Introduction

Under physiological conditions, proinflammatory cytokines and their receptors at low levels have been claimed to be important in neuronal development, controlling neurite outgrowth, neurogenesis, and cell survival [1]. However, neuroinflammation is a central nervous system (CNS) response to various insults, such as tissue damage, infection, autoimmune conditions, stress, and seizures [2]. This response involves the synthesis and release of proinflammatory molecules from glial cells, neurons, and endothelial cells (neurovascular unity) of the blood-brain barrier (BBB) [3, 4]. In addition, neuroinflammation appears to be an important component in epileptogenesis, reflecting complex crosstalk between microglia, astrocytes, and neurons [5]. Inflammation induces activation of the cerebrovasculature, affecting BBB integrity and increasing its permeability, as reported in several neurodegenerative diseases [6,7,8]. Activated leukocytes and CNS-resident cells secrete proinflammatory cytokines such as interleukin-1 beta (IL-1β), tumor necrosis factor-alpha (TNF-α), and interferon-gamma (IFN-γ), which can increase BBB permeability [3]. Microvasculature dysfunction and BBB disruption can contribute to a loop between seizure activity and neuroinflammatory responses [9,10,11].

Several studies have demonstrated that several proinflammatory pathways are activated in patients with drug-resistant epilepsy [2, 5]. Some proinflammatory molecules were reported to contribute to the mechanisms of seizure generation, drug resistance, and epileptogenesis [4, 12]. Expression of the proinflammatory mediators IL-1β and TNF-α and inducible nitric oxide synthase (iNOS), a marker of oxidative stress, has been observed in brain tissue obtained from patients with drug-resistant epilepsy [2, 13]. However, these factors could be expressed in the microvasculature of patients with temporal lobe epilepsy (TLE), which could be interesting and assist in a better understanding of the pathogenesis of epilepsy.

In experimental models, during epileptogenesis, IL-1β and IL-1R1 were expressed in both perivascular astrocytes and endothelial cells of the BBB, often in association with BBB leakage and neuronal damage [4, 7]. However, there is no evidence of the expression of proinflammatory mediators and oxidative stress in the BBB microvasculature of patients with TLE. The aim of this study was to evaluate expression levels of the IL-1β, TNF-α, TNF-R1, and iNOS proteins in the neocortical microvasculature of patients with TLE. Additionally, the obtained data were correlated with clinical parameters and analyzed the sex differences between groups. For the first time, this study demonstrates the expression of proinflammatory and oxidative stress mediators in the BBB microvasculature of patients with TLE and highlights the role of inflammation in this neurological disorder.

Materials and methods

Patient criteria and surgical samples

Samples were obtained from the temporal neocortex (T2 or T3) of 15 patients submitted to epilepsy surgery and with different diagnoses: 9 patients with TLE and mesial sclerosis, 2 patients with TLE secondary to tumor, 2 patients with cortical dysplasia, 1 patient with arteriovenous malformation, and 1 patient with dual pathology (cortical dysplasia and tumor). The clinical characteristics of the TLE patients are summarized in Table 1.

Table 1 Demographic and clinical data of TLE patients and autopsies
Full size table

Presurgical evaluation consisted of neurological assessment, electroencephalogram (EEG) and video-EEG recordings, neuropsychological and neuropsychiatric parameter assessment, and magnetic resonance imaging (MRI), the findings of which were in concordance with those of EEG recordings. The presurgical evaluation and epilepsy surgery were performed at the National Institute of Neurology and Neurosurgery “Manuel Velasco Suárez” (INNNMVS) in Mexico City. The epilepsy surgery consisted of unilateral amygdalohippocampectomy with resection of the T2 and T3 gyri. The dissected biopsies were immediately frozen and kept at − 70 °C until processing.

The Forensic Medical Service of Mexico City provided neocortical samples obtained from autopsies (n = 10) of people who died by diverse causes but without apparent clinical data of neurologic disease. The demographic characteristics of autopsies are summarized in Table 1.

All procedures were performed following the ethical principles of the Declaration of Helsinki for human research, and an informed consent form was signed by each patient. In addition, the full research protocol was approved by the Ethics Committee in Research of the INNNMVS (agreement No. CEI/058/16).

