‘Hippocampal innate inflammatory gliosis only’ – the future role of surgery in a novel temporal lobe epilepsy syndrome

Epilepsy is one of the most common neurological conditions affecting more than 50 million people worldwide (https://www.who.int/news-room/fact-sheets/detail/epilepsy). Despite numerous antiseizure medications (ASM), approximately 30% of all patients will develop drug-resistant epilepsy (DRE). DRE leads to devastating health and socio-economic consequences (Luoni et al., 2011).

Epilepsy surgery is a well-accepted treatment modality with proven success in drug-resistant mesial temporal lobe epilepsy (MTLE) (Wiebe et al., 2001), where hippocampal sclerosis is one of the most common causes of epilepsy. The work of Wiebe et al. is indeed one of the most important landmark studies setting indisputable evidence in favor of surgery for MTLE compared to medical treatment. Even so, real-world evidence, which has been accumulated over the last 20 years, suggests that resective surgery for MTLE achieves seizure freedom in approximately 60% of the cases, while the remaining patients continue to have seizures after surgery (Grote et al., 2016). Although both preoperative diagnostics and surgical techniques have evolved over the years, success rates after MTLE surgery have remained stable, raising the question, whether the underlying pathogenesis of MTLE is the same for all patients.

I²GO – novel MTLE disease entity limiting the surgical success: Following this line, Grote et al. (2023) aimed to understand the cause of failure after resective epilepsy surgery in MTLE. They assumed that the reason for this is a new MTLE entity hallmarked by the neuropathological finding of ‘gliosis only’ and coined the term of innate inflammatory gliosis only - I²GO. In their work, they found in a large cohort of 627 patients that a marked number of cases with hippocampus ‘gliosis only’, who underwent epilepsy surgery (either amygdalohippocampectomy or anterior temporal lobectomy) failed to achieve seizure freedom and showed more severe postoperative neuropsychological deterioration compared to patients with hippocampal sclerosis. Following this observation, they assumed that the underlying damage causing epilepsy in patients with ‘gliosis only’ might be less localized to a unilateral hippocampus, but rather more vastly spread. To prove this assumption, Grote et al. (2023) looked for clinical, neuropathological, and transcriptional differences between MTLE caused by hippocampal sclerosis or ‘gliosis only’. First, the neuropathological presentation of hippocampal gliosis (HS) and ‘gliosis only’ was completely different since ‘gliosis only’ showed excessive astrogliosis without marked neuronal loss in contrast to HS. The assumption that ‘gliosis only’ might represent a “precursor” state of HS was rebutted by the fact that ‘gliosis only’ patients were markedly younger and developed epilepsy markedly earlier compared to their HS counterparts.

Next, they used integrative transcriptomic profiling to reveal unique gene expression signatures in ‘gliosis only’ hallmarked by up-regulated genes from proteins with inflammatory relevant functions and complement factors. Using an unsupervised clustering approach, they showed that the transcriptional profile of astrocytes from HS clustered within a non-inflammatory state known from other diseases like stroke. On the contrary, the transcriptional profile of astrocytes state from ‘gliosis only’ was clustered together with astrocytes from other inflammatory diseases like multiple sclerosis. Finally, they showed that characteristic MRI alterations were absent or less obvious in ‘gliosis only’ patients and even affected the contralateral side after performing both visual inspection and volumetric analysis.

Perhaps the most important findings were the marked differences in postoperative seizure and neuropsychological outcomes. In detail, a marked number of patients with I²GO failed to become seizure-free after surgery, while this was associated with a greater risk for cognitive decline. Considering all findings, the authors hypothesized that I²GO presents a particular TLE syndrome characterized by an innate inflammatory pathogenesis, thus being more vastly spread and difficult to amend by a localized surgical resection.

Finally, the authors concluded that younger patients, who present with MTLE but show rather intact neuropsychological function, no clear MRI signs of HS, or even involvement of the contralateral side should be carefully considered for invasive diagnostic (possibly accompanied by a simultaneous biopsy) before recommendation for epilepsy surgery. In summary, this work suggests for the first time that damage caused by inflammation is less focal and therefore difficult to treat with local surgical resection.

