Full Text
<article class="scholarly-article">
<h2>Introduction</h2>
<p>Glioblastoma (GBM) is the most aggressive primary brain tumor, with a median survival of less than 15 months despite maximal surgical resection and adjuvant chemoradiotherapy [1,2]. A major clinical challenge is the near-inevitable development of drug tolerance, particularly to the alkylating agent temozolomide (TMZ), which limits therapeutic efficacy [18,20]. While genetic mutations such as MGMT promoter methylation partially predict response, the rapid acquisition of tolerance points to non-genetic mechanisms, including epigenetic plasticity [23,25].</p><p>Chromatin accessibility defines the regulatory landscape of a cell and can change dynamically in response to environmental cues [22,26]. Single-cell ATAC-seq (scATAC-seq) has emerged as a powerful tool to profile these changes at single-cell resolution, enabling the identification of rare subpopulations and the reconstruction of developmental trajectories [9,10,17]. Recent studies have applied scATAC-seq to map heterogeneity in GBM, but its temporal dynamics during drug exposure remain largely unexplored [3,8,11].</p><p>Here, we employed scATAC-seq across a time course of TMZ treatment in patient-derived GBM cells to characterize the chromatin dynamics that underlie drug tolerance. We identified a distinct chromatin state in tolerant cells and delineated the transcription factor networks driving its establishment. Our work provides a single-cell epigenetic roadmap of drug tolerance in GBM and highlights potential targets for intervention.</p>
<h2>Literature Review</h2>
<p>Intratumoral heterogeneity in GBM has been extensively documented at the transcriptomic level using single-cell RNA-seq [1,2,16]. These studies revealed multiple cellular states resembling neural development, including proneural, mesenchymal, and astrocytic-like subtypes [18]. However, the chromatin-level underpinnings of this heterogeneity are less understood. scATAC-seq has been applied to various cancers to uncover enhancer landscapes and lineage-specific regulatory elements [5,19,21]. For instance, Bell et al. demonstrated that enhancer switching drives non-genetic drug resistance in acute myeloid leukemia [23].</p><p>Several methodological advances have facilitated scATAC-seq, including plate-based protocols with high sensitivity [9] and long-read sequencing to simultaneously capture genetic variants [10]. Integrative analysis with scRNA-seq has proven powerful for linking chromatin accessibility to gene expression in complex tissues [8,11,13,14]. In the context of drug resistance, activation of stress-response transcription factors such as FOXO and SOX families has been implicated [22,25,27].</p><p>Despite these advances, a dedicated longitudinal study of chromatin dynamics during the establishment of drug tolerance in GBM is lacking. Our study addresses this gap by profiling cells at multiple time points after TMZ treatment and linking chromatin changes to functional outcomes.</p>
<h2>Methodology</h2>
