Mutational signature stratification of recurrent gliomas reveals distinct patterns of genomic traits

Mutational signature stratification of recurrent gliomas reveals distinct patterns of genomic traits | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Mutational signature stratification of recurrent gliomas reveals distinct patterns of genomic traits Joana Peixoto, Miguel Castanho, Bárbara Soares-Ferreira, Beatriz Esteves, and 10 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9093330/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Although temozolomide (TMZ) is widely used for glioma treatment, its therapeutic benefit is limited by acquired resistance and recurrence, facilitated by intratumor heterogeneity. Mutational signatures can inform tumor evolution and reveal alterations associated with treatment response. Methods: We performed a molecular analysis of 96 glioma recurrences with sufficient private single-base substitution (SBS) relative to their matched primary tumors, stratified by their dominant SBS. Results: Four groups were identified: SBS11 (TMZ, n=38), SBS1/5 (aging, n=32), SBS6/15/21/26 (microsatellite instability (MSI), n=13), and other dominant SBSs (n=13). Recurrences with dominant SBS11 showed markedly higher acquired mutational counts than the other groups (1338 vs 59 (aging) vs 57 (MSI) vs 57 (other); p<0.01). Mutations in SYNE2 , SZT2 , and FBN3 were exclusive to recurrences with dominant or second-dominant SBS11 (n=41), and 85% (35/41) harbored a mutation in at least one of these genes. Among TMZ-associated recurrences, SBS23 was frequent (44%, 18/41) and was associated with higher acquired mutational counts (2089 vs 1188; p=0.018) and more IDH-wildtype tumors (67% vs 30%; p=0.037) compared with cases lacking SBS23 dominance. MAPKBP1 mutations were enriched in SBS23-positive recurrences (56% (10/18) vs 0% (0/23); p<0.001). Aging-dominant recurrences showed more frequent acquired chromosome 16q losses (22% vs 8% (TMZ) vs 0% (MSI) vs 0% (other); p<0.05), which were associated with an increased fraction of the genome altered when compared with diploid chromosome 16q (15% vs 7%; p=0.01). Conclusions: Collectively, these findings show that signature-based grouping of recurrences can refine molecular characterization after therapy and nominate candidate biomarkers linked to resistance. Bioinformatics Oncology Neurology Bioinformatics Oncology Neurology Bioinformatics Oncology Neurology Recurrent gliomas temozolomide mutational signatures single-base substitutions Figures Figure 1 Figure 2 Figure 3 Figure 4 Key points Mutational-signature stratification of recurrent gliomas identifies distinct, therapy-associated genomic relapse patterns. Temozolomide-associated (single-base substitution (SBS11)) recurrences show markedly increased acquired mutational burden, with SYNE2 , SZT2 , and FBN3 mutations restricted to this subset . A SBS23-positive TMZ-associated subset is enriched for IDH-wildtype tumors and shows MAPKBP1 mutations restricted to this subset , while aging-dominant recurrences are associated with acquired 16q loss . Importance of the Study Recurrent gliomas remain clinically challenging, and molecular frameworks that capture post-therapy evolution are limited. By analyzing single-base substitutions (SBSs) private to recurrences relative to matched primary tumors in 96 glioma recurrences, we stratified tumors by dominant mutational signature and identified distinct relapse trajectories linked to therapy-associated mutational processes. Beyond confirming a temozolomide (TMZ)-associated SBS11 subgroup with markedly increased acquired mutational burden, we showed that SYNE2 , SZT2 , and FBN3 mutations were restricted to SBS11-associated recurrences (dominant or second-dominant SBS11). We also identified a SBS23-positive subset within TMZ-associated recurrences enriched for IDH-wildtype tumors, in which MAPKBP1 mutations were restricted to SBS23-positive cases . In contrast, aging-dominant recurrences were associated with acquired chromosome 16q loss and increased genomic disruption. These findings support mutational-signature stratification as a practical approach to refine post-therapy molecular characterization and prioritize resistance-linked biomarkers and candidate vulnerabilities for functional and translational follow-up. Introduction Gliomas are the most frequent primary malignant brain tumors in adults, comprising roughly 40%–50% of intracranial tumors 1 and more than 80% of malignancies arising in the central nervous system 2 . Within this spectrum, glioblastoma (GBM) represents the most lethal form, and outcomes remain poor, with fewer than 12% of patients surviving beyond 3 years 3,4 . Consistent with the 2021 WHO Classification of Tumors of the Central Nervous System grading framework 5 , GBM is now designated as grade 4. Even with advances in multimodal management, most diffuse gliomas ultimately relapse, commonly at or near the resection cavity 6 , 7 . Over the last two decades, large-scale sequencing studies have clarified key molecular events in gliomagenesis, including recurrent perturbations in phosphoinositide 3-kinase (PI3K), receptor tyrosine kinase (RTK), and mitogen-activated protein kinase (MAPK) signaling 8 . Certain molecular combinations define specific types of gliomas: simultaneous mutations in IDH1/2 , TP53 and ATRX typify IDH-mutant astrocytomas, while concomitant mutations in IDH1/2 , TERT and the chromosome 1p/19-codeletion define oligodendrogliomas IDH mutant and 1p/19q codeleted. On the other hand, TERT mutations (without concomitant IDH), EGFR alterations and deletion of chromosome 10 and amplification of chromosome 7 are typical features of GBM 9,8,10 . Despite this increasingly detailed molecular landscape, there is still no universally accepted standard approach for recurrent or progressive disease. Temozolomide (TMZ), aided by partial blood–brain barrier (BBB) penetration and convenient oral administration 11 , enhances the effect of radiotherapy and remains a cornerstone of first-line therapy for GBM 12-15 . However, durable benefit is frequently undermined by acquired resistance, a process shaped by pronounced intratumor heterogeneity that enables differential therapeutic responses and promotes recurrence. Longitudinal studies profiling primary tumors and matched recurrences by whole-exome (WES) and whole-genome sequencing (WGS) have highlighted several therapy-associated evolutionary patterns, including alkylating agent–related hypermutation, radiation-associated CDKN2A homozygous deletions, and links between MYC copy number (CN) gains and reduced overall survival 16-18 . Even so, additional mechanisms contributing to TMZ-acquired resistance remain insufficiently defined and warrant further investigation. Mutational signature analysis provides a complementary framework to infer the biological and environmental processes that generate somatic mutations in cancer. Across tumor genomes, single-base substitutions (SBSs), small insertions and deletions (indels), CN alterations, and structural rearrangements can arise through endogenous mechanisms—such as defective DNA repair, replication-associated errors, or oxidative damage—or through exogenous exposures, including ultraviolet radiation, tobacco carcinogens, and other agents 1 9 . In multiple cancer types, mutational signatures have been leveraged to reconstruct tumor evolution, connect specific mutational processes to gene-level alterations, and, importantly, support biomarker development for predicting therapeutic response 20-24 . To uncover additional genetic determinants associated with acquired treatment resistance in gliomas, we conducted an in-depth molecular analysis of 96 glioma recurrences from Barthel FP et al 16 and Varn FS et al 2 5 , stratified according to their inferred mutational signatures. Methods Case selection All patients with diffuse glioma under an institutional review board–approved protocol at Jackson Laboratory and whose tumors were subjected to WES and WGS 16,25 were identified (n=258 cases sequenced) and accessed through cBioportal. Clinicopathological data, including age at diagnosis, sample type (258 primary tumors and 282 matched recurrences), IDH mutational status, treatment, immune score, and past cancer history were also retrieved 16,25 . All diffuse gliomas were re-classified according to the latest WHO Classification of Tumors of the Central Nervous System grading standards 5 . In addition, 30 primary low-grade gliomas (LGGs) (all IDH1 mutants) and matched recurrences whose tumors were also subjected to WES were identified from Johnson BE et al 17 . Besides this, 26 primary high-grade gliomas and matched recurrences subjected to WGS were identified and retrieved from The Cancer Genome Atlas (TCGA) 9 . Massively Parallel Sequencing Analysis Relevant genomic data derived from WES and WGS included somatic mutations, CN alterations (CN gains and losses, gene amplifications, and homozygous deletions), tumor mutation counts (coding mutations only) and fraction of the genome altered (FGA) 16, 25 . The CN segment files were also retrieved to determine whether genes harboring somatic mutations were targeted by loss of heterozygosity (LOH) 26 . CN alterations were estimated by taking the median of the log ratios per chromosomal arm. The absolute CN values were then calculated according to the tumor purity of a given sample, as previously described 26 . The cancer cell fractions (CCFs) of selected somatic mutations were computed using ABSOLUTE 27 , taking into account the variant allele frequency (VAF) and the ploidy status of each somatic variant. Solutions from ABSOLUTE were manually reviewed. A mutation was classified as clonal if its probability of being clonal was >50% or if the lower bound of the 95% confidence interval of its CCF was >90%, as previously described 26 . Mutations that did not meet the above criteria were considered subclonal. Relevant genomic data from Johnson BE et al 17 and TCGA included somatic mutations only 9 . Mutational signature inferences from single-base substitutions Given that the primary tumors and matched recurrences were previously found to be clonally-related 16,25 , we categorized all genetic alterations (somatic mutations and CN alterations) into shared or private to a particular component. For instance, a given genetic alteration was considered "shared" if it was present in both the primary tumor and matched recurrence. We defined alterations "private to the primary lesion" and "private to the recurrence" as those present only in the primary tumor or in the matched recurrence, respectively (Supplementary Figure. 1A). Mutational signatures were defined by deconstructSigs 28 using all coding single-base substitutions (SBSs) (missense, splice-site, nonsense and silent mutations) at default parameters and based on the set of mutational signatures represented in COSMIC v2.0 21 , as previously described 29 , for samples with ≥50 somatic SBSs. Genetic comparisons between glioma recurrences according to the different mutational signature exposures All recurrences with private and sufficient SBSs for mutational signature analysis were stratified into four groups, according to the presence of the most dominant SBS, including those associated with TMZ treatment (SBS11), aging (SBS1 & 5), microsatellite instability (MSI, SBS6, 15, 21 & 26) and other dominant SBSs. The frequencies of somatic genetic alterations "private to the recurrence" of each group were compared with each other in a two-by-two comparison. Considering the extremely high number of somatic mutations observed in the recurrences displaying a dominant, or a second dominant SBS11 (TMZ, n=41), we ought to infer which mutated genes were specific to such group of samples. For this, we excluded all mutated genes present in less than 49% of cases (20/41), and those also mutated (shared or private mutations) in any sample not displaying the SBS11 (n=55) (Supplementary Figure 1B). To evaluate whether recurrences displaying a dominant, or a second dominant SBS11 was linked to mutations in selected genes, we stratified samples into two groups based on the presence of SBS11, using a prespecified cutoff: TMZ− (SBS11 low) for SBS11=0 and TMZ+ (SBS11 high) for SBS11>0. In parallel, we defined a composite binary mutation variable indicating whether a sample harbored at least one mutation affecting SYNE2 , SZT2 and FBN3 . We summarized the joint distribution of TMZ group and the composite mutation status by computing 2×2 contingency table counts and within-group prevalences. In addition, and within the TMZ context (n=41), the frequencies of somatic genetic alterations of the recurrences displaying a dominant, or a second dominant SBS23 (n=18), were compared with those not displaying such mutational signature (n=23). Statistical analysis Statistical analyses were performed using R v3.1.2. Fisher's exact tests were employed for comparisons between categorical variables, whereas Mann–Whitney U tests were used for continuous variables. All tests were two-sided, and a p-value <0.05 was considered statistically significant. Survival analyses were performed using univariate Cox regressions, and Kaplan–Meier curves were displayed using the R package survival 30 . For a proper visualization, only recurrently altered genes per sample type are represented in the heatmaps. Results Genetic characterization of glioma recurrences subjected to mutational signature inferences After having categorized all genetic alterations (somatic mutations and CN alterations) into shared or private alterations to the primary tumor or matched recurrence, 198 components had sufficient SBSs for mutational signature analysis, including 45 primary tumors with private SBSs, 57 primary tumors and recurrences with shared SBSs, and 96 recurrences with private SBSs, for a total of 146 glioma patients (Figure 1A). When looking at the distribution of the dominant mutational signatures of gliomas with private SBSs to the primary tumor and with shared SBSs, the majority of these components displayed a dominant signature associated with aging (SBS1 & 5) (Figure. 