The Effect of ARHGEF15 Knockout and Forebrain-Specic Human ARHGEF15 Knock-In on Seizure Susceptibility in Mice

The ARHGEF15 gene encodes the Rho guanine nucleotide exchange factor 15. Although multiple evidence indicates that ARHGEF15 may be related to epilepsy, it is not clear what role it plays. Methods Subjects were homozygous ARHGEF15 knockout (E5 −/− ) mice, wild-type (WT) mice, Cre-positive homozygous human ARHGEF15 conditional knock-in (E5 CKI/CKI •Cre+) mice which can overexpress human Ephexin5 in the forebrain, and Cre-positive (E5 WT/WT •Cre+) mice. Models with epileptic seizures were established by intraperitoneal injection of pentylenetetrazol (PTZ) in 60 mg/kg, and seizures were recorded by video. Then 7 indexes were counted, including “maximum seizure level”, “seizure level classication”, “tonic-clonic seizure or not”, “latency of tonic-clonic seizure”, “number of times of tonic-clonic seizure”, “total duration of tonic-clonic seizure”, and “dead or not”. Western blot was used to detect Ephexin5, RhoA, p-RhoA, ROCK2, and p-ROCK2 in WT mice and E5 −/− mice.


Background
In recent years, the research on the genetic mechanism of epilepsy increased rapidly, exploring new insights for the pathogenesis of epilepsy and providing evidence for identifying potential therapeutic targets [1]. With the application of high throughput sequencing in clinical practice, in addition to epilepsy-associated genes that have been identi ed, some genes that probably be related to epilepsy are found [2].
The ARHGEF15 gene is located on human chromosome 17 and encodes the Rho guanine nucleotide exchange factor 15, also known as Ephexin5, Vsm-RhoGEF, which has guanine nucleotide exchange activity and can regulate the conversion of Rho GTPase from inactive type to active type [3], so that Rho GTPase can play a molecular switch role in cell signal transduction pathways [4].
Previous studies have shown that Ephexin5 plays an important role in the central nervous system. Sahin et al. proved that Ephexin5 is highly expressed in the brain and it has a high homology with the encoding gene of Ephexin1 in mammals [5]. Margolis et al. found that Ephexin5 could limit the formation of excitatory synapses by activating RhoA, a kind of Rho GTPases [6]. Additionally, Hamilton et al. revealed that Ephexin5 plays a dual role in spinogenesis, which is not only a brake on overall spine outgrowth, but also an essential component in the site-speci c formation of new spines [7]. Moreover, the study of Ephexin5 in Alzheimer's disease has achieved an initial result that reducing the expression of the synapse inhibitory protein Ephexin5 can ameliorate Alzheimer's-like impairment in mice [8].
In 2013, a possible link between the ARHGEF15 gene and epilepsy was reported for the rst time [9]. Researchers found a de novo mutation in the ARHGEF15 gene in a patient with epileptic encephalopathy by the whole-exome sequencing (WES). In vitro analyses using cell-based assays revealed that the mutation in ARHGEF15 reduced the guanine nucleotide exchange activity of Ephexin5 by about 50% [9]. In a study of copy number variation (CNV) risk factors for rolandic epilepsy, duplicate copies of ARHGEF15 were found in one patient. Then the researchers performed a network analysis of all the variants detected and found that the variants were signi cantly enriched in guanine exchange factor activation pathways [10]. Besides, Ephexin5 is ubiquitinated with the assistant of E6AP, the protein product of the UBE3A gene, and then degraded by the proteasome. The de ciency of E6AP leads to Angelman syndrome (AS), which is clinically characterized by a lack of speech, abnormal motor development, and epileptic seizures [11]. Although multiple evidence indicates that the function of the ARHGEF15 gene may be related to epilepsy, it is not clear what role it plays.
To study whether the ARHGEF15 gene can participate in epilepsy, we generated ARHGEF15 knockout (E5 −/− ) mice and Cre-positive homozygous human ARHGEF15 conditional knock-in (E5 CKI/CKI •Cre+) mice to explore the effect of ARHGEF15 knockout and forebrain-speci c human ARHGEF15 knock-in on seizure susceptibility in mice. Considering Ephexin5 can activate RhoA speci cally [6], and that its downstream signaling molecule ROCK2 [12], was increased in epileptic foci of temporal lobe epilepsy (TLE) patients [13], we also investigated the expression of the RhoA-ROCK2 pathway in the E5 −/− mice.