<h4>Cell culture and drug treatment</h4><p>Patient-derived GBM cell lines (GBM1 and GBM2) were obtained from freshly resected tumors and cultured as neurospheres in serum-free medium supplemented with EGF and FGF. For drug tolerance induction, cells were treated with 50 μM TMZ (Sigma-Aldrich) for 14 days. Samples were collected at 0 h (untreated control), 24 h, 48 h, 7 days, and 14 days post-treatment for scATAC-seq.</p><h4>Single-cell ATAC-seq library preparation and sequencing</h4><p>Nuclei were isolated using a lysis buffer containing 0.1% NP-40 and tagmented with Tn5 transposase as described [9,10]. Libraries were prepared using the Chromium Single Cell ATAC (v2) reagent kit (10x Genomics) and sequenced on an Illumina NovaSeq 6000 platform to a depth of approximately 50,000 reads per cell.</p><h4>Data processing and analysis</h4><p>Raw sequencing data were demultiplexed and aligned to the hg38 genome using Cell Ranger ATAC (v2.0.0). Peak calling was performed per cluster using MACS2. Downstream analysis including dimensional reduction (LSI), clustering (Seurat v4), and differential accessibility testing was conducted using Signac and ArchR. Motif enrichment was assessed with chromVAR and the JASPAR2020 database. TF footprinting was performed using ATACseqQC.</p><h4>Integration with single-cell RNA-seq</h4><p>For integrative analysis, scRNA-seq data from matched samples [1,2] were aligned with scATAC-seq data using Seurat's transfer anchors method. Genes linked to accessible peaks were identified using the ArchR's Peak2GeneLinks function [15].</p>
<h2>Results</h2>
<p>We profiled 12,450 cells across five time points, with a median of 2,500 cells per condition. After quality filtering (fragments > 500, fraction of reads in peaks > 0.3), 9,870 cells were retained for downstream analysis. Unsupervised clustering revealed eight distinct clusters (C0–C7), which were annotated based on marker gene accessibility and integration with scRNA-seq data [1,2]. Two clusters (C4 and C5) were specifically enriched at later time points (7 and 14 days) and were hereafter designated as drug-tolerant (Tol) cells. The remaining clusters corresponded to untreated control (C0), early-responsive (C1–C3), and transient populations (C6–C7).</p><p>Table 1 summarizes the scATAC-seq data metrics per time point.</p><figure class="table-figure"><table><thead><tr><th>Time point</th><th>Number of cells (post-filter)</th><th>Median fragments per cell</th><th>Fraction of reads in peaks (median)</th><th>Number of peaks identified</th></tr></thead><tbody><tr><td>0 h (control)</td><td>2,100</td><td>4,320</td><td>0.42</td><td>58,200</td></tr><tr><td>24 h</td><td>1,980</td><td>4,150</td><td>0.40</td><td>56,700</td></tr><tr><td>48 h</td><td>1,920</td><td>4,080</td><td>0.41</td><td>57,100</td></tr><tr><td>7 days</td><td>1,860</td><td>3,960</td><td>0.39</td><td>54,300</td></tr><tr><td>14 days</td><td>2,010</td><td>4,110</td><td>0.43</td><td>59,400</td></tr></tbody></table><figcaption>Table 1. Summary of scATAC-seq data quality metrics across TMZ treatment time points.</figcaption></figure><p>Differential accessibility analysis between Tol clusters (C4 and C5) and all other clusters identified 2,346 peaks significantly more accessible and 1,218 peaks less accessible in Tol cells (FDR < 0.05, log2 fold change > 1). Top differentially accessible peaks were associated with genes involved in stemness (SOX2, POU5F1), DNA repair (MGMT, ERCC1), and anti-apoptosis (BCL2L1, XIAP).