Deep brain stimulation suppresses epileptic seizures in rats via inhibition of adenosine kinase and activation of adenosine A1 receptors

Abstract Aims Deep brain stimulation (DBS) of the anterior nucleus of the thalamus, is an effective therapy for patients with drug‐resistant epilepsy, yet, its mechanism of action remains elusive. Adenosine kinase (ADK), a key negative regulator of adenosine, is a potential modulator of epileptogenesis. DBS has been shown to increase adenosine levels, which may suppress seizures via A1 receptors (A1Rs). We investigated whether DBS could halt disease progression and the potential involvement of adenosine mechanisms. Methods Control group, SE (status epilepticus) group, SE‐DBS group, and SE‐sham‐DBS group were included in this study. One week after a pilocarpine‐induced status epilepticus, rats in the SE‐DBS group were treated with DBS for 4 weeks. The rats were monitored by video‐EEG. ADK and A1Rs were tested with histochemistry and western blot, respectively. Results Compared with the SE group and SE‐sham‐DBS group, DBS could reduce the frequency of spontaneous recurrent seizures (SRS) and the number of interictal epileptic discharges. The DPCPX, an A1R antagonist, reversed the effect of DBS on interictal epileptic discharges. In addition, DBS inhibited the overexpression of ADK and the downregulation of A1Rs. Conclusion The findings indicate that DBS can reduce SRS in epileptic rats via inhibition of ADK and activation of A1Rs. A1Rs might be a potential target of DBS for the treatment of epilepsy.


| INTRODUC TI ON
Deep brain stimulation (DBS) is an effective neuromodulation therapy for patients with drug-resistant epilepsy (DRE). [1][2][3] The anterior nucleus of the thalamus (ANT) is one of the most common targets for DBS in the treatment of DRE, and its effectiveness has been proven. 4,5 ANT is a crucial node in the Papez circuit and has abundant projections to the frontal and temporal cortex. 6,7 This explains why ANT-DBS most effectively reduces seizures originating in the temporal and frontal lobes. 8 Pilocarpine injections in rodents provide a suitable model for studying DRE. Pilocarpine epileptic rat models have long-term seizures and show extensive cell loss, which resembles what has been discovered in patients with temporal lobe epilepsy. [9][10][11][12] ANT-DBS reduced the seizure rate and retarded the development of seizures in pilocarpine rats. [13][14][15] However, the underlying mechanisms remain unclear.
Adenosine performs its anticonvulsant effects largely mediated via activation of adenosine A1 receptors (A 1 Rs), 16,17 whereas it acts as a pro-epileptic signal through adenosine A2A receptors (A 2A Rs). 18,19 The extracellular levels of adenosine in the brain are mainly curtailed by its major metabolic enzyme, adenosine kinase (ADK), 20 which is mainly located in astrocytes in the adult brain. 21 Overexpression of ADK in the epileptic hippocampus contributes to the development and progression of seizures. 22,23 Measured via microdialysis, DBS has been indicated to elevate adenosine levels in the hippocampus. 24 In addition, ADK has been highly recognized as a hallmark of epileptogenesis. 22,23 We hypothesized that DBS might prevent epileptogenesis through the augmentation of adenosine signaling. In this study, a pilocarpine rat model was established and treated with ANT-DBS to assess whether DBS can halt disease progression via potential underlying adenosine mechanisms.

| Animals
Adult Sprague-Dawley rats (male, 260-300 g; Beijing Vital River Laboratory Animal Technology Co., Ltd.) were used for the experiments. Animals were housed individually in a controlled animal facility (

| Tissue preparation
Half of the rats were perfused with 4% paraformaldehyde transcardially after being deeply anesthetized. The brains were submerged in 30% sucrose solution and then cut in a cryostat (thickness: 15 μm).
After deep anesthetization, the remaining rats were killed to obtain hippocampi. All tissues were preserved at −80°C.

| Histopathological assessment
Nissl staining was performed in sections containing electrode tracks to determine whether the electrodes were successfully implanted into the ANT (Figure 2). Nissl staining of sections containing the hippocampus was used to assess the condition of neurons.
Double-label immunofluorescence was performed as described previously. 29,32 After incubation with anti-GFAP (monoclonal mouse,

| Western blot
The experimental process of western blot was carried out as described previously 32  Antiβ-actin (monoclonal mouse, 1:1000; Cell Signaling Technology) was used to normalize results. After incubating with HRP-conjugated secondary antibodies (1:5000; Applygen Technologies Inc.), proteins were visualized using enhanced chemiluminescence reagents (Applygen, Technologies Inc.). Autoradiography films were scanned with a luminescent image analyzer (LAS-3000; Fujifilm). The quantitative analysis process of protein strips was performed using ImageJ software.