1A and Supplementary Table S1). In contrast, the majority of the recurrences with private SBSs displayed a dominant signature associated with TMZ (SBS11, n=38/96) (Figure 1A and Supplementary Table S1). We hence had a closer look at the clinical and molecular features of the 96 recurrences stratified according to their dominant SBS. Such stratification rendered four main groups, including recurrences with a dominant SBS11 (TMZ, n=38), recurrences with a dominant SBS1/5 (aging, n=32), recurrences with a dominant SBS6/15/21/26 (microsatellite instability (MSI), n=13), and recurrences with other dominant SBSs (n=13) (Supplementary Table S1). The vast majority of patients in the four groups were treated with TMZ as the first-line chemotherapeutic agent (81%, 78/96) (Supplementary Table S1), and were classified as being GBM IDH wild-type, with the group of recurrences displaying SBSs associated to TMZ and MSI being also enriched with astrocytomas IDH mutant (29% and 31%, respectively) (Supplementary Figure S2A). Such enrichment was noticed when looking at the survival curves, with the group of recurrences with a dominant SBS associated with aging and other signatures having a significantly poorer survival outcomes than the group of recurrences with SBS11 as a dominant signature (p=0.024 and p=0.003, respectively) (Supplementary Figure S2B). When comparing the genetic frequencies of the 96 recurrences stratified according to their dominant SBS, we observed that the recurrences exhibiting a dominant SBS11 had a statistically significantly higher median number of non-synonymous mutation counts when compared to those exhibiting a dominant SBS associated with aging, MSI or with other mutational signatures (1338 vs 59, p<0.001, 1338 vs 57, p<0.01, 1338 vs 57, p<0.001, respectively) (Figure 1B). No statistically significant median values of the acquired FGA were observed between the four groups, although the recurrences with a dominant SBS associated with aging were found to display the highest median of the acquired FGA (9.6%) (Figure 1C). When looking at the acquired genetic alterations in all the 96 recurrences, several genes were found to harbor significantly higher number of mutations and/or homozygous deletions in the group with a dominant SBS11 when compared with the other group of recurrences, such as KMT2D (58%), CREBBP (50%), AKAP9 (50%) and BRCA1 (32%) (Figure 1D). The group of recurrences displaying a dominant SBS associated with aging was found to have acquired genetic alterations mainly affecting TP53 (32%), RYR2 (28%) and PTEN (22%), and the group of recurrences displaying a dominant SBS associated with MSI had acquired genetic alterations in the MUC16 (46%), IL27RA (31%) and PLEKHG5 (31%) genes (Figure 1D). In the group displaying other dominant SBSs, MUC16 , FLG, PKD1L1 and AGAP2 were found to be the most altered genes (all at 31%) (Figure 1D). We also stratified the group of recurrences with a dominant SBS associated with TMZ and aging according to the IDH status (wild-type vs mutated). In the former group, the IDH wild-type cases had a significantly higher number of acquired mutations affecting KIF5A than the IDH mutated ones (47% vs 0%, p<0.01) (Supplementary Figure D2C). Although no significance was reached, all acquired TP53 genetic alterations in the group of recurrences with a dominant SBS1 & 5 were found to affect IDH wild-type cases only (40% vs 0%, p=0.069) (Supplementary Figure S2C). SYNE2 , SZT2 and FBN3 are exclusively mutated in the recurrences associated with the single-base substitution 11 (temozolomide) Considering the extremely high number of somatic mutations observed in the recurrences displaying a mutational signature associated with TMZ (SBS11), we next determined which genes were significantly and exclusively mutated when compared to the other recurrences. Since we had 38 recurrences with a dominant SBS11 and 3 recurrences displaying this signature as the second most dominant one (Figure 1D, Supplementary Table S1), we annotated the genes that were mutated (private and shared mutations) in ≥49% of samples (n=20/41), and without being altered in any other recurrences from the remaining cohort (n=55) (Supplementary Figure S1B, Supplementary Table S2). We ended up annotating three genes that were specifically mutated in the TMZ-associated recurrences, namely SYNE2 , SZT2 and FBN3 . The SZT2 gene was found to be mutated in 61% of cases, followed by SYNE2 (59%) and FBN3 (49%) (Supplementary Figure 3A). When looking at the three genes together, we observed that 35/41 (85%) recurrences had at least one mutation affecting these genes (Supplementary Figure 3A). Considering such findings, we looked at all of the 258 cases sequenced (primary tumors and matched recurrences) that would harbor a mutation in these three genes (Figure 2A). We found two cases with the primary tumor harboring two subclonal mutations affecting SZT2 that were not treated with TMZ and were hence lost in the matched recurrence (Figure 2A and Supplementary Table S3). We also observed 7 primary tumors and matched recurrences with shared SYNE2 (n=3), SZT2 (n=1) and FBN3 (n=4) mutations (6 cases treated with TMZ), with the primary tumor of four cases harboring subclonal mutations affecting SYNE2 (n=2 cases) and FBN3 (n=2 cases), and becoming fully clonal in the matched recurrence (Figure 2A and Supplementary Table S3). Besides such observations, 31 cases treated with TMZ were found to harbor SYNE2 , SZT2 and FBN3 mutations only in the recurrence counterpart (Figure 2A). While the majority of the mutations affecting SYNE2 and SZT2 were deemed subclonal, the majority of FBN3 mutations were deemed clonal (Figure 2A and Supplementary Table S3). In order to verify if mutations affecting these three genes were also significantly enriched in independent cohorts, we accrued 30 primary, IDH-mutant LGGs and matched recurrences from Johnson BE et al 17 , and 26 primary high-grade diffuse gliomas and matched recurrences from TCGA 9 . We were able to infer the mutational signatures in 9 recurrences from Johnson BE et al 17 and in 7 recurrences from TCGA 9 . When looking at both cohorts together, mutations affecting FBN3 , SYNE2 , SZT2 were found to be altered in 67%, 56% and 44% of the recurrences displaying a dominant SBS11, with no mutations affecting these genes being observed in the recurrences not displaying such signature (Supplementary Figure 3B). We also decided to look at the prevalence of mutations affecting these three genes in any glioma recurrence displaying the SBS11, regardless of its dominance. Of the 42 recurrences exhibiting SBS11 (Supplementary Table S1), the prevalence of such mutations was found to be of 83.3% (35/42), a statistically significant prevalence of mutations in this group when compared with the group not displaying the SBS11 (0% vs 83%, p<0.001) (Figure 2B, Supplementary Table S3). With the notion that these mutations were specifically found in the recurrences with a dominant or second dominant SBS11, we had a closer look at the SBS of all the mutations affecting SYNE2 , SZT2 and FBN3 , in an attempt of determining if these mutations would fall within the context of SBS11 21 . Of the 139 SBSs affecting the three genes, 88% (n=122) were found to be associated within the context of the SBS11 (Figure 2C and Supplementary Table S3), thus enhancing the effect of the TMZ treatment in the emergence of de novo mutations occurring in these three genes. MAPKBP1 is exclusively mutated in the recurrences displaying both the single-base substitutions 11 and 23 Within the context of the mutational SBS11 (n=41 recurrences), we next looked at other mutational signatures that could be substantially enriched. We observed that 44% (n=18) of these recurrences also displayed the SBS23, as the second most dominant signature (n=17), or as the most dominant one (n=1) (Figure 3 and Supplementary Table S1). When comparing the genetic repertoire of these recurrences with those not displaying the SBS23 (n=23), we found that the former had a statistically significantly higher median number of acquired mutations than the latter (2089 vs 1188, p=0.018) (Figure 3A). In addition, and although no statistically significant differences were observed regarding the median acquired FGA, the group of recurrences displaying a mutational SBS23 was found to have a statistically significantly higher number of recurrences deemed IDH wild-type than those not displaying such mutational signature (67% vs 30%, p=0.0369) (Figure 3B and C). When looking at the genetic alterations, we found that MAPKBP1 mutations were exclusively present in the former (56% vs 0%, p<0.001) (Figure 3D), while UPF2 and CBFA2T2 genetic alterations exclusively affected the latter (0% vs 26%, p<0.05, for both genes) (Figure 3D). Of note, the frequency of chromosome 4q losses were also found to be private events in the group of recurrences not displaying the SBS23 (0% vs 30%, p<0.05) (Figure 3D). Akin to SZT2 , SYNE2 and FBN3 mutated genes, and after looking at the 258 cases sequenced, we found that mutations affecting MAPKBP1 were exclusively present in the recurrences displaying the SBS23 as the second most dominant one (n=10), with 70% (7/10) of these mutations being subclonal (Figure 3E and Supplementary Table S4). In addition, and of the 3 recurrences displaying a dominant SBS11 and a second dominant SBS23 taken from Johnson BE et al 17 and from TCGA 9 , one was found to harbor a MAPKBP1 mutation with a 5% VAF, thus being most likely subclonal (Supplementary Figure 3B). When looking at the distribution of the SBSs associated with signature 23, we also found that all of the 11 MAPKBP1 SBSs present in the 10 recurrences (one recurrence harbored two MAPKBP1 SBSs) (Figure 3E) were compatible with the mutational context of such signature (Figure 3F). Such observation underlies the impact of MAPKBP1 mutations in the emergence of the SBS23 in a subset of gliomas treated with TMZ. Acquired copy number losses of chromosome 16q in recurrences associated with the single-base substitution 1 (aging) display high fractions of the genome altered Besides the recurrences associated with the SBS11, we turned our attention at additional acquired genetic traits that could be enriched in the recurrences displaying other SBSs. When looking at acquired CN gains and losses, we found that the group of recurrences with a dominant SBS associated with aging had statistically significantly higher frequencies of acquired CN alterations of chromosomes 7p (25%), 7q (22%), 14q (22%), 16q (25%), 18q (25%), 19p (22%), and 19q (25%, all with a p<0.05)) when compared to the other three groups (Figure 4A). The enrichment of acquired CN losses of chromosome 16q (22%) was found to be particularly interesting, as such losses had a median FGA significantly higher when compared to the median FGA of recurrences with no CN alterations affecting such chromosomal arm (15% vs 7%, p=0.011) (Figure 4B). This statistical difference was found to be even higher when adding the acquired CN losses of chromosome 16q observed in the group of recurrences associated with the SBS11 (n=3) (median FGA 15% vs 3%, p= 0.001) (Figure 4B). Interestingly, we also observed a statistical difference when looking at the immune score 16 between recurrences harboring CN losses of chromosome 16q and recurrences with no CN changes of such chromosomal arm, with the former displaying a significantly lower median immune score than the latter (-162.8 vs 341.4, p=0.042) (Figure 4C). Taken together, the acquisition of CN losses of chromosome 16q may confer a higher genomic instability in a subset of glioma recurrences, mainly in those associated with SBS1. Discussion To date, there is no comprehensive clinical prognostic or predictive framework for glioma, and survival for these aggressive tumors has changed only modestly over recent decades. As a disease group marked by substantial molecular diversity and pronounced intratumor heterogeneity, gliomas still lack effective therapies and robust prognostic indicators that integrate histology with outcome-relevant tumor biomarkers. A long-standing therapeutic paradigm in oncology relies on DNA-damaging agents and ionizing radiation to exploit limitations in tumor DNA repair capacity. Because this pressure can produce a mutator phenotype, we asked whether mutational signatures present at recurrence could serve as markers of treatment resistance, with a particular focus on TMZ-associated resistance. To address this, we stratified 96 recurrent gliomas from Barthel FP et al 16 and Varn FS et al 2 5 that harbored sufficient private SBSs relative to their matched primary tumor, by their dominant SBS, defining four groups: TMZ-associated (SBS11, n=38), aging-associated (SBS1/5, n=32), MSI-associated (SBS6/15/21/26, n=13), and other signatures (n=13). As expected, TMZ-associated recurrences displayed substantially higher acquired mutational counts than the remaining groups. This pattern is consistent with prior observations in TMZ-resistant gliomas 31 , 32 , and recent studies have further suggested that a subset of recurrent gliomas with high mutational burdens may be candidates for immunotherapy, with encouraging outcomes reported in selected patients 33 , 34 .Given the magnitude of acquired mutational burden in TMZ-associated recurrences, we next sought genes that might be selectively altered in tumors exhibiting SBS11 (dominant or second-dominant ; n=41 ). We identified SYNE2 , SZT2 , and FBN3 as being exclusively mutated in this subset, with 85% of cases harboring a mutation in at least one of these genes and an 83% prevalence among samples exhibiting SBS11 irrespective of dominance. Prior work in hypermutant recurrent GBM highlighted sets of genes that were selectively mutated at recurrence (including LRP1A , PCNX1 , KMT2D , DST , SYNE2 , and NEB ), enabling identification of hypermutated tumors with high sensitivity 33 . While 3/6 of those genes were also significantly altered among TMZ-associated recurrences in our cohort (Supplementary Table 2 ), SYNE2 emerged as the only gene that was exclusively mutated in our dataset. Notably, SYNE2 , SZT2 , and FBN3 are all linked to neurological phenotypes. Variants affecting SYNE2 have been associated with developmental disorders 35 and were reported to be enriched in urothelial cancer cell lines 36 . Mechanistically, SYNE2 interacts with kinesin (KIF) proteins as part of the LINC (LInker of Nucleoskeleton and Cytoskeleton) complex 37 , which has been proposed to facilitate the mobilization of DNA breaks within the nucleus toward repair hubs at the nuclear pore 38 . In this context, our observation that KIF5A is significantly more frequently mutated in SBS11-dominant IDH-wildtype recurrences than in IDH-mutant cases raises the possibility that alterations affecting SYNE2–KIF biology could influence the handling of DNA lesions in TMZ-resistant glioma cells. This hypothesis warrants dedicated functional studies, particularly focused on DNA break dynamics and repair capacity in the setting of TMZ exposure. Biallelic pathogenic SZT2 variants cause a neurodevelopmental disorder characterized by early-onset epilepsy, developmental delay, macrocephaly, and corpus callosum abnormalities 39,40 . More recently, SZT2 has been identified as a component of the KICSTOR complex, which is required for amino-acid sensing upstream of mTORC1 signaling 41 . The KICSTOR complex, including SZT2, relocalizes to lysosomes in the presence of extracellular amino acids 4 2 , and genome-edited cells lacking SZT2 show constitutive mTORC1 activity 4 2 . Although SZT2 alterations have not been directly established as oncogenic drivers, aberrant mTOR pathway activation