Epilepsy gene therapy using non-integrating lentiviral delivery of an engineered potassium channel gene

Refractory focal neocortical epilepsy is a devastating disease for which there is frequently no effective treatment. Gene therapy represents a promising alternative, but treating epilepsy in this way involves irreversible changes to brain tissue, so vector design must be carefully optimized to guarantee safety without compromising efficacy. We set out to develop an epilepsy gene therapy vector optimized for clinical translation. The gene encoding the voltage-gated potassium channel Kv1.1, KCNA1, was codon-optimized for human expression and mutated to accelerate the channels’ recovery from inactivation. For improved safety, this engineered potassium channel (EKC) gene was packaged into a non-integrating lentiviral vector under the control of a cell type-specific CAMK2A promoter. In a blinded, randomized, placebo-controlled pre-clinical trial, the EKC lentivector robustly reduced seizure frequency in a rat model of focal neocortical epilepsy characterized by discrete spontaneous seizures. This demonstration of efficacy in a clinically relevant setting, combined with the improved safety conferred by cell type-specific expression and integration-deficient delivery, identify EKC gene therapy as ready for clinical translation in the treatment of refractory focal epilepsy.

Introduction 44 45 Epilepsy affects over 60 million people worldwide (Ngugi et al., 2010). Even with optimal treatment 46 approximately 30% remain resistant to pharmacotherapy (Kwan et al., 2011;Picot et al., 2008). The 47 development of new anti-epileptic drugs in the last 20 years has had little impact on refractory 48 epilepsy; people with inadequately controlled seizures continue to experience major co-morbidities, 49 social exclusion, and an annual rate of sudden unexpected death in epilepsy (SUDEP) of 0.5-1% 50 (Devinsky, 2011;Hoppe and Elger, 2011). Although surgical resection of the epileptogenic zone can 51 result in seizure freedom, it is unsuitable for over 90% of refractory epilepsy patients (Lhatoo et al., 52 2003). Surgical intervention in focal neocortical epilepsy (FNE) is further complicated by the high risk 53 of damage to eloquent regions of the cortex involved in functions such as memory, language, vision 54 or fine motor control (Schuele and Lüders, 2008). People with FNE are therefore often left with very 55 few, usually palliative, treatment options, and there is an urgent need to develop alternative 56 therapies. 57 58 Gene therapy is one promising option , but major hurdles remain in achieving 59 stable, predictable and safe transgene expression with viral vectors. Because focal seizures often 60 arise from brain areas very close to eloquent cortex, lentiviral vectors, which generally lead to rapid, 61 stable and, most importantly, spatially-restricted transgene expression (Lundberg et al., 2008), are 62 an attractive delivery tool. In addition, the large packaging capacity of lentivectors allows a wide 63 choice of promoter-transgene combinations (Kantor et al., 2014), which can further increase the 64 specificity of expression. Hitherto, clinical trials with lentivectors for CNS disorders have been mainly 65 restricted to ex-vivo treatment of hematopoietic stem cells (Biffi et al., 2013(Biffi et al., , 2013Cartier et al., 66 2009). However, a recent trial using a lentivector injected directly into the striatum has 67 demonstrated safety and tolerability in Parkinson's disease, with evidence of decreased L-DOPA 68 requirement (Palfi et al., 2014). induced by tetanus neurotoxin (TeNT) injection into the rat motor cortex (Kätzel et al., 2014;Wykes 78 et al., 2012). In this model pathological high-frequency electrocorticographic (ECoG) activity is 79 prominent, but discrete seizures lasting over 20 seconds are rare. Lentiviral overexpression of the 80 human potassium channel Kv1.1, encoded by KCNA1, was highly effective at reducing pathological 81 high frequency activity (Wykes et al., 2012 We first asked whether the CMV-driven KCNA1 lentivector (CMV-KCNA1) used previously in a model 116 of EPC (Wykes et al., 2012) was also effective in an epilepsy model characterized by discrete 117 seizures. Epilepsy ( Fig. 1A; Supp. Fig. 1) was induced in adult rats with a single injection of TeNT into 118 the primary visual cortex. Seizures in this model typically last between 50 and 200 s, are 119 accompanied by unilateral, bilateral or generalized convulsions, and evolve over several weeks 120 before fading (Chang et al., 2018). To monitor local electrographic activity, a wireless ECoG 121 transmitter was implanted with a subdural intracranial recording electrode positioned above the 122 injection site. Two weeks after TeNT administration, following the establishment of epilepsy, animals 123 were randomized into two groups and injected via a pre-implanted cannula with either the CMV-124 KCNA1 lentivector or a control vector expressing only green fluorescent protein (GFP). Injections 125 were delivered directly into the seizure focus and followed by a further 4 weeks of ECoG recording 126 ( Fig. 1B). 127

