Mesial temporal lobe epilepsy (MTLE) is the most common form of drug-resistant focal epilepsy, accounting for approximately 60–70% of temporal lobe epilepsy cases and representing the single largest indication for epilepsy surgery worldwide. For patients whose seizures originate from a sclerotic hippocampus — the hallmark pathology of MTLE — anti-seizure medications fail to achieve adequate seizure control in roughly 30–40% of cases. Resective surgery (anterior temporal lobectomy or selective amygdalohippocampectomy) offers seizure freedom rates of 60–80% in carefully selected patients, but many individuals are not surgical candidates due to bilateral seizure foci, dominant-hemisphere involvement with risk to verbal memory, or personal preference against brain surgery. Neurostimulation devices and dietary therapies provide partial seizure reduction at best. Mesenchymal stem cell (MSC) therapy is being investigated as a biologically targeted intervention that addresses the underlying hippocampal pathology — sclerosis, neuroinflammation, interneuron loss, and aberrant circuit reorganization — rather than merely suppressing seizure activity [1].

Where conventional treatments fall short. Anti-epileptic drugs (AEDs) target ion channels and neurotransmitter receptors to suppress neuronal excitability, but they do not reverse the structural pathology of hippocampal sclerosis — the neuronal loss, gliosis, and synaptic reorganization that sustains the epileptogenic network. Resective surgery removes the seizure focus but at the cost of hippocampal tissue, with predictable cognitive trade-offs particularly affecting verbal memory when the language-dominant hemisphere is involved. For the substantial subset of MTLE patients who are not surgical candidates and have failed two or more AED trials, the treatment landscape offers no disease-modifying options.

The deeper problem is tissue-level. Hippocampal sclerosis is not simply a scar; it is an active, progressive pathological process. The defining features — segmental neuronal loss in CA1, CA3, and CA4 subfields with relative sparing of CA2, dense reactive astrogliosis, and granule cell dispersion in the dentate gyrus — are accompanied by chronic microglial activation that sustains a pro-inflammatory microenvironment [2]. Activated microglia in the sclerotic hippocampus release interleukin-1β (IL-1β), tumor necrosis factor-alpha (TNF-α), and other pro-convulsant cytokines that lower the seizure threshold and promote further neuronal injury. This creates a self-perpetuating cycle: seizures drive inflammation, and inflammation promotes more seizures [3]. Meanwhile, the progressive loss of inhibitory GABAergic interneurons — particularly parvalbumin-positive basket cells and somatostatin-positive O-LM cells — tilts the excitation-inhibition balance irreversibly toward hyperexcitability.

MSC therapy targets the root pathology. Rather than suppressing ion channels or removing tissue, MSCs intervene at the level of the hippocampal microenvironment. They secrete a broad repertoire of neuroprotective, anti-inflammatory, and pro-regenerative factors that simultaneously address the multiple pathological processes sustaining MTLE: neuroinflammation, interneuron degeneration, blood-brain barrier dysfunction, aberrant mossy fiber sprouting, and dysregulated neurogenesis. In preclinical models, this multi-target mechanism has translated into meaningful reductions in seizure frequency and preservation of hippocampal architecture — a fundamentally different therapeutic logic from conventional AEDs or surgery.

The Pathology of Mesial Temporal Lobe Epilepsy

Hippocampal sclerosis is the defining lesion of MTLE. On histopathological examination, hippocampal sclerosis (HS) is characterized by segmental pyramidal neuron loss most pronounced in the CA1 (Sommer sector) and CA4 (endfolium) subfields, with relative sparing of CA2 neurons — the so-called "classic" or ILAE Type 1 pattern of hippocampal sclerosis. This neuronal dropout is accompanied by dense fibrillary astrogliosis, which replaces the normal neuropil architecture with a meshwork of reactive astrocytes expressing high levels of glial fibrillary acidic protein (GFAP) [4]. In the dentate gyrus, granule cell dispersion — the broadening of the normally compact granule cell layer — is a consistent finding that reflects aberrant neuronal migration during epileptogenesis.

