Mesial Temporal Lobe Epilepsy (MTLE) Drug-Refractoriness Is Associated With P2X7 Receptors Overexpression in the Human Hippocampus and Temporal Neocortex and May Be Predicted by Low Circulating Levels of miR-22

Abstract

Objective: ATP-gated ionotropic P2X7 receptors (P2X7R) actively participate in epilepsy and other neurological disorders. Neocortical nerve terminals of patients with Mesial Temporal Lobe Epilepsy with Hippocampal Sclerosis (MTLE-HS) express higher P2X7R amounts. Overexpression of P2X7R bolsters ATP signals during seizures resulting in glial cell activation, cytokines production, and GABAergic rundown with unrestrained glutamatergic excitation. In a mouse model of status epilepticus, increased expression of P2X7R has been associated with the down-modulation of the non-coding micro RNA, miR-22. MiR levels are stable in biological fluids and normally reflect remote tissue production making them ideal disease biomarkers. Here, we compared P2X7R and miR-22 expression in epileptic brains and in the serum of patients with MTLE-HS, respectively. Methods: Quantitative RT-PCR was used to evaluate the expression of P2X7R in the hippocampus and anterior temporal lobe of 23 patients with MTLE-HS and 10 cadaveric controls. Confocal microscopy and Western blot analysis were performed to assess P2X7R protein amounts. MiR-22 expression was evaluated in cell-free sera of 40 MTLE-HS patients and 48 healthy controls. Results: Nerve terminals of the hippocampus and neocortical temporal lobe of MTLE-HS patients overexpress (p < 0.05) an 85 kDa P2X7R protein whereas the normally occurring 67 kDa receptor protein dominates in the brain of the cadaveric controls. Contrariwise, miR-22 serum levels are diminished (p < 0.001) in MTLE-HS patients compared to age-matched control blood donors, a situation that is more evident in patients requiring multiple (>3) anti-epileptic drug (AED) regimens. Conclusion: Data show that there is an inverse relationship between miR-22 serum levels and P2X7R expression in the hippocampus and neocortex of MTLE-HS patients, which implies that measuring serum miR-22 may be a clinical surrogate of P2X7R brain expression in the MTLE-HS. Moreover, the high area under the ROC curve (0.777; 95% CI 0.629-0.925; p = 0.001) suggests that low miR-22 serum levels may be a sensitive predictor of poor response to AEDs among MTLE-HS patients. Results also anticipate that targeting the miR-22/P2X7R axis may be a good strategy to develop newer AEDs.

Keywords: P2X7 purinoceptor; hippocampus; mesotemporal lobe epilepsy; miR-22; microRNAs; refractory epilepsy.

Figure 1

The P2X7R mRNA is overexpressed in the hippocampus and temporal neocortex of drug-refractory MTLE-HS patients submitted to amygdalohippocampectomy compared to non-epileptic cadaveric controls (CTR). qPCR data are expressed as mean ± SD; n numbers inside each bar represent the number of individuals among 10 controls and 23 MTLE-HS patients in which quality assessment of retrieved RNA samples was suitable for quantification. ^*p < 0.05 (two-way ANOVA followed by Holm-Šídák’s multiple comparisons test was used) represents significant differences when compared to control individuals.

Figure 2

Representative confocal micrographs of different sub-regions of the human hippocampus showing that the P2X7R immunoreactivity (green) is higher in the hippocampus of MTLE-HS patients than of non-epileptic cadaveric controls. A negative control resulting from incubation of the DG/CA4 hippocampal region of an MTLE-HS patient with the anti-rabbit secondary antibody without previous addition of the rabbit anti-P2X7R primary antibody (#APR-004) and differential interference contrast (DIC) images are shown for comparison; three confocal micrographs were obtained per individual; three individuals from each group (control and MTLE-HS) were analyzed showing similar results; scale bars = 300 μm.

Figure 3

No significant correlation was observed between the P2X7R expression and the age of epileptic patients (panels B and D) and non-epileptic cadaveric controls (panels A and C), both in the hippocampus (panels A and B) and in the temporal neocortex (panels C and D). Each point represents an individual sample among 10 controls and 23 MTLE-HS patients in which quality assessment of retrieved RNA samples was suitable for quantification. Spearman’s correlation coefficients and significance p values (two-tailed) are shown inside each graph.