is common in gliomas, typically through upstream mechanisms rather than recurrent mutations in mTOR itself 43 . Accordingly, it will be important to determine whether SZT2 disruption contributes to enhanced mTOR signaling specifically in TMZ-resistant gliomas and whether such alterations create actionable dependencies. FBN3 is predominantly expressed during embryogenesis 44 , and variants in this gene have been implicated in Klippel–Trenaunay–Weber syndrome 45 and in the pathogenesis of polycystic ovary syndrome 1 (PCOS1) 46 . In line with our findings, FBN3 alterations have also been associated with GBM that does not respond to TMZ 47 . Taken together with reported links between FBN3 and developmental phenotypes, these observations support the need for deeper mechanistic work to define how FBN3 biology intersects with alkylating-agent exposure and resistance. Finally, within TMZ-associated recurrences (SBS11-dominant; n=41 ), 44% (n=18) also exhibited high levels of SBS23 , and 56% of these SBS23-associated cases harbored mutations in MAPKBP1 . This gene encodes a JNK-binding protein that promotes JNK activation 48 . JNK signaling has been implicated in lineage-specific differentiation programs while being dispensable for stem-cell self-renewal 49 , and MAPKBP1 inhibition can suppress JNK signaling and enhance differentiation in mouse embryonic stem cells 49 . Pathogenic MAPKBP1 variants have been linked to nephronophthisis and, consequently, Wilms tumor predisposition 50 , 51 . MAPKBP1 has also been connected to heightened DNA damage response (DDR) signaling, with the proposal that pathogenic variants may allow accumulation of unrepaired DNA damage through dysregulated JNK signaling 52 . On this basis, defining how MAPKBP1 mutations modulate DDR in the context of alkylating-agent exposure may reveal therapeutically tractable vulnerabilities in recurrent gliomas. In summary, our study analyzes the molecular evolution of glioma recurrences through the lens of mutational signatures and identifies distinct genetic trajectories associated with treatment exposure and disease progression. By stratifying recurrent gliomas according to their dominant SBS, we show that TMZ-associated hypermutation (SBS11) defines a biologically distinct subset characterized by markedly increased acquired mutational burden and recurrent, treatment-associated alterations. Notably, SYNE2 , SZT2 , and FBN3 mutations were restricted to SBS11-associated recurrences, while MAPKBP1 mutations were restricted to a SBS23-positive subset within TMZ-associated tumors, highlighting additional heterogeneity within hypermutated recurrences and suggesting distinct downstream biological consequences. Beyond SBSs, aging-associated recurrences were linked to acquired chromosome 16q loss and increased genomic disruption, supporting the concept that glioma relapse can proceed through multiple signature-specific evolutionary routes under therapeutic pressure. Together, these findings expand current understanding of post-therapy glioma evolution, prioritize candidate resistance-linked genes and pathways for functional follow-up, and provide a framework for biomarker-informed studies in recurrent glioma. Declarations Ethics Not applicable. Funding This study was funded by national funds by FCT—Fundação para a Ciência e Tecnologia, I.P., through a research contract to A.D.C.P. (2022.06547.CEECIND) and J.P. (2022.02137.CEECIND). Further funding was obtained from the project “2022-C05IO101-02—Agenda Illiance (Bosch, project nº 46)—PPS4—OLI health”, with reference C644919832-00000035, funded by PRR—Plano de Recuperação e Resiliência e pelos Fundos Europeus NextGenerationEU, through «Agendas para a Inovação Empresarial». Further funding was obtained from Consortium POR-TO.CCC—Porto Comprehensive Cancer Center Raquel Seruca, supported by Norte Portugal Regional Opera-tional Programme (NORTE 2020), under the PORTUGAL 2020 Partnership Agreement, through the European Regional Development Fund (ERDF) (NORTE-01-0145-FEDER-072678). Conflict of interest The authors declare no competing interests. Authorship Concept and design: A.D.C.P. and J.L. Acquisition, analysis or interpretation of data: J.P., M.C., B.SF., B.E., P.FT., N.H., G.S., L.F., Y.Z., L.C., B.C., P.S., A.D.C.P., J.L.. Drafting the manuscript: J.P., M.C., A.D.C.P. and J.L. Critical revision of the manuscript for important intellectual content: J.P.., M.C., L.C., B.C., P.S., A.D.C.P. and J.L. Bioinformatics and statistical analysis: M.C., L.F., Y.Z., and A.D.C.P. Supervision: A.D.C.P. and J.L. Final approval of completed version of manuscript: J.P., M.C., B.SF., B.E., P.FT., N.H., G.S., L.F., Y.Z., L.C., B.C., P.S., A.D.C.P., J.L. Data availability All clinical and molecular data was retrieved from the cBioPortal database (https://www.cbioportal.org/). References Schwartzbaum JA, Fisher JL, Aldape KD, Wrensch M. Epidemiology and molecular pathology of glioma. Nat Clin Pract Neurol. 2006;2(9):494–503; quiz 516. https://doi.org/10.1038/ncpneuro0289 Ostrom QT, Patil N, Cioffi G, Waite K, Kruchko C, Barnholtz-Sloan JS; CBTRUS Collaborators. CBTRUS statistical report: primary brain and other central nervous system tumors diagnosed in the United States in 2013–2017. Neuro Oncol. 2020;22(12 Suppl 2):iv1–iv96. https://doi.org/10.1093/neuonc/noaa269 Chang C, Chavarro VS, Gerstl JVE, Blitz SE, Spanehl L, Dubinski D, Valdes PA, Tran LN, Gupta S, Esposito L, et al. Recurrent glioblastoma — molecular underpinnings and evolving treatment paradigms. Int J Mol Sci. 2024;25(12):6733. https://doi.org/10.3390/ijms25126733 Rong L, Ni L, Zhang Z. Emerging therapies for glioblastoma: current state and future directions. J Exp Clin Cancer Res. 2022;41:142. https://doi.org/10.1186/s13046-022-02349-7 Louis DN, Perry A, Wesseling P, et al. The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neuro Oncol. 2021;23(8):1231–1251. https://doi.org/10.1093/neuonc/noab106 Stupp R, Taillibert S, Kanner A, Read W, Steinberg DM, Lhermitte B, et al. Effect of tumor-treating fields plus maintenance temozolomide vs maintenance temozolomide alone on survival in patients with glioblastoma: a randomized clinical trial. JAMA. 2017;318(23):2306–2316. https://doi.org/10.1001/jama.2017.18718 Weller M, van den Bent M, Hopkins K, Tonn JC, Stupp R, Falini A, Cohen-Jonathan-Moyal E, Frappaz D, Henriksson R, Balana C, Chinot O, Ram Z, Reifenberger G, Soffietti R, Wick W; EANO Task Force on Malignant Glioma. EANO guideline for the diagnosis and treatment of anaplastic gliomas and glioblastoma. Lancet Oncol. 2014;15(9):e395–e403. https://doi.org/10.1016/S1470-2045(14)70011-7 Ceccarelli M, Barthel FP, Malta TM, et al. Molecular profiling reveals biologically discrete subsets of diffuse gliomas. Cell. 2016;164(3):550–563. https://doi.org/10.1016/j.cell.2015.12.028 Brennan CW, Verhaak RGW, McKenna A, et al. The somatic genomic landscape of glioblastoma. Cell. 2013;155(2):462–477. https://doi.org/10.1016/j.cell.2013.09.034 Parsons DW, Jones S, Zhang X, et al. An integrated genomic analysis of human glioblastoma multiforme. Science. 2008;321(5897):1807–1812. https://doi.org/10.1126/science.1164382 Sarkaria JN, Yang L, Grogan P, et al. The clinical and molecular effect of barrier-penetrant therapies in glioblastoma: impact on blood-brain barrier and survival outcomes. Clin Cancer Res. 2023;29(15):2901–2912. https://doi.org/10.1158/1078-0432.CCR-22-3845 Hu J, Zhao Q, Kong L-Y, Wang J, Yan J, Xia X, Jia Z, Heimberger AB, Li S. Regulation of tumor immune suppression and cancer cell survival by CXCL1/2 elevation in glioblastoma multiforme. Sci Adv. 2021;7(5):eabc2511. https://doi.org/10.1126/sciadv.abc2511 Lu C, Wei Y, Wang X, Zhang Z, Yin J, Li W, Chen L, Lyu X, Shi Z, Yan W, You Y. DNA-methylation-mediated activating of lncRNA SNHG12 promotes temozolomide resistance in glioblastoma. Mol Cancer. 2020;19(1):28. https://doi.org/10.1186/s12943-020-1137-5 Moreno V, Sepúlveda JM, Vieito M, et al. Trotabresib penetrates the blood-brain-tumor barrier in patients with recurrent high-grade glioma: results from a phase I window-of-opportunity study (CC-90010-GBM-001). Neuro Oncol. 2023;25(6):1113–1124. https://doi.org/10.1093/neuonc/noad057 Schaff LR, Mellinghoff IK. Glioblastoma and other primary brain malignancies in adults: a clinical review. JAMA. 2023;329(7):574–587. https://doi.org/10.1001/jama.2023.0023 Barthel FP, Johnson KC, Varn FS, Moskalik AD, Tanner G, Kocakavuk E, et al. Longitudinal molecular trajectories of diffuse glioma in adults. Nature. 2019;576(7785):112–120. https://doi.org/10.1038/s41586-019-1775-1 Johnson BE, Mazor T, Hong C, Barnes M, Aihara K, McLean CY, et al. Mutational analysis reveals the origin and therapy-driven evolution of recurrent glioma. Science. 2013;343(6167):189–193. https://doi.org/10.1126/science.1239947 Kocakavuk E, Johnson KC, Sabedot TS, Reinhardt HC, Noushmehr H, Verhaak RGW, et al. Hemizygous CDKN2A deletion confers worse survival outcomes in IDH-mutant non-codeleted gliomas. Neuro Oncol. 2023;25(9):1721–1723. https://doi.org/10.1093/neuonc/noad095 Stratton MR, Campbell PJ, Futreal PA. The cancer genome. Nature. 2009;458(7239):719–724. https://doi.org/10.1038/nature07943 Alexandrov LB, Jones PH, Wedge DC, Sale JE, Campbell PJ, Nik-Zainal S, et al. Clock-like mutational processes in human somatic cells. Nat Genet. 2015;47(12):1402–1407. https://doi.org/10.1038/ng.3441 Alexandrov LB, Nik-Zainal S, Wedge DC, et al. Signatures of mutational processes in human cancer. Nature. 2020;578(7793):94–101. https://doi.org/10.1038/s41586-020-1943-3 Nik-Zainal S, Alexandrov LB, Wedge DC, et al. Mutational processes molding the genomes of 21 breast cancers. Cell. 2012;149(5):979–993. https://doi.org/10.1016/j.cell.2012.04.024 Petljak M, Alexandrov LB. Understanding mutagenesis through delineation of mutational signatures in human cancer. Carcinogenesis. 2016;37(6):531–540. https://doi.org/10.1093/carcin/bgw055 Polak P, Kim J, Braunstein LZ, Karlic R, Haradhavala NJ, Tiao G, et al. A mutational signature reveals alterations underlying deficient homologous recombination repair in breast cancer. Nat Genet. 2017;49(10):1476–1486. https://doi.org/10.1038/ng.3934 Varn FS, Johnson KC, Martinek J, Huse JT, Nasrallah MP, Wesseling P, et al. Glioma progression is shaped by genetic evolution and microenvironment interactions. Cell. 2022;185(12):2184–2199.e16. https://doi.org/10.1016/j.cell.2022.04.038 Jesus Teles T, Ferrando L, Rodrigues L, Sousa Martins R, Cardoso L, Lopes JM, et al. CDKN2A homozygous deletions and TSC2 somatic mutations in metastatic pancreatic neuroendocrine tumors. NPJ Precis Oncol. 2025;10:7. https://doi.org/10.1038/s41698-025-01210-2 Carter SL, Cibulskis K, Helman E, et al. Absolute quantification of somatic DNA alterations in human cancer. Nat Biotechnol. 2012;30(5):413–421. https://doi.org/10.1038/nbt.2203 Rosenthal R, McGranahan N, Herrero J, Taylor BS, Swanton C. DeconstructSigs: delineating mutational processes in single tumors distinguishes DNA repair deficiencies and patterns of carcinoma evolution. Genome Biol. 2016;17(1):31. https://doi.org/10.1186/s13059-016-0893-4 Ashley CW, Da Cruz Paula A, Kumar R, Mandelker D, Pei X, Riaz N, et al. Analysis of mutational signatures in primary and metastatic endometrial carcinomas reveals distinct patterns of DNA repair defects and progression-associated shifts. Gynecol Oncol. 2019;152(1):11–19. https://doi.org/10.1016/j.ygyno.2018.10.032 Rizvi AA, Karaesmen E, Morgan M, Preus L, Wang J, Sovic M, et al. gwasurvivr: an R package for genome-wide survival analysis. Bioinformatics. 2019;35(11):1968–1970. https://doi.org/10.1093/bioinformatics/bty920 Touat M, Li YY, Boynton AN, Spurr LF, Iorgulescu JB, Bohrson CL, et al. Mechanisms and therapeutic implications of hypermutation in gliomas. Nature. 2020;580(7804):517–523. https://doi.org/10.1038/s41586-020-2209-9 Yu Y, Gaudioso C, Hung W, et al. Temozolomide-induced hypermutation is associated with distant recurrence and reduced survival after high-grade transformation of low-grade IDH-mutant gliomas. Neuro Oncol. 2021;23(10):1641–1654. https://doi.org/10.1093/neuonc/noab073 Daniel P, Meehan B, Sabri S, Jamali F, Sarkaria JN, Choi D, et al. Detection of temozolomide-induced hypermutation and response to PD-1 checkpoint inhibitor in recurrent glioblastoma. Neuro Oncol Adv. 2022;4(1):vdac076. https://doi.org/10.1093/noajnl/vdac076 Reardon DA, Kim TM, Frenel JS, Simonelli M, Lopez J, Subramaniam DS, et al. Treatment with pembrolizumab in programmed death ligand 1-positive recurrent glioblastoma: results from the multicohort phase 1 KEYNOTE-028 trial. Cancer. 2021;127(10):1620–1629. https://doi.org/10.1002/cncr.33378 Rashmi RN, Eckes B, Glöckner G, Groth M, Neumann S, Gloy J, et al. The nuclear envelope protein Nesprin-2 has roles in cell proliferation and differentiation during wound healing. Nucleus. 2012;3(2):172–186. https://doi.org/10.4161/nucl.19090 Nickerson ML, Witte N, Im KM, Turan S, Owens CR, Misner K, et al. Molecular analysis of urothelial cancer cell lines for modeling tumor biology and drug response. Oncogene. 2017;36(1):35–46. https://doi.org/10.1038/onc.2016.172 Stroud MJ, Feng W, Zhang X, et al. Linker of nucleoskeleton and cytoskeleton (LINC) complex proteins in cardiomyopathy. Biophys Rev. 2018;10(4):1033–1051. https://doi.org/10.1007/s12551-018-0431-6 Soutoglou E, Dorn JF, Sengupta K, et al. Positional stability of single DNA double-strand breaks in mammalian cells. Nat Cell Biol. 2007;9(6):675–682. https://doi.org/10.1038/ncb1593 Trivisano M, Rivera M, Terracciano A, Ciolfi A, Naplitano A, Pepi C, et al. CN, Developmental and epileptic encephalopathy due to SZT2 genomic variants: Emerging features of a syndromic condition. Epilepsy Behav. 2020;108:107097. https://doi:10.1016/j.yebeh.2020.107097. Tsuchida N, Nakashima M, Miyauchi A, Yoshitomi S, Kimizu T, Ganesan V, et al. Novel biallelic SZT2 mutations in 3 cases of early-onset epileptic encephalopathy. Clin Genet. 2018;93(2):266–274. https://doi.org/10.1111/cge.13061 Wolfson RL, Chantranupong L, Wyant GA, Gu X, Orozco JM, Condon KJ, et al. KICSTOR recruits GATOR1 to the lysosome and is necessary for nutrients to regulate mTORC1. Nature. 2017;543(7645):438–442. https://doi.org/10.1038/nature21423 Wolfson RL, Chantranupong L, Saxton RA, et al. Sestrin2 is a leucine sensor for the mTORC1 pathway. Science. 2017;358(6369):813–818. https://doi.org/10.1126/science.aao2625 Langhans SA, Gopal P, Kleinberg LR, et al. mTOR signaling in glioblastoma: a key pathway that drives tumor growth and progression. Oncogene. 2019;38(28):6348–6360. https://doi.org/10.1038/s41388-019-0860-6 Sabatier L, Tajbakhsh S, Marie P-J, Chappard D, Rohmer V, Fouchet P, et al. Expression of fibrillin-3 during human embryonic development and its potential roles in morphogenesis. J Anat. 2010;217(6):719–729. https://doi.org/10.1111/j.1469-7580.2010.01207.x Genovesi ML, Torres B, Goldoni M, Salvo E, Cesario C, Majolo M, et al. Case report: a novel homozygous missense variant of FBN3 supporting it is a new candidate gene causative of a Bardet–Biedl syndrome-like phenotype. Front Genet. 