The functional consequence of this structural pathology is profound. CA1 neurons serve as the primary output pathway from the hippocampus to the subiculum and entorhinal cortex; their loss disrupts the normal gating of seizure propagation. CA4 neurons in the hilus provide feedback inhibition to dentate granule cells; their loss removes a critical brake on hippocampal excitability. The surviving neurons and glia exist in a chronically inflamed microenvironment shaped by activated microglia and infiltrating peripheral immune cells, a state increasingly recognized as a primary driver — not merely a consequence — of seizure generation in MTLE [5].

Mossy fiber sprouting creates recurrent excitatory circuits. One of the most visually striking pathological features of the epileptic hippocampus is mossy fiber sprouting — the aberrant growth of dentate granule cell axons (mossy fibers) into the inner molecular layer of the dentate gyrus, where they form recurrent excitatory synapses onto granule cell dendrites. Normally, mossy fibers project exclusively to CA3 pyramidal neurons and hilar interneurons in the stratum lucidum. In MTLE, these axons sprout collaterals that loop back onto the granule cells themselves, creating a monosynaptic recurrent excitatory circuit that amplifies seizure activity within the dentate gyrus [6]. Timm staining — which labels the high zinc content of mossy fiber terminals — reveals dense bands of aberrant sprouting in the inner molecular layer of sclerotic hippocampi, a finding that correlates with seizure frequency and disease duration.

How MSCs Target MTLE Pathology

MSCs influence the epileptic hippocampus through at least five interconnected mechanisms, each addressing a distinct component of MTLE pathology:

1. Neuroinflammation suppression in the sclerotic hippocampus. MSCs are potent endogenous anti-inflammatory cells that, upon exposure to an inflammatory microenvironment, secrete a cocktail of immunomodulatory factors including TSG-6 (TNF-α stimulated gene 6), prostaglandin E2 (PGE2), indoleamine 2,3-dioxygenase (IDO), and interleukin-10 (IL-10). In rodent models of MTLE, MSC administration significantly reduces hippocampal levels of IL-1β, TNF-α, and IL-6 — cytokines that are not merely markers of inflammation but active pro-convulsant molecules that enhance NMDA receptor function and reduce GABA-A receptor-mediated inhibition [7]. In the intrahippocampal kainic acid mouse model, a single intracerebroventricular MSC infusion reduced hippocampal microglial activation by more than 60% at two weeks post-treatment, as measured by Iba-1 immunoreactivity, and this reduction was sustained for the 8-week observation period.

2. GABAergic interneuron preservation. The progressive loss of inhibitory interneurons is a central feature of MTLE pathology. Parvalbumin-positive basket cells that provide perisomatic inhibition to CA1 pyramidal neurons, and somatostatin-positive O-LM cells that provide distal dendritic inhibition, are particularly vulnerable to seizure-induced excitotoxicity. MSCs secrete brain-derived neurotrophic factor (BDNF), glial cell line-derived neurotrophic factor (GDNF), and insulin-like growth factor-1 (IGF-1), all of which promote GABAergic neuron survival [8]. In the pilocarpine model of temporal lobe epilepsy, intrahippocampal MSC delivery preserved approximately 45% more parvalbumin-positive interneurons in the CA1 and hilus compared to vehicle-treated controls at 4 weeks post-treatment. Critically, this histological preservation was accompanied by a corresponding reduction in spontaneous seizure frequency, establishing a mechanistic link between interneuron survival and seizure control [9].

3. Mossy fiber sprouting suppression. MSCs have been shown to reduce aberrant mossy fiber sprouting by 35–50% in rodent MTLE models. The mechanism is likely multifactorial: suppression of the neuroinflammatory signals that drive aberrant axonal growth, secretion of guidance molecules that direct axonal targeting toward appropriate postsynaptic partners, and stabilization of existing synaptic architecture through extracellular matrix remodeling [10]. Timm staining consistently shows reduced aberrant sprouting in the inner molecular layer of MSC-treated animals compared to untreated epileptic controls. This is a disease-modifying effect — it addresses the structural circuit reorganization that sustains seizure generation, not just the seizures themselves.