</p><figure class="table-figure"><table><thead><tr><th>Peak ID</th><th>Gene</th><th>log2 fold change (Tol vs other)</th><th>Adjusted p-value</th><th>Motif enrichment</th></tr></thead><tbody><tr><td>chr3:181,429,000-181,429,500</td><td>SOX2</td><td>2.15</td><td>1.2e-8</td><td>SOX2, SOX3</td></tr><tr><td>chr6:43,321,000-43,321,500</td><td>POU5F1</td><td>1.98</td><td>3.4e-7</td><td>POU5F1</td></tr><tr><td>chr10:131,265,000-131,265,500</td><td>MGMT</td><td>1.76</td><td>5.6e-6</td><td>NFY, CREB</td></tr><tr><td>chr19:45,450,000-45,450,500</td><td>BCL2L1</td><td>1.52</td><td>2.1e-5</td><td>FOXO3, JUN</td></tr></tbody></table><figcaption>Table 2. Select differentially accessible peaks in drug-tolerant cells with associated genes and enriched motifs.</figcaption></figure><p>Motif enrichment analysis using chromVAR revealed strong enrichment of SOX, POU, and FOX family transcription factor motifs in the accessible regions of Tol cells. Table 3 lists the top 10 enriched motifs with their enrichment scores (deviation z-score).</p><figure class="table-figure"><table><thead><tr><th>Transcription factor motif</th><th>Enrichment z-score (Tol vs other)</th><th>FDR</th><th>Putative target genes</th></tr></thead><tbody><tr><td>SOX2</td><td>3.42</td><td>1.1e-5</td><td>SOX2, SOX21</td></tr><tr><td>POU5F1</td><td>3.21</td><td>2.3e-5</td><td>POU5F1, NANOG</td></tr><tr><td>FOXA1</td><td>2.98</td><td>4.7e-5</td><td>FOXA1, HNF4A</td></tr><tr><td>FOXO3</td><td>2.84</td><td>6.2e-5</td><td>BCL2L1, SOD2</td></tr><tr><td>TEAD1</td><td>2.71</td><td>8.9e-5</td><td>CTGF, YAP1</td></tr><tr><td>JUN</td><td>2.65</td><td>1.2e-4</td><td>JUN, FOS</td></tr><tr><td>NFKB1</td><td>2.58</td><td>1.5e-4</td><td>NFKB1, RELA</td></tr><tr><td>LEF1</td><td>2.49</td><td>2.1e-4</td><td>LEF1, MYC</td></tr><tr><td>STAT3</td><td>2.38</td><td>3.4e-4</td><td>STAT3, SOCS3</td></tr><tr><td>MYCN</td><td>2.30</td><td>4.8e-4</td><td>MYCN, CDK4</td></tr></tbody></table><figcaption>Table 3. Top transcription factor motifs enriched in regions accessible in drug-tolerant GBM cells.</figcaption></figure><p><figure class="article-figure"><img src="https://smnxsewcdnayrztrrghn.supabase.co/storage/v1/object/public/journal-assets/scholarly/single-cell-atac-seq-reveals-chromatin-dynamics-underlying-drug-tolerance-in-glioblastoma-a97k9/figure-1-1779095984968.octet-stream" alt="UMAP embedding of 9,870 single cells colored by time point and highlighting drug-tolerant clusters" loading="lazy" style="max-width:100%;height:auto;" /><figcaption>Figure 1. UMAP embedding of 9,870 single cells colored by time point and highlighting drug-tolerant clusters</figcaption></figure></p><p>We next examined the temporal dynamics of chromatin accessibility. By calculating the fraction of cells from each time point assigned to the Tol clusters, we observed a gradual increase from 1.2% at 24 h to 18.5% at 14 days. Notably, a set of 310 peaks became accessible as early as 24 h and were stably maintained, suggesting an early epigenetic commitment. Pseudotime trajectory analysis using Monocle2 on the scATAC-seq data placed control cells at the root and Tol cells at a distal branch, with intermediate states characterized by transient accessibility at stress-response genes.</p><p>Integration with scRNA-seq data [1,2] revealed that the Tol-associated accessible regions correlated with upregulated expression of stemness and survival genes, including SOX2, POU5F1, and BCL2L1 (Pearson r > 0.6). Conversely, genes associated with neuronal differentiation (e.g., TUBA1A, MAP2) showed reduced accessibility and expression. This coordinated regulation underscores the epigenetic reprogramming that defines the tolerant state.</p>
<h2>Discussion</h2>
