How a Shifting Cast of Transcription Factors Rewires Androgen Signaling in the Aging Brain

Testosterone has long been known to do more than build muscle and bone. It also helps keep the male brain healthy, and clinical evidence has repeatedly linked falling androgen levels with cognitive decline. Men undergoing androgen deprivation therapy for prostate cancer, for example, show a higher incidence of dementia and Alzheimer’s disease. Yet the molecular machinery that allows the androgen receptor, the protein that reads testosterone’s instructions in the genome, to function inside human neurons has remained largely mysterious. A new study published in iScience by Kenichi Takayama of the Tokyo Metropolitan Institute for Geriatrics and Gerontology and colleagues, including Satoshi Inoue and Yutaka Suzuki, now maps that machinery in unprecedented detail, and its findings suggest that Alzheimer’s disease may sabotage androgen signaling not simply by draining the brain of hormone, but by dismantling the very transcription factor teams that the receptor depends on.

The androgen receptor belongs to the family of nuclear hormone receptors that bind DNA only when their ligand is present. Once activated by dihydrotestosterone, the potent form of testosterone, the receptor docks onto thousands of genomic sites called androgen receptor binding sites and switches nearby genes on or off. Where the receptor lands is not random. In prostate cancer, a so-called pioneer factor called FOXA1 prises open the chromatin and guides the receptor to its targets. But whether the same rulebook applies in neurons was unknown. To find out, the team turned to SH-SY5Y cells, a human neuroblastoma-derived line widely used as a model of neuronal biology, and used chromatin immunoprecipitation sequencing, or ChIP-seq, to chart every place the androgen receptor bound across the genome.

The results were striking. In these neuronal cells, the genes nearest to androgen receptor binding sites were overwhelmingly involved in axonogenesis, forebrain development and synapse organization, and androgen treatment shifted the expression of genes tied to nervous system and synaptic function. When the researchers scanned the DNA sequences underneath the binding peaks for recognizable motifs, the strongest hit was not a familiar nuclear receptor pattern but the binding motif of the GATA family of transcription factors, with GATA3 standing out. GATA3 is a lineage-defining factor in the immune system and in neurons, and the team confirmed that it physically associates with the androgen receptor through co-immunoprecipitation and immunofluorescence. ChIP-seq showed that GATA3’s own binding signal was enriched precisely at the receptor’s peaks, a molecular handshake suggesting the two proteins operate as a collaborative complex.

The collaboration proved functionally essential. When the researchers knocked down GATA3 with siRNA, androgen-driven induction of neuronal genes collapsed, and the receptor’s own protein levels dropped as well, though overexpression of the receptor could partially rescue the loss. Reporter assays drove the point home: a luciferase construct carrying a neuronal androgen receptor binding enhancer from the HAND2 gene lost its androgen responsiveness when the GATA3 motif within it was mutated. Perhaps most tellingly, when SH-SY5Y cells were coaxed to differentiate with retinoic acid, androgen treatment and receptor overexpression both lengthened the cells’ neurites and boosted NeuN, a canonical neuronal marker, but silencing GATA3 blunted these effects. In other words, GATA3 is not a passive bystander; it helps maintain androgen receptor protein and enables the hormone to push cells toward a mature neuronal identity.

Digging deeper into the epigenetic landscape, the team found that androgen receptor binding sites in the neuronal cells overlapped heavily with super-enhancers, the dense clusters of acetylated histone H3K27 that mark the genes most critical to a cell’s identity. Among the receptor-regulated super-enhancer genes were NFIX and HAND2, transcription factors that govern inhibitory interneuron function and neurogenesis, as well as KLF7 and NEFL, both required for axonal growth. The picture that emerges is of the androgen receptor plugging into the very enhancer network that defines what a neuron is, amplifying the transcriptional programs that keep neurons wired and communicative.

