Reversal of behavioural phenotype by the cannabinoid-like compound VSN16R in fragile X syndrome mice

Abstract

Fragile X syndrome is the most common inherited intellectual disability and mono-genetic cause of autism spectrum disorder. It is a neurodevelopmental condition occurring due to a CGG trinucleotide expansion in the FMR1 gene. Polymorphisms and variants in large-conductance calcium-activated potassium channels are increasingly linked to intellectual disability and loss of FMR protein causes reduced large-conductance calcium-activated potassium channel activity leading to abnormalities in synapse function.

Using the cannabinoid-like large-conductance calcium-activated potassium channel activator VSN16R we rescued behavioural deficits such as repetitive behaviour, hippocampal dependent tests of daily living, hyperactivity and memory in a mouse model of fragile X syndrome.

VSN16R has been shown to be safe in a phase 1 study in healthy volunteers and in a phase 2 study in patients with multiple sclerosis with high oral bioavailability and no serious adverse effects reported. VSN16R could therefore be directly utilized in a fragile X syndrome clinical study. Moreover, VSN16R showed no evidence of tolerance, which strongly suggests that chronic VSN16R may have great therapeutic value for fragile X syndrome and autism spectrum disorder.

This study provides new insight into the pathophysiology of fragile X syndrome and identifies a new pathway for drug intervention for this debilitating disorder.

Fragile X syndrome is a genetic condition characterized by autism-like features. Hurley et al. report that the cannabinoid-like compound VSN16R can rescue behavioural deficits in an animal model of the disorder through activation of an ion channel that is located in parts of the brain affected by fragile X syndrome.

Introduction

Fragile X syndrome (FXS) is an inherited neurodevelopmental condition characterized by intellectual disability, hyperactivity/hyperkinesis, attention deficits, social difficulties, anxiety, depression, irritability, mania, obsessive-compulsive behaviour, aggression, self-injury and autistic-like behaviours.¹ It is caused by a CGG trinucleotide expansion in the fragile X mental retardation 1 (FMR1) gene on the X chromosome at locus Xq27.3 that leads to lowered expression of fragile X mental retardation (FMRP) protein.^2–5 FMRP is essential in neurodevelopment and for normal synaptic function in the adult brain.⁶ There is currently no specific cure or effective disease-modifying treatment for FXS.

FMRP interacts with large-conductance voltage and calcium-activated potassium (BK) channels to control their expression and function.⁷ BK channels consist of alternatively spliced isoforms of four pore-forming, voltage- and calcium-sensing α-subunits (KCNMA1) encoded by a single gene usually in association with accessory transmembrane β- (KCNMB1–4) and γ-subunits, which change the kinetic, pharmacological and membrane-trafficking properties of the BK channel complex.^8,9 BK channels are widely expressed on neuronal and smooth muscle cells and undergo complex post-translational modification.¹⁰

Laumonnier et al.¹¹ first reported reduced BK channel function in autism and intellectual disability in a subject with a haploinsufficiency in KCNMA1-α. In FXS, the first indication of a reduction of BK channel activity was reported by Tang et al.¹² who showed that protein levels of Kcnma1-α in Fmr1 knockout (KO) synaptic fractions were reduced by ∼50%, resembling the haploinsufficiency reported by Laumonnier et al.¹¹ Hébert et al.¹³ showed that the BK channel opener BMS-204352 rescued the FXS phenotype in Fmr1 KO mice. The range of BK channel polymorphisms affecting intellectual disability and FXS have recently been expanded and reported by many different groups.⁷

VSN16R is an orally active cannabinoid-like compound that activates BK channels in the brain and is well tolerated, has high oral bioavailability and no adverse neurobehavioural effects in rodents or humans.¹⁴ VSN16R was shown to be safe in a phase 1 study of healthy individuals¹⁵ and completed a phase II trial (NCT02542787) for spasticity.¹⁶ VSN16R showed no evidence of tolerance and may have therapeutic value for FXS and autism spectrum disorder, so this study sought to evaluate the effect of chronic treatment with VSN16R on the FXS-like behavioural phenotype of Fmr1 KO2 mice that exhibit many of symptoms found in patients with FXS.¹⁷

VSN16R rescued behavioural deficits caused by FMRP loss such as repetitive behaviour, hippocampal dependent tests of daily living, hyperactivity, memory impairment, stereotypy and aggression. In addition, we collated the existing data on polymorphisms and intellectual disability in the BK channel and mapped these onto the BK channel cryo-electron microsopy (cryo-EM) crystal structure. Published gene expression data for the BK channel show that the isoforms of the BK channel activated by VSN16R are found in areas of the brain that are associated with FXS. This supports the role of BK channels in the cause of FXS symptoms and modulation of BK channel function as a therapy for treatment of FXS.

