Dados do Trabalho

Título

Hippocampal T2 relaxation correlates with the levels of functional astroglial proteins in patients with mesial temporal lobe epilepsy.

Resumo

Introduction: Hippocampal sclerosis is the most frequent etiology of pharmacoresistant epilepsy in adults. For the surgical treatment of these patients, an extensive pre-surgical evaluation is performed, including electroencephalogram (EEG) and video-EEG analysis, neuropsychological tests, psychiatric evaluation, and magnetic resonance imaging (MRI). The MRI evaluation in most patients with hippocampal sclerosis shows a reduced volume and increased T2 signal, although in some only a T2 increase is seen. The presurgical pathological evaluation frequently reveals severe neuron loss and astrogliosis throughout the hippocampal subfields. Astrocytes, crucial in controlling tissue excitability and oxidative stress, have several specific proteins for this purpose. Among them, metallothioneins I and II (MT-I/II), astroglial proteins released in the extracellular space, control zinc homeostasis. This metal acts as an important modulator of glutamatergic neurotransmission in the hippocampus, which concentrates zinc in glutamatergic vesicles of granule cells and pyramidal neurons CA3 and CA1. With seizures, excessive zinc release contributes to glutamatergic excitotoxicity and oxidative stress. Besides the increase in glutamate release during seizures, in patients with hippocampal sclerosis, there is a reduction in the astrocytic glutamate transporter EAAT2, increasing glutamate remotion from the synaptic cleft. Of note, more than 90% of the glutamate clearance is done by the EAAT2 transporter. Thus, changes in the pattern of glial protein expression can directly contribute to the excitability of the region and, therefore, to the genesis of seizures. Astrocytes are also key elements for presurgical evaluation, as qualitative MRI inspection indicates a link between the T2 signal increase seen in hippocampal sclerosis and the presence of astrogliosis. However, quantitative studies failed to show strong correlations between the intensity of astrogliosis and the quantification of T2 relaxation. We hypothesize that functional glial markers such as MT-I/II and EAAT2 have a stronger correlation with the T2 signal than the expression of the structural protein GFAP, the classical marker of tissue astrogliosis in grey matter structures. Objective: To compare the correlations between T2 relaxation and the proteins MT-I/II and EAAT2 with the correlations with the structural astroglial protein GFAP in patients with hippocampal sclerosis. Methods: Patients with temporal lobe epilepsy and hippocampal sclerosis (n=42) who underwent axial multiecho T2 MRI scan at a 3 Tesla machine (TEs = 20, 40, 60, 80, and 100 ms; TR = 3000 ms; voxel = 1x1x3 mm) and epilepsy surgery, were selected. Paraffin-embedded hippocampal sections were submitted to immunohistochemistry for the detection of EAAT2, MT-I/II, and GFAP proteins. Whole hippocampus T2 relaxation time and immunostaining intensity for the selected glial proteins in all hippocampal subfields were quantified. Patients were divided in cases with increased T2 relaxation and those with T2 relaxation within the range seen in a sex and age matched population without neurological diseases already published. Mann-Whitney’s U test was employed to compare the immunostaining intensity of glial proteins between these groups, Spearman’s correlation test and stepwise multiple linear regression were employed to investigate the association between glial proteins and T2 relaxation time. Results: Among our cases, 20 presented T2 relaxation within normal range (107.4 ± 6.9 ms) and 22 had increased hippocampal T2 relaxation (132.6 ± 13.4 ms). Cases with increased T2 relaxation had higher GFAP staining in CA2 (p = 0.03) and a trend towards reduced EAAT2 staining in CA4 (p = 0.05) and increased MT-I/II expression in CA2 (p = 0.09). Hippocampal T2 relaxation time correlates positively with the immunostaining for MT-I/II (​r = 0.507, p = 0.005), and GFAP (r = 0.371, p = 0.043) in CA2 and negatively with EAAT2 in CA4 (r = -0.373, p = 0.016). The immunostaining for MT-I/II in the subiculum presented a trend towards a positive correlation with T2 relaxation time (r = 0.306, p = 0,055). A stepwise linear regression indicated that the combination of MT-I/II immunostaining in all CA subfields, EAAT2 immunostaining in the granule cell layer, CA4-CA2, and the subiculum, and GFAP immunostaining in CA4 explain 77% of the hippocampal T2 relaxation (r = 0.88, p < 0.0001). Conclusion: Despite the common view that the increase in T2 signal observed in TLE is related to gliosis, few studies have attempted to correlate astrogliosis degree in the hippocampal regions with the T2 quantification. Most found no direct relation between structural markers of gliosis and T2 signal, or only mild associations. Our study revealed that proteins related to astroglial function such as MT-I/II and EAAT2 reflect better the effects of astrogliosis on T2 signal changes than the expression of structural glial proteins such as the classical marker GFAP. Considering only the subfields that contribute more for the hippocampal volume (i.e., CA4, CA1, and the subiculum) and, thus, would have a stronger impact on the whole hippocampal T2, only the functional proteins EAAT2 and MT-I/II in CA4 and the subiculum correlated with T2 relaxation, while none of the GFAP staining in these areas correlated with T2 relaxation. Furthermore, among the 10 most significant areas indicated by the stepwise linear regression as correlates to T2 relaxation, staining for GFAP was significant only in CA2, possibly explaining the difficulty of previous studies in correlating GFAP staining/positive cell count with T2 signal. Since hundreds of astroglial proteins are up- or downregulated in the hippocampus of TLE patients, the inclusion of more astrocyte exclusive proteins may improve the correlation between T2 signal and tissue changes seen in hippocampal sclerosis. Additionally, the individual assessment of T2 relaxation in each hippocampal subfield could improve the match between MRI and the pathological changes observed in TLE, as shown in our previous study. Overall, we observed that the expression levels of astroglial proteins related to tissue functionality, represented in this study by EAAT2 and MT-I/II, showed stronger correlations with the T2 relaxation time than the structural marker of astrogliosis GFAP. A better understanding of the glial mechanisms associated with neuronal excitability and MRI changes may contribute not only to understanding the tissue correlates of radiological and electrographic changes observed in the epileptogenic hippocampus, but also to indicate new avenues of treatment.