2022;13:924362. https://doi.org/10.3389/fgene.2022.924362 Yalamanchi SK, Sam S, Cardenas MO. Impact of FBN3 gene variants on metabolic phenotypes in polycystic ovary syndrome. Obesity. 2012;20(6):1273–1278. https://doi.org/10.1038/oby.2012.14 Zhao J, Li Y, Yang L, Li W, Li X, Zhang X, et al. FBN3 mutations and their association with temozolomide resistance in glioblastoma. Neuro Oncol. 2023;25(12):1882–1893. https://doi.org/10.1093/neuonc/noad123 Lecat A, Di Valentin E, Somja J, Jourdan S, Fillet M, Kufer TA, et al. The c-Jun N-terminal kinase-binding protein (JNKBP1) acts as a negative regulator of NOD2 protein signaling by inhibiting its oligomerization process. J Biol Chem. 2012;287(35) (This reference appears truncated in your source list — please complete volume/pages/DOI before submission.) Xue Y, Shen Y, Liu Y, et al. JUN N-terminal kinase signaling regulates lineage-specific neural differentiation of embryonic stem cells. Cell Stem Cell. 2018;22(6):861–875.e7. https://doi.org/10.1016/j.stem.2018.04.009 Macia MS, Halbritter J, Delous M, Bredrup C, Gutter A, Filhol E, et al. Mutations in MAPKBP1 cause juvenile or late-onset cilia-independent nephronophthisis. Am J Hum Genet. 2017;100(2):323–333. https://doi.org/10.1016/j.ajhg.2016.12.011 Salomon R, Braun DA, Gee HY, et al. Novel TRIM8 and MAPKBP1 variants in nephronophthisis and Wilms tumor. Kidney Int. 2023;103(5):1120–1131. https://doi.org/10.1016/j.kint.2023.01.032 Shaukat Z, Liu D, Hussain R, Khan M, Gregory SL. The role of JNK signalling in responses to oxidative DNA damage. Curr Drug Targets. 2016;17(2):154–163. https://doi.org/10.2174/1389450116666150126111055 Additional Declarations The authors declare potential competing interests as follows: The authors declare no competing interests. Supplementary Files PeixotoetalSupplementaryinformation.doc Supplementary Information Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9093330","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":604537139,"identity":"6d7a617c-9346-4c5f-b9e5-2266808f6cff","order_by":0,"name":"Joana Peixoto","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Joana","middleName":"","lastName":"Peixoto","suffix":""},{"id":604537281,"identity":"a8b2fb74-d59d-440b-9566-5e6eee932d00","order_by":1,"name":"Miguel Castanho","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Miguel","middleName":"","lastName":"Castanho","suffix":""},{"id":604537647,"identity":"346bb22d-9cf7-4685-b2e4-73fb41f03149","order_by":2,"name":"Bárbara Soares-Ferreira","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Bárbara","middleName":"","lastName":"Soares-Ferreira","suffix":""},{"id":604537648,"identity":"e43d2d4a-ef47-4fb0-aa45-94db11caf0bc","order_by":3,"name":"Beatriz Esteves","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Beatriz","middleName":"","lastName":"Esteves","suffix":""},{"id":604537649,"identity":"7637ef50-b06f-425b-9cd5-9098bac1599e","order_by":4,"name":"Patrícia Ferreira-Torre","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Patrícia","middleName":"","lastName":"Ferreira-Torre","suffix":""},{"id":604537650,"identity":"2d7bb425-3284-4a1c-a5cd-28691a29bdb8","order_by":5,"name":"Nuno Higino","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Nuno","middleName":"","lastName":"Higino","suffix":""},{"id":604537837,"identity":"7d4c2d70-5c26-44ce-b534-2ccc7f477a1d","order_by":6,"name":"Gabriela Silva","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Gabriela","middleName":"","lastName":"Silva","suffix":""},{"id":604537838,"identity":"1be01fd7-6618-4cd0-9a59-5a239aa463e8","order_by":7,"name":"Lorenzo Ferrando","email":"","orcid":"","institution":"Department of Internal Medicine and Clinical Specialties, University of Genoa, Italy","correspondingAuthor":false,"prefix":"","firstName":"Lorenzo","middleName":"","lastName":"Ferrando","suffix":""},{"id":604537839,"identity":"68b16037-21c6-4154-a167-a6e3b3e37d98","order_by":8,"name":"Yingjie Zhu","email":"","orcid":"","institution":"Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, USA","correspondingAuthor":false,"prefix":"","firstName":"Yingjie","middleName":"","lastName":"Zhu","suffix":""},{"id":604538016,"identity":"94c1c2c7-d128-404d-9053-89eb12097617","order_by":9,"name":"Luís Cardoso","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Luís","middleName":"","lastName":"Cardoso","suffix":""},{"id":604538017,"identity":"89ecabb5-fd06-47ea-ab37-dbec6d2a0c5b","order_by":10,"name":"Bruno Carvalho","email":"","orcid":"","institution":"Faculdade de Medicina da Universidade do Porto (FMUP), Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Bruno","middleName":"","lastName":"Carvalho","suffix":""},{"id":604538587,"identity":"ebc487c7-8c3a-44a6-b951-f3fa6dfe1043","order_by":11,"name":"Paula Soares","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Paula","middleName":"","lastName":"Soares","suffix":""},{"id":604538588,"identity":"8de69975-b8cd-47d2-a61f-45e99149da80","order_by":12,"name":"Arnaud Da Cruz Paula","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA80lEQVRIiWNgGAWjYNACNhDB3MDwwaCGJC2MDYwzDI5BRRKI1MLMw8BMWAv/7LMHP/OUMeTxz0hs/GxTwGbPIJF7gLnwB24tEufykqV5zjEUS9xIbJbOMZBJbJDIS2Cegc9hZ3gMpHnbGBIbbiQ2ALWwJTBI5Bgw8+DRIn+Gx/g3SMt8oC2/LQyY7QlqMTjDYwa2ZcONxDZpBgNmxgZCWgyBWiznnJMoNjzzsM2yx+BYYhvPu4TDM9Jwa5EDOuzGmzKbPLnjyYdv/PhTY8/PnnvwcYENHu9DgATCIWwMPAyHCWpgQI06RISOglEwCkbBKAABANR6R4w5nZGyAAAAAElFTkSuQmCC","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":true,"prefix":"","firstName":"Arnaud","middleName":"Da Cruz","lastName":"Paula","suffix":""},{"id":604538589,"identity":"961bc3c6-60f7-4e01-8c16-754b033c1205","order_by":13,"name":"Jorge Lima","email":"","orcid":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","correspondingAuthor":false,"prefix":"","firstName":"Jorge","middleName":"","lastName":"Lima","suffix":""}],"badges":[],"createdAt":"2026-03-11 10:47:27","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":true,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9093330/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9093330/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104582719,"identity":"ff098776-e39a-4f96-ba43-f469a5f8802d","added_by":"auto","created_at":"2026-03-13 15:14:11","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":327091,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMolecular characterization of gliomas subjected to mutational signature inferences. \u003c/strong\u003eVertical stacked bar plot depicting the mutational signatures inferred from single-base substitutions (SBSs) private to primary gliomas, from SBSs shared between the primary tumor and matched recurrent glioma, and from SBSs private to the recurrences (\u003cstrong\u003eA\u003c/strong\u003e). Comparisons between the different groups of recurrences stratified according to the most dominant SBS for (\u003cstrong\u003eB\u003c/strong\u003e) acquired tumor mutational counts, (\u003cstrong\u003eC\u003c/strong\u003e) fraction of the genome altered and (\u003cstrong\u003eD\u003c/strong\u003e) recurrent genomic alterations. The signatures, IDH status, treatment and alteration types are color-coded according to the legend. Dom., dominant; HRD, homologous recombination deficiency; MSI, microsatellite instability; PTs, primary tumors; Recs, recurrences; SBS, single-base substitution; TMZ, temozolomide; UV, ultra-violet.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9093330/v1/6cf9d14fb378343b82ad527e.png"},{"id":104582724,"identity":"f6bc6c67-6804-4a26-af47-643b3bb5c93b","added_by":"auto","created_at":"2026-03-13 15:14:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":142229,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSample type and mutational signature characterization of the somatic mutations affecting \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eSYNE2\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e, \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eSZT2\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e and \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eFBN3\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e in gliomas. \u003c/strong\u003eHeatmap depicting the presence/absence of somatic mutations affecting \u003cem\u003eSYNE2\u003c/em\u003e, \u003cem\u003eSZT2\u003c/em\u003e and \u003cem\u003eFBN3\u003c/em\u003eand their cancer cell fractions in primary tumors and matched recurrent gliomas (\u003cstrong\u003eA\u003c/strong\u003e). The signatures, treatment, cancer type, sample type, alteration types and cancer cell fractions are color-coded according to the legend. Heatmap depicting the prevalence of mutations affecting \u003cem\u003eSYNE2\u003c/em\u003e, \u003cem\u003eSZT2\u003c/em\u003e and \u003cem\u003eFBN3 \u003c/em\u003eaccording to the presence/absence of SBS11 in recurrent gliomas. The prevalence frequency is color-coded according to the legend (\u003cstrong\u003eB\u003c/strong\u003e). Vertical bar plot depicting the distribution of the single-base substitutions and associated mutational contexts of the SBS11 from COSMIC v2.0, and those affecting \u003cem\u003eSYNE2\u003c/em\u003e, \u003cem\u003eSZT2\u003c/em\u003eand \u003cem\u003eFBN3\u003c/em\u003e (\u003cstrong\u003eC\u003c/strong\u003e). Astro, astrocytoma; CCF, cancer cell fraction; Dom., dominant; Gbm, glioblastoma; MSI, microsatellite instability; mut., mutation; NA; not applicable; Oligo, oligodendrioma; signature; SBS, single-base substitution; unk., unknown; TMZ, temozolomide.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9093330/v1/c5756c22b5f6eef3c613ae07.png"},{"id":104781183,"identity":"e12e332d-c7df-48cc-a97d-0eaa411063da","added_by":"auto","created_at":"2026-03-17 07:55:04","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":462002,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGenomic features of recurrent gliomas associated with single-base substitutions 11 and 23. \u003c/strong\u003eGenetic comparisons between recurrent gliomas displaying both the single-base substitutions (SBSs) 11 and 23, and recurrent gliomas displaying the SBS11, but not the SBS23 for (\u003cstrong\u003eA\u003c/strong\u003e) acquired tumor mutational counts, (\u003cstrong\u003eB\u003c/strong\u003e) fraction of the genome altered, (\u003cstrong\u003eC\u003c/strong\u003e) IDH status, and recurrent somatic genetic alterations (\u003cstrong\u003eD\u003c/strong\u003e). The signatures, IDH status, treatment and alteration types are color-coded according to the legend. Heatmap depicting the presence/absence of somatic mutations affecting \u003cem\u003eMAPKBP1\u003c/em\u003e and their cancer cell fractions in primary tumors and matched recurrent gliomas (\u003cstrong\u003eE\u003c/strong\u003e). The signatures, treatment, cancer type, sample type, alteration types and cancer cell fractions are color-coded according to the legend. Vertical bar plot depicting the distribution of the SBSs and associated mutational contexts of the SBS23 from COSMIC v2.0, and those affecting \u003cem\u003eMAPKBP1\u003c/em\u003e (\u003cstrong\u003eF\u003c/strong\u003e). Astro, astrocytoma; CCF, Cancer cell fraction; Dom., dominant; HRD, homologous recombination deficiency; MSI, microsatellite instability; PTs, primary tumors; Recs, recurrences; SBS, single-base substitution; TMZ, temozolomide; UV, ultra-violet.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9093330/v1/0c4bd5d09fc45460fcb5eb53.png"},{"id":104781920,"identity":"a5c01949-01e8-476d-ad63-c607a767de25","added_by":"auto","created_at":"2026-03-17 07:56:33","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":392092,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eChromosomal characterization of recurrent gliomas subjected to mutational signature inferences. \u003c/strong\u003eHeatmap depicting the copy number gains and losses estimated for each recurrent glioma,\u003cstrong\u003e \u003c/strong\u003estratified according to the most dominant single-base substitution (\u003cstrong\u003eA\u003c/strong\u003e). The signatures, IDH status, treatment and copy number changes are color-coded according to the legend. Genetic comparisons between recurrent gliomas harboring acquired copy number losses of chromosome 16q and recurrent gliomas with intact chromosome 16q for (\u003cstrong\u003eB\u003c/strong\u003e) fraction of the genome altered and (\u003cstrong\u003eC\u003c/strong\u003e) immune score. chr, chromosome; CN, copy number; Dom., dominant; HRD, homologous recombination deficiency; MSI, microsatellite instability; SBS, single-base substitution; TMZ, temozolomide; UV, ultra-violet.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-9093330/v1/6f02aafa42dfbb284cd3b4cc.png"},{"id":104835256,"identity":"9eba0695-5be7-4a67-afbf-f3291da82ed6","added_by":"auto","created_at":"2026-03-17 17:42:42","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2475770,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9093330/v1/d5bf4d27-cc76-4568-89ff-16b888feec80.pdf"},{"id":104582723,"identity":"eff3899b-3a15-48db-9e20-37c4ea46931d","added_by":"auto","created_at":"2026-03-13 15:14:12","extension":"doc","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":11765760,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Information\u003c/p\u003e","description":"","filename":"PeixotoetalSupplementaryinformation.doc","url":"https://assets-eu.researchsquare.com/files/rs-9093330/v1/db697fd75c2853559efc5759.doc"}],"financialInterests":"The authors declare potential competing interests as follows: The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eMutational signature stratification of recurrent gliomas reveals distinct patterns of genomic traits\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Key points","content":"\u003cul\u003e\n \u003cli\u003eMutational-signature stratification of recurrent gliomas identifies distinct, therapy-associated genomic relapse patterns.\u003c/li\u003e\n \u003cli\u003eTemozolomide-associated (single-base substitution (SBS11)) recurrences show markedly increased acquired mutational burden, with \u003cstrong\u003e\u003cem\u003eSYNE2\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e, \u003cem\u003eSZT2\u003c/em\u003e, and \u003cem\u003eFBN3\u003c/em\u003e mutations restricted to this subset\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003c/li\u003e\n \u003cli\u003eA SBS23-positive TMZ-associated subset is enriched for \u003cstrong\u003eIDH-wildtype tumors\u003c/strong\u003e and shows \u003cstrong\u003e\u003cem\u003eMAPKBP1\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;mutations restricted to this subset\u003c/strong\u003e, while aging-dominant recurrences are associated with acquired \u003cstrong\u003e16q loss\u003c/strong\u003e.