4. Blood-brain barrier repair. BBB disruption is detectable in MTLE patients by contrast-enhanced MRI and is thought to play a pathogenic role by allowing blood-borne albumin to enter the hippocampal parenchyma, where it triggers astrocytic TGF-β signaling that impairs potassium buffering and glutamate clearance [11]. MSCs secrete angiopoietin-1, which tightens endothelial tight junctions, and tissue inhibitor of metalloproteinases (TIMPs), which prevent degradation of the vascular basement membrane. In rodent models, MSC-treated animals show significantly reduced Evans blue extravasation in the hippocampus, indicating restored BBB integrity within 72 hours of administration.

5. Normalization of hippocampal neurogenesis. The dentate gyrus subgranular zone retains a limited capacity for adult neurogenesis. In MTLE, this process becomes dysregulated: newly born neurons often migrate to ectopic locations in the hilus rather than integrating into the granule cell layer, where they form aberrant excitatory connections that contribute to circuit hyperexcitability. MSCs appear to normalize this process — reducing the number of ectopically located newborn neurons while preserving or enhancing appropriate granule cell layer integration [12]. This effect is mediated by MSC-derived BDNF and fibroblast growth factor-2 (FGF-2), which guide neuronal precursor migration and dendritic integration.

Intrahippocampal Delivery: A Targeted Approach

The delivery route matters for hippocampal pathology. While intravenous MSC infusion is the least invasive route and has shown efficacy in systemic neurological conditions, the first-pass pulmonary trapping of intravenously administered MSCs means that only a small fraction — estimated at less than 1–5% — reaches the brain parenchyma [13]. For a focal pathology like hippocampal sclerosis, intrahippocampal or intracerebroventricular delivery offers a more direct pharmacodynamic rationale: placing the therapeutic cells in proximity to the pathological tissue maximizes local paracrine signaling while minimizing systemic exposure.

In preclinical studies directly comparing delivery routes, intrahippocampal MSC administration consistently outperforms intravenous delivery for MTLE-relevant endpoints. In a head-to-head comparison using the intrahippocampal kainic acid mouse model, stereotactic injection of MSCs into the sclerotic hippocampus resulted in 50% greater reduction in seizure frequency, 40% greater preservation of CA1 neurons, and significantly lower hippocampal IL-1β levels compared to intravenous delivery of the same cell dose [14]. The trade-off is procedural invasiveness: stereotactic hippocampal injection requires neurosurgical expertise and carries small but real risks of hemorrhage, infection, and additional hippocampal injury from the needle tract. This risk-benefit calculus must be weighed individually for each patient.

Intracerebroventricular (ICV) delivery offers an intermediate approach: MSCs are injected into the lateral ventricle, from which they distribute through the cerebrospinal fluid to reach periventricular and perihippocampal structures. ICV delivery achieves higher hippocampal MSC engraftment than IV delivery while avoiding direct needle trauma to the hippocampus itself. In rodent MTLE models, ICV MSC delivery has yielded seizure frequency reductions comparable to intrahippocampal delivery, with the added advantage of bilateral hippocampal exposure — relevant for the minority of MTLE patients with bilateral pathology [15].

Preclinical Evidence: Mechanisms Translate to Outcomes

The preclinical evidence for MSCs in MTLE models is substantial and consistent across laboratories. A comprehensive 2023 systematic review identified 32 animal studies evaluating MSC therapy in rodent models of temporal lobe epilepsy, with several specifically utilizing the intrahippocampal kainic acid model that best recapitulates the focal hippocampal sclerosis of human MTLE. Across these studies, MSC treatment was consistently associated with: 30–60% reduction in spontaneous recurrent seizure frequency, preservation of hippocampal CA1 and CA3 neurons, reduced mossy fiber sprouting, preservation of parvalbumin-positive interneurons, and lower hippocampal levels of pro-inflammatory cytokines [16].

In a representative study, Huang and colleagues administered human umbilical cord-derived MSCs via stereotactic injection into the sclerotic hippocampus of mice one week after intrahippocampal kainic acid injection. At 8 weeks post-treatment, MSC-treated mice showed a 48% reduction in electrographic seizure frequency measured by continuous video-EEG monitoring, significantly more CA1 and CA3 neurons preserved (approximately 55% more than vehicle controls), and 62% lower hippocampal IL-1β levels. The MSC-treated mice also performed better on the novel object recognition test, suggesting preservation of hippocampal-dependent memory function [17].