| Statistics analyses
Statistical analysis was performed by GraphPad Prism 8.4.2 software (GraphPad Software). Results are displayed as mean ± SEM. All data were subject to tests for normality using the Shapiro-Wilk test and all p-values were > 0.05. One-way analysis of variance (ANOVA) followed by Tukey's multiple comparisons test was used for multiple groups. p < 0.05 was considered statistically significant.

| DBS reduces the frequency of interictal epileptic discharge in epileptic rats, and DPCPX can block the antiepileptic effect of DBS
Interictal epileptic discharges were calculated to evaluate the therapeutic effect of DBS in epileptic rats. Interictal epileptic discharges in the EEG are shown in Figure 3B. The number of interictal epileptic discharges in the SE-DBS group was significantly lower than that in the SE-sham-DBS group ( Figure 3E,

| DBS decreases the loss of hippocampal neurons in epileptic rats
To assess the protective effect of DBS on hippocampal neuronal loss in pilocarpine rats during epileptogenesis, we counted the number of hippocampal neurons across each group. We found that there was no apparent neuronal loss in the CA1 and CA3 regions in the control group ( Figure 4A,a,e,i). Compared with control group, neuronal loss within CA1 and CA3 regions was remarkably observed in SE ( Figure 4A,b,f,j), SE-DBS ( Figure 4A

| DBS inhibits the overexpression of ADK
Similar to previous studies, 30,32 in control group, ADK immunoreactivity was present in sparse astroglial cells with only weak staining ( Figure 5A,a,e,i) and a lack of ADK immunoreactivity in neuronal cells. Compared with control group, overexpression of ADK was shown in the reactive astroglial cells within CA1 and CA3 regions of the hippocampus in SE ( Figure 5A,b,f,j, inset arrow), SE-DBS ( Figure 5A,c,g,k), and SE-sham-DBS ( Figure 5A Quantitative analysis of ADK confirmed that DBS treatment can significantly inhibit the increased density of ADK in the hippocampus of epileptic rats.

| DBS attenuates the downregulation of A 1 Rs
The immunostaining of A 1 Rs was studied by immunohistochemistry in brain specimens of rats in each group. In control group, A 1 R immunoreactivity was present exclusively in neuronal cells in the hippocampus ( Figure 6A,a,e,i, insect arrow). Compared with the control group, the levels of A 1 Rs were significantly downregulated within CA1 and CA3 regions of the hippocampus in SE ( Figure 6A,b,f,j), SE-DBS ( Figure 6A,c,g,k), and SE-sham-DBS ( Figure 6A The duration of SRS was compared between each group, demonstrating that the mean SRS duration of the SE-DBS group was significantly shorter than that of the SE-sham-DBS group (**p < 0.01, n = 8). (E) Quantification analysis showed that the number of interictal epileptic discharges in the SE-DBS group was significantly lower than that in the SE-sham-DBS group (**p < 0.01, n = 8). A total duration of 30 min after the DPCPX injection, the number of interictal epileptic discharges in the SE-DBS group increased significantly (**p < 0.01, n = 8).
that DBS therapy during epileptogenesis can attenuate the downregulation of A 1 Rs in epileptic rats.