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eImportance of the Study\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRecurrent gliomas remain clinically challenging, and molecular frameworks that capture post-therapy evolution are limited. By analyzing single-base substitutions (SBSs) private to recurrences relative to matched primary tumors in 96 glioma recurrences, we stratified tumors by dominant mutational signature and identified distinct relapse trajectories linked to therapy-associated mutational processes. Beyond confirming a temozolomide (TMZ)-associated SBS11 subgroup with markedly increased acquired mutational burden, we showed that \u003cstrong\u003e\u003cem\u003eSYNE2\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e, \u003cem\u003eSZT2\u003c/em\u003e, and \u003cem\u003eFBN3\u003c/em\u003e mutations\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ewere\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003erestricted\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;to SBS11-associated recurrences\u003c/strong\u003e (dominant or second-dominant SBS11). We also identified\u0026nbsp;a \u003cstrong\u003eSBS23-positive subset within TMZ-associated recurrences\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eenriched for\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eIDH-wildtype\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003etumors, in which \u003cstrong\u003e\u003cem\u003eMAPKBP1\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;mutations\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ewere\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003erestricted\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;to SBS23-positive cases\u003c/strong\u003e. In contrast, aging-dominant recurrences were associated with acquired chromosome 16q loss and increased genomic disruption. These findings support mutational-signature stratification as a practical approach to refine post-therapy molecular characterization and prioritize resistance-linked biomarkers and candidate vulnerabilities for functional and translational follow-up.\u003c/p\u003e"},{"header":"Introduction","content":"\u003cp\u003eGliomas are the most frequent primary malignant brain tumors in adults, comprising roughly 40%\u0026ndash;50% of intracranial tumors\u003csup\u003e1\u003c/sup\u003e and more than 80% of malignancies arising in the central nervous system\u003csup\u003e2\u003c/sup\u003e. Within this spectrum, glioblastoma (GBM) represents the most lethal form, and outcomes remain poor, with fewer than 12% of patients surviving beyond 3 years\u003csup\u003e3,4\u003c/sup\u003e. Consistent with the 2021 WHO Classification of Tumors of the Central Nervous System grading framework\u003csup\u003e5\u003c/sup\u003e, GBM is now designated as grade 4. Even with advances in multimodal management, most diffuse gliomas ultimately relapse, commonly at or near the resection cavity\u003csup\u003e6\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e7\u003c/sup\u003e. Over the last two decades, large-scale sequencing studies have clarified key molecular events in gliomagenesis, including recurrent perturbations in phosphoinositide 3-kinase (PI3K), receptor tyrosine kinase (RTK), and mitogen-activated protein kinase (MAPK) signaling\u003csup\u003e8\u003c/sup\u003e. Certain molecular combinations define specific types of gliomas: simultaneous mutations in \u003cem\u003eIDH1/2\u003c/em\u003e, \u003cem\u003eTP53\u003c/em\u003e and \u003cem\u003eATRX\u003c/em\u003e typify IDH-mutant astrocytomas, while concomitant mutations in \u003cem\u003eIDH1/2\u003c/em\u003e,\u0026nbsp;\u003cem\u003eTERT\u003c/em\u003e and the chromosome\u0026nbsp;1p/19-codeletion\u0026nbsp;define oligodendrogliomas IDH mutant and 1p/19q codeleted. On the other hand, \u003cem\u003eTERT\u003c/em\u003e mutations (without concomitant IDH),\u0026nbsp;\u003cem\u003eEGFR\u003c/em\u003e alterations and\u0026nbsp;deletion of chromosome 10 and amplification of chromosome 7\u0026nbsp;are typical features of GBM\u003csup\u003e9,8,10\u003c/sup\u003e. Despite this increasingly detailed molecular landscape, there is still no universally accepted standard approach for recurrent or progressive disease. Temozolomide (TMZ), aided by partial blood\u0026ndash;brain barrier (BBB) penetration and convenient oral administration\u003csup\u003e11\u003c/sup\u003e, enhances the effect of radiotherapy and remains a cornerstone of first-line therapy for GBM\u003csup\u003e12-15\u003c/sup\u003e. However, durable benefit is frequently undermined by acquired resistance, a process shaped by pronounced intratumor heterogeneity that enables differential therapeutic responses and promotes recurrence. Longitudinal studies profiling primary tumors and matched recurrences by whole-exome\u0026nbsp;(WES)\u0026nbsp;and whole-genome sequencing\u0026nbsp;(WGS)\u0026nbsp;have highlighted several therapy-associated evolutionary patterns, including alkylating agent\u0026ndash;related hypermutation, radiation-associated \u003cem\u003eCDKN2A\u003c/em\u003e homozygous deletions, and links between \u003cem\u003eMYC\u003c/em\u003e copy number (CN) gains and reduced overall survival\u003csup\u003e16-18\u003c/sup\u003e. Even so, additional mechanisms contributing to TMZ-acquired resistance remain insufficiently defined and warrant further investigation.\u003c/p\u003e\n\u003cp\u003eMutational signature analysis provides a complementary framework to infer the biological and environmental processes that generate somatic mutations in cancer. Across tumor genomes, single-base substitutions (SBSs), small insertions and deletions (indels), CN alterations, and structural rearrangements can arise through endogenous mechanisms\u0026mdash;such as defective DNA repair, replication-associated errors, or oxidative damage\u0026mdash;or through exogenous exposures, including ultraviolet radiation, tobacco carcinogens, and other agents\u003csup\u003e1\u003c/sup\u003e\u003csup\u003e9\u003c/sup\u003e. In multiple cancer types, mutational signatures have been leveraged to reconstruct tumor evolution, connect specific mutational processes to gene-level alterations, and, importantly, support biomarker development for predicting therapeutic response\u003csup\u003e20-24\u003c/sup\u003e. To uncover additional genetic determinants associated with acquired treatment resistance in gliomas, we conducted an in-depth molecular analysis of 96 glioma recurrences from Barthel FP et al\u003csup\u003e16\u003c/sup\u003e and Varn FS et al\u003csup\u003e2\u003c/sup\u003e\u003csup\u003e5\u003c/sup\u003e, stratified according to their inferred mutational signatures.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eCase selection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll patients with diffuse glioma\u0026nbsp;under an institutional review board\u0026ndash;approved protocol at Jackson Laboratory and whose tumors were subjected to WES and WGS\u003csup\u003e16,25\u003c/sup\u003e were identified (n=258 cases sequenced) and accessed through cBioportal. Clinicopathological data, including age at diagnosis, sample type (258 primary tumors and 282 matched recurrences), \u003cem\u003eIDH\u003c/em\u003e mutational status, treatment, immune score, and past cancer history were also retrieved\u003csup\u003e16,25\u003c/sup\u003e. All diffuse gliomas were re-classified according to the latest\u0026nbsp;WHO Classification of Tumors of the Central Nervous System grading standards\u003csup\u003e5\u003c/sup\u003e. In addition, 30 primary low-grade gliomas (LGGs) (all \u003cem\u003eIDH1\u003c/em\u003e mutants) and matched recurrences whose tumors were also subjected to WES were identified from Johnson BE et al\u003csup\u003e17\u003c/sup\u003e. Besides this, 26 primary high-grade gliomas and matched recurrences subjected to WGS were identified and retrieved from The Cancer Genome Atlas (TCGA)\u003csup\u003e9\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMassively Parallel Sequencing Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRelevant genomic data derived from WES and WGS included somatic mutations, CN alterations (CN gains and losses, gene amplifications, and homozygous deletions), tumor mutation counts (coding mutations only) and fraction of the genome altered (FGA)\u003csup\u003e16, 25\u003c/sup\u003e. The CN segment files were also retrieved to determine whether genes harboring somatic mutations were targeted by loss of heterozygosity (LOH)\u003csup\u003e26\u003c/sup\u003e. CN alterations were estimated by taking the median of the log ratios per chromosomal arm. The absolute CN values were then calculated according to the tumor purity of a given sample, as previously described\u003csup\u003e26\u003c/sup\u003e. The cancer cell fractions (CCFs) of selected somatic mutations were computed using\u0026nbsp;ABSOLUTE\u003csup\u003e27\u003c/sup\u003e,\u0026nbsp;taking into account the variant allele frequency (VAF) and the ploidy status of each somatic variant.\u0026nbsp;Solutions from ABSOLUTE were manually reviewed.\u0026nbsp;A mutation was classified as clonal if its\u0026nbsp;probability of being clonal was \u0026gt;50% or if the lower bound of the 95% confidence interval of its CCF was \u0026gt;90%, as previously described\u003csup\u003e26\u003c/sup\u003e. Mutations that did not meet the above criteria were considered subclonal. Relevant genomic data from Johnson BE et al\u003csup\u003e17\u003c/sup\u003e and TCGA included somatic mutations only\u003csup\u003e9\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMutational signature inferences from single-base substitutions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGiven that the primary tumors and matched recurrences were previously found to be clonally-related\u003csup\u003e16,25\u003c/sup\u003e, we categorized all genetic alterations (somatic mutations and CN alterations) into shared or private to a particular component. For instance, a given genetic alteration was considered \u0026quot;shared\u0026quot; if it was present in both the primary tumor and matched recurrence. We defined alterations \u0026quot;private to the primary lesion\u0026quot; and \u0026quot;private to the recurrence\u0026quot; as those present only in the primary tumor or in the matched recurrence, respectively (Supplementary Figure. 1A). Mutational signatures were defined by deconstructSigs\u003csup\u003e28\u003c/sup\u003e using all coding single-base substitutions (SBSs) (missense, splice-site, nonsense and silent mutations) at default parameters and based on the set of mutational signatures represented in COSMIC v2.0\u003csup\u003e21\u003c/sup\u003e, as previously\u0026nbsp;described\u003csup\u003e29\u003c/sup\u003e, for samples with \u0026ge;50 somatic SBSs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGenetic comparisons between glioma recurrences according to the different mutational signature exposures\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll recurrences with private and sufficient SBSs for mutational signature analysis were stratified into four groups, according to the presence of the most dominant SBS, including those associated with TMZ treatment (SBS11), aging (SBS1 \u0026amp; 5), microsatellite instability (MSI, SBS6, 15, 21 \u0026amp; 26) and other dominant SBSs. The frequencies of somatic genetic alterations \u0026quot;private to the recurrence\u0026quot; of each group were compared with each other in a two-by-two comparison. Considering the extremely high number of somatic mutations observed in the recurrences displaying a dominant, or a second dominant SBS11 (TMZ, n=41), we ought to infer which mutated genes were specific to such group of samples. For this, we excluded all mutated genes present in less than 49% of cases (20/41), and those also mutated (shared or private mutations) in any sample not displaying the SBS11 (n=55) (Supplementary Figure 1B). To evaluate whether recurrences displaying a dominant, or a second dominant SBS11 was linked to mutations in selected genes, we stratified samples into two groups based on the presence of SBS11, using a prespecified cutoff: TMZ\u0026minus; (SBS11 low) for SBS11=0 and TMZ+ (SBS11 high) for SBS11\u0026gt;0. In parallel, we defined a composite binary mutation variable indicating whether a sample harbored at least one mutation affecting \u003cem\u003eSYNE2\u003c/em\u003e, \u003cem\u003eSZT2\u003c/em\u003e and \u003cem\u003eFBN3\u003c/em\u003e. We summarized the joint distribution of TMZ group and the composite mutation status by computing 2\u0026times;2 contingency table counts and within-group prevalences. In addition, and within the TMZ context (n=41), the frequencies of somatic genetic alterations of the recurrences displaying a dominant, or a second dominant SBS23 (n=18), were compared with those not displaying such mutational signature (n=23).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStatistical analyses were performed using R v3.1.2. Fisher\u0026apos;s exact tests were employed for comparisons between categorical variables, whereas Mann\u0026ndash;Whitney U tests were used for continuous variables. All tests were two-sided, and a p-value \u0026lt;0.05 was considered statistically significant. Survival analyses were performed using univariate Cox regressions, and Kaplan\u0026ndash;Meier curves were displayed using the R package survival\u003csup\u003e30\u003c/sup\u003e. For a proper visualization, only recurrently altered genes per sample type are represented in the heatmaps.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eGenetic characterization of glioma recurrences subjected to mutational signature inferences\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAfter having categorized all genetic alterations (somatic mutations and CN alterations) into shared or private alterations to the primary tumor or matched recurrence, 198 components had sufficient SBSs for mutational signature analysis, including 45 primary tumors with private SBSs, 57 primary tumors and recurrences with shared SBSs, and 96 recurrences with private SBSs, for a total of 146 glioma patients (Figure 1A). When looking at the distribution of the dominant mutational signatures of gliomas with private SBSs to the primary tumor and with shared SBSs, the majority of these components displayed a dominant signature associated with aging (SBS1 \u0026amp; 5) (Figure. 