The crucial question was whether this architecture holds in actual human brains, and what happens to it with age and disease. Using postmortem temporal lobe gray matter from the Brain Bank of the Tokyo Metropolitan Institute for Geriatrics and Gerontology, the researchers performed ChIP-seq for the androgen receptor in five elderly male donors: two controls with no or minimal Alzheimer’s pathology and three Alzheimer’s patients spanning moderate to advanced disease. In the control brains, thousands of reproducible receptor binding sites were detected, validated with two different antibodies targeting the receptor’s N-terminal and C-terminal domains. Motif analysis revealed that, alongside the receptor’s own motif, the peaks were enriched for binding sites of NRF1 and SP1, neuron-associated factors, as well as AP-1 family motifs, and ChIP-seq confirmed that NRF1, SP1, GATA3 and the AP-1 component Fra-1 all co-occupied the receptor’s binding regions in vivo.

Then came the disease-related twist. Public transcriptome datasets and the team’s own clinical samples showed that GATA3 and GATA4 expression declines with aging and is further reduced in Alzheimer’s brains, while Fra-1, an AP-1 family factor, rises with both age and disease. In the two most advanced Alzheimer’s samples, the number of detectable androgen receptor binding sites plummeted, and the sites that did appear were largely divergent, found in regions where binding was negligible in control brains. Motif analysis of these disease-specific peaks pointed instead to the STAT family of transcription factors, and indeed STAT3 and STAT5 proteins were elevated in Alzheimer’s tissue and interacted more strongly with the androgen receptor. Because inflammatory response genes, including STAT3, were the most significantly upregulated category in the Alzheimer’s samples, the authors propose that inflammation-driven transcription factors hijack or displace the receptor’s normal collaborative partners, diverting its binding landscape in advanced disease.

The consequences for neuronal genes were severe. RNA sequencing showed that genes bound by the androgen receptor, whether alone or together with GATA3 or Fra-1, were significantly enriched among the genes downregulated in Alzheimer’s brains. Among these were genes tied to mitochondrial ATP synthesis and the proteasome, pathways with well-established protective roles against Alzheimer’s pathology. In control brains, receptor binding was enriched at the promoters of such genes; in advanced disease, that enrichment largely vanished. Notably, receptor binding was preserved in the brains of aged men without Alzheimer’s, despite their naturally reduced androgen levels, which suggests that the collapse of androgen signaling in disease is driven less by hormone depletion than by the disintegration of the transcription factor network that anchors the receptor to chromatin.

To test causality, the team forced expression of Fra-1 in the SH-SY5Y neuronal model. The effect was dramatic: the receptor’s binding profile reshaped itself around the new factor, Fra-1’s motif replacing GATA3’s as the dominant feature at binding sites, and histone acetylation at key neuronal enhancers, exemplified by the NEFL and KLF7 loci, was suppressed. Cells overexpressing Fra-1 lost their neuron-like morphology and failed to induce NeuN, and neuron-specific genes near the remodeled binding sites were downregulated. Reporter assays indicated that Fra-1 represses enhancer activity through a mechanism that depends on intact GATA3 motifs, hinting that Fra-1 displaces GATA3 from shared genomic sites. The authors are careful to note the study’s limitations, including the small number of brain donors, the possibility that neuronal loss contributes to reduced binding signals, and the need for single-cell analyses to confirm which cell types drive the observed shifts. Even so, the model they propose is compelling: the integrity of the androgen receptor’s collaborative transcription factor network, rather than hormone levels alone, may determine how well androgen signaling protects the aging brain, and restoring or protecting factors like GATA3, or countering the inflammatory STAT and AP-1 drift, could open new therapeutic avenues for slowing Alzheimer’s progression.

Subject of Research: Dysregulation of neuronal androgen receptor transcription factor signaling in aging and Alzheimer's disease

Article Title: Dysregulation of neuronal androgen receptor signaling by divergent collaborative transcription factors in Alzheimer’s disease

Article References: Takayama, K., Suzuki, Y., Saito, Y., Suzuki, T., Tsutsumi, S., Aburatani, H., & Inoue, S. (2026). Dysregulation of neuronal androgen receptor signaling by divergent collaborative transcription factors in Alzheimer’s disease. iScience, 29(10), Article 117731. https://doi.org/10.1016/j.isci.2026.117731

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117731

Keywords: androgen receptor, Alzheimer's disease, GATA3, transcription factors, super-enhancers, ChIP-seq, Fra-1, STAT3, neurons, aging brain, epigenetics, testosterone