128
The CMV-KCNA1 lentivector transduced neurons within a narrow column of the cortex (Fig. 1C). As is 129 typical of this model (Chang et al., 2018), the total number of seizures experienced by each animal 130 over the 6 weeks of recording was highly variable (Fig. 1D). Consequently, to compare seizure 131 frequency between the two treatment groups the numbers of seizures experienced each week were 132 normalized to the number experienced in the week preceding treatment (week −1, or baseline (Bl) 133 week). Despite the small sample size (6 treated vs. 5 controls), the CMV-KCNA1 lentivector 134 significantly reduced normalized seizure frequency compared to controls in the weeks following 135 treatment (generalized log-linear mixed model on weeks 0 -3, treatment*week interaction effect: 136 F(1,40) = 4.851, p = 0.033; Fig. 1E). The therapeutic effect emerged rapidly; plots of normalized 137 cumulative daily seizure frequency for the two groups diverged within 3 days of lentivector injection, 138 consistent with rapid transgene expression, as seen previously in the motor cortex model (Fig. 1F). 139 140 This pilot study strongly suggests that KCNA1 gene therapy can suppress spontaneous discrete 141 seizures. However, the CMV-KCNA1 lentivector tested is poorly suited for clinical translation. We 142 therefore set out to develop an optimized vector with improved safety and efficacy. The transfer plasmid used to synthesize the optimized lentivector differed from the original CMV-147 KCNA1 construct in several ways ( Fig. 2A). The non-cell type specific CMV promoter was replaced 148 with a 1.3 kb human CAMK2A promoter to bias expression to excitatory neurons (Dittgen et  To test the therapeutic efficacy of the EKC lentivector, we designed a blinded, randomized, placebo-174 controlled pre-clinical trial, and selected normalized seizure frequency as the primary outcome 175 measure. Eleven days after injection of TeNT into the visual cortex, 26 rats were randomized into 176 two groups and injected via a pre-implanted cannula with either the EKC lentivector or its dscGFP-177 only control. ECoG recordings were continued for a further 4 weeks. The timeline was altered from 178 that of the pilot study to treat after 11 days in order to capture the period when seizure activity is at 179 its highest (2 -4 weeks following TeNT injection) (Fig. 3A). 180 181 To minimize the confounding influence of animals that displayed a very low seizure frequency prior 182 to treatment, subjects were excluded if they exhibited fewer than five seizures in the week 183 preceding lentiviral delivery (the baseline week). This criterion, applied before unblinding, led to the 184 exclusion of eight animals (6 EKC, 2 control). Of the remaining 18, all but one survived for the 185 duration of recording. This rat (from the EKC group) was culled in the final week due to detachment 186 of its headpiece. However, because the subject had already passed through the period of peak 187 seizure activity, and in order to maximise the amount of data obtained from the study, this 188 incomplete dataset was included in the overall analysis. Again, this decision was made before 189 unblinding. We have previously shown that overexpression of Kv1.1 can reduce the frequency of brief (< 1 s), 210 high-frequency epileptiform discharges in a motor cortex TeNT model of EPC (Wykes et al., 2012). 211 However, this study did not investigate whether Kv1.1 overexpression could inhibit discrete seizures 212 lasting 1 -2 minutes, more typical of common forms of focal epilepsy. We show here, in two 213 independent trials, that Kv1.1 overexpression is indeed sufficient to reduce the frequency of 214 seizures, although interestingly seizure duration was not altered. This is most simply explained by 215 proposing that seizure initiation is rapidly accompanied by propagation to other brain areas, beyond 216 In the case of epilepsy, an additional safety concern is the possibility of potassium channel 239 overexpression in interneurons, which could aggravate seizure activity by exacerbating rather than 240 attenuating local excitability. To mitigate this risk we have used a human CAMK2A promoter that in 241 rats led to very little expression in GABAergic cells. Promoter specificity can differ between species 242 (Lerchner et al., 2014;Yaguchi et al., 2013), and the specificity of the human CAMK2A promoter for 243 excitatory glutamatergic neurons will ultimately need to be validated in the human brain. Evidently, 244 if EKC gene therapy is to progress to the clinic, such validation will need to be performed in the 245 absence of a fluorescent reporter. Data from the pilot study were used to determine sample sizes for the final pre-clinical trial. We 380 estimated that the maximal weekly seizure frequency would double from baseline, and we wished to 381 detect with 80% power a 40% reduction from this maximum at p < 0.05. Given a mean baseline 382 weekly seizure frequency of 5 or above, a modification of Lehr's formula (Lehr, 1992) for the Poisson 383 distribution suggested 7 -8 animals per group would be sufficient to detect a reduction in seizure 384 frequency from 10 to 6 per week. Our modified Lehr's formula is given by the following equation: 385 where n is the size of each sample (treatment group), λ 1 the mean weekly seizure frequency before 388 treatment, and λ 2 the mean weekly seizure frequency after treatment. 389 390 Efficacy of treatment data (Fig. 1E, 3B) were analysed using a generalized log-linear mixed model 391 with random effect of animal (autoregressive covariance) and fixed effects of treatment group, 392 week, and the interaction between treatment group and week. Seizure counts in the week preceding 393 treatment were compared using a Mann Whitney U test. Current densities at +20 mV (Fig. 2Bii) were 394 compared using a Welch's one-way ANOVA followed by Games-Howell post-hoc tests.  . Normalized cumulative seizure frequency (per day) in panels B and C are presented as mean ± SEM.

KC gene therapy robustly reduces seizure frequency in a blinded
A. Timeline highlighting key experimental milestones. Note the injection of lentiviral vectors 11 days Normalized seizure frequency (per week) for red; n = 11).
. Normalized cumulative seizure frequency (per day) are presented as mean ± SEM.

KC gene therapy robustly reduces seizure frequency in a blinded
A. Timeline highlighting key experimental milestones. Note the injection of lentiviral vectors 11 days Normalized seizure frequency (per week) for animals treated with the . Normalized cumulative seizure frequency (per day).
are presented as mean ± SEM.