MTLE within the landscape of inflammation: The evidence that chronic inflammation is the driver of at least some types of DRE epilepsy is increasing (Campos-Bedolla et al., 2022). A recent work by Tröscher et al. (2021) revealed that T cells can infiltrate the hippocampus in TLE depending on the underlying pathology and both CD3⁺ T cells and CD8⁺ cytotoxic T cells correlate with the extent of neuronal loss. Furthermore, Tröscher et al. (2022) showed that in MTLE with GAD antibodies, neuronal death is caused by the T cells. Similar findings were reported by Tezer et al., who reported a prominent CD3⁺ and CD8⁺ T-cell infiltration in patients with HS (ILAE Type 1 HS). Taken together it seems feasible that T-cell-mediated processes may cause neuronal damage, which can disturb neuronal networks and thus play a role in epileptogenesis. Of note, those studies investigated patients with TLE due to HS which is characterized by significant neuronal loss (Blumcke et al., 2013). The involvement of the immune system in HS with a largely preserved neuronal density as in ‘gliosis only’ may be linked to reactively transformed astrocytes showing an inflammatory signature (Blumcke et al., 2013). A recent laboratory work by Henning et al. (2023) suggested the role of reactive microglia to contribute to astrocytic dysfunction and acute seizures without affecting neurodegeneration. They showed an activation of the innate immune system over time, which was transiently observed in the contralateral hippocampus as well. This resembles to some extent the clinical presentation of I²GO but how the innate immune system activation takes place remains an enigmatic question for further investigation. The complexity of the underlying pathophysiology involving the innate immune system has been further highlighted by the work of Waltl et al. (2018) who showed that viral-induced damage of the hippocampus could induce seizures and the depletion of the macrophages can reduce the seizure rate but did not prevent further hippocampal damage, suggesting that the underlying mechanisms are still unknown.

Surgery and inflammation in epileptogenesis – “The king is dead, long live the king!”: More than 20 years after the work of Wiebe et al. (2001), surgery for MTLE plays an important role in the treatment of ‘classical’ HS. This resembles however only one side of the medal. The other side can be referred to as “the quest” for understanding why surgery fails, and recent evidence suggests a crucial role of inflammation in this matter.

I²GO and other inflammatory syndromes causing MTLE might develop more vastly spread inflammatory damage, which is difficult to be controlled by focal surgical resections. Since the number of diagnosed HS decreases, while the number of MRI-negative MTLE increases (Muhlhofer et al., 2017), it is reasonable to discuss the future role of surgery in the treatment of temporal lobe epilepsy.

Resective surgery seems to be a less suitable treatment option for MTLE caused by inflammation, which moves the focus to other surgical techniques (Figure 1). One important future perspective, which has already become an indispensable part of epilepsy treatment, is neuromodulation. Neuromodulation techniques like vagal nerve stimulation and deep brain stimulation of the anterior nucleus of the thalamus (ANT-DBS) develop treatment effects, which influence the epileptogenic potential of the whole brain or both hippocampi simultaneously as in case of responsive neurostimulation (Ryvlin et al., 2021; Figure 1). These specific characteristics as well as the recent development towards machine-learning based algorithms for close-loop systems, which can both sense and stimulate in a patient-specific pattern, allow the vision that surgery, including both resections and neuromodulation, will continue to play an indispensable part in the treatment of MTLE.

Figure 1.

Schematic image of different underlying pathologies in mesial temporal lobe epilepsy.

Image (A) suggests that inflammatory processes affect larger and less focal epileptogenic zones, which reduces the success rate of resective surgery but may present a good opportunity for neuromodulation procedures. Image (B) shows the classic pattern of a unilateral pathology focused to the mesial structures with a favorable outcome after resective epilepsy surgery. This figure was created with BioRender.com.