Animals
Mice were sourced from Cyagen Biosciences Inc. (Guangzhou, China), and housed at the Laboratory Animal Center of Jining Medical University. All animal experiments comply with the REACT guidelines and ARRIVE guidelines. All experiments are conducted in accordance with the UK Animal (Scienti c Procedures) Act (1986) and relevant guidelines. Heterozygous ARHGEF15 knockout (E5 +/− ) mice, with a C57BL/6 background, were constructed by knocking out the sequence containing exon 2 to exon 8 in the ARHGEF15 gene causing a frameshift mutation and an early termination, and then bred to obtain homozygous ARHGEF15 knockout (E5 −/− ) mice and wild-type (WT) mice which were used as the control.
Heterozygous human ARHGEF15 conditional knock-in (E5 CKI/WT ) mice containing the "CAG promoter-loxP-stop-loxP-kozak-human ARHGEF15 CDS-polyA" in the genome, with a C57BL/6 background, were constructed and bred to obtain homozygous human ARHGEF15 conditional knock-in (E5 CKI/CKI ) mice. At this time, the human ARHGEF15 sequence in E5 CKI/WT mice and E5 CKI/CKI mice can not be transcribed, because the "loxP-stop-loxP" was inserted to separate the CAG promoter and the kozak-human ARHGEF15 CDS. Then E5 CKI/CKI mice were bred with Camk2a-cre mice which can drive Cre recombinase expression in the forebrain, including the CA1 pyramidal cell layer of the hippocampus (http://jaxmice.jax.org/strain/005359.html) to obtain Cre-positive heterozygous human ARHGEF15 conditional knock-in (E5 CKI/WT •Cre+) mice. Inbreeding of E5 CKI/WT •Cre + mice produced Cre-positive homozygous human ARHGEF15 conditional knock-in (E5 CKI/CKI •Cre+) mice and Cre-positive (E5 WT/WT •Cre+) mice without human ARHGEF15 conditional knock-in. E5 CKI/WT •Cre + mice and E5 CKI/CKI •Cre + mice expressed the Cre recombinase acting on loxP sites to knock out "loxP-stop-loxP", so the CAG promoter was connected with the kozak-human ARHGEF15 CDS and the mice can express human ARHGEF15 protein (human Ephexin5) in the forebrain speci cally.
To identify genotypes of offspring mice, the nails were cut at 10 days after birth, and the DNA was extracted, and then PCR and agarose gel electrophoresis were performed ( Fig. S1-3). Testing Seizure Susceptibility Models with epileptic seizures were established by intraperitoneal injection of pentylenetetrazol (PTZ, Sigma) in 60 mg/kg [14]. Behavior changes in mice of different genotypes were recorded by video within half an hour after drug injection. The seizures of the mice were evaluated according to a revised Racine's scale as follows: stage 0, behavioral arrest; stage 1, mouth clonus; stage 2, head bobbing; stage 3, unilateral forelimb clonus; stage 3.5, alternating forelimb clonus; stage 4, bilateral forelimb clonus with rearing; stage 5, bilateral forelimb clonus with rearing and falling over; stage 6, wild running and jumping with vocalization; stage 7, tonic-clonic seizure [15]. Then we counted 7 indexes according to the video, including "maximum seizure level", "seizure level classi cation (severe seizure: maximum seizure level ≥ stage 4, mild seizure: maximum seizure level < stage 4)", "tonic-clonic seizure or not", "latency of tonicclonic seizure", "number of times of tonic-clonic seizure", "total duration of tonic-clonic seizure", and "dead or not". Spontaneous epileptic seizures were not observed in all mice. Mice were euthanized by cervical dislocation.