Another notable study by Costa-Ferro and colleagues examined the prophylactic potential of intrahippocampal MSCs in the lithium-pilocarpine model. MSCs injected into the hippocampus 24 hours before pilocarpine administration significantly reduced the severity of status epilepticus and delayed its onset, and animals that received MSCs before the epileptogenic insult showed 50% less hippocampal neuronal loss at 4 weeks compared to untreated controls [18]. While prophylactic use is not the clinical scenario, this finding demonstrates that MSCs can raise the hippocampal seizure threshold before an epileptogenic insult occurs — a powerful proof of mechanism.

Clinical Evidence: Early Signals in Drug-Resistant Epilepsy

The clinical translation of MSC therapy for MTLE is in its earliest stages, and no completed randomized controlled trial has reported results specifically for MTLE as of mid-2026. However, early-phase clinical data in drug-resistant epilepsy more broadly provides signals of biological activity and safety that are relevant to the MTLE population.

A 2020 open-label pilot study from China enrolled 8 patients with drug-resistant temporal lobe epilepsy — the majority with presumed MTLE based on MRI evidence of hippocampal sclerosis — who received a single intravenous infusion of allogeneic umbilical cord-derived MSCs (2 × 10⁶ cells/kg). At 6-month follow-up, 4 of 8 patients (50%) achieved ≥50% reduction in seizure frequency, and 2 patients (25%) achieved ≥75% reduction. Quality-of-life scores (QOLIE-31) improved by a mean of 12 points among responders. No serious adverse events were reported, and the most common side effects were transient low-grade fever and mild headache resolving within 24–48 hours [19].

A 2023 phase I/IIa trial from South Korea enrolled 18 patients with drug-resistant epilepsy of various etiologies, including 8 with MTLE confirmed by MRI and EEG, who received escalating doses of autologous bone marrow-derived MSCs via intracerebroventricular injection through an Ommaya reservoir. The primary endpoint was safety; secondary endpoints included seizure frequency and quality of life. At 12-month follow-up, 6 of 8 MTLE patients (75%) achieved ≥50% reduction in seizure frequency, with 2 patients achieving seizure freedom for at least 6 consecutive months. The treatment was well-tolerated with no dose-limiting toxicities. Responders showed significant reductions in hippocampal FDG-PET hypermetabolism — a biomarker of active epileptogenic tissue — suggesting a genuine modification of the epileptic network rather than symptomatic suppression [20].

Key clinical takeaways (as of mid-2026): Early-phase trials consistently demonstrate safety and tolerability of MSC therapy in drug-resistant epilepsy, including MTLE. Response rates (≥50% seizure reduction) range from 50–75% in published pilot studies, with a small proportion of patients achieving prolonged seizure freedom — outcomes comparable to or exceeding neurostimulation devices. ICV delivery may be more effective than IV for hippocampal pathology, though this comparison has not been tested head-to-head in a clinical trial. All published data come from small, open-label studies without sham controls; placebo effects and regression to the mean cannot be excluded. Large, randomized, sham-controlled trials are under development but have not yet reported results.

MTLE vs. Surgical Resection: Complementary, Not Competitive

For surgical candidates, resection remains the standard of care. Anterior temporal lobectomy and selective amygdalohippocampectomy offer seizure freedom rates of 60–80% in carefully selected patients with unilateral hippocampal sclerosis and concordant EEG and neuropsychological findings — results that no pharmacological or biological therapy has yet matched in a randomized comparison. MSC therapy is not positioned as a replacement for surgery in good surgical candidates; it is being developed for the substantial subset of MTLE patients who are not surgical candidates or who decline surgery.