Materials and methods

Animals

Fmr1 KO2 mice supplied by the FRAXA Research Foundation were used in this study. Experiments were conducted in line with the requirements of the UK Animals (Scientific Procedures) Act, 1986 and were approved by the ethics committee of the Institute of Ecology and Biodiversity, Faculty of Sciences, University of Chile.

Animal housing

Mice were housed in groups of five in plastic cages (35 × 30 × 12 cm) in a controlled environment [21 ± 2°C, relative humidity (55 ± 5%), 12-h light–dark cycle (lights on 7 a.m. to 7 p.m.) and air exchange 16 times per h] with free access to commercial food pellets and water.

Treatment groups

Before chronic dosing with VSN16R (R, Z)-3–[6-(dimethyl-amino)-6-oxohex-1-en-1-yl)-N-(1-hydroxypropan-2-yl)-benzamide (Canbex Therapeutics Ltd), 2 mg/kg i.v., once a day, in 0.9% NaCl, 4 weeks], mice were challenged with a single intravenous dose of VSN16R and monitored for adverse behavioural or physiological effects. VSN16R was well tolerated by Fmr1 KO2 and wild-type mice consistent with other toxicological studies.¹⁵ Following chronic dosing, 14-week-old Fmr1 KO2 and wild-type mice were tested within 30 min of VSN16R administration in the following groups: wild-type + vehicle; wild-type + VSN16R; Fmr1 KO2 + vehicle; and Fmr1 KO2 + VSN16R (n = 10).

Behavioural tests

Behavioural testing was conducted between 8 a.m. and 4 p.m. as previously described.¹⁸ Hence any effects of the BK channel on circadian pacemaker control would be the same for all animals used in the study. Mice were randomly assigned to treatment groups. The behavioural experimenter was blind to genotype, drug treatment and subsequent data analysis. Mice were tested in one behavioural task on each experimental day and each behavioural test was separated by 3 days.

Open field test

The open field test was used to determine hyperactivity and habituation to a novel environment—one of the most elementary forms of learning—in which decreased exploration as a function of repeated exposure to the same environment (an enclosed arena 50 × 30 cm divided into 10 cm squares) was taken as an index of memory. Testing consisted of an initial exposure (T1), then at 10 min (T2) to test short-term memory and after 24 h (T3) to assess long-term memory.

Contextual fear conditioning

Fear conditioning to a context tests associative learning. Mice were given a 120-s habituation period in the apparatus before the first of two identical trials (210 s apart) to allow exploration of the chamber. An 80 dB auditory cue was then presented (15–30 s) with a mild foot shock (0.6 mA, 1 s) administered during the last 2 s of the tone presentation, that co-terminated with the tone. Memory was tested 24 h after training for 5 min.

Marble burying

Transparent plastic cages were filled (10 cm deep) with sawdust, on top of which 10 glass marbles were placed in two rows. Each animal was left undisturbed in the cage for 30 min, after which the number of marbles buried to at least two-thirds of their depth was recorded.

Self-grooming

Mice were placed individually in a cage (46 × 23.5 × 20 cm) illuminated at 40 lx. After a 5-min habituation period the time spent grooming was recorded for 3 min.

Aggression

An experimental ‘test’ and wild-type ‘control’ mouse were placed in the testing cage simultaneously. The latency to attack was recorded.

BK channel and FMRP modelling and polymorphism mapping

Numerous BK channel polymorphisms are linked with intellectual disability (Supplementary Table 1). A molecular operating environment modelling package (Chemical Computing Group) was used to analyse the BK channel cryo-EM structure¹⁹ and visualize the location of polymorphisms.

