOD scid gamma (NSG) major histocompatibility complex (MHC) I/II knockout (KO) (NSG MHC I/II KO) mice bearing NLGN4X-expressing experimental gliomas.RESULTS:An HLA-A*02-restricted vaccine-induced T cell receptor specifically binding NLGN4X131-139 was applied for preclinical therapeutic use. Reactivity, cytotoxicity, and polyfunctionality of this NLGN4X-specific TCR are demonstrated in various cellular models. Intracerebroventricular administration of NLGN4X-TCR-T prolongs survival and leads to an objective response rate of 44.4% in experimental glioma-bearing NSG MHC I/II KO mice compared to 0.0% in control groups.CONCLUSION:NLGN4X-TCR-T demonstrate efficacy in a preclinical glioblastoma model. On a global scale, we provide the first evidence for the therapeutic retrieval of vaccine-induced human TCRs for the off-the-shelf treatment of glioblastoma patients.Keywords cell therapy | glioblastoma | T cell receptor | tumor antigen.

Abstract BACKGROUND Neurolymphomatosis (NL) refers to lymphomatous infiltration of the peripheral nervous system (PNS). NL diagnosis and treatment are challenging given the broad differential diagnosis of peripheral neuropathy, the lack of larger cohorts and the subsequent unavailability of prognostic factors or consensus therapy. This study aimed to define characteristics and prognostic factors of NL. METHODS A systematic review of the literature (2004-2023) was performed using PubMed and Scopus databases and reported following PRISMA guidelines. Clinical, radiological, pathological and outcome information were extracted. Multivariable survival analyses were performed using Cox proportional hazard models. RESULTS A total of 459 NL cases from 264 studies were accumulated. NL was the first manifestation of malignancy (primary NL) in 197 patients. PNS relapse of known non-Hodgkin lymphoma (secondary NL) occurred in 262 cases after a median 12 months. NL predominantly presented with rapidly deteriorating, asymmetric painful polyneuropathy. Infiltrated structures included peripheral nerves (56%), nerve roots (52%), plexus (33%) and cranial nerves (32%). Diagnosis was established at a median of 3 months after symptom onset with substantial delays in primary NL. It mainly relied on PNS biopsy or FDG-PET, which carried high diagnostic yields (> 90%). Post mortem diagnoses were rare (3%). Most cases were classified as B-cell (90%) lymphomas. Tumor-directed therapy was administered in 96% of patients and typically consisted of methotrexate or rituximab-based polychemotherapy. Median overall survival was 18 months. Primary NL without concurrent systemic disease outside the nervous system (hazard ratio [HR]: 0.44; 95% confidence interval (CI): 0.25-0.78; p = 0.005), performance status (ECOG < 2, HR: 0.30; 95% CI: 0.18-0.52; p < 0.0001), and rituximab-based treatment (HR: 0.46; 95% CI: 0.28-0.73; p = 0.001) were identified as favorable prognostic markers on multivariable analysis when adjusting for clinical and sociodemographic parameters. CONCLUSIONS Advances in neuroimaging modalities, particularly FDG-PET, facilitate NL diagnosis and offer a high diagnostic yield. Rituximab-based therapy improves NL outcome. Findings may assist clinicians in early recognition, prognostic stratification, and treatment of NL.

This study investigates the biological effect of Tumor Treating Fields (TTFields) on key drivers of glioblastoma’s malignancy—tumor microtube (TM) formation—and on the function and overall integrity of the tumor cell network. Using a two-dimensional monoculture GB cell network model (2DTM) of primary glioblastoma cell (GBC) cultures (S24, BG5 or T269), we evaluated the effects of TTFields on cell density, interconnectivity and structural integrity of the tumor network. We also analyzed calcium (Ca2+) transient dynamics and network morphology, validating findings in patient-derived tumoroids and brain tumor organoids. In the 2DTM assay, TTFields reduced cell density by 85–88