Isolation of human neocortex microvessels

Preliminary experiments using rats were carried out to establish the experimental conditions for isolation of the BBB from frozen tissue according to a procedure described previously [14]. Frozen cortical samples were thawed on ice, homogenized with Ringer-Hepes (RH) buffer (150 mM NaCl, 2.2 mM CaCl2, 0.2 mM MgCl2, 5.2 mM KCl, 2.8 mM sucrose, 5 mM Hepes, and 6 mM NaHCO3, pH 7.4), and centrifuged at 1000×g for 15 min at 4 °C. The resulting pellet was suspended in 17.5% dextran-RH buffer and centrifuged at 1500×g for 15 min at 4 °C. The resulting pellet was suspended in 2 mL of RH buffer containing 1% bovine serum albumin (BSA), while the supernatant was transferred to a fresh tube and centrifuged twice more at 1500×g for 15 min each at 4 °C in 17.5% dextran-RH buffer. The pellets obtained from the three centrifugation procedures were pooled, passed through a 100-μm nylon mesh filter, and then passed through a 40-μm nylon mesh filter to strain the microvessels. The microvessels were centrifuged at 1000×g for 10 min at 4 °C. Then, the recovered pellet was washed with RH buffer, centrifuged at 700×g for 5 min at 4 °C, resuspended in RH buffer, and centrifuged at 14,000×g for 2 min at 4 °C. The final pellet (purified microvessels) was frozen in 1 mL of RH buffer at − 70 °C until processing.

Western blotting

Purified microvessel samples were homogenized in RIPA buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, and 0.1% Triton X-100) with protease inhibitor cocktail (Roche Diagnostics GmbH, Germany) in a cold bath at 4 °C. Then, the homogenates were centrifuged at 14,000×g for 10 min at 4 °C, and the supernatant (total protein extract) was immediately collected, aliquoted, and frozen at − 70 °C. The protein concentrations of the extracts were determined according to the Bradford method (Bio-Rad Laboratories, USA) using BSA as an external standard.

Samples were boiled for 5 min at 95 °C in Laemmli buffer (500 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 10% β-mercaptoethanol, and 0.1% bromophenol blue), and samples containing 20 μg of protein were separated in each lane of a 10 and 12% polyacrylamide gel by electrophoresis at 95 V for 30 min and 100 V for 90 min using Tris/glycine/SDS buffer (25 mM Tris, 192 mM glycine and 0.1% SDS, pH 8.3, Bio-Rad Laboratories, USA). The proteins were subsequently electroblotted onto a 0.45-μm nitrocellulose membrane (GE Healthcare Life Science, Germany) at 110 V for 30 min using transfer buffer (25 mM Tris, 250 mM glycine, 0.1% SDS and 20% methanol, pH 8.3). Then, the membranes were incubated for 1 h at 4 °C with a 5% blocking solution (Blot-QuickBlocker, EMD Millipore, USA) in TBS-0.1% Tween (20 mM Tris, 500 mM NaCl, 0.1% Tween 20, pH 7.5). The membranes were then incubated overnight at 4 °C with the following antibodies: mouse monoclonal anti-IL-1β (1:250; cat. sc-52,018, lot I1817, Santa Cruz Biotechnology, USA), rabbit polyclonal anti-TNF-α (1:250; cat. ab66579, lot GR63398-1, Abcam, USA), mouse monoclonal anti-TNF-R1 (1:500; cat. sc-8436, lot G0218, Santa Cruz Biotechnology, USA), rabbit polyclonal anti-iNOS (1:500; cat. ab15323, lot GR67797-1, Abcam, USA), and rabbit polyclonal anti-β-actin (1:5000; cat. ab8227, lot GR2782191, Abcam, USA) as a protein loading control. Next, membranes were incubated with the corresponding secondary antibodies HRP-goat anti-rabbit IgG (1:10000; cat. 926-80011, lot C70207-01, LI-COR Bioscience, USA) and HRP-horse anti-mouse IgG (1:10000; cat. PI-2000, lot ZC1212, Vector Laboratories, USA) for 2 h at 4 °C. Finally, the membranes were incubated in Clarity Western ECL substrate (Bio-Rad Laboratories, USA) at room temperature for 5 min. The chemiluminescent signal was acquired through a C-DiGit Blot Scanner (LI-COR Bioscience, USA) and analyzed with free Image Studio Lite Software v.3.1.4 (LI-COR Bioscience, USA). The data were normalized using β-actin as control protein, resulting in relative expression ratios, with determinations performed in duplicate.

