When Stem Cells Go Rogue: How “Cholesterol-Stuffed” Brain Cells May Drive Progressive MS

Progressive multiple sclerosis (PMS) is the phase of MS where disability quietly but relentlessly worsens, even when obvious relapses have faded. Therapies that work well in early, inflammatory MS often do very little here, suggesting that new, more “inside-the-brain” mechanisms are at play. Pathology studies have hinted that some neural stem/progenitor–like cells in PMS brain lesions look old and stressed: they carry markers of cellular senescence and sit in chronically inflamed, slowly expanding lesions. But whether these “aged” stem-like cells actually contribute to neurodegeneration, or are just bystanders, has been unclear. In a new Cell Stem Cell paper, Ionescu and colleagues tackle this head-on using a patient stem cell–derived model of PMS. They show that neural stem cells directly reprogrammed from the skin of PMS patients are not only senescent and inflamed but also hypermetabolic, channeling glucose into cholesterol synthesis and storing it as lipid droplets. That metabolic state, in turn, drives a neurotoxic secretome that can kill neurons in vitro — and, crucially, can be pharmacologically “reprogrammed” with a statin.

Building an “aged” neural stem cell model from patient skin
Instead of using classic iPSCs (which tend to “rejuvenate” cells epigenetically), the team used direct reprogramming to convert fibroblasts from PMS patients and unaffected controls into induced neural stem/progenitor cells (iNSCs), preserving age-related features. Both PMS and control iNSCs expressed canonical NSC markers like NESTIN and SOX2 at similar levels, and they lacked pluripotency markers, confirming that reprogramming worked as intended.

The differences emerged when they looked at senescence. PMS iNSCs showed increased expression of classic senescence genes (CDKN2A, CDKN1A, TP53) and higher senescence-associated β-galactosidase activity, while the original fibroblasts did not differ between groups. That means the senescent phenotype is closely tied to acquiring an NSC-like identity in PMS cells. Proteomics and metabolomics revealed a parallel shift in function: PMS iNSCs upregulated inflammatory and interferon signaling and showed a coordinated boost in metabolic pathways linked to energy production, lipid synthesis, and anabolic/catabolic balance — a textbook senescence-associated hypermetabolic state.

Glucose in, cholesterol out: how PMS stem cells become lipid-laden
To understand what this hypermetabolism was actually doing, the authors traced 13C-labeled glucose through PMS and control iNSCs. PMS cells consumed more glucose, shunted more of it through glycolysis and the pentose phosphate pathway, and fed more labeled carbon into the TCA cycle. At 24 hours, they showed increased labeling of fatty acids and, strikingly, cholesterol and its precursors — meaning that glucose was being actively turned into lipid precursors, not just burned for ATP.

Lipidomics put a structural face on this shift. Inside PMS iNSCs, multiple lipid classes were elevated, but cholesteryl esters (ChEs) stood out with a ~12-fold increase and involvement of most species in that class. These lipids accumulated as perinuclear lipid droplets, clearly visible as dense puncta in microscopy images on page 3–4 of the paper.

The overall lipidome became more saturated, with shorter acyl chains — a pattern consistent with a cell that is channeling building blocks into storage. Inhibiting triglyceride synthesis partially reduced droplet accumulation, but blocking HMG-CoA reductase (HMGCR) (the rate-limiting enzyme in cholesterol synthesis) with simvastatin almost completely normalized lipid droplets, pointing to overactive cholesterol synthesis as the core driver.

From fat storage to toxic talk: the senescence-associated secretory phenotype
Senescent, lipid-laden cells are notorious for developing a senescence-associated secretory phenotype (SASP) — a cocktail of cytokines, proteases, growth factors, and matrix proteins that can reshape their microenvironment, often in damaging ways. When the authors profiled proteins secreted into the conditioned medium of PMS vs control iNSCs, they found exactly that: PMS cells released higher levels of SASP-like factors such as TIMP1, fibronectin (FN1), and MMP2, along with broader changes in pathways tied to neurite dynamics, metabolism, IGF signaling, inflammatory cascades, and extracellular matrix (ECM) degradation.

A multi-omics covariance network overlaid intracellular lipids, metabolites, and proteins with the secretome. Two “hubs” emerged: lipid droplet biosynthesis and the secretory phenotype, with dense connectivity between them. Many of the transcription factors predicted to orchestrate the PMS iNSC gene and protein expression signatures — including E2F1, EGR1, WT1, SP1, c-JUN, and JUNB — are known to be modulated by cholesterol or pathways of cholesterol synthesis, esterification, and fatty acid saturation.

