Glioblastoma is the most common and aggressive primary malignant brain tumor in adults. Even with the full modern arsenal of surgery, radiation, and the chemotherapy drug temozolomide, median survival hovers around fifteen months, and the disease has long defied the wave of therapeutic advances that have transformed many other cancers. One reason is the tumor’s extraordinary adaptability: glioblastoma cells rewire their own metabolism, recruit and corrupt surrounding cells, and sculpt the local immune environment into one that shelters rather than attacks them. The new study points to a specific molecular conversation between the tumor microenvironment and the brain’s own signaling chemistry as a driver of that malignant orchestration — and identifies sex as a decisive variable in the conversation.

For decades, neuroscience and cancer biology operated on largely separate tracks. That began to change when researchers demonstrated that neurons can directly stimulate tumor growth, most famously in prostate, breast, gastric, and pancreatic cancers, and later in gliomas themselves, where neuronal activity was shown to promote tumor proliferation through activity-dependent secretion of factors such as neuroligin-3 and the signaling molecule BDNF. GABA, however, occupies a peculiar place in this emerging neuro-oncology landscape. As the brain’s dominant inhibitory neurotransmitter, it is released in vast quantities by interneurons throughout gray matter. Earlier work had suggested that some glioma cells could even metabolize GABA as a fuel source, and that GABAergic signaling might influence tumor cell proliferation. But the new study reframes the question: rather than acting directly on tumor cells, GABA signaling appears to exert its pro-tumor influence by reshaping the immune landscape of the tumor itself.

The research team set out to dissect this relationship using genetically engineered mouse models of glioblastoma, in which tumors arise spontaneously in the brain and recapitulate many features of the human disease, including its cellular heterogeneity and its infiltration by diverse immune populations. By manipulating GABA signaling in these models — enhancing it in some animals and dampening it in others — the investigators could trace its effects on tumor growth and survival. The results were unambiguous: activating GABA signaling accelerated glioblastoma progression, while interfering with that signaling slowed tumor growth. Female mice, in particular, bore the brunt of the effect, with GABA activation driving markedly more aggressive disease than in their male counterparts — a sex-specific pattern that mirrors, intriguingly, epidemiological data in humans, where glioblastoma incidence and certain molecular features of the disease differ between men and women.

The mechanism, however, proved to be the study’s most consequential revelation. When the researchers profiled the immune composition of tumors exposed to heightened GABA signaling, they found a striking expansion of myeloid-derived suppressor cells — a heterogeneous population of immature myeloid cells that, as their name suggests, suppress the activity of T cells and other antitumor immune players. In glioblastoma, where the tumor is infiltrated by an extraordinary abundance of myeloid-lineage cells that often make up the majority of its cellular mass, these suppressor cells are already recognized as central architects of the tumor’s immunosuppressive fortress. What the new study demonstrates is that GABA signaling acts, in effect, as a recruitment and activation signal for this immunosuppressive armada.

Delving into the cellular details, the team found that GABA exerts its influence on myeloid cells through GABA receptors expressed on their surface. Engagement of these receptors triggers intracellular signaling cascades that reprogram the cells’ behavior, skewing them toward a potently immunosuppressive state. The consequence is a dampening of cytotoxic T-cell activity within the tumor — the very immune cells that immunotherapies such as immune checkpoint inhibitors depend upon to eliminate cancer. In other words, GABA does not make tumor cells grow faster so much as it blinds the immune system to their presence. This distinction matters therapeutically: targeting a signaling pathway that acts on host immune cells rather than on genetically unstable tumor cells may offer a more durable intervention, one less prone to the rapid evolution of resistance that plagues treatments aimed directly at cancer cells.

The sex specificity of the effect adds a further layer of biological interest. Sex differences in cancer have long been catalogued but poorly explained, spanning differences in incidence, molecular subtype distribution, immune infiltration, and treatment response. In glioblastoma, males are diagnosed at somewhat higher rates, yet the underlying biology of sex dimorphism remains murky. The new findings suggest that one axis of this dimorphism may run through neurotransmitter signaling and its downstream immunological consequences. Whether the difference in mice reflects differences in GABAergic tone, hormone-dependent modulation of GABA receptor expression on myeloid cells, or sex-linked variation in the myeloid compartment’s responsiveness remains a question the authors and the field will now pursue. The implication, however, is clear: preclinical studies that use only male animals — a historical default in much of biomedical research — risk missing mechanisms of genuine clinical relevance.

