Pathological brain compression drives neuronal apoptosis, impairs synaptic activity, and promotes glial neuroinflammation in mice and humans

Introduction: : Several central nervous system pathologies—including traumatic brain injury (TBI), multiple sclerosis (MS), and glioblastoma (GBM)—present with mass effect, where expanding lesions compress and deform surrounding tissue. This deformation, visible on MRI as ventricular compression or midline shift, arises from edema and/or the accumulation of cells and extracellular matrix in a confined space. Unlike spinal or peripheral nerve compression, intracranial mass effect is compounded by the skull’s rigidity, often resulting in debilitating neurological symptoms.

In GBM, mass effect stems from both peritumoral edema and solid stress generated by tumor and stromal components. These forces reduce cerebral blood flow, damage neurons, and predict poor clinical outcomes. While peritumoral edema has prognostic value, its direct contribution to neurological impairment remains unclear. Solid stress, however, has been more clearly linked to functional decline. Our prior in vivo studies demonstrated that chronic mechanical compression alone, even in the absence of a tumor, can lead to vessel collapse, neuronal loss, and behavioral deficits.

Similar mechanical deformations also occur in tumefactive MS and severe TBI, where acute injury causes neuroinflammation, excitotoxicity, and axonal degeneration. However, the effects of chronic compression on neurons and glia remain poorly defined.

Here, we use in vitro compression systems to model mass effect and examine its direct impact on neurons and glia. We show that chronic compression impairs neuronal network function, reduces synapse density, and induces neuroinflammatory gene expression in astrocytes and microglia. These findings shed light on how solid stress contributes to neurodegeneration in GBM and other space-occupying lesions.

Materials and

Methods: : Human iPSCs from three healthy donors were differentiated into induced neurons (iNs) via doxycycline-inducible Ngn2 overexpression. Cells were cultured on Matrigel-coated transwells, transduced with lentivirus (rtTA, Ngn2, GCaMP8m, and/or NLS-GFP), and co-cultured with mouse glia for 60 days.

To simulate mass effect, a 1% agarose cushion and 3D-printed PLA weight (0.15 kPa) were placed on iNs, astrocytes, or microglia for 24 hours. Controls received agarose only. Cell viability was assessed using Hoechst and PI staining; fluorescent images were analyzed in ImageJ to quantify nuclear labeling and membrane integrity.

Calcium imaging was performed using GCaMP8m fluorescence and time-lapse microscopy. Neurons were imaged, and recorded videos were analyzed in MATLAB to extract amplitude, frequency, and synchronicity of events.

For immunocytochemistry, cells were fixed, blocked, and stained with fluorescently labeled antibodies (MAP2, synapsin, FOSB, HIF1A). Hypoxia was evaluated using HypoxiTRAK dye and flow cytometry.

RNA-seq was performed on iNs and human astrocytes. Reads were aligned using STAR, and differential expression was computed with DESeq2. Cell-type deconvolution used NNLS and BayesPrism with scRNA-seq references. Motif enrichment was done with HOMER.

In vivo compression was modeled using a cranial compression window in adult mice, applying gradual compression via a screw mechanism over three weeks. Paraformaldehyde-fixed brain sections were processed for immunohistochemistry and stained for HIF1A, NeuN, and GFAP. Confocal images were analyzed to assess neuronal loss and gliosis.

Results, Conclusions, and Discussions:: Using an in vitro compression system, we modeled the mechanical stress generated by glioblastoma (GBM) and found that chronic solid stress induces neuronal death, synapse loss, and impaired network activity in human iPSC-derived neurons. Applying 0.15 kPa of compressive force—similar to that measured in GBM—we observed a >2-fold increase in PI+ apoptotic neurons, reduced synaptic puncta density, and diminished calcium activity, indicating disrupted synapse connectivity and function. These effects were consistent across multiple iPSC donors, although neurons from one line (iPSC #3) exhibited a less pronounced response, suggesting possible genetic modifiers of injury susceptibility.

RNA-seq revealed that compression activates hypoxia signaling, apoptosis, and pro-inflammatory transcriptional programs in both neurons and co-cultured murine glia. In neurons, HIF-1α target genes such as BNIP3 and PGK1 were upregulated alongside apoptotic mediators, suggesting HIF-1α–mediated mitophagy may contribute to cell death. In glia, compression triggered NF-κB and AP-1 signaling, with elevated expression of cytokines including IL1B and TNF, indicative of reactive gliosis. Human astrocytes and microglia subjected to compression in monoculture also showed direct, sustained activation of reactive transcriptional programs—independent of hypoxia—highlighting a mechanosensitive glial response.

In vivo, chronic cortical compression in mice reproduced these effects, with increased HIF1A+ neurons, neuronal loss, and GFAP+ astroglial scarring. Similar molecular signatures were identified in GBM patient samples from the Ivy Glioblastoma Atlas Project. Stratifying patients by midline shift adjusted for edema, we found increased expression of hypoxia and inflammatory genes at tumor margins, particularly in neurons and astrocytes.

Together, our findings demonstrate that solid stress activates HIF-1–driven metabolic stress in neurons and NF-κB–mediated inflammation in glia, both contributing to neurodegeneration. These results suggest that mechanical compression is a key driver of peritumoral neurotoxicity in GBM and may represent a therapeutic target in other mass effect–associated neurological diseases.

Acknowledgements and/or References (Optional):: Funding for this work was provided by the National Institutes of Health (NIH/NIGMS R35GM151041, to M.D.), and the Harper Cancer Research Institute (Cancer Cure Venture grant, to M.D. and C.P.). Other funding support included the Berthiaume Institute for Precision Health Graduate Fellowship and the Scientific Artificial Intelligence Fellowship (both to M.Z.).