Statistical analysis

Western blotting data were analyzed using Student’s t test with Holm-Sidak post hoc correction. The results obtained are expressed as the mean ± standard deviation (SD). Differences for which p ≤ 0.05 were considered significant. Spearman’s test was used to identify significant correlations between the obtained data and clinical data. Statistical analyses and graph construction were achieved using GraphPad Prism software v.6.01 (GraphPad Software, Inc., USA).

Results

Clinical data

The patients with TLE comprised 4 males and 11 females. The clinical data from the patients were as follows (mean ± SD): age, 29.20 ± 10.15 years; age of seizure onset, 12.55 ± 11.46 years; epilepsy duration, 16.65 ± 13.80 years; and frequency of seizures, 9.60 ± 7.92 per month. The autopsies comprised 5 males and 5 females. The autopsies had an age of 32.50 ± 12.74 years (p > 0.05 when compared with patients) and a postmortem interval of 16.00 ± 2.83 h.

Expression of proinflammatory cytokines

Expression of the proinflammatory mediators IL-1β, TNF-α and its receptor, TNF-R1, was evaluated in microvessels isolated from the neocortex of patients with TLE and autopsy samples. Specific bands corresponding to molecular weights of 35 kDa (IL-1β), 25 kDa (TNF-α), and 55 kDa (TNF-R1) were observed by immunolabeling. These data agree with the information from the antibody manufacturer.

The autopsy samples showed protein expression ratios for IL-1β, TNF-α, and TNF-R1 of 0.35 ± 0.12, 0.48 ± 0.26, and 0.34 ± 0.19, respectively. The microvasculature of patients with TLE showed high IL-1β expression (40%), although it was not significantly different from that of autopsy samples (Fig. 1b). On the other hand, the levels of TNF-α and its receptor, TNF-R1, were higher in the microvasculature of patients with TLE than in the autopsy group (144%, p < 0.01 and 224%, p < 0.01, respectively; Figs. 2b and 3b). Interestingly, in patients with TLE, males showed an increase of IL-1β and TNF-R1 protein level compared with females (p < 0.05; Table 2), suggesting a sex-dependent mechanism in the immune response in TLE.

Fig. 1
figure 1

Protein expression level of IL-1β in the neocortical microvasculature of autopsies and TLE patients. Panel (a) shows representative images of western-blot for IL-1β (35 kDa) and its respective loading control protein β-actin (43 kDa) for both groups. Expression ratio of IL-1β normalized to β-actin is shown in panel (b) for autopsies (n = 10) and TLE samples (n = 15). Data represent the mean ± SD from each sample performed in duplicate. Statistical analysis was conducted by Student’s t test with Holm-Sidak post hoc correction and the analysis showed no significant differences between groups

Full size image
Fig. 2
figure 2

Protein expression level of TNF-α in the neocortical microvasculature of autopsies and TLE patients. Panel (a) shows representative images of western-blot for TNF-α (25 kDa) and its respective loading control protein β-actin (43 kDa) for both groups. Expression ratio of TNF-α normalized to β-actin is shown in panel (b) for autopsies (n = 10) and TLE samples (n = 15). Data represent the mean ± SD from each sample performed in duplicate. Statistical analysis was conducted by Student’s t test with Holm-Sidak post hoc correction and the statistically significant differences are represented by **p < 0.01, compared with autopsies