| DBS therapy delays disease development during epileptogenesis
In recent decades, research on electrical stimulation of specific regions of the brain has been conducted. 33 There is accumulating evidence that ANT-DBS has therapeutic effects on DRE. 1,3,13 In this study, we adopted the pilocarpine epilepsy rat model with similar neuropathological and electrophysiological characteristics of DRE patients. Epileptogenesis is a process in which a healthy brain becomes an epileptic brain. 34 Via video-EEG monitoring, we found that high-frequency ANT-DBS therapy during epileptogenesis can not only reduce the frequency of SRS, but also shorten the duration of seizures (decrease the severity of seizures) (Figure 3). In addition, the number of interictal epileptic discharges of the epileptic rats was significantly reduced as well after DBS treatment ( Figure 3), which is in accordance with the previous research. 35,36 DBS delays epileptogenesis by interfering with the propagation of epileptiform activity when applied during the interictal period and alleviates seizures by delaying the modification of microcircuit recruitment. 37 Epileptogenesis can result in remarkable neuronal loss in the hippocampus through an array of complex pathological alterations. [38][39][40] DBS can reduce neuronal loss in the hippocampus of epileptic rats. 41,42 The inhibition of apoptosis in the hippocampus mediates the neuroprotective effect. 41 In this study, F I G U R E 4 Effect of DBS on neuronal loss in the hippocampus. Nissl-stained sections through the hippocampus were used to assess pilocarpine-induced neuronal loss at 5 weeks (Day 29-35) after SE. (A) No apparent neuronal loss was discovered in the CA1 and CA3 regions in control group (a, e, and i, respectively); Obvious neuronal loss within CA1 and CA3 regions were observed in SE (b, f, and j, respectively), SE-DBS (c, g, and k, respectively), and SE-sham-DBS (d, h, and l, respectively) groups, compared with control group. (B and C) Compared to the SE-sham-DBS group, significantly less neuronal loss was observed within hippocampal CA1 and CA3 regions in the SE-DBS group (**p < 0.01, n = 4). Scale bars = 500 μm (a-d) and 50 μm (e-l).

F I G U R E 5
Effect of DBS on ADK levels in the hippocampus. Immunohistochemical staining of ADK, double-label immunofluorescence staining of ADK and GFAP, and analysis of ADK by western blotting were evaluated in the hippocampus at 5 weeks after SE. (A) ADK immunoreactivity was present in sparse astroglial cells with only a weak staining and lack of ADK immunoreactivity in neuronal cells in control group (a, e, and i, respectively); Compared with control group, overexpression of ADK was shown in the reactive astroglial cells within CA1 and CA3 regions of the hippocampus in SE group (b, f, and j, respectively), SE-DBS group (c, g, and k, respectively), and SE-sham-DBS group (d, h, and l, respectively); (f) Overlay of GFAP (green) and ADK (red) immunofluorescence(inset); (j) ADK-positive cells (arrow, inset). (B and C) The number of ADK-positive cells within CA1 and CA3 regions of the hippocampus in the SE-DBS group was fewer than that in the SE-sham-DBS group (**p < 0.01, n = 4). (D) Analysis of ADK in the hippocampus by western blot (original, uncropped image see Figure S1). (E) In SE and SE-sham-DBS groups, ADK levels in the hippocampus were elevated compared with the control group (**p < 0.01, n = 4); In SE-DBS group, less ADK were observed compared with the SE-sham-DBS group (*p < 0.05, n = 4). Scale bars = 500 μm (a-d) and 50 μm (e-l).

| DBS therapy increases the adenosine signaling via inhibition of ADK
Adenosine is mainly removed by ADK, 20 which is principally expressed in astrocytes in the adult brain. 21 Astrocyte-expressed ADK, a key negative regulator of the brain inhibitory molecule adenosine, is a potential predictor and modulator of epileptogenesis. 29 During epileptogenesis, ADK overexpression in the hippocampus promotes the development and progression of seizures. 22,23 Upregulation of ADK has been highly regarded as a common pathologic hallmark of epilepsy. 23,29,43,44 In addition, DBS can promote ATP release and induce adenosine accumulation in the brain. 24,[45][46][47] In this study, DBS therapy exerts remarkable inhibition of ADK upregulation in epileptic rats ( Figure 5). Given that ADK is the main scavenging enzyme of adenosine, inhibiting its overexpression interferes with adenosine elimination, which is a crucial factor promoting adenosine accumulation. It has been demonstrated that therapeutic adenosine augmentation can prevent epileptogenesis. 48 The long-term accumulation of adenosine may inhibit epilepsy progression in epileptic rats.
In addition, ADK is involved in epileptogenesis by regulating the DNA methylation pathway. 49 DNA methylation in the central nervous system participates in epilepsy formation by regulating neuronal networks and synaptic plasticity. 48,50,51 Locally increasing the level of adenosine can reduce or reverse abnormal methylation of DNA and inhibit the development of epilepsy. Thus, the decrease in ADK levels or an increase in adenosine in the brain reverses the equilibrium of the methylation pathway, which provides therapeutic value as regards the treatment of epileptogenesis. 48 Therefore, we hypothesized that DBS therapy might retard epileptogenesis via the inhibition of ADK.
It is worth noting that the efficacy of DBS gradually increases with therapeutic time and is far more remarkable in the long term. 3 DBS provides long-term adenosine enhancement in the brain, which might explain the characteristic of long-term DBS efficacy.

