Glioblastoma does not choose its victims by sex, but a striking new study suggests it may wage its war on treatment differently in men and women. Researchers at Johns Hopkins University and the Harvard T.H. Chan School of Public Health report in the journal Biology of Sex Differences that women with glioblastoma carry a distinctive rewiring of the genetic circuitry that governs how their tumors respond to low oxygen — a rewiring that links the oxygen-sensing machinery of the tumor to the metabolic, immune and structural programs that make this cancer so difficult to eradicate. By constructing mathematical maps of gene regulation from tumor data spanning more than a thousand patients across two independent cohorts, the team found that hypoxia-responsive pathways, anchored by the master regulator HIF1A, are placed under markedly stronger transcriptional control in female glioblastoma than in male glioblastoma or in slower-growing low-grade gliomas. The finding offers one of the clearest mechanistic accounts yet of why the deadliest brain tumor may resist therapy through different routes in each sex, and it points toward drug strategies that could be tailored accordingly.
The stakes of the question could hardly be higher. Gliomas, tumors that arise from the glial cells that support and protect neurons, account for roughly 26 percent of all brain tumors and 80 percent of malignant ones. Under the World Health Organization’s classification system, grades 1 and 2 are considered low-grade gliomas, while grade 4 corresponds to glioblastoma, IDH wild-type — the aggressive, treatment-refractory form. The clinical gulf between the two categories is enormous. Patients with glioblastoma who receive the full arsenal of surgery, radiation and the chemotherapy drug temozolomide survive a median of only about 17 months; without any treatment, survival drops to roughly three months. Low-grade glioma, by contrast, carries a median survival of about seven years. This stark difference is precisely what made low-grade glioma such a valuable comparator in the new study: it shares an organ and a cell of origin with glioblastoma but responds far better to treatment, and it shows no meaningful sex difference in outcome.
That sex difference — or the absence of one — sits at the heart of the puzzle the researchers set out to solve. Overall outcomes in glioblastoma are grim for both men and women, yet the picture is not symmetrical. Women respond better to temozolomide, the standard of care, than men do; even so, the five-year survival rate among females with glioblastoma remains below 10 percent. In low-grade glioma, by contrast, sex is not associated with survival at all, despite a slightly higher incidence among males. Known sex-specific molecular features, such as the higher frequency of MGMT promoter hypermethylation in female tumors, explain only a fraction of the gap. The team therefore reasoned that comparing glioblastoma against low-grade glioma within each sex — rather than against healthy brain tissue — would isolate the regulatory changes specific to the resistant, late-stage disease, and reveal whether those changes themselves differ between men and women.
To do this, the group turned to gene regulatory networks: computational models that map which transcription factors — the proteins that switch genes on or off by binding to regulatory DNA — exert control over which target genes, with weighted connections reflecting the strength of evidence for each relationship. The researchers used an algorithm called PANDA, or Passing Attributes between Networks for Data Assimilation, which employs message passing to reconcile three distinct data sources: a prior network built from transcription factor binding motifs cataloged in the CIS-BP database and mapped onto regulatory regions of the human genome; protein-protein interaction data from the GRAND database; and gene co-expression measured within each patient group. Starting from RNA sequencing data in The Cancer Genome Atlas — 153 glioblastoma samples, 53 from women and 100 from men, and 652 low-grade glioma samples, 273 from women and 379 from men — the team built four separate networks, one each for female and male glioblastoma and female and male low-grade glioma, normalizing the raw counts with the voom method into log2-counts per million and restricting the analysis to protein-coding genes. In the female networks, prior weights for Y-chromosome transcription factors were set to zero.