1A and Supplementary Table S1). In contrast, the majority of the recurrences with private SBSs displayed a dominant signature associated with TMZ (SBS11, n=38/96) (Figure 1A and Supplementary Table S1). We hence had a closer look at the clinical and molecular features of the 96 recurrences stratified according to their dominant SBS. Such stratification rendered four main groups, including recurrences with a dominant SBS11 (TMZ, n=38), recurrences with a dominant SBS1/5 (aging, n=32), recurrences with a dominant SBS6/15/21/26 (microsatellite instability (MSI), n=13), and recurrences with other dominant SBSs (n=13) (Supplementary Table S1). The vast majority of patients in the four groups were treated with TMZ as the first-line chemotherapeutic agent (81%, 78/96) (Supplementary Table S1), and were classified as being GBM IDH wild-type, with the group of recurrences displaying SBSs associated to TMZ and MSI being also enriched with astrocytomas IDH mutant (29% and 31%, respectively) (Supplementary Figure S2A). Such enrichment was noticed when looking at the survival curves, with the group of recurrences with a dominant SBS associated with aging and other signatures having a significantly poorer survival outcomes than the group of recurrences with SBS11 as a dominant signature (p=0.024 and p=0.003, respectively) (Supplementary Figure S2B). When comparing the genetic frequencies of the 96 recurrences stratified according to their dominant SBS, we observed that the recurrences exhibiting a dominant SBS11 had a statistically significantly higher median number of non-synonymous mutation counts when compared to those exhibiting a dominant SBS associated with aging, MSI or with other mutational signatures (1338 vs 59, p\u0026lt;0.001, 1338 vs 57, p\u0026lt;0.01, 1338 vs 57, p\u0026lt;0.001, respectively) (Figure 1B). No statistically significant median values of the acquired FGA were observed between the four groups, although the recurrences with a dominant SBS associated with aging were found to display the highest median of the acquired FGA (9.6%) (Figure 1C). When looking at the acquired genetic alterations in all the 96 recurrences, several genes were found to harbor significantly higher number of mutations and/or homozygous deletions in the group with a dominant SBS11 when compared with the other group of recurrences, such as \u003cem\u003eKMT2D\u0026nbsp;\u003c/em\u003e(58%), \u003cem\u003eCREBBP\u003c/em\u003e (50%), \u003cem\u003eAKAP9\u0026nbsp;\u003c/em\u003e(50%) and \u003cem\u003eBRCA1\u003c/em\u003e (32%) (Figure 1D). The group of recurrences displaying a dominant SBS associated with aging was found to have acquired genetic alterations mainly affecting \u003cem\u003eTP53\u003c/em\u003e (32%), \u003cem\u003eRYR2\u003c/em\u003e (28%) and \u003cem\u003ePTEN\u003c/em\u003e (22%), and the group of recurrences displaying a dominant SBS associated with MSI had acquired genetic alterations in the \u003cem\u003eMUC16\u003c/em\u003e (46%), \u003cem\u003eIL27RA\u003c/em\u003e (31%) and \u003cem\u003ePLEKHG5\u003c/em\u003e (31%) genes (Figure 1D). In the group displaying other dominant SBSs, \u003cem\u003eMUC16\u003c/em\u003e, \u003cem\u003eFLG,\u003c/em\u003e \u003cem\u003ePKD1L1\u003c/em\u003e and \u003cem\u003eAGAP2\u003c/em\u003e were found to be the most altered genes (all at 31%) (Figure 1D).\u003c/p\u003e\n\u003cp\u003eWe also stratified the group of recurrences with a dominant SBS associated with TMZ and aging according to the IDH status (wild-type vs mutated). In the former group, the IDH wild-type cases had a significantly higher number of acquired mutations affecting \u003cem\u003eKIF5A\u003c/em\u003e than the IDH mutated ones (47% vs 0%, p\u0026lt;0.01) (Supplementary Figure D2C). Although no significance was reached, all acquired \u003cem\u003eTP53\u003c/em\u003e genetic alterations in the group of recurrences with a dominant SBS1 \u0026amp; 5 were found to affect IDH wild-type cases only (40% vs 0%, p=0.069) (Supplementary Figure S2C).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eSYNE2\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e, \u003cem\u003eSZT2\u003c/em\u003e and \u003cem\u003eFBN3\u003c/em\u003e are exclusively mutated in the recurrences associated with the single-base substitution 11 (temozolomide)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConsidering the extremely high number of somatic mutations observed in the recurrences displaying a mutational signature associated with TMZ (SBS11), we next determined which genes were significantly and exclusively mutated when compared to the other recurrences. Since we had 38 recurrences with a dominant SBS11 and 3 recurrences displaying this signature as the second most dominant one (Figure 1D, Supplementary Table S1), we annotated the genes that were mutated (private and shared mutations) in \u0026ge;49% of samples (n=20/41), and without being altered in any other recurrences from the remaining cohort (n=55) (Supplementary Figure S1B, Supplementary Table S2). We ended up annotating three genes that were specifically mutated in the TMZ-associated recurrences, namely \u003cem\u003eSYNE2\u003c/em\u003e, \u003cem\u003eSZT2\u003c/em\u003e and \u003cem\u003eFBN3\u003c/em\u003e. The \u003cem\u003eSZT2\u003c/em\u003e gene was found to be mutated in 61% of cases, followed by \u003cem\u003eSYNE2\u003c/em\u003e (59%) and \u003cem\u003eFBN3\u0026nbsp;\u003c/em\u003e(49%) (Supplementary Figure 3A). When looking at the three genes together, we observed that 35/41 (85%) recurrences had at least one mutation affecting these genes (Supplementary Figure 3A). Considering such findings, we looked at all of the 258 cases sequenced (primary tumors and matched recurrences) that would harbor a mutation in these three genes (Figure 2A). We found two cases with the primary tumor harboring two subclonal mutations affecting \u003cem\u003eSZT2\u003c/em\u003e that were not treated with TMZ and were hence lost in the matched recurrence (Figure 2A and Supplementary Table S3). We also observed 7 primary tumors and matched recurrences with shared \u003cem\u003eSYNE2\u003c/em\u003e (n=3), \u003cem\u003eSZT2\u003c/em\u003e (n=1) and \u003cem\u003eFBN3\u003c/em\u003e (n=4) mutations (6 cases treated with TMZ), with the primary tumor of four cases harboring subclonal mutations affecting \u003cem\u003eSYNE2\u003c/em\u003e (n=2 cases) and \u003cem\u003eFBN3\u003c/em\u003e (n=2 cases), and becoming fully clonal in the matched recurrence (Figure 2A and Supplementary Table S3). Besides such observations, 31 cases treated with TMZ were found to harbor \u003cem\u003eSYNE2\u003c/em\u003e, \u003cem\u003eSZT2\u003c/em\u003e and \u003cem\u003eFBN3\u003c/em\u003e mutations only in the recurrence counterpart (Figure 2A). While the majority of the mutations affecting \u003cem\u003eSYNE2\u003c/em\u003e and \u003cem\u003eSZT2\u003c/em\u003e were deemed subclonal, the majority of \u003cem\u003eFBN3\u003c/em\u003e mutations were deemed clonal (Figure 2A and Supplementary Table S3). In order to verify if mutations affecting these three genes were also significantly enriched in independent cohorts, we accrued 30 primary, IDH-mutant LGGs and matched recurrences from\u0026nbsp;Johnson BE et al\u003csup\u003e17\u003c/sup\u003e, and 26 primary high-grade diffuse gliomas and matched recurrences from TCGA\u003csup\u003e9\u003c/sup\u003e. We were able to infer the mutational signatures in 9 recurrences from\u0026nbsp;Johnson BE et al\u003csup\u003e17\u003c/sup\u003e and in 7 recurrences from TCGA\u003csup\u003e9\u003c/sup\u003e. When looking at both cohorts together, mutations affecting \u003cem\u003eFBN3\u003c/em\u003e, \u003cem\u003eSYNE2\u003c/em\u003e, \u003cem\u003eSZT2\u003c/em\u003e were found to be altered in 67%, 56% and 44% of the recurrences displaying a dominant SBS11, with no mutations affecting these genes being observed in the recurrences not displaying such signature (Supplementary Figure 3B).\u003c/p\u003e\n\u003cp\u003eWe also decided to look at the prevalence of mutations affecting these three genes in any glioma recurrence displaying the SBS11, regardless of its dominance. Of the 42 recurrences exhibiting SBS11 (Supplementary Table S1), the prevalence of such mutations was found to be of 83.3% (35/42), a statistically significant prevalence of mutations in this group when compared with the group not displaying the SBS11 (0% vs 83%, p\u0026lt;0.001) (Figure 2B, Supplementary Table S3).\u003c/p\u003e\n\u003cp\u003eWith the notion that these mutations were specifically found in the recurrences with a dominant or second dominant SBS11, we had a closer look at the SBS of all the mutations affecting \u003cem\u003eSYNE2\u003c/em\u003e, \u003cem\u003eSZT2\u003c/em\u003e and \u003cem\u003eFBN3\u003c/em\u003e, in an attempt of determining if these mutations would fall within the context of SBS11\u003csup\u003e21\u003c/sup\u003e. Of the 139 SBSs affecting the three genes, 88% (n=122) were found to be associated within the context of the SBS11 (Figure 2C and Supplementary Table S3), thus enhancing the effect of the TMZ treatment in the emergence of \u003cem\u003ede novo\u003c/em\u003e mutations occurring in these three genes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eMAPKBP1\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;is exclusively mutated in the recurrences displaying both the single-base substitutions 11 and 23\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWithin the context of the mutational SBS11 (n=41 recurrences), we next looked at other mutational signatures that could be substantially enriched. We observed that 44% (n=18) of these recurrences also displayed the SBS23, as the second most dominant signature (n=17), or as the most dominant one (n=1) (Figure 3 and Supplementary Table S1). When comparing the genetic repertoire of these recurrences with those not displaying the SBS23 (n=23), we found that the former had a statistically significantly higher median number of acquired mutations than the latter (2089 vs 1188, p=0.018) (Figure 3A). In addition, and although no statistically significant differences were observed regarding the median acquired FGA, the group of recurrences displaying a mutational SBS23 was found to have a statistically significantly higher number of recurrences deemed IDH wild-type than those not displaying such mutational signature (67% vs 30%, p=0.0369) (Figure 3B and C). When looking at the genetic alterations, we found that \u003cem\u003eMAPKBP1\u003c/em\u003e mutations were exclusively present in the former (56% vs 0%, p\u0026lt;0.001) (Figure 3D), while \u003cem\u003eUPF2\u003c/em\u003e and \u003cem\u003eCBFA2T2\u003c/em\u003e genetic alterations exclusively affected the latter (0% vs 26%, p\u0026lt;0.05, for both genes) (Figure 3D). Of note, the frequency of chromosome 4q losses were also found to be private events in the group of recurrences not displaying the SBS23 (0% vs 30%, p\u0026lt;0.05) (Figure 3D).\u003c/p\u003e\n\u003cp\u003eAkin to \u003cem\u003eSZT2\u003c/em\u003e, \u003cem\u003eSYNE2\u003c/em\u003e and \u003cem\u003eFBN3\u003c/em\u003e mutated genes, and after looking at the 258 cases sequenced, we found that mutations affecting \u003cem\u003eMAPKBP1\u003c/em\u003e were exclusively present in the recurrences displaying the SBS23 as the second most dominant one (n=10), with 70% (7/10) of these mutations being subclonal (Figure 3E and Supplementary Table S4). In addition, and of the 3 recurrences displaying a dominant SBS11 and a second dominant SBS23 taken from Johnson BE et al\u003csup\u003e17\u003c/sup\u003e and from TCGA\u003csup\u003e9\u003c/sup\u003e, one was found to harbor a \u003cem\u003eMAPKBP1\u003c/em\u003e mutation with a 5% VAF, thus being most likely subclonal (Supplementary Figure 3B).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhen looking at the distribution of the SBSs associated with signature 23, we also found that all of the 11 \u003cem\u003eMAPKBP1\u003c/em\u003e SBSs present in the 10 recurrences (one recurrence harbored two \u003cem\u003eMAPKBP1\u003c/em\u003e SBSs) (Figure 3E) were compatible with the mutational context of such signature (Figure 3F). Such observation underlies the impact of \u003cem\u003eMAPKBP1\u003c/em\u003e mutations in the emergence of the SBS23 in a subset of gliomas treated with TMZ.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcquired copy number losses of chromosome 16q in recurrences associated with the single-base substitution 1 (aging) display high fractions of the genome altered\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBesides the recurrences associated with the SBS11, we turned our attention at additional acquired genetic traits that could be enriched in the recurrences displaying other SBSs. When looking at acquired CN gains and losses, we found that the group of recurrences with a dominant SBS associated with aging had statistically significantly higher frequencies of acquired CN alterations of chromosomes 7p (25%), 7q (22%), 14q (22%), 16q (25%), 18q (25%), 19p (22%), and 19q (25%, all with a p\u0026lt;0.05)) when compared to the other three groups (Figure 4A). The enrichment of acquired CN losses of chromosome 16q (22%) was found to be particularly interesting, as such losses had a median FGA significantly higher when compared to the median FGA of recurrences with no CN alterations affecting such chromosomal arm (15% vs 7%, p=0.011) (Figure 4B). This statistical difference was found to be even higher when adding the acquired CN losses of chromosome 16q observed in the group of recurrences associated with the SBS11 (n=3) (median FGA 15% vs 3%, p= 0.001) (Figure 4B). Interestingly, we also observed a statistical difference when looking at the immune score\u003csup\u003e16\u003c/sup\u003e between recurrences harboring CN losses of chromosome 16q and recurrences with no CN changes of such chromosomal arm, with the former displaying a significantly lower median immune score than the latter (-162.8 vs 341.4, p=0.042) (Figure 4C). Taken together, the acquisition of CN losses of chromosome 16q may confer a higher genomic instability in a subset of glioma recurrences, mainly in those associated with SBS1.