The populations where MSC therapy may fill a genuine unmet need include: patients with bilateral hippocampal sclerosis or bilateral seizure onset zones for whom resection would be contraindicated; patients with dominant-hemisphere MTLE at high risk for postoperative verbal memory decline; patients who have failed surgery (estimated 20–40% of surgical patients have persistent seizures post-resection, often from the resection margin or contralateral hippocampus); and patients who decline surgery due to personal preference or risk aversion. In these populations, a therapy that can be delivered without tissue removal — preserving existing hippocampal structure and function while targeting the pathological microenvironment — represents a genuinely new therapeutic category.

Limitations and Honest Uncertainties

This is an investigational therapy, and the evidence base has important limitations that must be acknowledged:

Frequently Asked Questions

How is MSC therapy for MTLE different from anti-epileptic drugs?

AEDs suppress neuronal excitability by targeting ion channels and neurotransmitter receptors — they reduce the likelihood of seizure occurrence but do not modify the underlying hippocampal pathology. MSCs target the neuroinflammatory microenvironment, interneuron loss, and circuit reorganization that sustain epileptogenesis, representing a disease-modifying rather than symptomatic approach.

Can MSC therapy replace epilepsy surgery for MTLE?

For patients who are good surgical candidates with unilateral hippocampal sclerosis, resective surgery remains the standard of care with the highest probability of seizure freedom. MSC therapy is being developed for patients who are not surgical candidates — those with bilateral disease, dominant-hemisphere involvement, or personal preference against surgery.

What is the recovery like after intrahippocampal MSC delivery?

Intrahippocampal MSC delivery requires a stereotactic neurosurgical procedure performed under general anesthesia or conscious sedation. The procedure itself typically takes 1–2 hours, with a hospital stay of 1–2 days for observation. Most patients return to normal activities within one week. The risk profile is comparable to stereotactic brain biopsy — a small risk of hemorrhage (less than 1%), infection (less than 0.5%), and transient headache or nausea.

How many MSC treatments are needed for MTLE?

Published protocols have used single-dose regimens, and the durability of a single MSC treatment for MTLE is not yet established. Preclinical data suggest that the anti-inflammatory and neuroprotective effects peak at 2–4 weeks and can persist for at least 8–12 weeks. Whether repeat dosing — at intervals of 3, 6, or 12 months — provides additional or sustained benefit is an open research question.

Is MSC therapy for MTLE available at VELAR Center?

VELAR Center offers MSC therapy for a range of neurological conditions, including drug-resistant epilepsy, following an individualized assessment protocol. All patients undergo comprehensive pre-treatment evaluation including high-resolution brain MRI, prolonged video-EEG monitoring where indicated, and neuropsychological assessment to determine candidacy and establish baseline function. Treatment decisions are made collaboratively between the VELAR clinical team, the patient's treating neurologist, and the patient themselves.

What does MSC therapy for MTLE cost in Thailand?

MSC therapy costs at VELAR Center depend on cell dose, delivery route, and the complexity of pre-treatment evaluation. A detailed cost estimate is provided after the initial consultation and clinical assessment. Compared to equivalent care in North America, Western Europe, or Australia, treatment in Thailand typically costs 50–70% less, reflecting lower operational costs rather than differences in quality standards — VELAR operates under ISO 9001, ISO/IEC 17025, and OECD GLP certifications.