From these networks the researchers computed a targeting score for every gene — the sum of all incoming regulatory edge weights — and ranked the genes within each group. Differences between those rankings, a quantity the team calls differential targeting, revealed which genes come under new or intensified transcription factor control in one condition compared with another. Pathway enrichment was then assessed with the fgsea algorithm against the Molecular Signatures Database canonical pathway collection, using a false discovery rate threshold of 0.05. A pathway was counted as female-specific to glioblastoma only if it was enriched in the comparisons of female glioblastoma against both male glioblastoma and female low-grade glioma, but not in the sex comparison within low-grade glioma, where no clinical sex difference exists. Two additional tools deepened the analysis. MONSTER models the transition between two networks as a matrix equation, estimating a transition matrix whose largest off-diagonal weights identify the transcription factors driving the rewiring between disease states. BLOBFISH searches the networks for subnetworks in which the same regulators jointly control multiple groups of genes, retaining only statistically supported connections. Every key finding was then re-tested in an independent cohort, REMBRANDT.
The results in women were strikingly coherent. Female glioblastoma showed increased targeting of hypoxia pathways and of the programs hypoxia is known to drive: aerobic glycolysis and glycan biosynthesis within carbohydrate metabolism, powered by genes such as LDHA, PGK1, GAPDH, HK3 and PKM; immune processes including neutrophil degranulation and lysosomal pathways; and extracellular matrix and collagen remodeling. One enriched pathway initially appeared puzzling — renal cell carcinoma — but its gene content resolved the mystery. The signal was driven by the angiogenesis genes VEGFA and PDGFRA, an entire glycolysis program, and VHL, the tumor suppressor responsible for tagging HIF1A for destruction. In other words, the brain tumor had mobilized the very circuitry that kidney cancers use to thrive in low oxygen. MONSTER analysis reinforced the picture: 249 transcription factors had rewired their targeting between low-grade glioma and glioblastoma in women, compared with only 12 in men. Of these, 240 were exclusive to females, and among them were three regulators of the hypoxia response — CXXC5, EGFR and, most consequentially, HIF1A itself — alongside more than a dozen controllers of innate immunity.
Men told a completely different story. The pathways placed under altered control in male glioblastoma were the spliceosome — the molecular machine that cuts and rejoins RNA transcripts — and androgen receptor signaling. The differential targeting touched genes across the spliceosomal E, pre-B, B, Bact, B, C and P complexes, indicating a broad reorganization of RNA processing rather than a change at a single point. This finding dovetails with earlier observations that different transcription factors target the spliceosome in the healthy brain tissue of men and women, and with evidence that alternative splicing can undermine treatment response independent of hypoxia by reshaping transcription factor activity, fueling angiogenesis and easing tumor cell migration. In male glioblastoma, in other words, resistance appears to route through RNA processing and androgen signaling rather than through oxygen sensing at all.
The most distinctive result, however, lay in how these pathways talk to each other. In the female glioblastoma networks, the same transcription factors were found to co-regulate hypoxia genes together with genes in carbohydrate metabolism, the extracellular matrix and immune processes — an interlocking regulatory web in which a single controller reaches across multiple disease programs at once. Male networks showed co-regulation too, but of a different architecture: mRNA splicing was linked with metabolism, extracellular matrix and immune pathways, while hypoxia sat entirely outside the web. When the team repeated the analysis in the REMBRANDT cohort, the canonical hypoxia result did not fully replicate, but the female-specific pattern re-emerged through a back door: genes targeted by HIF1A within the renal cell carcinoma pathway, particularly BIRC7, PDGFRA and VEGFA — all established downstream effectors that drive tumor progression under hypoxia — were co-regulated with the same pathway categories in women only, while the androgen receptor co-regulation seen in TCGA males did not reappear.
The biology behind these patterns is well supported by prior work. Hypoxia is a known driver of chemotherapy resistance: low oxygen forces metabolic reprogramming toward glycolysis, and in glioblastoma specifically HIF1A promotes angiogenesis, glucose metabolism and cell migration. Hypoxia also enables immune evasion through multiple mechanisms, and the neutrophil degranulation pathway flagged in the female tumors has been implicated in the necrosis and poorer survival that characterize aggressive disease. The new findings mesh neatly with earlier sex-difference research. One prior study found that temozolomide response in women, but not men, was linked to reduced integrin signaling — and integrins are known to regulate HIF1A expression and to remodel the extracellular matrix under hypoxic conditions. Another reported that granulocytic myeloid-derived suppressor cells, immunosuppressive cells that blunt anti-tumor immunity, are elevated in female mice with glioblastoma but not in males. A third showed that male mice respond better to anti-PD-1 immunotherapy, possibly because PD-L1, the molecular brake that the therapy targets, is overexpressed in hypoxic tumor environments.