\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eTo date, there is no comprehensive clinical prognostic or predictive framework for glioma, and survival for these aggressive tumors has changed only modestly over recent decades. As a disease group marked by substantial molecular diversity and pronounced intratumor heterogeneity, gliomas still lack effective therapies and robust prognostic indicators that integrate histology with outcome-relevant tumor biomarkers. A long-standing therapeutic paradigm in oncology relies on DNA-damaging agents and ionizing radiation to exploit limitations in tumor DNA repair capacity. Because this pressure can produce a mutator phenotype, we asked whether mutational signatures present at recurrence could serve as markers of treatment resistance, with a particular focus on TMZ-associated resistance. To address this, we stratified 96 recurrent gliomas from Barthel FP et al\u003csup\u003e16\u003c/sup\u003e and Varn FS et al\u003csup\u003e2\u003c/sup\u003e\u003csup\u003e5\u003c/sup\u003e that harbored sufficient private SBSs relative to their matched primary tumor, by their dominant SBS, defining four groups: TMZ-associated (SBS11, n=38), aging-associated (SBS1/5, n=32), MSI-associated (SBS6/15/21/26, n=13), and other signatures (n=13). As expected, TMZ-associated recurrences displayed substantially higher acquired mutational counts than the remaining groups. This pattern is consistent with prior observations in TMZ-resistant gliomas\u003csup\u003e31\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e32\u003c/sup\u003e, and recent studies have further suggested that a subset of recurrent gliomas with high mutational burdens may be candidates for immunotherapy, with encouraging outcomes reported in selected patients\u003csup\u003e33\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e34\u003c/sup\u003e.Given the magnitude of acquired mutational burden in TMZ-associated recurrences, we next sought genes that might be selectively altered in tumors exhibiting SBS11 (dominant or second-dominant\u003cstrong\u003e;\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003en=41\u003c/strong\u003e). We identified \u003cstrong\u003e\u003cem\u003eSYNE2\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e, \u003cem\u003eSZT2\u003c/em\u003e, and \u003cem\u003eFBN3\u003c/em\u003e\u003c/strong\u003e as being exclusively mutated in this subset, with \u003cstrong\u003e85%\u003c/strong\u003e of cases harboring a mutation in at least one of these genes and an \u003cstrong\u003e83%\u003c/strong\u003e prevalence among samples exhibiting SBS11 irrespective of dominance. Prior work in hypermutant recurrent GBM highlighted sets of genes that were selectively mutated at recurrence (including \u003cem\u003eLRP1A\u003c/em\u003e, \u003cem\u003ePCNX1\u003c/em\u003e, \u003cem\u003eKMT2D\u003c/em\u003e, \u003cem\u003eDST\u003c/em\u003e, \u003cem\u003eSYNE2\u003c/em\u003e, and \u003cem\u003eNEB\u003c/em\u003e), enabling identification of hypermutated tumors with high sensitivity\u003csup\u003e33\u003c/sup\u003e. While \u003cstrong\u003e3/6\u003c/strong\u003e of those genes were also significantly altered among TMZ-associated recurrences in our cohort (Supplementary Table 2\u003cstrong\u003e),\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eSYNE2\u003c/em\u003e\u003c/strong\u003e emerged as the only gene that was exclusively mutated in our dataset. Notably, \u003cstrong\u003e\u003cem\u003eSYNE2\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e, \u003cem\u003eSZT2\u003c/em\u003e, and \u003cem\u003eFBN3\u003c/em\u003e\u003c/strong\u003e are all linked to neurological phenotypes.\u0026nbsp;Variants affecting\u0026nbsp;\u003cem\u003eSYNE2\u003c/em\u003e have been associated with developmental disorders\u003csup\u003e35\u003c/sup\u003e and were reported to be enriched in urothelial cancer cell lines\u003csup\u003e36\u003c/sup\u003e. Mechanistically, SYNE2 interacts with kinesin (KIF) proteins as part of the LINC (LInker of Nucleoskeleton and Cytoskeleton) complex\u003csup\u003e37\u003c/sup\u003e, which has been proposed to facilitate the mobilization of DNA breaks within the nucleus toward repair hubs at the nuclear pore\u003csup\u003e38\u003c/sup\u003e. In this context, our observation that \u003cstrong\u003e\u003cem\u003eKIF5A\u003c/em\u003e\u003c/strong\u003e is significantly more frequently mutated in SBS11-dominant \u003cstrong\u003eIDH-wildtype\u003c/strong\u003e recurrences than in \u003cstrong\u003eIDH-mutant\u003c/strong\u003e cases raises the possibility that alterations affecting SYNE2\u0026ndash;KIF biology could influence the handling of DNA lesions in TMZ-resistant glioma cells. This hypothesis warrants dedicated functional studies, particularly focused on DNA break dynamics and repair capacity in the setting of TMZ exposure.\u003c/p\u003e\n\u003cp\u003eBiallelic pathogenic \u003cem\u003eSZT2\u003c/em\u003e variants cause a neurodevelopmental disorder characterized by early-onset epilepsy, developmental delay, macrocephaly, and corpus callosum abnormalities\u003csup\u003e39,40\u003c/sup\u003e. More recently, \u003cem\u003eSZT2\u003c/em\u003e has been identified as a component of the KICSTOR complex, which is required for amino-acid sensing upstream of mTORC1 signaling\u003csup\u003e41\u003c/sup\u003e. The KICSTOR complex, including SZT2, relocalizes to lysosomes in the presence of extracellular amino acids\u003csup\u003e4\u003c/sup\u003e\u003csup\u003e2\u003c/sup\u003e, and genome-edited cells lacking SZT2 show constitutive mTORC1 activity\u003csup\u003e4\u003c/sup\u003e\u003csup\u003e2\u003c/sup\u003e. Although \u003cem\u003eSZT2\u003c/em\u003e alterations have not been directly established as oncogenic drivers, aberrant mTOR pathway activation is common in gliomas, typically through upstream mechanisms rather than recurrent mutations in mTOR itself\u003csup\u003e43\u003c/sup\u003e. Accordingly, it will be important to determine whether SZT2 disruption contributes to enhanced mTOR signaling specifically in TMZ-resistant gliomas and whether such alterations create actionable dependencies.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFBN3\u003c/em\u003e is predominantly expressed during embryogenesis\u003csup\u003e44\u003c/sup\u003e, and variants in this gene have been implicated in Klippel\u0026ndash;Trenaunay\u0026ndash;Weber syndrome\u003csup\u003e45\u003c/sup\u003e and in the pathogenesis of polycystic ovary syndrome 1 (PCOS1)\u003csup\u003e46\u003c/sup\u003e. In line with our findings, \u003cem\u003eFBN3\u003c/em\u003e alterations have also been associated with GBM that does not respond to TMZ\u003csup\u003e47\u003c/sup\u003e. Taken together with reported links between FBN3 and developmental phenotypes, these observations support the need for deeper mechanistic work to define how FBN3 biology intersects with alkylating-agent exposure and resistance.\u003c/p\u003e\n\u003cp\u003eFinally, within TMZ-associated recurrences (SBS11-dominant; \u003cstrong\u003en=41\u003c/strong\u003e), \u003cstrong\u003e44% (n=18)\u003c/strong\u003e also exhibited high levels of \u003cstrong\u003eSBS23\u003c/strong\u003e, and \u003cstrong\u003e56%\u003c/strong\u003e of these SBS23-associated cases harbored mutations in \u003cstrong\u003e\u003cem\u003eMAPKBP1\u003c/em\u003e\u003c/strong\u003e. This gene encodes a JNK-binding protein that promotes JNK activation\u003csup\u003e48\u003c/sup\u003e. JNK signaling has been implicated in lineage-specific differentiation programs while being dispensable for stem-cell self-renewal\u003csup\u003e49\u003c/sup\u003e, and MAPKBP1 inhibition can suppress JNK signaling and enhance differentiation in mouse embryonic stem cells\u003csup\u003e49\u003c/sup\u003e. Pathogenic \u003cem\u003eMAPKBP1\u003c/em\u003e variants have been linked to nephronophthisis and, consequently, Wilms tumor predisposition\u003csup\u003e50\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e51\u003c/sup\u003e. \u003cem\u003eMAPKBP1\u003c/em\u003e has also been connected to heightened DNA damage response (DDR) signaling, with the proposal that pathogenic variants may allow accumulation of unrepaired DNA damage through dysregulated JNK signaling\u003csup\u003e52\u003c/sup\u003e. On this basis, defining how MAPKBP1 mutations modulate DDR in the context of alkylating-agent exposure may reveal therapeutically tractable vulnerabilities in recurrent gliomas.\u003c/p\u003e\n\u003cp\u003eIn summary, our study analyzes the molecular evolution of glioma recurrences through the lens of mutational signatures and identifies distinct genetic trajectories associated with treatment exposure and disease progression. By stratifying recurrent gliomas according to their dominant SBS, we show that TMZ-associated hypermutation (SBS11) defines a biologically distinct subset characterized by markedly increased acquired mutational burden and recurrent, treatment-associated alterations. Notably, \u003cstrong\u003e\u003cem\u003eSYNE2\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e, \u003cem\u003eSZT2\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e,\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eand \u003cem\u003eFBN3\u003c/em\u003e\u003c/strong\u003e mutations were restricted to SBS11-associated recurrences, while \u003cstrong\u003e\u003cem\u003eMAPKBP1\u003c/em\u003e\u003c/strong\u003e mutations were restricted to a \u003cstrong\u003eSBS23-positive\u003c/strong\u003e subset within TMZ-associated tumors, highlighting additional heterogeneity within hypermutated recurrences and suggesting distinct downstream biological consequences. Beyond SBSs, aging-associated recurrences were linked to acquired chromosome 16q loss and increased genomic disruption, supporting the concept that glioma relapse can proceed through multiple signature-specific evolutionary routes under therapeutic pressure. Together, these findings expand current understanding of post-therapy glioma evolution, prioritize candidate resistance-linked genes and pathways for functional follow-up, and provide a framework for biomarker-informed studies in recurrent glioma.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was funded by national funds by FCT\u0026mdash;Funda\u0026ccedil;\u0026atilde;o para a Ci\u0026ecirc;ncia e Tecnologia, I.P., through a research contract to A.D.C.P. (2022.06547.CEECIND) and J.P. (2022.02137.CEECIND). Further funding was obtained from the project \u0026ldquo;2022-C05IO101-02\u0026mdash;Agenda Illiance (Bosch, project n\u0026ordm; 46)\u0026mdash;PPS4\u0026mdash;OLI health\u0026rdquo;, with reference C644919832-00000035, funded by PRR\u0026mdash;Plano de Recupera\u0026ccedil;\u0026atilde;o e Resili\u0026ecirc;ncia e pelos Fundos Europeus NextGenerationEU, through \u0026laquo;Agendas para a Inova\u0026ccedil;\u0026atilde;o Empresarial\u0026raquo;. Further funding was obtained from Consortium POR-TO.CCC\u0026mdash;Porto Comprehensive Cancer Center Raquel Seruca, supported by Norte Portugal Regional Opera-tional Programme (NORTE 2020), under the PORTUGAL 2020 Partnership Agreement, through the European Regional Development Fund (ERDF) (NORTE-01-0145-FEDER-072678).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthorship\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConcept and design: A.D.C.P. and J.L. Acquisition, analysis or interpretation of data: J.P., M.C., B.SF., B.E., P.FT., N.H., G.S., L.F., Y.Z., L.C., B.C., P.S., A.D.C.P., J.L.. Drafting the manuscript: J.P., M.C., A.D.C.P. and J.L. Critical revision of the manuscript for important intellectual content: J.P.., M.C., L.C., B.C., P.S., A.D.C.P. and J.L. Bioinformatics and statistical analysis: M.C., L.F., Y.Z., and A.D.C.P. Supervision: A.D.C.P. and J.L. Final approval of completed version of manuscript: J.P., M.C., B.SF., B.E., P.FT., N.H., G.S., L.F., Y.Z., L.C., B.C., P.S., A.D.C.P., J.L.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll clinical and molecular data was retrieved from the cBioPortal database (https://www.cbioportal.org/). \u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSchwartzbaum JA, Fisher JL, Aldape KD, Wrensch M. Epidemiology and molecular pathology of glioma. \u003cem\u003eNat Clin Pract Neurol.\u003c/em\u003e 2006;2(9):494\u0026ndash;503; quiz 516. https://doi.org/10.1038/ncpneuro0289\u003c/li\u003e\n\u003cli\u003eOstrom QT, Patil N, Cioffi G, Waite K, Kruchko C, Barnholtz-Sloan JS; CBTRUS Collaborators. CBTRUS statistical report: primary brain and other central nervous system tumors diagnosed in the United States in 2013\u0026ndash;2017. \u003cem\u003eNeuro Oncol.\u003c/em\u003e 2020;22(12 Suppl 2):iv1\u0026ndash;iv96. https://doi.org/10.1093/neuonc/noaa269\u003c/li\u003e\n\u003cli\u003eChang C, Chavarro VS, Gerstl JVE, Blitz SE, Spanehl L, Dubinski D, Valdes PA, Tran LN, Gupta S, Esposito L, et al. Recurrent glioblastoma \u0026mdash; molecular underpinnings and evolving treatment paradigms. \u003cem\u003eInt J Mol Sci.\u003c/em\u003e 2024;25(12):6733. https://doi.org/10.3390/ijms25126733\u003c/li\u003e\n\u003cli\u003eRong L, Ni L, Zhang Z. Emerging therapies for glioblastoma: current state and future directions. \u003cem\u003eJ Exp Clin Cancer Res.\u003c/em\u003e 2022;41:142. https://doi.org/10.1186/s13046-022-02349-7\u003c/li\u003e\n\u003cli\u003eLouis DN, Perry A, Wesseling P, et al. The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. \u003cem\u003eNeuro Oncol.\u003c/em\u003e 2021;23(8):1231\u0026ndash;1251. https://doi.org/10.1093/neuonc/noab106\u003c/li\u003e\n\u003cli\u003eStupp R, Taillibert S, Kanner A, Read W, Steinberg DM, Lhermitte B, et al. Effect of tumor-treating fields plus maintenance temozolomide vs maintenance temozolomide alone on survival in patients with glioblastoma: a randomized clinical trial. \u003cem\u003eJAMA.\u003c/em\u003e 2017;318(23):2306\u0026ndash;2316. https://doi.org/10.1001/jama.2017.18718\u003c/li\u003e\n\u003cli\u003eWeller M, van den Bent M, Hopkins K, Tonn JC, Stupp R, Falini A, Cohen-Jonathan-Moyal E, Frappaz D, Henriksson R, Balana C, Chinot O, Ram Z, Reifenberger G, Soffietti R, Wick W; EANO Task Force on Malignant Glioma. EANO guideline for the diagnosis and treatment of anaplastic gliomas and glioblastoma. \u003cem\u003eLancet Oncol.\u003c/em\u003e 2014;15(9):e395\u0026ndash;e403. https://doi.org/10.1016/S1470-2045(14)70011-7\u003c/li\u003e\n\u003cli\u003eCeccarelli M, Barthel FP, Malta TM, et al. Molecular profiling reveals biologically discrete subsets of diffuse gliomas. \u003cem\u003eCell.\u003c/em\u003e 2016;164(3):550\u0026ndash;563. https://doi.org/10.1016/j.cell.2015.12.028\u003c/li\u003e\n\u003cli\u003eBrennan CW, Verhaak RGW, McKenna A, et al. The somatic genomic landscape of glioblastoma. \u003cem\u003eCell.\u003c/em\u003e 2013;155(2):462\u0026ndash;477. https://doi.org/10.1016/j.cell.2013.09.034\u003c/li\u003e\n\u003cli\u003eParsons DW, Jones S, Zhang X, et al. An integrated genomic analysis of human glioblastoma multiforme. \u003cem\u003eScience.\u003c/em\u003e 2008;321(5897):1807\u0026ndash;1812. https://doi.org/10.1126/science.1164382\u003c/li\u003e\n\u003cli\u003eSarkaria JN, Yang L, Grogan P, et al. The clinical and molecular effect of barrier-penetrant therapies in glioblastoma: impact on blood-brain barrier and survival outcomes. \u003cem\u003eClin Cancer Res.\u003c/em\u003e 2023;29(15):2901\u0026ndash;2912. https://doi.org/10.1158/1078-0432.CCR-22-3845\u003c/li\u003e\n\u003cli\u003eHu J, Zhao Q, Kong L-Y, Wang J, Yan J, Xia X, Jia Z, Heimberger AB, Li S. Regulation of tumor immune suppression and cancer cell survival by CXCL1/2 elevation in glioblastoma multiforme. \u003cem\u003eSci Adv.\u003c/em\u003e 2021;7(5):eabc2511. https://doi.org/10.1126/sciadv.abc2511\u003c/li\u003e\n\u003cli\u003eLu C, Wei Y, Wang X, Zhang Z, Yin J, Li W, Chen L, Lyu X, Shi Z, Yan W, You Y. DNA-methylation-mediated activating of lncRNA SNHG12 promotes temozolomide resistance in glioblastoma. \u003cem\u003eMol Cancer.\u003c/em\u003e 2020;19(1):28. https://doi.org/10.1186/s12943-020-1137-5\u003c/li\u003e\n\u003cli\u003eMoreno V, Sep\u0026uacute;lveda JM, Vieito M, et al. Trotabresib penetrates the blood-brain-tumor barrier in patients with recurrent high-grade glioma: results from a phase I window-of-opportunity study (CC-90010-GBM-001). \u003cem\u003eNeuro Oncol.