Figure 4

Representative confocal micrographs showing that the P2X7R immunoreactivity is located predominantly in nerve terminals, but not in glial cells, of all sub-regions of the hippocampus of MTLE-HS patients. Synaptic nerve terminals were identified with an antibody against the vesicle-associated membrane protein 1 (VAMP-1 or synaptobrevin 1), whereas astrocytes were stained with an antibody against the glial fibrillary acidic protein (GFAP). Note that VAMP-1-positive nerve terminals (red) are endowed with the P2X7R (green; panel A), but no significant co-localization was observed between P2X7R (green) and GFAP (red; panel B); scale bars = 50 μm. Data in panels (C) and (D) correspond to staining overlap and Pearson’s Coefficient (ρ) parameters calculated from three to four confocal micrographs per individual; at least three individuals from each group, control, and MTLE-HS, were analyzed. These parameters were automatically calculated per image with Olympus Fluoview 4.2 Software (Olympus FV1000, Tokyo, Japan) and were used to estimate the co-localization of P2X7R and type-specific cell markers (yellow staining). Overlap between two colors gives values between +1 (total overlap) and 0 (no overlap); the Pearson’s Coefficient (ρ) is a measure of the linear correlation between two variables (stainings), giving values between +1 and −1 inclusive, where 1 is total positive correlation, 0 is no correlation, and −1 is total negative correlation. Higher magnification images show that the P2X7R immunoreactivity also co-localizes with the synaptic vesicle glicoprotein synaptophysin (Synapt), which is one of the most commonly used neuronal cell markers in neuropathology (Panel E), but not with GFAP (Panel F). Nuclei are stained with DAPI; cross-reactivity for the secondary antibodies was tested in control experiments in which primary antibodies were omitted (negative controls).

Figure 5

The P2X7R protein is upregulated in nerve terminals of the hippocampus of drug-refractory MTLE-HS human patients. Panel (A) shows that using our methodology nerve terminals isolated from the hippocampus of MTLE-HS patients exhibit a higher density of the synaptic vesicle marker, synaptophysin (~34 kDa), compared to the astrocytic cell marker, GFAP, whilst the opposite was observed in total hippocampal lysates. In panel (B) are shown representative Western blots of the P2X7R immunoreactivity in total lysates and nerve terminals isolated from the human hippocampus of control individuals and MTLE-HS patients; gels were loaded with 100 μg of protein. Two protein species were recognized by the P2X7R antibody from Alomone (#APR-004, Jerusalem, Israel) corresponding to the naturally occurring 67 kDa receptor isoform and to a higher molecular mass (~85 kDa) P2X7R isotype; the latter is highly enriched in nerve terminals of the hippocampus of MTLE-HS patients compared to non-epileptic controls. Please note that the two bands corresponding to the P2X7R protein disappeared after pre-adsorption of the primary antibody with a control antigen peptide equivalent to the amino-acid residues 576–595 of the intracellular C-terminus of the P2X7R (negative control); β-Actin (38–41 kDa) was used as a reference protein. Panel (C) shows computed data obtained from immunoblot experiments; data are expressed as mean ± SD; each individual sample was processed in duplicate; at least three individuals from each group (control and MTLE-HS) were analyzed.

Figure 6

The inverse relationship between miR-22 serum levels and the P2X7R expression in the hippocampus (A) and temporal neocortex (B) from drug-refractory MTLE-HS patients submitted to amygdalohippocampectomy. Ordinates represent ΔCt variation of P2X7R and miR-22 expression in epileptic patients compared to the corresponding median of non-epileptic controls; 0 represents null variation; positive and negative values correspond to increases and decreases relative to the control population, respectively. Please note that there is a mismatch between miR-22 serum quantifications and P2X7R mRNA determinations in the hippocampus (A, two patients missing) and temporal neocortex (B, one patient missing) because RNA samples were insufficient or did not pass the quality assessment.

Figure 7

Patients with drug-refractory MTLE-HS exhibit lower miR-22 serum levels than the control population (panel A). This difference is exaggerated when patients with poor response to medication (requirement of three or more AEDs) are considered (panel B). Boxes and whiskers (Tukey method) represent pooled data from 48 blood donor controls (CTRL) and 40 MTLE-HS patients (panel A); in panel (B) epileptic patients were subdivided into drug-resistant (≥3 AE Drugs) and non-drug-resistant (<3 AE Drugs). The values plotted individually are outliers according to the Grubbs test. In panel (A), ^***p < 0.001 (unpaired Student’s t-test with Welch correction) represent significant differences compared to the control population; in panel (B), ^****p < 0.0001 (one-way ANOVA with Dunnett’s multiple comparison test) represent significant differences compared to the control population. ns = non-significant. In panels (C) and (D), represented are the receiver-operator characteristics (ROC) curves of miR-22 serum levels considering the whole MTLE-HS patients’ population (panel C) or only patients from this cohort taking more than three AEDs (panel D). Please note that the area under de curve (AUC) increases considering MTLE-HS patients with poorer response to medication.