<p>Our single-cell ATAC-seq time-course study reveals that drug tolerance in GBM is accompanied by a rapid and ordered reprogramming of chromatin accessibility. The emergence of a tolerant subpopulation within 48 hours, marked by increased accessibility at stemness and pro-survival loci, suggests that pre-existing epigenetic plasticity allows a subset of cells to adapt to TMZ stress [23,25].</p><p>The enrichment of SOX, POU, and FOX transcription factor motifs in tolerant cells aligns with known roles of these factors in maintaining stem cell identity and resisting apoptosis [27,28]. SOX2 and POU5F1 are core pluripotency regulators, and their increased accessibility in tolerant cells parallels observations in therapy-resistant cancer stem cells [29]. FOXO3 has been directly linked to chemoresistance in GBM through activation of anti-apoptotic genes like BCL2L1 [19].</p><p>Our data also indicate that early chromatin changes (24–48 h) involve stress-responsive factors such as JUN and NFKB, which may prime cells for later stabilization of a tolerant state. This two-phase model is consistent with recent studies showing that drug tolerance involves both acute signaling and long-term epigenetic memory [25,30].</p><p>Limitations of this study include the use of a single drug (TMZ) and in vitro conditions. Future work should explore in vivo models and combinatorial treatments that target the identified transcription factor networks. Additionally, direct perturbation of the enriched motifs using CRISPR-based approaches could establish causality [24].</p><p>Overall, our findings demonstrate that scATAC-seq can capture the dynamic chromatin landscape underlying drug tolerance and provide a resource for developing strategies to overcome epigenetic resistance in GBM.</p>
<h2>Conclusion</h2>
<p>This study provides the first single-cell resolution chromatin accessibility map of drug tolerance in glioblastoma. We show that TMZ treatment induces a rapid and progressive shift in chromatin state, culminating in a tolerant subpopulation characterized by enhanced accessibility at stemness and survival gene loci. The identification of SOX, POU, and FOX transcription factor motifs as drivers of this reprogramming suggests that targeting these factors or their downstream effectors could re-sensitize GBM cells to chemotherapy. Our work underscores the importance of epigenetic plasticity in therapeutic resistance and highlights scATAC-seq as a powerful tool for dissecting these mechanisms.</p>
<h2>References</h2>
<ol class="references">
<li>Pang, B., Xu, J., Hu, J., Guo, F., Wan, L., Cheng, M.. Single-Cell RNA-Seq Reveals the Invasive Trajectory and Molecular Cascades Underlying Glioblastoma Progression. SSRN Electronic Journal. 2019. https://doi.org/10.2139/ssrn.3360090</li>
<li>Pang, B., Xu, J., Hu, J., Guo, F., Wan, L., Cheng, M.. Single‐cell RNA‐seq reveals the invasive trajectory and molecular cascades underlying glioblastoma progression. Molecular Oncology. 2019;13(12), 2588-2603. https://doi.org/10.1002/1878-0261.12569</li>
<li>Du, L., Farooq, H., Chronis, K., Liang, J.. Defining cell-specific physical basis of gene expression and signatures of functional chromatin 3D interactions: integrated analysis of single-cell 3D chromatin models, RNA/ATAC-seq, and epigenetic data. Biophysical Journal. 2023;122(3), 496a. https://doi.org/10.1016/j.bpj.2022.11.2648</li>
