5q-associated spinal muscular atrophy (SMA) is a monogenic disease causing progressive alpha motor neuron degeneration, muscle atrophy, and weakness. Intrathecal therapy with the antisense oligonucleotide nusinersen modifies the disease course. However, biomarkers for understanding underlying molecular pathomechanisms and monitoring therapy are not yet known. A total of 130 cerebrospinal fluid (CSF) samples from 24 adult patients with SMA type 2 or 3 were collected over 3.5 years, and CSF proteome was analyzed using mass spectrometry (MS). By applying two complementary MS protein quantification methods, label-free quantification (LFQ) and tandem mass tag (TMT) isotopic labeling, specific protein patterns reflecting changes in the CSF in response to nusinersen therapy were identified. These results were combined with cellular and metabolic profiles. Nusinersen therapy led to a median motor function improvement of 2.2 Hammersmith Functional Motor Scale-Expanded points after 10 months and 2.6 points after 34 months. CSF macrophages increased in number and showed an altered morphology. Albumin quotient (qAlb), glucose, and lactate concentrations were inversely correlated with clinical improvement. MS analysis of CSF identified 1,674 (TMT) and 441 (LFQ) proteins. Protein profiles reflected reduced inhibition of “nervous system development” and “axogenesis” pathways under therapy. In addition, clinical improvement was associated with upregulation of the interacting proteins α-dystroglycan and beta-1,4-glucuronyltransferase 1, reduction of complement factors, negative correlation in immunoglobulin- and B cell-related pathways, and reduction of cellular mediators such as lymphocytes. The present multi-proteomic analysis contributes to the understanding of the molecular mechanisms underlying nusinersen’s therapeutic effects and offers potential biomarkers for monitoring treatment response in SMA.

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.

(A) IFN-𛾠ELISpot raw data of all HLA-A*02:01 patients analyzed (n = 7) showing the IFN-𛾠spots / 1x105 T-cells for patient cells against TAAs vs. patient cells against the negative control (human immunodeficiency virus (HIV) gag/pol, 9 amino acids). Grey bars indicate a significantly increased immunogenicity compared to negative control (solid line). (B) Homogenous levels of stimulation while comparing the total immune response by mean of IFN-𛾠spot numbers among astrocytomas (red) and oligodendrogliomas (blue) to a certain reactive epitope. (n.s., not significant, *, p < 0.05; **, p < 0.01)

(A) Significantly increased immune responses in IDHmut lower-grade glioma patients (P) vs. healthy donors (HD) seen for the potential T-cell target antigens CRKII, CFL1, CNTN1, NME2, and TKT. Homogenous levels of stimulation while comparing IFN-𛾠spot numbers among astrocytomas (red) and oligodendrogliomas (blue) to a certain TAA as well as (B) the total immune response. (C) IFN-𛾠ELISpot raw data showing the spot count / 1x105 T-cells upon stimulation with autologous dendritic cells loaded with antigens (CRKII, CFL1, CNTN1, NME2, or TKT, black bar) vs. negative control (IgG1, grey bar) for each patient (P) (n = 16). Highlighted in green: significant immune responses (p < 0.05) with an > 2-fold increased spot count compared to the negative control (IgG) (*, p < 0.05; **, p < 0.01; ***, p < 0.001)

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.

Representative images of immunohistochemical stainings of the isotype controls IgG and IgG1 in normal brain (NB) tissues, astrocytomas (WHO{degree sign}II: n = 10; WHO{degree sign}III: n = 10), and oligodendrogliomas (WHO{degree sign}II: n = 10; WHO{degree sign}III: n = 10). Scale bar: 50 µm.

Supplementary Figure Legends

Peptide information of potential T cell target antigens, tested by IFN-γ ELISpot

(A) Applied gating strategy (one representative patient, NCH645) for the expression analysis of tumor-associated antigens in IDHmut lower-grade glioma and secondary glioblastoma GSCs, analyzed by flow cytometry in regard to an adequate isotype control. (B) Fluorescence intensity measurements for the isotype control and the antigen expressing (positive) GSC population, showing counts in percentage [%] over the mean PE-intensity.(C) Gating strategy, including the markers pacific orange (PO), CD3, and CD8, used for quantification of antigen-specific peripheral cytotoxic CD3+CD8+ T-cells by flow cytometry and (D) antigen-loaded pMHC I-tetramers (PE-conjugated) and the respective positive (Influenza M1) as well as negative control HLA-A*02-tetramers. GSCs: glioma stem-like cells, HIV: human immunodeficiency virus, HLA: human leukocyte antigen, MHC: major histocompatibility complex, PE: Phycoerythrin, pos: positive, US: unstained.

(A) Semi-quantitative evaluation on immunohistochemically stained sections of the frequency of tumor cells expressing the potential T-cell target antigens CRKII, CFL1, NME2, and CNTN1 in astrocytoma (WHO{degree sign}II: n = 10, WHO{degree sign}III: n = 10) and (B) oligodendroglioma (WHO{degree sign}II: n = 10, WHO{degree sign}III: n = 10) tumor tissues.

Gating strategy to identify cytokine-secreting cells by cytokine capture assay. The acquired cells were gated based on their size and granularity as lymphocytes excluding cell debris. Then we gated on single cells to remove cell duplexes and clusters. Apoptotic cells were gated out next by gating on pacific orange (PO)-negative cells (live cells). The respective T-cell (TC) population was identified by surface staining with anti-CD4 PerCP-Cy5.5, and anti-CD8-V450 antibodies. The cytokine-positive cell gate was placed according to a FMO control (exemplarily shown for IFN-ð›¾) and the frequency of cytokine-secreted cells in samples stimulated with IgG1 peptide (negative control) or the antigen of interest (exemplarily shown for CFL1) were determined based on that gate.