The picture that emerges is elegant and slightly sinister: a cholesterol-driven, HMGCR-dependent lipogenic state rewires transcriptional networks to enforce a neurotoxic SASP.

Statin intervention: simvastatin reshapes the secretome without “unsenescing” the cells
Given that the central metabolic node is HMGCR, the team deployed simvastatin, a widely used HMGCR inhibitor. Treating PMS iNSCs for 48 hours didn’t erase senescence: senescence marker expression and β-gal activity stayed unchanged, and the global hypermetabolic profile by Seahorse analysis persisted. What simvastatin did do was redistribute lipids and remodel signaling. Lipid droplets shrank to near-control levels, ChEs dropped, and specific fatty acyl groups previously tied up in ChEs were now used in other lipid classes, shifting the lipidome toward longer, less saturated chains. On the secretome side, simvastatin had a much stronger impact on PMS iNSCs than on controls. Proteomics of conditioned media revealed new differentially secreted proteins emerging only in PMS + simvastatin, while some of the previously abnormal factors (like EPHA4, TIMP2, PRDX2) moved toward control levels. Enrichment analyses showed partial correction of pathways related to ECM degradation, ephrin signaling, ROS detoxification, and apoptosis, and a tilt toward more cytoprotective SASP components (e.g., oxidoreductase activity, cell differentiation) and away from cytokine-associated responses. Transcription factor enrichment again pointed to cholesterol-responsive TFs as likely mediators of these changes. Taken together, simvastatin doesn’t rejuvenate the cells, but it does retune their “language” to the surrounding brain.

Putting it to the test: how PMS stem cell secretions affect neurons
Of course, the crucial question is whether this altered secretome actually harms neurons. To test this, the authors exposed differentiated human SH-SY5Y neurons to conditioned media from control or PMS iNSCs. Conditioned media from PMS iNSCs caused neurite retraction and increased cleaved caspase-3 staining, a marker of apoptosis — a clear sign that the PMS secretome is neurotoxic. By contrast, media from control iNSCs were comparatively benign. When PMS iNSCs were pretreated with simvastatin, their conditioned media no longer induced neurite loss or caspase-3 activation; neuronal morphology and survival looked much closer to the control condition. Blocking a single candidate molecule, the EphA4 receptor, was not sufficient to rescue neurite length or prevent cell death, implying that neurotoxicity arises from a complex mix of reduced protective factors and increased inflammatory/toxic ones, rather than any single villain. These functional assays tie the whole story together: cholesterol-driven metabolic rewiring in senescent PMS iNSCs leads to a toxic SASP that can kill neurons, and simvastatin can flip that secretome into a more protective mode.

Why this matters for MS — and what’s next
This work has several important implications. First, it strengthens the idea that neural stem/progenitor–like cells in PMS are not just passive victims of the disease environment but active players whose senescence, metabolism, and secretory behavior can drive neurodegeneration. That adds a new dimension to how we think about brain aging in MS and raises caution about autologous NSC-based therapies: if patient-derived stem cells carry intrinsic, disease-linked dysfunctions, simply transplanting them back may not be harmless.

Second, the study provides a plausible, CNS-intrinsic mechanism for the clinical benefits seen with high-dose simvastatin in the MS-STAT trial, where reduced brain atrophy and slower progression could not be fully explained by lowering blood cholesterol. Here, simvastatin acts not as a generic anti-inflammatory, but as a metabolic reprogrammer of lipid-laden, senescent neural stem cells, dampening their neurotoxic SASP without needing to reverse senescence itself.

There are caveats: the number of donors is modest, no isogenic controls are available (a common limitation in MS), and the system is in vitro, based on a single neuronal cell line. The in-brain context, with myelin, microglia, and complex circuitry, may modify or amplify these effects. Still, the multi-omics depth, functional assays, and clear mechanistic link between cholesterol synthesis, transcriptional control, and neurotoxic signaling make this a compelling model. Going forward, similar approaches in organoids or co-culture systems, and correlation with in vivo biomarkers of lipid metabolism and SASP in PMS patients, could help translate these insights into targeted neuroprotective strategies — whether via statins, more selective HMGCR modulators, or drugs that specifically disrupt the cholesterol-SASP axis in senescent neural cells.

Disclaimer: This blog post is based on the provided research article and is intended for informational purposes only. It is not intended to provide medical advice. Please consult with a healthcare professional for any health concerns.

References:
Ionescu, R. B., Nicaise, A. M., Reisz, J. A., Williams, E. C., Prasad, P., Willis, C. M., ... & Pluchino, S. (2024). Increased cholesterol synthesis drives neurotoxicity in patient stem cell-derived model of multiple sclerosis. Cell Stem Cell, 31(11), 1574-1590.