Western Blot
All the mice applied to detect the expression level of the Ephexin5, RhoA-Rock2 pathway were not injected with PTZ. Mice were anesthetized by intraperitoneal injection of 1.25% tribromoethanol (Sigma) in 0.2ml/10g. Then the brain was taken and the hippocampus was detached. A mixture of 495 µl RIPA lysis buffer and 5 µl PMSF (Beyotime Biotechnology) was added to every 100 mg tissue and then homogenized by an ultrasonic processor. EP tubes with tissue suspension were placed on ice for 30 min, and then centrifuged at 14,000 RPM, 4 ℃ for 20 min. Then the supernatant was transferred into a new EP tube. The protein concentration was measured using BCA Protein Assay Kit (Beyotime Biotechnology). After SDS-PAGE, proteins were transferred to PVDF membranes (Millipore). 5% milk-TBST was used to block the membranes. Then the membranes were incubated with primary antibodies overnight at 4°C. Primary antibodies were as the following: anti-Ephexin5 antibody (Rabbit, Abclonal, raised against a recombinant fragment corresponding to amino acids 140-380 of mouse Ephexin5), anti-human Ephexin5 antibody (Rabbit, Abcam-ab221376), anti-RhoA antibody (Rabbit, Abclonal-A13947), anti-p-RhoA antibody (Rabbit, Abcam-ab41435), anti-ROCK2 antibody (Rabbit, Abclonal-A5698), anti-p-ROCK2 antibody (Rabbit, Abcam-ab228008), anti-ACTB antibody (Rabbit, Abclonal-AC026). HRP Goat Anti-Rabbit IgG (Abclonal-AS014) was used as the secondary antibody.

Statistics
The data of seizure susceptibility were analyzed with IBM SPSS Statistics 25 software. The independentsamples T-test was used to compare the measurement data ("maximum seizure level", "latency of tonicclonic seizure", "number of times of tonic-clonic seizure", "total duration of tonic-clonic seizure"), and the homogeneity of variance was detected by Levene's tests. Chi-square test was used to compare the counting data ("seizure level classi cation", "tonic-clonic seizure or not", "dead or not"). The level of signi cance was preset to P < .05.
Image J software was used to calculate the results of Western blot, and GraphPad Prism 8 software was used to make statistical graphs. The independent-samples T-test was used to compare the expression of proteins between two groups. The level of signi cance was preset to P < .05.

Results
The seizure susceptibility of the ARHGEF15 knockout mice The experimental group was E5 −/− mice, and the control group was WT mice. There were 6 males and 6 females in each group, and all were 8-10 weeks old. Compared with WT mice, E5 −/− mice had a shorter "total duration of tonic-clonic seizure" (P < .05). There were no signi cant differences in "maximum seizure level", "latency of tonic-clonic seizure", "number of times of tonic-clonic seizure", "seizure level classi cation", "tonic-clonic seizure or not", and "dead or not" between the two groups ( Table 1, Table 2). The seizure susceptibility of the forebrain-speci c human ARHGEF15 knock-in mice The experimental group was E5 CKI/CKI •Cre + mice, and the control group was E5 WT/WT •Cre + mice. There were 6 males and 6 females in each group, and all were 8-10 weeks old. There were no signi cant differences between the two groups in all the 7 indexes including "maximum seizure level", "latency of tonic-clonic seizure", "number of times of tonic-clonic seizure", "total duration of tonic-clonic seizure", "seizure level classi cation", "tonic-clonic seizure or not", and "dead or not" ( Table 3, Table 4).  The expression of Ephexin5 in the ARHGEF15 knockout mice and the forebrain-speci c human ARHGEF15 knock-in mice We used western blot to detect the expression level of Ephexin5 in E5 −/− mice and WT mice. At 2 days after birth, the expression of Ephexin5 was detected in the whole brain tissue of WT mice, but almost no Ephexin5 was expressed in the brain of E5 −/− mice at the same age. In adult mice aged 8-10 weeks, Ephexin5 was not expressed in the hippocampus of WT mice and E5 −/− mice (Fig. 1a). Western blot was also used to detect the expression of human Ephexin5 in the hippocampus of E5 WT/WT •Cre + mice, E5 CKI/WT •Cre + mice, E5 CKI/CKI •Cre + mice, and E5 CKI/CKI mice aged 8-10 weeks. As a result, E5 CKI/WT •Cre + mice and E5 CKI/CKI •Cre + mice carrying both the human ARHGEF15 gene and the Cre gene expressed human Ephexin5 in the hippocampus, while E5 WT/WT •Cre + mice with the Cre gene only and E5 CKI/CKI mice with the human ARHGEF15 gene only expressed no human Ephexin5 in the hippocampus (Fig. 1b).

The expression of the RhoA-ROCK2 pathway in the ARHGEF15 knockout mice
The expression of the RhoA-ROCK2 pathway in the hippocampus of 8-10 weeks mice was detected by western blot. As a result, the levels of RhoA, p-RhoA, ROCK2, and p-ROCK2 were not signi cantly different between WT mice and E5 −/− mice (Fig. 2).