Representative multicolor stainings of CFL1, CRKII, NME2, and CNTN1 on acetone-fixed cryosections with markers for common cell types in IDHmut LGGs: anti-GFAP (tumor cells, #Z0334, DAKO), anti-CD68 (microglia/ macrophages, #M0718, DAKO), anti-CD31 (endothelial cells, #223609, BD Pharmingen). Detection was performed by using fluorochrom-conjugated secondary antibodies (anti-mouse AF647 (#A-21463, Invitrogen), anti-rabbit AF555 (#A-21428, Invitrogen), DAPI (#D1306, ThermoFisher), and the Zenon AF488 mouse IgG1 Labeling Kit (#Z25001, ThermoFisher)) according to the manufacturer`s protocol (scale bar: 50 µM, scale bar zoom: 10 µM).

Graphical representation of IFN-𛾠ELISpots using peripheral blood mononuclear T cells, co-cultivated with dendritic cells, pulsed with autologous tumor lysate fractions of (A) the first (1st) PF2D dimension and corresponding (B) second (2nd) PF2D dimension of patients NCH1390, NCH612, and NCH519a. Solid line represents mean background of negative control (PBL). F: fraction, PBL: peripheral blood lymphocytes (*, p < 0.05; **, p < 0.01; ***, p < 0.001)

Apoptosis is a tightly controlled cell death program executed by proteases, the so-called caspases. It plays an important role in tissue homeostasis and is often dysregulated in cancer. Here, we identified FYCO1, a protein that promotes microtubule plus end-directed transport of autophagic and endosomal vesicles as a molecular interaction partner of activated CASP8 (caspase 8). The absence of FYCO1 sensitized cells to basal and TNFSF10/TRAIL-induced apoptosis by receptor accumulation and stabilization of the Death Inducing Signaling Complex (DISC). Loss of FYCO1 resulted in impaired transport of TNFRSF10B/TRAIL-R2/DR5 (TNF receptor superfamily member 10b) to the lysosomes in TNFSF10/TRAIL-stimulated cells. More in detail, we show that FYCO1 interacted via its C-terminal GOLD domain with the CCZ1-MON1A complex, which is necessary for RAB7A activation and for the fusion of autophagosomal/endosomal vesicles with lysosomes. We demonstrated that FYCO1 is a novel and specific CASP8 substrate. The cleavage at aspartate 1306 resulted in the release of the C-terminal GOLD domain, inactivating FYCO1 function, and allowing for the progression of apoptosis. Furthermore, the lack of FYCO1 resulted in a stronger and prolonged formation of the TNFRSF1A/TNF-R1 signaling complex. Thus, FYCO1 limits the ligand-induced and steady-state signaling of TNFR-superfamily members, providing a control mechanism that fine-tunes both apoptotic and inflammatory answers.Abbreviations: AP: affinity purification; CHX: cycloheximide; co-IP: co-immunoprecipitation; CRISPR: clustered regularly interspaced short palindromic repeats; DISC: death-inducing signaling complex; DR: death receptors; doxy: doxycycline; GEF: guanine nucleotide exchange factor; ind: inducible; KD: knockdown; KO: knockout; MS: mass spectrometry; shRNA: short hairpin RNA; siRNA: small interfering RNA; TIP: two-step co-immunoprecipitation; WB: western blot.

Clinical data of IDHmut LGG patients

Usage of RNA Seq TCGA data of IDHmut LGG tumors to compare mRNA expression levels with T-cell infiltration. Significant increased expression of NME2 in the TCGA IDHmut LGG dataset was shown to be positively correlated with the expression of CD3E (encoding CD3-ε as part of the CD3 T-cell receptor complex) (r = 0.2352). Stronger correlation in the histology of astrocytomas (r = 0.3693**) than in oligodendrogliomas (r = 0.1919**). (*, p < 0.05; **, p < 0.01; ***, p < 0.001)

Tumor cell extensions called tumor microtubes (TMs) in glioma resemble neurites during neurodevelopment and connect glioma cells to a network that has considerable relevance for tumor progression and therapy resistance. The determination of interconnectivity in individual tumors has been challenging and the impact of tumor cell connectivity on patient survival remained unresolved so far. Here, a connectivity signature from single-cell RNA-sequenced (scRNA-Seq) xenografted primary glioblastoma (GB) cells was established and clinically validated. Thirty-four of 40 connectivity genes were related to neurogenesis, neural tube development or glioma progression, including the TM-network-relevant GAP43 gene. Astrocytic-like and mesenchymal-like GB cells had the highest connectivity signature scores in scRNA-Seq data of patient-derived xenografts and patient samples. In 230 human GBs, high connectivity correlated with the mesenchymal expression subtype, TP53 wildtype, and with dismal patient survival. CHI3L1 was identified as a robust molecular marker of connectivity. Thus, the connectivity signature allows novel insights into brain tumor biology, provides a proof-of-principle that tumor cell connectivity is relevant for patients’ prognosis, and serves as a robust biomarker that can be used for future clinical trials. Statement of significance Integration of GB cells into functional networks drives tumor progression and resistance. Here, we established and validated a novel connectivity gene expression signature of single GB cells and whole tumors that can be easily applied to clinical and preclinical samples. It is shown that connectivity is determining prognosis combining molecular, functional and clinical insights into the disease.