The practical implication is that drugs already built to block HIF signaling in kidney cancer deserve testing as sex-specific therapy for women with glioblastoma, while men may require an entirely different strategy aimed at RNA splicing or androgen signaling. The authors are careful about the limits of their approach: the network method cannot say whether a regulatory relationship is activating or inhibitory, the data were collected at a single time point so causality must rest on prior knowledge rather than causal models, the tumors were classified under older WHO schemes that may have misclassified some cases, and both cohorts skew heavily toward white participants, with female glioblastoma samples forming the smallest group. Still, the study — funded by the National Institutes of Health and published open access — reframes glioblastoma treatment resistance as a problem of regulatory architecture that differs by sex, and it hands oncologists a concrete, testable shortlist of targets, with the oxygen sensor HIF1A at the top of the list for women.
Subject of Research: Sex-specific gene regulatory network rewiring in glioblastoma, focusing on female-specific dysregulation of hypoxia pathways and their co-regulation with carbohydrate metabolism, immune, and extracellular matrix programs driving treatment resistance
Subject of Research: Biology
Article Title: Gene regulatory network analysis identifies dysregulation of hypoxia pathways as contributing to glioblastoma treatment resistance in females
Article References: Adebari, T., Fanfani, V., Guebila, M. B., DeConti, D., Shutta, K. H., Lopes-Ramos, C. M., Hsu, L., DeMeo, D. L., Quackenbush, J., & Eicher, T. (2026). Gene regulatory network analysis identifies dysregulation of hypoxia pathways as contributing to glioblastoma treatment resistance in females. Biology of Sex Differences, 17(1), Article 134. https://doi.org/10.1186/s13293-026-00927-4
Image Credits: AI Generated
DOI: 10.1186/s13293-026-00927-4
Keywords: Glioblastoma, Low-grade glioma, Gene regulatory network, Transcription factor, Gene expression, Hypoxia, Treatment resistance
Cite Scienmag News
APA
MLA
Chicago
Juliet Wilcox. (August 30, 2026). Study links hypoxia pathway dysregulation to glioblastoma treatment resistance in women. Scienmag. https://scienmag.com/study-links-hypoxia-pathway-dysregulation-to-glioblastoma-treatment-resistance-in-women/
Juliet Wilcox. “Study links hypoxia pathway dysregulation to glioblastoma treatment resistance in women.” Scienmag, 30 August 2026, https://scienmag.com/study-links-hypoxia-pathway-dysregulation-to-glioblastoma-treatment-resistance-in-women/. Accessed 30 August 2026.
Juliet Wilcox. “Study links hypoxia pathway dysregulation to glioblastoma treatment resistance in women.” Scienmag. August 30, 2026. https://scienmag.com/study-links-hypoxia-pathway-dysregulation-to-glioblastoma-treatment-resistance-in-women/
Copy citation
Download RIS
Tags: gender-specific gene regulation in brain tumorsgene regulation mapping in glioblastoma patientsgenetic circuitry in glioblastomagenetic circuitry rewiring in female glioblastomaglioblastoma treatment resistanceHIF1A role in glioblastomahypoxia pathway dysregulation in glioblastomahypoxia pathway dysregulation in womenimpact of hypoxia on glioma progressionimpact of oxygen-sensing machinery on tumor progressionmetabolic and immune pathways in glioblastoma resistancemetabolic and immune programs in glioblastomamolecular mechanisms of treatment resistancepersonalized glioblastoma therapy strategiespersonalized treatment strategies for glioblastomarole of HIF1A in glioblastomasex differences in brain tumor biologysex differences in tumor hypoxia pathwayssex differences in tumor hypoxia responsesex-based differences in tumor microenvironmentsex-specific tumor microenvironment in gliomastailored drug development for glioblastoma based on sextailored drug therapies