| DBS therapy inhibits seizures via A 1 Rs
The actions of adenosine in the brain are mainly mediated by A 1 Rs and A 2A Rs, while the A 2B Rs and A 3 Rs are still poorly studied. 52 As an endogenous substance, adenosine acts as an anticonvulsant mostly through A 1 Rs. 16,17 Activation of the presynaptic A 1 Rs reduces the release of excitatory neurotransmitters, especially glutamate, 53 while the postsynaptic A 1 Rs hyperpolarize neurons through G proteincoupled potassium channels, which inhibits the excitability of the hippocampus. 54 In this study, downregulation of hippocampal A 1 Rs was observed in epileptic rats. DBS therapy attenuates the downregulation of hippocampal A 1 Rs ( Figure 6). It is worth mentioning that the staining shows a predominant A 1 R staining in the cell bodies ( Figure 6), while A 1 Rs are very well established to be mainly synaptically located. 55 It is probably because antibodies in our study do not reach synaptic epitopes in the process of immunohistochemical staining.
In a rat kindling model, the reduction of A 1 R levels and the loss of endogenous adenosine-based seizure control indicated a failure of endogenous adenosine-mediated seizure control mechanisms. 56 A 1 R knockout mice developed spontaneous electrographic seizures and lethal SE following injection with kainic hippocampal acid in the hippocampus or traumatic brain injury. 57 In pilocarpine-induced epileptic rats, the levels of A 1 Rs in the hippocampus showed a significant decrease in epileptogenesis. 58 Actually, in several other previous studies, A 1 Rs density in the hippocampus is decreased in kainic acid-induced and kindling models. 56,59,60 These studies suggest that the pathophysiology of epilepsy is associated with abnormal A 1 R signaling.
Stimulation-induced adenosine release in ferret slices was able to abolish spontaneous spindle oscillations. 46 In pilocarpine-induced epileptic rats, DBS therapy increased adenosine levels in the hippocampus. 24 In vitro study showed that the excitability reduction of DBS therapy in rats pretreated with A 1 R antagonists was completely eliminated and was strongly enhanced by A 1 R agonists. 24 To confirm the effect of A 1 Rs on the therapeutic effect of DBS therapy in vivo, the rats in the SE-DBS group were administrated with an A 1 R antagonist DPCPX in this study. The frequency of interictal epileptic discharges in the SE-DBS group was increased and restored to the level of SE-group following injection of DPCPX ( Figure 3B,E), thereby revealing that the inhibition of interictal epileptic discharges in the hippocampus of epileptic rats by DBS treatment was completely eliminated by the loss of A 1 R signal. Therefore, DBS therapy might suppress seizures in epileptic rats at least partially via A 1 Rs.

| CON CLUS ION
Our data reveal that DBS therapy initiating in epileptogenesis can dampen disease development, inhibit hippocampal neuron loss, and F I G U R E 6 Effect of DBS on A 1 R levels in the hippocampus. Immunohistochemical staining of A 1 Rs and analysis of A 1 Rs by western blotting were evaluated in the hippocampus at 5 weeks after SE. (A) In control group, A 1 R immunoreactivity was present exclusively in neuronal cells in the hippocampus (a, e, and i, respectively, arrow, insect); Compared with the control group, the levels of A 1 Rs were significantly downregulated within CA1 and CA3 regions of the hippocampus in SE group (b, f, and j), SE-DBS group (c, g, and k, respectively) and SE-sham-DBS group (d, h, and l, respectively). (B and C) Quantitative analysis revealed a significant increase in the number of A 1 R-positive cells within CA1 and CA3 regions of the hippocampus in the SE-DBS group, compared with the SE-sham-DBS group (**p < 0.01, n = 4). (D) Analysis of A 1 Rs in the hippocampus by western blot (original, uncropped image see Figure S2). (E) Compared with control group, there were significantly lower levels of A 1 Rs in SE, SE-DBS, and SE-sham-DBS groups (**p < 0.01, n = 4); The levels of A 1 Rs within the hippocampus in the SE-DBS group were significantly higher than that in the SE-sham-DBS group (**p < 0.01, n = 4). Scale bars = 500 μm (a-d) and 50 μm (e-l).
inhibit ADK overexpression and A 1 R downregulation in epileptic rats. A 1 Rs may be potential targets of DBS for the treatment of epilepsy. The effectiveness of DBS therapy initiating in chronic stage of epilepsy needs further study in the future.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors have nothing to disclose and declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.