Tumor microtubes (TMs) connect glioma cells to a network with considerable relevance for tumor progression and therapy resistance. However, the determination of TM-interconnectivity in individual tumors is challenging and the impact on patient survival unresolved. Here, we establish a connectivity signature from single-cell RNA-sequenced (scRNA-Seq) xenografted primary glioblastoma (GB) cells using a dye uptake methodology, and validate it with recording of cellular calcium epochs and clinical correlations. Astrocyte-like and mesenchymal-like GB cells have the highest connectivity signature scores in scRNA-sequenced patient-derived xenografts and patient samples. In large GB cohorts, TM-network connectivity correlates with the mesenchymal subtype and dismal patient survival. CHI3L1 gene expression serves as a robust molecular marker of connectivity and functionally influences TM networks. The connectivity signature allows insights into brain tumor biology, provides a proof-of-principle that tumor cell TM-connectivity is relevant for patients’ prognosis, and serves as a robust prognostic biomarker.

Purpose: Primary central nervous system (CNS) gliomas can be classified by characteristic genetic alterations. In addition to solid tissue obtained via surgery or biopsy, cell-free DNA (cfDNA) from cerebrospinal fluid (CSF) is an alternative source of material for genomic analyses.Experimental Design: We performed targeted next-generation sequencing of CSF cfDNA in a representative cohort of 85 patients presenting at two neurooncological centers with suspicion of primary or recurrent glioma. Copy-number variation (CNV) profiles, single-nucleotide variants (SNV), and small insertions/deletions (indel) were combined into a molecular-guided tumor classification. Comparison with the solid tumor was performed for 38 cases with matching solid tissue available.Results: Cases were stratified into four groups: glioblastoma (n = 32), other glioma (n = 19), nonmalignant (n = 17), and nondiagnostic (n = 17). We introduced a molecular-guided tumor classification, which enabled identification of tumor entities and/or cancer-specific alterations in 75.0% (n = 24) of glioblastoma and 52.6% (n = 10) of other glioma cases. The overlap between CSF and matching solid tissue was highest for CNVs (26%-48%) and SNVs at predefined gene loci (44%), followed by SNVs/indels identified via uninformed variant calling (8%-14%). A molecular-guided tumor classification was possible for 23.5% (n = 4) of nondiagnostic cases.Conclusions: We developed a targeted sequencing workflow for CSF cfDNA as well as a strategy for interpretation and reporting of sequencing results based on a molecular-guided tumor classification in glioma. See related commentary by Abdullah, p. 2860Conclusions: We developed a targeted sequencing workflow for CSF cfDNA as well as a strategy for interpretation and reporting of sequencing results based on a molecular-guided tumor classification in glioma. See related commentary by Abdullah, p. 2860