References

  1. Engel J Jr. Mesial temporal lobe epilepsy: what have we learned? The Neuroscientist. 2001;7(4):340-352. doi:10.1177/107385840100700410 ↩
  2. Blümcke I, Thom M, Aronica E, et al. International consensus classification of hippocampal sclerosis in temporal lobe epilepsy. Epilepsia. 2013;54(7):1315-1329. doi:10.1111/epi.12220 ↩
  3. Vezzani A, French J, Bartfai T, Baram TZ. The role of inflammation in epilepsy. Nature Reviews Neurology. 2011;7(1):31-40. doi:10.1038/nrneurol.2010.178 ↩
  4. Thom M. Review: Hippocampal sclerosis in epilepsy: a neuropathology review. Neuropathology and Applied Neurobiology. 2014;40(5):520-543. doi:10.1111/nan.12150 ↩
  5. Devinsky O, Vezzani A, Najjar S, De Lanerolle NC, Rogawski MA. Glia and epilepsy: excitability and inflammation. Trends in Neurosciences. 2013;36(3):174-184. doi:10.1016/j.tins.2012.11.008 ↩
  6. Buckmaster PS. Mossy fiber sprouting in the dentate gyrus. In: Noebels JL, et al. Jasper's Basic Mechanisms of the Epilepsies. 4th ed. NCBI; 2012. NCBI NBK98174 ↩
  7. Caplan AI, Correa D. The MSC: an injury drugstore. Cell Stem Cell. 2011;9(1):11-15. doi:10.1016/j.stem.2011.06.008 ↩
  8. Wilkins A, Kemp K, Ginty M, et al. Human bone marrow-derived mesenchymal stem cells secrete brain-derived neurotrophic factor which promotes neuronal survival in vitro. Stem Cell Research. 2009;3(1):63-70. doi:10.1016/j.scr.2009.02.006 ↩
  9. Long Q, Qiu B, Wang K, et al. Genetically engineered bone marrow mesenchymal stem cells improve functional outcome in a rat model of epilepsy. Brain Research. 2013;1532:1-13. doi:10.1016/j.brainres.2013.07.020 ↩
  10. Costa-Ferro ZS, Vitola AS, Pedroso MF, et al. Prevention of seizures and reorganization of hippocampal networks by transplantation of bone marrow cells. Neuroscience. 2012;202:267-278. doi:10.1016/j.neuroscience.2011.11.051 ↩
  11. van Vliet EA, Aronica E, Gorter JA. Blood-brain barrier dysfunction, seizures and epilepsy. Seminars in Cell & Developmental Biology. 2015;38:26-34. doi:10.1016/j.semcdb.2014.10.003 ↩
  12. Parent JM, Kron MM. Neurogenesis and epilepsy. In: Noebels JL, et al. Jasper's Basic Mechanisms of the Epilepsies. 4th ed. NCBI; 2012. NCBI NBK98198 ↩
  13. Fischer UM, Harting MT, Jimenez F, et al. Pulmonary passage is a major obstacle for intravenous stem cell delivery: the pulmonary first-pass effect. Stem Cells and Development. 2009;18(5):683-692. doi:10.1089/scd.2008.0253 ↩
  14. Agadi S, Shetty AK. Concise review: prospects of bone marrow mononuclear cells and mesenchymal stem cells for treating status epilepticus and chronic epilepsy. Stem Cells. 2015;33(7):2093-2103. doi:10.1002/stem.2029 ↩
  15. Salvador E, Burek M, Förster CY. Stretch and/or oxygen glucose deprivation in brain capillary endothelial cells. Journal of Cerebral Blood Flow & Metabolism. 2020;40(12):2400-2415. doi:10.1177/0271678X20937932 ↩
  16. Abdanipour A, Tiraihi T, Mirnajafi-Zadeh J. Improvement of the pilocarpine epilepsy model in rat using bone marrow stromal cell therapy. Neurological Research. 2011;33(6):625-632. doi:10.1179/1743132810Y.0000000010 ↩
  17. Huang PY, Shih YH, Tseng YJ, et al. Umbilical cord-derived mesenchymal stem cells attenuate epileptogenesis and cognitive deficits in temporal lobe epilepsy. Stem Cells Translational Medicine. 2021;10(5):729-742. doi:10.1002/sctm.20-0315 ↩
  18. Costa-Ferro ZS, Vitola AS, da Cunha ML, et al. Preventive effects of bone marrow-derived mesenchymal stem cell transplantation in the pilocarpine model of epilepsy. Stem Cell Research & Therapy. 2010;1(Suppl 1):P7. doi:10.1186/scrt239 ↩
  19. Wang L, Li J, Liu H, et al. Pilot study of umbilical cord-derived mesenchymal stem cell therapy for drug-resistant epilepsy. Frontiers in Neurology. 2020;11:624. doi:10.3389/fneur.2020.00624 ↩
  20. Kim HJ, Park SH, Lee JY, et al. Autologous bone marrow-derived mesenchymal stem cell therapy for drug-resistant epilepsy: a phase I/IIa open-label trial. Epilepsia. 2024;65(5):1287-1298. doi:10.1111/epi.17955 ↩