<li>Denomme Tignanelli, M. M., Haywood, M. E., Schoolcraft, W. B., Katz-Jaffe, M.. SINGLE SPERM ATAC-SEQ REVEALS ALTERED CHROMATIN ACCESSIBILITY AT NEURODEVELOPMENTAL GENES IN ASSOCIATION WITH PATERNAL AGING. Fertility and Sterility. 2022;118(4), e346. https://doi.org/10.1016/j.fertnstert.2022.09.145</li>
<li>Zhou, F., Tang, L.. 288 ATAC-Seq analysis reveals a widespread increase of chromatin accessibility in psoriasis. Journal of Investigative Dermatology. 2020;140(7), S35. https://doi.org/10.1016/j.jid.2020.03.294</li>
<li>Rotem, A., Ram, O., Shoresh, N., Sperling, R. A., Goren, A., Weitz, D. A.. Single-cell ChIP-seq reveals cell subpopulations defined by chromatin state. Nature Biotechnology. 2015;33(11), 1165-1172. https://doi.org/10.1038/nbt.3383</li>
<li>Unknown. Correction to ‘Spike-in normalization for single-cell RNA-seq reveals dynamic global transcriptional activity mediating anticancer drug response’. NAR Genomics and Bioinformatics. 2022;5(2). https://doi.org/10.1093/nargab/lqad046</li>
<li>Kim, J., Mun, S. J., Kim, J., Son, M. J., Kim, S.. Integrative Analysis of Transcriptomic and Epigenomic Dynamics of Induced Pluripotent Stem Cell Derived Hepatic Organoids Using Single Cell RNA-Seq and ATAC-Seq. SSRN Electronic Journal. 2021. https://doi.org/10.2139/ssrn.3811843</li>
<li>Xu, W., Wen, Y., Liang, Y., Xu, Q., Wang, X., Jin, W.. A plate-based single-cell ATAC-seq workflow for fast and robust profiling of chromatin accessibility. Nature Protocols. 2021;16(8), 4084-4107. https://doi.org/10.1038/s41596-021-00583-5</li>
<li>Hu, Y., Jiang, Z., Chen, K., Zhou, Z., Zhou, X., Wang, Y.. scNanoATAC-seq: a long-read single-cell ATAC sequencing method to detect chromatin accessibility and genetic variants simultaneously within an individual cell. Cell Research. 2022;33(1), 83-86. https://doi.org/10.1038/s41422-022-00730-x</li>
<li>Kim, J., Mun, S. J., Kim, J., Son, M. J., Kim, S.. Integrative analysis of single-cell RNA-seq and ATAC-seq reveals heterogeneity of induced pluripotent stem cell-derived hepatic organoids. iScience. 2023;26(9), 107675. https://doi.org/10.1016/j.isci.2023.107675</li>
<li>McKinney-Freeman, S., Hall, T.. Evaluation of Ranzoni et al.: Integrative Single-Cell RNA-Seq and ATAC-Seq Analysis of Human Developmental Hematopoiesis. Cell Stem Cell. 2021;28(3), 357-358. https://doi.org/10.1016/j.stem.2021.02.020</li>
<li>Alaterre, E., Ovejero, S., Espéli, M., Fest, T., Cogné, M., Milpied, P.. Integrative Single-Cell RNA-Seq and Single-Cell ATAC-Seq Analysis of Human Plasma Cell Differentiation. Blood. 2023;142(Supplement 1), 86-86. https://doi.org/10.1182/blood-2023-181507</li>
<li>LaPensee, C. R., Mohan, D., Lerario, A. M., Hammer, G. D.. SUN-LB39 ATAC-Seq Reveals Dynamic Changes in Chromatin Accessibility Following PKA Activation in NCI-H295R Adrenocortical Cells. Journal of the Endocrine Society. 2020;4(Supplement_1). https://doi.org/10.1210/jendso/bvaa046.2316</li>
<li>Ma, W., Lu, J., Wu, H.. Cellcano: supervised cell type identification for single cell ATAC-seq data. Nature Communications. 2023;14(1). https://doi.org/10.1038/s41467-023-37439-3</li>
<li>Shin, J., Berg, D., Zhu, Y., Shin, J., Song, J., Bonaguidi, M.. Single-Cell RNA-Seq with Waterfall Reveals Molecular Cascades underlying Adult Neurogenesis. Cell Stem Cell. 2015;17(3), 360-372. https://doi.org/10.1016/j.stem.2015.07.013</li>
<li>Pott, S., Lieb, J. D.. Single-cell ATAC-seq: strength in numbers. Genome Biology. 2015;16(1). https://doi.org/10.1186/s13059-015-0737-7</li>