Full size image
Fig. 3
figure 3

Protein expression level of TNF-R1 in the neocortical microvasculature of autopsies and TLE patients. Panel (a) shows representative images of western-blot for TNF-R1 (55 kDa) and its respective loading control protein β-actin (43 kDa) for both groups. Expression ratio of TNF-R1 normalized to β-actin is shown in panel (b) for autopsies (n = 10) and TLE samples (n = 15). Data represent the mean ± SD from each sample performed in duplicate. Statistical analysis was conducted by Student’s t test with Holm-Sidak post hoc correction and the statistically significant differences are represented by **p < 0.01, compared with autopsies

Full size image
Table 2 Sex differences in autopsies and patients with TLE
Full size table

In patients with TLE, the protein expression of TNF-R1 correlated positively with the duration of epilepsy (r = 0.656, p < 0.05; Table 3), indicating that a longer duration of epilepsy is associated with higher neuroinflammation. No further correlations were found (Table 3).

Table 3 Correlation between proteins evaluated and epilepsy clinical variable
Full size table

Expression of an oxidative stress mediator

The expression of iNOS in microvessels isolated from the neocortical tissue of patients with TLE and autopsy samples was evaluated. iNOS immunolabeling revealed one band corresponding to the molecular weight of this protein in the proximity of 130 kDa.

In the microvessels of the autopsy samples, the expression ratio of iNOS was 0.48 ± 0.17. Patients with TLE presented higher iNOS expression (210%, p < 0.001; Fig. 4b). These results did not correlate with the clinical patient data (Table 3). Our results showed no differences between the expression of proinflammatory and stress oxidative markers and the type of epilepsy (data not shown).

Fig. 4
figure 4

Protein expression level of iNOS in the neocortical microvasculature of autopsies and TLE patients. Panel (a) shows representative images of western-blot for iNOS (130 kDa) and its respective loading control protein β-actin (43 kDa) for both groups. Expression ratio of iNOS normalized to β-actin is shown in panel (b) for autopsies (n = 10) and TLE samples (n = 15). Data represent the mean ± SD from each sample performed in duplicate. Statistical analysis was conducted by Student’s t test with Holm-Sidak post hoc correction and the statistically significant differences are represented by ***p < 0.001, compared with autopsies

Full size image

Discussion

Several studies have shown an increase in proinflammatory cytokines in whole neocortex and hippocampus tissues of patients with TLE and frontal lobe epilepsy (FLE) [5, 12, 13, 15]. Given the proximity of BBB microvessels to neurons, failure of the BBB can cause abnormal excitability of the neurons, establishing epilepsy [11]. For the first time, this study has shown an increase in the proinflammatory cytokines IL-1β, TNF-α and its receptor, TNF-R1, and an oxidative stress mediator (iNOS) in the neocortical microvasculature of patients with TLE.

Sex differences in immune responses have been described widely [16]; however, there are no pieces of evidence about sex differences in the expression of proinflammatory cytokines in patients with TLE. Although several studies have shown sex differences in epilepsy [17, 18], we showed for the first time that in male patients with TLE, IL-1β and TNF-R1 were higher than in TLE female patients. Our findings are interesting since in TLE, gender appears to be predominant in the primary projections from the epileptogenic region, which could reflect innate sex differences in networks. Such differences might support seizure propagation and reorganization of networks due to seizures themselves [19].

The BBB is an immunologic barrier that blocks leukocyte migration under normal conditions. However, the BBB is altered in many CNS pathologies, and these alterations include the upregulation of luminal adhesion molecules, increased adhesion and transmigration of leukocytes, and increased leakiness of tight junction proteins and the extravasation of plasma proteins [20]. Furthermore, seizures could modulate BBB functions through the induction of albumin extravasation and activation of the innate immune system and eventually generate modifications of neuronal networks [21]. Moreover, disruption of the BBB drives endothelial cells to secrete proinflammatory mediators, such as TNF-α, IL-1β, and IL-6 cytokines, as well as prostaglandins, among others [22]. The microvasculature of the neurovascular unit plays an active role in the mediation of the neuroimmune response, by either the production of proinflammatory mediators or the expression of adhesion molecules [23]. Modulation of BBB function may involve complex molecular and cellular alterations that trigger the generation and propagation of seizures, as well as the establishment of epilepsy [21].