Translating the finding to the clinic will require careful groundwork, and the authors and outside experts caution that mouse models, however sophisticated, capture only part of human glioblastoma biology. Still, the therapeutic logic is compelling. Drugs that modulate GABA signaling already exist in abundance, developed over decades for epilepsy, anxiety, and other neurological conditions. Gabapentinoids, benzodiazepines, GABA reuptake inhibitors, and receptor-selective modulators constitute a mature pharmacopoeia with well-characterized safety profiles and, in several cases, proven ability to cross the blood–brain barrier. Repurposing or carefully redeploying members of this pharmacological family to blunt the pro-tumor, immunosuppressive effects of GABA signaling in glioblastoma — particularly in female patients whose tumors may be more dependent on this pathway — represents an unusually direct route from mechanism to potential clinical trial.

The study also resonates with a broader and rapidly expanding body of work on the neuro-immune axis in cancer. It is now well established that the nervous system innervates tumors and their microenvironments, and that neural signals can regulate everything from cancer stem cell function to angiogenesis to immune surveillance. Within tumors, neurotransmitters act less as long-range wires than as chemical signals exchanged among neurons, glia, tumor cells, and immune cells. GABA itself has been implicated in other cancers: in some solid tumors of the pancreas and breast, GABAergic signaling has been reported to promote invasion and stem-like behavior. The new glioblastoma study extends this theme into the one organ where GABA is most abundant, and identifies the myeloid immune compartment — rather than the tumor cell itself — as the critical sensor. That reframing may resolve some of the confusion in earlier literature, where direct effects of GABA on tumor cells appeared modest or inconsistent.

For patients, the immediate significance lies less in a new treatment — none is yet available — than in a changed understanding of what glioblastoma is. A brain tumor is not simply a mass of dividing cells; it is an ecosystem, wired into the electrical and chemical circuitry of the organ it invades. Each new node of that circuitry that science maps is a potential point of intervention. The identification of GABA-driven, myeloid-mediated immunosuppression as a sex-biased engine of tumor progression provides both a mechanistic target and a biomarker opportunity: if human glioblastomas can be stratified by their dependence on GABAergic signaling — and by the sex of the patient — trials of GABA-targeted immunomodulation could be designed with a precision that glioblastoma therapy has rarely enjoyed.

The study, published in Nature Cancer, was conducted by Ashish Pathak, Sravya P., B. Colon, and colleagues, who combined spontaneous and transplant-based mouse glioblastoma models with immunophenotyping, receptor-level perturbation, and mechanistic dissection of myeloid cell function. Their demonstration that a core neurotransmitter of the healthy brain can be co-opted to disarm antitumor immunity is likely to spur a wave of follow-up work, from human tissue analyses to drug repurposing efforts. As the neuro-oncology field continues to dismantle the boundary between brain and tumor, the message of this study is both sobering and energizing: the very chemistry that lets the brain think may also, under the wrong circumstances, help a tumor hide — and that hiddenness, at last, is something science can begin to target.

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.

The research team embarked on an ambitious journey to decrypt the transcriptional regulatory networks operational within hindbrain GPs. Utilizing advanced genomic and bioinformatic approaches, they meticulously mapped the top regulons—clusters of transcription factors (TFs) and their downstream targets—that govern cellular differentiation, proliferation, and lineage specification. Strikingly, the analyses revealed a remarkable dichotomy in the regulatory landscape between male and female GPs.

In female GPs, the dominant regulons were emblematic of differentiation processes. Key regulators such as Neurod1, implicated in neuronal differentiation, along with Rara, responsive to retinoic acid signals, and Nr2e1, an orphan nuclear receptor pivotal in neural specification, were highly enriched. This cadre of TFs primes female progenitors towards a more differentiated, lineage-committed state, suggesting an inherent bias favoring developmental progression.

