Somatic Depdc5 deletion recapitulates electroclinical features of human focal cortical dysplasia type IIA

Abstract

Epileptogenic mechanisms in focal cortical dysplasia (FCD) remain elusive, as no animal models faithfully recapitulate FCD seizures, which have distinct electrographic features and a wide range of semiologies. Given that DEPDC5 plays significant roles in focal epilepsies with FCD, we used in utero electroporation with clustered regularly interspaced short palindromic repeats gene deletion to create focal somatic Depdc5 deletion in the rat embryonic brain. Animals developed spontaneous seizures with focal pathological and electroclinical features highly clinically relevant to FCD IIA, paving the way toward understanding its pathogenesis and developing mechanistic-based therapies. Ann Neurol 2018;83:140-146.

Figure 1. Depdc5 CRISPR-IUE leads to focal cortical mTOR hyperactivation

a) Two short guide RNAs (sgRNAs) are used to target rat Depdc5 either at exon 8 or exon 12. b) CRISPR-IUE experiments are performed at E13–14. Animals are euthanized at P21 for pathology evaluation or are treated with Everolimus for 11 days beginning at P10. Video-EEG monitoring starts at P60. c-d) Co-electroporation of pCAG-GFP and Depdc5 CRISPR significantly increases pS6 immunoreactivity in both GFP+ and GFP− cells. Stars indicate cell non-autonomous mTOR hyperactivation in the overlying (c) and adjacent cortices (d). To show that increased p-S6 in GFP-negative cells is not due to dilution of episomal GFP transfection by dividing cells, Depdc5 CRISPR was co-electroporated with a Piggybac transposon system encoding GFP (see method) that integrates into the genome of neural progenitors. e) in control CRISPR-transfected brains, baseline pS6 immunoreactivity is minimal while f) In Depdc5 CRISPR-transfected brains, pS6 expression still markedly increases in both GFP-positive and GFP-negative cells. e′) higher magnification images of e). GFP labeled glial cells (arrows) suggest successful genome integration of GFP plasmid. f′) higher magnification images of f). Many GFP-negative cells show strong pS6 immunoreactivity (arrowheads), suggesting cell non-autonomous mTOR hyperactivation. g) Because plasmid co-transfection efficiency is not 100%, GFP-/pS6+ positive neurons could be neurons transfected with only Depdc5 CRISPR plasmid. To ensure that all neurons undergoing Depdc5 CRISPR deletion are labeled with GFP, we electroporated a single vector plasmid that encodes eCAS9, Depdc5 gRNA and GFP. With this approach, increased pS6 is still observed in GFP-negative cells (arrowheads), indicating cell non-autonomous mTOR hyperactivation. h) Antibody to a non-phosphorylated neurofilament protein (SMI-32) reveal aberrant cytoplasmic accumulation in both GFP(arrows) positive and negative neurons (arrowheads). i) Positive filamentous Nestin (arrows) staining, a typical pattern in reactive astrocytes, is seen in the dysplastic cortex. However, balloon cells, a pathological hallmark of FCD IIB, that show patchy and accumulated Nestin staining in the cell body, is never seen in Depdc5-CRISP transfected brains. An arrowhead here indicates that a hypertrophic GFP positive cell is negative for Nestin staining. j) The soma size of Depdc5 CRISPR-transfected neurons (N=3 rats, n=412 cells) is nearly doubled vs. control cells (N=3, n= 715). Two-tailed t test, p=0.002. k) Postnatal administration of Everolimus partially rescues cytomegalic neurons. Drug group: N=3, n=661; No drug vehicle group: N=3, n=1096. Two-tailed t test, p=0.005. l) All cortical neurons with Depdc5 deletion fire initial doublet action potentials (none in control), suggesting increased intrinsic excitability, and m) show decreased input resistance. N=3 rats, two-tail t-test. p<0.05. Scale bar: 100 μm.

Figure 2. Interictal discharges highly resemble human FCD

For all rat EEG recordings, the upper trace is from electrode implanted in the right parietal head region and the lower trace is from the left (see method). Interictal EEG recordings from rats with Depdc5 CRISPR-IUE, including PFA (a), polyspikes (b), rhythmic spike discharges (c), and PEDs (d) are lateralized to the lesional hemisphere, and are highly similar to intracranial subdural or depth EEG recordings from patient with FCD IIa including one with a DEPDC5 mutation (p.Tyr281Phe)[22] (a′-d′). For all human intracranial recordings, shown is one electrode recording that captures the most representative EEG discharges. Other typical focal IED examples including e) single focal spikes, f) spikes with overriding fast activity (stars) with minimal electrical field to the opposite hemisphere. g) rapid bilaterally synchronous-appearing IEDs are seen occasionally but they display lateralized onset based upon asymmetric amplitude and lead-in time (red vertical line). h) ripples associated with a spike are displayed with 70–250 Hz frequency band. i) Ripples that are not associated with spikes. j) fast ripples associated with a train of spike-and-wave discharges are displayed with 250–1000 Hz frequency band. These HFOs are also lateralized to the lesional hemisphere. k) Shown is a spectrogram of oscillatory power (color) versus time (horizontal) across various frequencies at each time point (vertical). IEDs, shown as purple lines in the lower power spectral image occur predominatly in NREM sleep, recognized by increased lower frequency band power and lower motion scores - see upper panel for fully visible spectrogram, and see methods. To quantify IEDs during sleep vs. wakefulness, IEDs were counted during one hour of sleep and wakefulness for 3 days from 3 rats with Depdc5 CRISPR-IUE. The average number of IEDs is 32.89 ± 6.326 during sleep, and 7 ± 3 during wakefulness (Two tailed t-test, p=0.002).

Figure 3. Electroclinical seizures highly resemble human FCD

a) shown is an entire seizure from a rat with Depdc5 CRISPR-IUE. The top EEG traces have a compressed time-base of 5 mm/s, and the lower ones have a standard time-base of 30 mm/s. The seizure starts with fast rhythmic spikes at low voltage fast activity (LVFA) (stars) followed by 2–3 Hz spike and wave discharges (arrowheads). b) intracranial recordings from patients with FCD II show common ictal onset patterns including low voltage fast, background suppression and burst of rhythmic spikes. c–d) two additional representative seizures recorded from rats with Depec5 CRISRP-IUE show typical seizure onset patterns in c) rhythmic spikes and d) background attenuation. e–h) 39 seizures from 3 different animals were analyzed for seizure duration (e; 35.97 ± 22.85s, median=26s), frequency (f; 13 ± 2.5/24 hr), EEG onset patterns (g) and semiology (h). The majority of seizures started with rhythmic spike discharges (RSW: 53.8%), background suppression (25.0%) and low voltage fast activity (LVFA: 21.2%). Behaviorally, most seizures showed clonus/myoclonus (30.1%) or behavioral arrest (e.g., freezing, stopping exploration) (23.1%). Other types included tonic posturing (12.8%) and arousal without other clear clinical signs (15.1%). 15.1% of electrographic seizures were subclinical. Notably, only one seizure progressed to a generalized tonic-clonic seizure. i) representative reconstructed brain sections showing the size and position of Depdc5 CRISPR transfection (each black dot is a transfected neuron) at 5 rostro-caudal levels in 5 rats after EEG recording was completed at P70–90. Arrowheads indicate the section in each series with the most extensive Depdc5 CRISPR transfection.