The induction of inflammatory signaling pathways mostly involves activated microglia, astrocytes, neurons, and BBB cell components [11, 12]. In addition, inflamed brain endothelial cells show increased expression of adhesion molecules and chemokines, which are essential for leukocyte migration into the CNS [3]. Several studies have shown elevated levels of proinflammatory mediators, such as caspases, interleukins, and TNF-α, in both experimental models and TLE patients [5, 6, 12, 15, 24,25,26]. Interestingly, in another type of epilepsy, FLE, overexpression of the proinflammatory mediators IL-1β, IL-6, TNF-α, and iNOS, a marker of oxidative stress, was demonstrated in the olfactory bulb of patients with this neurological disorder, suggesting that these patients exhibit olfactory dysfunction [13].

The cytokine IL-1β has proconvulsant properties in limbic seizures [12]. The activation of IL-1R1 in endothelial cells of the BBB and perivascular astrocytes may alter the permeability function of the BBB by inducing the breakdown of tight junction proteins or enhancing the transcytosis of macromolecules in endothelial cells [12, 27]. In addition, IL-1β and TNF-α have been shown to increase oxidative stress, inflammation, and BBB permeability and promote angiogenesis [28, 29]. Furthermore, IL-1β and TNF-α affect neuronal excitability by increasing the extracellular glutamate concentration, which contributes to neuronal excitotoxicity [27]. Interestingly, a recent study showed that inflammatory stress (IL-17, IL-6, and TNF-α) led to BBB disruption with a significant increase in permeability, which was correlated with a decrease in tight junction protein expression [30]. In addition, IL-1β was shown to promote pericyte modifications and pericyte-microglia clustering, modifying BBB permeability and exacerbating epileptic pathology [31]. In this context, our study has revealed a correlation between the level of TNF-R1 expression and the duration of epilepsy, which could suggest chronic inflammation through the signaling of this receptor at the level of the BBB microvasculature. On the other hand, our results showed no differences between the expression of proinflammatory markers and the type of epilepsy. This finding suggests that inflammation at the level of the BBB as a consequence of seizure activity is independent of the type of epilepsy.

Several studies have shown that the expression of HIF-1α and TNF-α is upregulated in the temporal cortex and hippocampus of patients with mesial TLE, suggesting that these factors contribute to the pathogenesis of TLE [32,33,34]. Interestingly, Nikolic et al. (2018) demonstrated that the blockade of TNF-α-driven astrocytic purinergic signaling restored normal excitatory synaptic activity in the inflamed hippocampus in an experimental model [35]. Selective blockade of IL-1β/IL-1R1 signaling reduced symptomatic seizures or established chronic recurrent seizures in rodents [27]. In this context, inflammation may promote changes in the structure and function of the BBB in the human epileptic brain.

Oxidative stress is induced by the production of reactive oxygen species and reactive nitrogen species due to mitochondrial dysfunction and the increased activities of NADPH oxidase, xanthine oxidase, and iNOS. Oxidative stress persists during epileptogenesis and in chronic epilepsy, as shown by increased markers of oxidative stress, such as iNOS, in animal models. In addition, these markers are increased in both the brain and blood in human epilepsy [27]. We showed an increase in iNOS on the microvasculature of the BBB in TLE patients, indicating enhanced oxidative stress in the endothelial cells of patients with TLE, a condition that may affect the function of the BBB.

Based on the above, in our study, we evaluated several proinflammatory mediators, IL-1β, TNF-α, and TNF-R1, and a mediator of oxidative stress, iNOS, which showed an increase in these proinflammatory cytokines in the microvasculature (mainly endothelial cells) of the BBB. Our results suggest a link between inflammation and damage to endothelial cells of the BBB in patients with TLE.

Conclusion

Our study indicates a close relationship between neuroinflammation and BBB microvessels in epilepsy. The overexpression of proinflammatory molecules and oxidative stress mediators in the cortical microvasculature in TLE patients can promote BBB dysfunction, affecting the local vascular network, which triggers neuroinflammatory factors and seizure progression. More studies are needed to clarify the mechanisms that regulate inflammation and oxidative stress and identify a therapeutic target in epilepsy.