Conversely, male GPs exhibited enrichment of regulons associated with maintenance of stemness and proliferative capacity. Notably, Six2 and Max, factors linked to neural progenitor identity, were prominent alongside Sp1, a transcriptional regulator implicated in sustaining proliferation and undifferentiated status. Additionally, male-specific networks incorporated hormone-responsive and neuroendocrine pathways, with TFs such as Etv2, Rax, and Mafk exerting influence. This constellation underscores a sex-specific regulatory program that preserves progenitor plasticity and delays differentiation.

Despite these sex-biased distinctions, certain core regulatory elements were consistently active across both sexes. Regulons governed by Ezh2, Sox2, Sox9, and Msx1—well-established master regulators of glial progenitors—exhibited comparable activity, affirming a shared foundational framework essential for GP identity and function. This observation highlights that while male and female GPs operate under a common genetic scaffold, sex-specific overlays fine-tune developmental outcomes.

Delving deeper, differential gene expression analyses between male and female GPs further delineated the molecular underpinnings driving these phenotypic divergences. Female GPs showcased activation of genes integral to translational initiation, including Eif3j1 and Eif5b, along with regulators of RNA metabolism such as Igf2bp1 and Elavl4. These molecular signatures emphasize a robust engagement in protein synthesis and RNA processing, consistent with a commitment to differentiation and specialized function.

In stark contrast, male GPs were marked by elevated expression of genes associated with neural precursor proliferation, including Fabp7 and Ptn—markers known to sustain progenitor cell cycling. Moreover, Y-chromosome-linked genes like Ddx3y and Kdm5d were uniquely expressed, reflecting the male genomic context. Genes implicated in gliogenesis, such as Nfib and Plpp3, were also enriched, further suggesting a predilection for maintaining progenitor pools and lineage plasticity.

Collectively, these findings portray male GPs as comparatively less differentiated than their female counterparts, hinting at a delayed or protracted maturation program. This sex-biased developmental tempo may have significant repercussions, potentially predisposing males to distinct neurodevelopmental vulnerabilities or differential responses to embryonic stresses.

The investigation gains additional gravity when considered in the context of pathological outcomes such as posterior fossa ependymoma (PFA), a lethal tumor type predominantly affecting males. By elucidating how androgen-driven transcriptional programs remodel GP states, this study provides a molecular framework to understand sex disparities in tumor susceptibility and aggressiveness.

Furthermore, the interplay between androgen activity and the identified male-biased regulatory networks could offer novel therapeutic targets. Modulating key TFs or pathways responsible for the progenitor state might recalibrate aberrant developmental cues and mitigate tumorigenic risk.

This research exemplifies the power of integrative genomic analyses coupled with sex as a biological variable, underscoring the necessity to consider sexual dimorphism in developmental neuroscience. It sets a precedent for extending such frameworks to other neural cell types and regions, fostering a more nuanced comprehension of brain development.

In summary, the study delineates a sophisticated balance of conserved and sex-specific transcriptional programs within embryonic hindbrain glial progenitors. Female GPs are skewed towards differentiation, driven by networks primed for specialization and maturation, whereas male GPs retain a more proliferative, stem-like identity governed by androgen-influenced regulators. This duality not only enhances our understanding of normal hindbrain development but also illuminates sex-linked susceptibilities underlying pediatric brain tumors.

The implications resonate beyond developmental biology, suggesting that therapeutic strategies for neuro-oncological conditions must integrate sex-specific molecular contexts to optimize efficacy and precision. Future studies will undoubtedly build upon this foundation, exploring the mechanistic pathways in greater depth and translating these insights into clinical interventions tailored by sex.

Subject of Research: Molecular mechanisms of glial differentiation in the embryonic hindbrain with a focus on sex-specific transcriptional regulatory networks.

Article Title: Androgen activity in the male embryonic hindbrain drives lethal PFA ependymoma.

Article References: Zhang, J., Ong, W., Rasnitsyn, A. et al. Androgen activity in the male embryonic hindbrain drives lethal PFA ependymoma. Nature (2026). https://doi.org/10.1038/s41586-026-10264-6

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41586-026-10264-6