Abstract Despite the remarkable success of chimeric antigen receptor (CAR)-T cell treatment for patients with hematologic malignancies, this method has yet failed to confer meaningful survival benefits to patients suffering from glioblastoma (GB), the most lethal type of brain tumor. This highlights the need to develop novel CAR-T cell approaches for this disease. CD70 is a member of the tumor necrosis factor receptor (TNFR) superfamily. Although absent on normal brain tissue, it is ectopically expressed in a substantial fraction of GB patients, indicating its suitability as a CAR-T cell therapy candidate. In this study, we generated CD70-targeting CAR-T cells and tested their cytotoxicity in vitro and in vivo. First, we detected CD70 in a panel of primary GB cell lines and to investigate its role, it was overexpressed in human and murine GB cells. In a syngeneic glioma model, C57BL/6J mice injected with CD70-overexpressing GL261 cells developed significantly larger tumors compared to the control counterparts. Additionally, RNA-sequencing revealed that CD70-overexpression in the same cells led to higher expression levels of immunosuppressive marker CD200 and lower levels of tumor inhibition genes Prkg2 and Sh3bgrl2, suggesting a tumor-promoting role in GB. For our immunotherapeutic intervention, we designed CD70-targeting CAR-T cell constructs featuring different co-stimulatory domains (CD27, CD28 or 4-1BB) and successfully transduced primary T-Cells from healthy donors. In an in vitro co-culture, all CAR-T cells recognized and eliminated primary cancer cells in a target-dependent and donor-independent manner, while secreting high levels of Granzyme B. The killing capacity of these cells was further highlighted in a 3D system in which in vitro-generated cortical organoids were treated with CAR-T cells after being infiltrated by CD70-expressing or control cells. Immunofluorescence (IF) staining and enzyme-linked immunosorbent assay (ELISA) revealed increased levels of Granzyme B and IFN-γ in all treated organoids previously infiltrated by CD70+ tumor cells. Importantly, generated CAR-T cells showed high specificity and efficiency in killing CD70+ tumors in vivo in the brains of immunodeficient mice. Namely, 80% of NSG mice orthotopically implanted with CD70+ GB cells and subsequently treated by CARs but not mock-transduced T-Cells showed complete tumor remission by the end of the experimental endpoint, determined by bioluminescence imaging (BLI). IF analysis of these brains showed high levels of apoptotic marker cleaved caspase-3, enhanced effector cell presence and significantly lower tumor cell occupancy compared to control-treated animals. In conclusion, we provide evidence in favor of utilization of CAR-T cells against CD70-expressing gliomas. Based on these findings, a phase-I clinical trial to assess the safety and efficacy of autologous CD70-specific CAR-T cells for relapsed CD70+ GB is being planned. Citation Format: Alexandros Kourtesakis, Hiu Nam Hannah Chow, Dennis Alexander Agardy, Eileen Bailey, Sandra Horschitz, Ammar Jabali, Rainer Will, Denise Reibold, Sonja Pusch, Christoph Schifflers, Manuel Fischer, Ling Hai, Dirk C. Hoffmann, Yu-Chan Chih, Robin Wagener, Leon Kaulen, Philipp Koch, Michael Breckwoldt, Michael Schmitt, Wolfgang Wick, Tim Sauer, Tobias Kessler. CD70-specific CAR-T cell therapy for the treatment of glioblastoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 55.

Recordings of the physiological history of cells provide insights into biological processes, yet obtaining such recordings is a challenge. To address this, we introduce a method to record transient cellular events for later analysis. We designed proteins that become labeled in the presence of both a specific cellular activity and a fluorescent substrate. The recording period is set by the presence of the substrate, whereas the cellular activity controls the degree of the labeling. The use of distinguishable substrates enabled the recording of successive periods of activity. We recorded protein-protein interactions, G protein–coupled receptor activation, and increases in intracellular calcium. Recordings of elevated calcium levels allowed selections of cells from heterogeneous populations for transcriptomic analysis and tracking of neuronal activities in flies and zebrafish.

Abstract Recently, we have shown that primary brain tumors form highly organized, small-world scale-free networks, reminiscent of early networks in neurodevelopment and highly resistant to therapy. Such putative cancer-cell intrinsic neural mechanisms are becoming an increasing focus in the field of Cancer Neuroscience, but it has not yet been shown whether tumors of non-CNS origin also have the ability to form multicellular networks which sustain proliferative signaling and increase resistance to therapy. We observed synchronized calcium activity using longitudinal microscopy both in vivo and in vitro that were reminiscent of gap junction coupled networks. Indeed, dye transfer experiments verified functional coupling of brain metastases cells in vitro. Similarly, inhibition of gap junctions with different pharmacological agents significantly reduced calcium oscillations and tumour proliferation in vitro. Furthermore, gap junctions are significantly upregulated in brain-tropic sublines compared to parental cells. To understand the functional relevance of these gap-junction coupled tumor cell networks, we performed bulk RNA-Sequencing of two melanoma metastasis models in monoculture under gap junction inhibition and control: We observed reduced calcium communication and a concordant downregulation of neurodevelopmental and synaptic pathways within the tumor cells upon network disconnection, highlighting a potential recapitulation of neural-like features in metastases to the brain. To investigate the therapeutic potential of targeting these networks, we treated mice with brain metastases with two brain-penetrant gap junction blockers and observed significantly reduced tumor burden. Finally, we investigated the presence of heterotypic networks between brain metastases cells and the brain microenvironment with our SR101 dye transfer pipeline. Contrary to our findings in primary brain tumors, no gap junction connections to cells of the brain microenvironment in those models with reduced tumor growth upon network inhibition could be detected, supporting a cancer cell intrinsic gap junction mediated network and highlighting a selectivity of these homogenous tumor networks and potential therapeutic vulnerability.