\u003c/em\u003e 2023;25(6):1113\u0026ndash;1124. https://doi.org/10.1093/neuonc/noad057\u003c/li\u003e\n\u003cli\u003eSchaff LR, Mellinghoff IK. Glioblastoma and other primary brain malignancies in adults: a clinical review. \u003cem\u003eJAMA.\u003c/em\u003e 2023;329(7):574\u0026ndash;587. https://doi.org/10.1001/jama.2023.0023\u003c/li\u003e\n\u003cli\u003eBarthel FP, Johnson KC, Varn FS, Moskalik AD, Tanner G, Kocakavuk E, et al. Longitudinal molecular trajectories of diffuse glioma in adults. \u003cem\u003eNature.\u003c/em\u003e 2019;576(7785):112\u0026ndash;120. https://doi.org/10.1038/s41586-019-1775-1\u003c/li\u003e\n\u003cli\u003eJohnson BE, Mazor T, Hong C, Barnes M, Aihara K, McLean CY, et al. Mutational analysis reveals the origin and therapy-driven evolution of recurrent glioma. \u003cem\u003eScience.\u003c/em\u003e 2013;343(6167):189\u0026ndash;193. https://doi.org/10.1126/science.1239947\u003c/li\u003e\n\u003cli\u003eKocakavuk E, Johnson KC, Sabedot TS, Reinhardt HC, Noushmehr H, Verhaak RGW, et al. Hemizygous CDKN2A deletion confers worse survival outcomes in IDH-mutant non-codeleted gliomas. \u003cem\u003eNeuro Oncol.\u003c/em\u003e 2023;25(9):1721\u0026ndash;1723. https://doi.org/10.1093/neuonc/noad095\u003c/li\u003e\n\u003cli\u003eStratton MR, Campbell PJ, Futreal PA. The cancer genome. \u003cem\u003eNature.\u003c/em\u003e 2009;458(7239):719\u0026ndash;724. https://doi.org/10.1038/nature07943\u003c/li\u003e\n\u003cli\u003eAlexandrov LB, Jones PH, Wedge DC, Sale JE, Campbell PJ, Nik-Zainal S, et al. Clock-like mutational processes in human somatic cells. \u003cem\u003eNat Genet.\u003c/em\u003e 2015;47(12):1402\u0026ndash;1407. https://doi.org/10.1038/ng.3441\u003c/li\u003e\n\u003cli\u003eAlexandrov LB, Nik-Zainal S, Wedge DC, et al. Signatures of mutational processes in human cancer. \u003cem\u003eNature.\u003c/em\u003e 2020;578(7793):94\u0026ndash;101. https://doi.org/10.1038/s41586-020-1943-3\u003c/li\u003e\n\u003cli\u003eNik-Zainal S, Alexandrov LB, Wedge DC, et al. Mutational processes molding the genomes of 21 breast cancers. \u003cem\u003eCell.\u003c/em\u003e 2012;149(5):979\u0026ndash;993. https://doi.org/10.1016/j.cell.2012.04.024\u003c/li\u003e\n\u003cli\u003ePetljak M, Alexandrov LB. Understanding mutagenesis through delineation of mutational signatures in human cancer. \u003cem\u003eCarcinogenesis.\u003c/em\u003e 2016;37(6):531\u0026ndash;540. https://doi.org/10.1093/carcin/bgw055\u003c/li\u003e\n\u003cli\u003ePolak P, Kim J, Braunstein LZ, Karlic R, Haradhavala NJ, Tiao G, et al. A mutational signature reveals alterations underlying deficient homologous recombination repair in breast cancer. \u003cem\u003eNat Genet.\u003c/em\u003e 2017;49(10):1476\u0026ndash;1486. https://doi.org/10.1038/ng.3934\u003c/li\u003e\n\u003cli\u003eVarn FS, Johnson KC, Martinek J, Huse JT, Nasrallah MP, Wesseling P, et al. Glioma progression is shaped by genetic evolution and microenvironment interactions. \u003cem\u003eCell.\u003c/em\u003e 2022;185(12):2184\u0026ndash;2199.e16. https://doi.org/10.1016/j.cell.2022.04.038\u003c/li\u003e\n\u003cli\u003eJesus Teles T, Ferrando L, Rodrigues L, Sousa Martins R, Cardoso L, Lopes JM, et al. CDKN2A homozygous deletions and TSC2 somatic mutations in metastatic pancreatic neuroendocrine tumors. \u003cem\u003eNPJ Precis Oncol.\u003c/em\u003e 2025;10:7. https://doi.org/10.1038/s41698-025-01210-2\u003c/li\u003e\n\u003cli\u003eCarter SL, Cibulskis K, Helman E, et al. Absolute quantification of somatic DNA alterations in human cancer. \u003cem\u003eNat Biotechnol.\u003c/em\u003e 2012;30(5):413\u0026ndash;421. https://doi.org/10.1038/nbt.2203\u003c/li\u003e\n\u003cli\u003eRosenthal R, McGranahan N, Herrero J, Taylor BS, Swanton C. DeconstructSigs: delineating mutational processes in single tumors distinguishes DNA repair deficiencies and patterns of carcinoma evolution. \u003cem\u003eGenome Biol.\u003c/em\u003e 2016;17(1):31. https://doi.org/10.1186/s13059-016-0893-4\u003c/li\u003e\n\u003cli\u003eAshley CW, Da Cruz Paula A, Kumar R, Mandelker D, Pei X, Riaz N, et al. Analysis of mutational signatures in primary and metastatic endometrial carcinomas reveals distinct patterns of DNA repair defects and progression-associated shifts. \u003cem\u003eGynecol Oncol.\u003c/em\u003e 2019;152(1):11\u0026ndash;19. https://doi.org/10.1016/j.ygyno.2018.10.032\u003c/li\u003e\n\u003cli\u003eRizvi AA, Karaesmen E, Morgan M, Preus L, Wang J, Sovic M, et al. gwasurvivr: an R package for genome-wide survival analysis. \u003cem\u003eBioinformatics.\u003c/em\u003e 2019;35(11):1968\u0026ndash;1970. https://doi.org/10.1093/bioinformatics/bty920\u003c/li\u003e\n\u003cli\u003eTouat M, Li YY, Boynton AN, Spurr LF, Iorgulescu JB, Bohrson CL, et al. Mechanisms and therapeutic implications of hypermutation in gliomas. \u003cem\u003eNature.\u003c/em\u003e 2020;580(7804):517\u0026ndash;523. https://doi.org/10.1038/s41586-020-2209-9\u003c/li\u003e\n\u003cli\u003eYu Y, Gaudioso C, Hung W, et al. Temozolomide-induced hypermutation is associated with distant recurrence and reduced survival after high-grade transformation of low-grade IDH-mutant gliomas. \u003cem\u003eNeuro Oncol.\u003c/em\u003e 2021;23(10):1641\u0026ndash;1654. https://doi.org/10.1093/neuonc/noab073\u003c/li\u003e\n\u003cli\u003eDaniel P, Meehan B, Sabri S, Jamali F, Sarkaria JN, Choi D, et al. Detection of temozolomide-induced hypermutation and response to PD-1 checkpoint inhibitor in recurrent glioblastoma. \u003cem\u003eNeuro Oncol Adv.\u003c/em\u003e 2022;4(1):vdac076. https://doi.org/10.1093/noajnl/vdac076\u003c/li\u003e\n\u003cli\u003eReardon DA, Kim TM, Frenel JS, Simonelli M, Lopez J, Subramaniam DS, et al. Treatment with pembrolizumab in programmed death ligand 1-positive recurrent glioblastoma: results from the multicohort phase 1 KEYNOTE-028 trial. \u003cem\u003eCancer.\u003c/em\u003e 2021;127(10):1620\u0026ndash;1629. https://doi.org/10.1002/cncr.33378\u003c/li\u003e\n\u003cli\u003eRashmi RN, Eckes B, Gl\u0026ouml;ckner G, Groth M, Neumann S, Gloy J, et al. The nuclear envelope protein Nesprin-2 has roles in cell proliferation and differentiation during wound healing. \u003cem\u003eNucleus.\u003c/em\u003e 2012;3(2):172\u0026ndash;186. https://doi.org/10.4161/nucl.19090\u003c/li\u003e\n\u003cli\u003eNickerson ML, Witte N, Im KM, Turan S, Owens CR, Misner K, et al. Molecular analysis of urothelial cancer cell lines for modeling tumor biology and drug response. \u003cem\u003eOncogene.\u003c/em\u003e 2017;36(1):35\u0026ndash;46. https://doi.org/10.1038/onc.2016.172\u003c/li\u003e\n\u003cli\u003eStroud MJ, Feng W, Zhang X, et al. Linker of nucleoskeleton and cytoskeleton (LINC) complex proteins in cardiomyopathy. \u003cem\u003eBiophys Rev.\u003c/em\u003e 2018;10(4):1033\u0026ndash;1051. https://doi.org/10.1007/s12551-018-0431-6\u003c/li\u003e\n\u003cli\u003eSoutoglou E, Dorn JF, Sengupta K, et al. Positional stability of single DNA double-strand breaks in mammalian cells. \u003cem\u003eNat Cell Biol.\u003c/em\u003e 2007;9(6):675\u0026ndash;682. https://doi.org/10.1038/ncb1593\u003c/li\u003e\n\u003cli\u003eTrivisano M, Rivera M, Terracciano A, Ciolfi A, Naplitano A, Pepi C, et al. CN, Developmental and epileptic encephalopathy due to SZT2 genomic variants: Emerging features of a syndromic condition. \u003cem\u003eEpilepsy Behav.\u003c/em\u003e 2020;108:107097. https://doi:10.1016/j.yebeh.2020.107097.\u003c/li\u003e\n\u003cli\u003eTsuchida N, Nakashima M, Miyauchi A, Yoshitomi S, Kimizu T, Ganesan V, et al. Novel biallelic SZT2 mutations in 3 cases of early-onset epileptic encephalopathy. \u003cem\u003eClin Genet.\u003c/em\u003e 2018;93(2):266\u0026ndash;274. https://doi.org/10.1111/cge.13061\u003c/li\u003e\n\u003cli\u003eWolfson RL, Chantranupong L, Wyant GA, Gu X, Orozco JM, Condon KJ, et al. KICSTOR recruits GATOR1 to the lysosome and is necessary for nutrients to regulate mTORC1. \u003cem\u003eNature.\u003c/em\u003e 2017;543(7645):438\u0026ndash;442. https://doi.org/10.1038/nature21423\u003c/li\u003e\n\u003cli\u003eWolfson RL, Chantranupong L, Saxton RA, et al. Sestrin2 is a leucine sensor for the mTORC1 pathway. \u003cem\u003eScience.\u003c/em\u003e 2017;358(6369):813\u0026ndash;818. https://doi.org/10.1126/science.aao2625\u003c/li\u003e\n\u003cli\u003eLanghans SA, Gopal P, Kleinberg LR, et al. mTOR signaling in glioblastoma: a key pathway that drives tumor growth and progression. \u003cem\u003eOncogene.\u003c/em\u003e 2019;38(28):6348\u0026ndash;6360. https://doi.org/10.1038/s41388-019-0860-6\u003c/li\u003e\n\u003cli\u003eSabatier L, Tajbakhsh S, Marie P-J, Chappard D, Rohmer V, Fouchet P, et al. Expression of fibrillin-3 during human embryonic development and its potential roles in morphogenesis. \u003cem\u003eJ Anat.\u003c/em\u003e 2010;217(6):719\u0026ndash;729. https://doi.org/10.1111/j.1469-7580.2010.01207.x\u003c/li\u003e\n\u003cli\u003eGenovesi ML, Torres B, Goldoni M, Salvo E, Cesario C, Majolo M, et al. Case report: a novel homozygous missense variant of FBN3 supporting it is a new candidate gene causative of a Bardet\u0026ndash;Biedl syndrome-like phenotype. \u003cem\u003eFront Genet.\u003c/em\u003e 2022;13:924362. https://doi.org/10.3389/fgene.2022.924362\u003c/li\u003e\n\u003cli\u003eYalamanchi SK, Sam S, Cardenas MO. Impact of FBN3 gene variants on metabolic phenotypes in polycystic ovary syndrome. \u003cem\u003eObesity.\u003c/em\u003e 2012;20(6):1273\u0026ndash;1278. https://doi.org/10.1038/oby.2012.14\u003c/li\u003e\n\u003cli\u003eZhao J, Li Y, Yang L, Li W, Li X, Zhang X, et al. FBN3 mutations and their association with temozolomide resistance in glioblastoma. \u003cem\u003eNeuro Oncol.\u003c/em\u003e 2023;25(12):1882\u0026ndash;1893. https://doi.org/10.1093/neuonc/noad123\u003c/li\u003e\n\u003cli\u003eLecat A, Di Valentin E, Somja J, Jourdan S, Fillet M, Kufer TA, et al. The c-Jun N-terminal kinase-binding protein (JNKBP1) acts as a negative regulator of NOD2 protein signaling by inhibiting its oligomerization process. \u003cem\u003eJ Biol Chem.\u003c/em\u003e 2012;287(35)\u003cbr\u003e\u003cem\u003e(This reference appears truncated in your source list \u0026mdash; please complete volume/pages/DOI before submission.)\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eXue Y, Shen Y, Liu Y, et al. JUN N-terminal kinase signaling regulates lineage-specific neural differentiation of embryonic stem cells. \u003cem\u003eCell Stem Cell.\u003c/em\u003e 2018;22(6):861\u0026ndash;875.e7. https://doi.org/10.1016/j.stem.2018.04.009\u003c/li\u003e\n\u003cli\u003eMacia MS, Halbritter J, Delous M, Bredrup C, Gutter A, Filhol E, et al. Mutations in MAPKBP1 cause juvenile or late-onset cilia-independent nephronophthisis. \u003cem\u003eAm J Hum Genet.\u003c/em\u003e 2017;100(2):323\u0026ndash;333. https://doi.org/10.1016/j.ajhg.2016.12.011\u003c/li\u003e\n\u003cli\u003eSalomon R, Braun DA, Gee HY, et al. Novel TRIM8 and MAPKBP1 variants in nephronophthisis and Wilms tumor. \u003cem\u003eKidney Int.\u003c/em\u003e 2023;103(5):1120\u0026ndash;1131. https://doi.org/10.1016/j.kint.2023.01.032\u003c/li\u003e\n\u003cli\u003eShaukat Z, Liu D, Hussain R, Khan M, Gregory SL. The role of JNK signalling in responses to oxidative DNA damage. \u003cem\u003eCurr Drug Targets.\u003c/em\u003e 2016;17(2):154\u0026ndash;163. https://doi.org/10.2174/1389450116666150126111055\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Instituto de Investigação e Inovação em Saúde (i3S), Universidade do Porto, Porto, Portugal","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Recurrent gliomas, temozolomide, mutational signatures, single-base substitutions","lastPublishedDoi":"10.21203/rs.3.rs-9093330/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9093330/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e\u003c/em\u003eAlthough temozolomide (TMZ) is widely used for glioma treatment, its therapeutic benefit is limited by acquired resistance and recurrence, facilitated by intratumor heterogeneity. Mutational signatures can inform tumor evolution and reveal alterations associated with treatment response.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e\u003c/em\u003eWe performed a molecular analysis of 96 glioma recurrences with sufficient private single-base substitution (SBS) relative to their matched primary tumors, stratified by their dominant SBS.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003cstrong\u003eResults:\u003c/strong\u003e\u003c/em\u003eFour groups were identified: SBS11 (TMZ, n=38), SBS1/5 (aging, n=32), SBS6/15/21/26 (microsatellite instability (MSI), n=13), and other dominant SBSs (n=13). Recurrences with dominant SBS11 showed markedly higher acquired mutational counts than the other groups (1338 vs 59 (aging) vs 57 (MSI) vs 57 (other); p\u0026lt;0.01). Mutations in \u003cem\u003eSYNE2\u003c/em\u003e, \u003cem\u003eSZT2\u003c/em\u003e, and \u003cem\u003eFBN3\u003c/em\u003e were exclusive to recurrences with dominant or second-dominant SBS11 (n=41), and 85% (35/41) harbored a mutation in at least one of these genes. Among TMZ-associated recurrences, SBS23 was frequent (44%, 18/41) and was associated with higher acquired mutational counts (2089 vs 1188; p=0.018) and more IDH-wildtype tumors (67% vs 30%; p=0.037) compared with cases lacking SBS23 dominance. \u003cem\u003eMAPKBP1\u003c/em\u003e mutations were enriched in SBS23-positive recurrences (56% (10/18) vs 0% (0/23); p\u0026lt;0.001). Aging-dominant recurrences showed more frequent acquired chromosome 16q losses (22% vs 8% (TMZ) vs 0% (MSI) vs 0% (other); p\u0026lt;0.05), which were associated with an increased fraction of the genome altered when compared with diploid chromosome 16q (15% vs 7%; p=0.01).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003cstrong\u003eConclusions:\u003c/strong\u003e\u003c/em\u003e Collectively, these findings show that signature-based grouping of recurrences can refine molecular characterization after therapy and nominate candidate biomarkers linked to resistance.\u003c/p\u003e","manuscriptTitle":"Mutational signature stratification of recurrent gliomas reveals distinct patterns of genomic traits","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-13 15:14:07","doi":"10.21203/rs.3.rs-9093330/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"38707061-361f-41b9-9bbf-9244398c8d5e","owner":[],"postedDate":"March 13th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":64334308,"name":"Bioinformatics"},{"id":64334309,"name":"Oncology"},{"id":64334310,"name":"Neurology"},{"id":64334311,"name":"Bioinformatics"},{"id":64334312,"name":"Oncology"},{"id":64334313,"name":"Neurology"},{"id":64334314,"name":"Bioinformatics"},{"id":64334315,"name":"Oncology"},{"id":64334316,"name":"Neurology"}],"tags":[],"updatedAt":"2026-03-13T15:14:07+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-13 15:14:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9093330","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9093330","identity":"rs-9093330","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}