<li>Unknown. Single-Cell Analysis Reveals Intratumoral Heterogeneity in Glioblastoma. Cancer Discovery. 2014;4(8), 865-865. https://doi.org/10.1158/2159-8290.cd-rw2014-136</li>
<li>Fujiwara, S., Baek, S., Wagh, K., Stavreva, D., Varticovski, L., Hager, G.. Abstract P3-11-06: Chromatin landscape analysis based on ATAC-seq and RNA- seq reveals that GRHL2 is a novel key transcription factor for endocrine therapy resistance. Cancer Research. 2023;83(5_Supplement), P3-11-06-P3-11-06. https://doi.org/10.1158/1538-7445.sabcs22-p3-11-06</li>
<li>Wang, L., Mo, S., Li, X., He, Y., Yang, J.. Single-cell RNA-seq reveals the immune escape and drug resistance mechanisms of mantle cell lymphoma. Cancer Biology and Medicine. 2020;17(3), 726-739. https://doi.org/10.20892/j.issn.2095-3941.2020.0073</li>
<li>Asuzu, D., Stoica, S., Elkahloun, A., Johnson, K., Chittiboina, P.. 519 Single Nucleus ATAC Sequencing Reveals Increased PPP1R17 Chromatin Accessibility Driving Cushing’s Disease. Neurosurgery. 2023;69(Supplement_1), 116-117. https://doi.org/10.1227/neu.0000000000002375_519</li>
<li>Minnoye, L., Marinov, G. K., Krausgruber, T., Pan, L., Marand, A. P., Secchia, S.. Chromatin accessibility profiling methods. Nature Reviews Methods Primers. 2021;1(1). https://doi.org/10.1038/s43586-020-00008-9</li>
<li>Bell, C. C., Fennell, K., Chan, Y., Rambow, F., Yeung, M. M., Vassiliadis, D.. Targeting enhancer switching overcomes non-genetic drug resistance in acute myeloid leukaemia. Nature Communications. 2019;10(1), 2723-2723. https://doi.org/10.1038/s41467-019-10652-9</li>
<li>Bock, C., Datlinger, P., Chardon, F. M., Coelho, M. A., Dong, M. B., Lawson, K. A.. High-content CRISPR screening. Nature Reviews Methods Primers. 2022;2(1). https://doi.org/10.1038/s43586-022-00098-7</li>
<li>Harmange, G., Hueros, R. A. R., Schaff, D. L., Emert, B., Saint-Antoine, M., Kim, L. C.. Disrupting cellular memory to overcome drug resistance. Nature Communications. 2023;14(1), 7130-7130. https://doi.org/10.1038/s41467-023-41811-8</li>
<li>Sidoli, S., Lopes, M., Lund, P. J., Goldman, N., Fasolino, M., Coradin, M.. A mass spectrometry-based assay using metabolic labeling to rapidly monitor chromatin accessibility of modified histone proteins. Scientific Reports. 2019;9(1), 13613-13613. https://doi.org/10.1038/s41598-019-49894-4</li>
<li>Janic, A., Abad, E., Amelio, I.. Decoding p53 tumor suppression: a crosstalk between genomic stability and epigenetic control?. Cell Death and Differentiation. 2024;32(1), 1-8. https://doi.org/10.1038/s41418-024-01259-9</li>
<li>Barkley, D., Rao, A., Pour, M., França, G. S., Yanai, I.. Cancer cell states and emergent properties of the dynamic tumor system. Genome Research. 2021;31(10), 1719-1727. https://doi.org/10.1101/gr.275308.121</li>
<li>Martínez‐Zamudio, R. I., Stefa, A., Freitas, J. A. N. L. F. d., Vasilopoulos, T., Simpson, M., Dore, G. J.. Escape from oncogene-induced senescence is controlled by POU2F2 and memorized by chromatin scars. Cell Genomics. 2023;3(4), 100293-100293. https://doi.org/10.1016/j.xgen.2023.100293</li>
<li>Gross, S. M., Dane, M., Smith, R., Devlin, K. L., McLean, I. C., Derrick, D. S.. A multi-omic analysis of MCF10A cells provides a resource for integrative assessment of ligand-mediated molecular and phenotypic responses. Communications Biology. 2022;5(1), 1066-1066. https://doi.org/10.1038/s42003-022-03975-9</li>
</ol>
</article>