Discussion
In this study, to explore whether the ARHGEF15 gene can participate in epilepsy, the ARHGEF15 knockout mice were constructed. In the seizure susceptibility experiment, because none of the mice had spontaneous epileptic seizures, the PTZ-induced epileptic animal model was selected, which is used widely in investigating the pathophysiology of epilepsy and related genes [16]. Compared to WT mice, E5 −/− mice had no other signi cant difference in seizures induced by PTZ, except a shorter "total duration of tonic-clonic seizure", indicating that ARHGEF15 knockout can decrease "duration of tonic-clonic seizure" in adult mice with epileptic seizures induced by PTZ, but it cannot in uence other indexes including "maximum seizure level", "latency of tonic-clonic seizure", "number of times of tonic-clonic seizure", "seizure level classi cation", "tonic-clonic seizure or not", and "dead or not".
The level of Ephexin5 was high in the whole brain tissue of 2-day-old WT mice, and an extremely low expression of Ephexin5 was detected in the brain tissue of E5 −/− mice of the same age, but no expression of Ephexin5 was found in the adult hippocampus of the two genotypes, illuminating that the level of Ephexin5 in the brain tissue of mice is age-dependent, which was consistent with the study of Margolis et al. [6]. In the brain of WT mice, Ephexin5 was expressed during 12 days of the embryonic period to 21 days after birth [6].
In previous studies, Ephexin5 can activate RhoA speci cally [6], a kind of Rho GTPases, and the expression of ROCK2, a downstream signaling molecule of RhoA [12], was increased in epileptic foci of temporal lobe epilepsy (TLE) patients [13]. In this study, we investigated the expression of the RhoA-ROCK2 pathway in the hippocampus of WT mice and E5 −/− mice. However, there was no signi cant difference in levels of RhoA, p-RhoA, ROCK2, and p-ROCK2 in the adult hippocampus between WT and E5 −/− , which probably attributes to no expression of Ephexin5 in the hippocampus of adult mice. Considering ARHGEF15 knockout did not affect the expression of the RhoA-ROCK2 pathway in the hippocampus of adult mice, although Ephexin5 was demonstrated to be a RhoA guanine nucleotide exchange factor [6], we could not conclude that the effect of ARHGEF15 knockout on "duration of tonicclonic seizure" in adult mice was mediated by the RhoA-ROCK2 pathway. Accordingly, we speculated that during the embryonic period and early childhood when Ephexin5 can express in the brain, ARHGEF15 knockout may cause some changes in the brain of mice, thus affecting the seizure susceptibility in adulthood. In addition to Ephexin5, other members of the Ephexin family such as Ephexin1 can also activate RhoA in neurons [5]. Perhaps after the ARHGEF15 gene knockout, the body's negative feedback regulatory function could maintain the RhoA activation at a normal level by enhancing the expression of other Ephexins.

Conclusions
In conclusion, the effect of the ARHGEF15 knockout and the effect of the forebrain-speci c human ARHGEF15 knock-in on the seizure susceptibility in mice still need to be veri ed by electrophysiological activity tests, such as the video electroencephalography. The mechanism for the differences in "duration of tonic-clonic seizure" induced by PTZ between WT mice and E5 −/− mice also needs to be further explored.

Figure 1
The expression of Ephexin5 in the ARHGEF15 knockout mice and forebrain-speci c human ARHGEF15 knock-in mice. a: The level of Ephexin5 was high in the whole brain tissue of 2-day-old WT mice, and an extremely low expression of Ephexin5 was detected in the brain tissue of E5-/-mice of the same age. No expression of Ephexin5 was found in the hippocampus of the two genotypes aged 8-10 weeks. b: The 8-10 weeks E5CKI/CKI•Cre+ mice and E5CKI/WT•Cre+ mice expressed human Ephexin5 in the hippocampus, and E5WT/WT•Cre+ mice and E5CKI/CKI mice expressed no human Ephexin5 in the hippocampus.

Figure 2
The expression of the RhoA-ROCK2 pathway in the ARHGEF15 knockout mice. a: The expression of RhoA, p-RhoA, ROCK2, and p-ROCK2 in the hippocampus of WT mice and E5-/-mice aged 8-10 weeks detected by western blot. b: There was no signi cant difference between WT mice (n=6) and E5-/-mice (n=6) in the levels of RhoA, p-RhoA, ROCK2, and p-ROCK2 by the independent-samples T-test (P > .05).