Abstract Chitinase 3-like 1 (CHI3L1) is a secreted glycoprotein and its RNA expression elevated in glioblastoma (GB) compared to other tumor types and related normal tissues. Furthermore, transcript levels in GB dictate aggressiveness through modulating stemness, proliferation and tumor microenvironment. This ultimately influences patient survival. We here provide evidence that the pathogenic relevance of CHI3L1 expression is associated with the extent of tumor microtubes (TMs) - ultralong membrane tubes that connect GB cells (GBCs) to a network with considerable relevance for tumor progression and therapy resistance. Single cell RNA profiling of xenografted GBCs with different degrees of morphological and functional TMs identified CHI3L1 as a prognostic marker for TM network extent. We demonstrate that both RNA and protein expression levels are suitable markers in preclinical in vitro systems modeling TM connectivity as well as in clinical specimens. Genetic perturbation of CHI3L1 influenced GBC network integration, caused a shift of the dominant cell state and altered the phosphorylation status of the TM-driver GAP43. Pharmacological blocking of CHI3L1 with an antibody reduced TM networks, thus providing a handle for future clinical translation. Together, these data identify a functional and upstream role of CHI3L1 in governing tumor cell connectivity, CHI3L1 RNA and protein expression as a novel way to determine overall GBC connectivity for future trials, and finally a new therapeutic target for tumor network-disrupting strategies.

Diffuse infiltrating gliomas, especially glioblastomas, are the most common incurable primary brain tumors in adults. The aggressive growth and high resistance against cytotoxic agents are mediated by ultra-long tubular membrane protrusions, Tumor Microtubes (TMs), that interconnect single glioblastoma cells via gap junctions to multicellular communicating and resistant networks. Early studies showed that targeting these malignant networks increases response to therapy. Thus, the disconnection of such networks has emerged as a new therapeutic strategy to decrease radioresistance. However, the development of drugs that interfere with TM- and network formation is compromised by the lack of established drug screening pipelines to study a compound‘s specific anti-TM and network-inhibiting activity in combination with radiotherapy. To identify compounds with anti-TM activity, we first developed an in vitro medium-throughput imaging-based drug screening. Human-derived primary glioblastoma cells were seeded in a newly developed 2D monolayer where tumor cells form TMs and networks. The cells were treated with compounds with potential anti-TM activity. Using high-resolution laser scanning microscopy, morphological readouts were obtained and analyzed. The most promising compounds were then evaluated in a xenograft chronic cranial window mouse model. Treatments were administered with concomitant radiotherapy (6Gy per day for 3 days). By observing TM and tumor cell network changes via longitudinal in vivo 2-photon microscopy, tumor cell count, growth, and TM parameters were measured in the live mouse over up to 12 weeks. Among the 87 compounds tested in vitro, 42 significantly inhibited TMs. Two of these hits showed a higher count of dead cells in combination with radiotherapy as opposed to glioblastoma cells treated with radiotherapy alone and were therefore studied in the in vivo pipeline. One of these two agents showed reductions of TM length and striking responses in a combinatorial treatment regime with radiotherapy in individual mice that were not observed in control animals treated with radiotherapy only; however, the responses observed were heterogeneous between regions and mice. Further studies revealed this compound directly phosphorylates the gap junction protein Connexin 43 on S368, which contributes to a decrease in intercellular communication of glioblastoma cells. By applying different imaging techniques and AI algorithms we successfully established a novel in-vitro-to-in-vivo drug screening pipeline for the development of disconnecting agents. Most importantly, our data shows that disconnecting multicellular brain tumor networks increases response to radiotherapy. Further studies will be needed to develop reliable anti-TM agents as novel co-treatment options to sensitize glioblastomas to cytotoxic therapy. Citation Format: Daniel D. Azorín, Sophie Weil, Dirk C. Hoffmann, David Hausmann, Erik Jung, Matthias Osswald, Jill Reckless, Nigel Ramsden, Simon Thompson, David Grainger, Wolfgang Wick, Frank Winkler. AI-aided drug development for disconnecting glioma tumor microtube networks [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1084.