BK channel expression in rodent and human brain

Expression data were obtained from public repositories to compare gene expression of FRMP and the BK channel isotypes in human and rodent brain (Supplementary Figs 1 and 2).

Data analysis

Data were analysed using SPSS™ statistics software v.27 (IBM) and visualized with Prism™ v.9 (GraphPad).

Data availability

All data generated or analysed and used in this study are available upon reasonable request.

Results

Behavioural tests

Open field

The open field trial 1 (T1) was performed to characterize hyperactivity in Fmr1 KO2 and wild-type littermates habituated to daily handling under novelty and familiar conditions. Fmr1 KO2 mice injected with vehicle displayed a significant increase in total distance travelled (a parameter for hyperactivity) compared to the wild-type vehicle-control group. In contrast, the Fmr1 KO2 mice injected with VSN16R had a significantly reduced total distance travelled compared to the Fmr1 KO2 mice injected with vehicle and displayed similar activity to the wild-type control group treated with vehicle or VSN16R that indicated chronic VSN16R treatment decreased spontaneous hyperactivity occurring in Fmr1 KO2 mice (Fig. 1A).

Figure 1.

Behavioural effects of VSN16R on Fmr1 KO2 mice. Wild-type (WT) or Fmr1 KO2 (KO) adult mice were treated with vehicle (V) or VSN16R for 4 weeks and then subjected to a range of behavioural tests. VSN16R treatment caused significant improvement in (A) hyperactivity, (B) short term memory and (C) long-term memory in Fmr1 KO2 mice when compared to wild-type littermates in the open field test. (D) VSN16R rescued impaired learning and memory of the Fmr1 KO2 mice in the contextual fear conditioning test. (E) VSN16R reversed impaired marble burying activity, a behaviour analogous to activities of daily living in humans, of the Fmr1 KO2 mice when compared to wild-type littermates. (F) Self-grooming, a measure of stereotypy that occurs in Fmr1 KO2 mice, was restored to levels found in wild-type animals by chronic VSN16R treatment and (G) the increased aggression that occurs in Fmr1 KO2 mice was reversed by chronic VSN16R treatment. Data were visualized as box plots with interquartile ranges, presenting all the data-points. The distribution of behavioural data was assessed with a normality test. Data with a normal distribution (Shapiro-Wilk, P > 0.05 = normal distribution) were analysed by univariate general linear model one-way ANOVA (F) followed, where appropriate, by a Bonferroni corrected multiple pair-wise comparison. Marble burying failed normality testing and was analysed by the Kruskal-Wallis one-way ANOVA by ranks (H) with Bonferroni corrected multiple pair-wise comparisons. n = 10 mice per group, *P < 0.05; ***P < 0.001.

The open field trial 2 (T2) test assessed short-term memory. All groups other than the vehicle-treated Fmr1 KO2 had a lower activity, indicating habituation (i.e. memory of the environment), than was observed in the T1 test. Vehicle-treated Fmr1 KO2 mice exhibited significantly increased locomotor activity compared to wild-type vehicle-treated mice indicating that they did not remember having explored the open field before and therefore had a short-term memory problem. The increased locomotor activity was significantly reduced (i.e. short-term memory deficit was corrected) in the Fmr1 KO2 mice injected with VSN16R compared to the vehicle-treated Fmr1 KO2 mice to a level similar to that occurring in the wild-type animals (Fig. 1B).

The open field trial 3 (T3) test assessed long-term memory of the open-field environment. The exploratory behaviour was again lower in all groups apart from the vehicle-treated group Fmr1 KO2 group that showed significantly less habituation (i.e. more activity), indicative of reduced long-term memory of the environment (Fig. 1C).

Contextual fear conditioning

Freezing in response to an aversive stimulus (a measure of associative learning) was significantly lower in Fmr1 KO2 mice treated with vehicle compared to wild-type animals treated with vehicle or VSN16R. Chronic treatment with VSN16R significantly reduced the fear conditioning response in Fmr1 KO2 mice, although it was still significantly less than the response seen in wild-type animals (Fig. 1D).

Marble burying

Marble burying (analogous to activities of daily living in humans) was significantly reduced in Fmr1 KO2 compared to wild-type animals and chronic VSN16R treatment significantly restored this behaviour (Fig. 1E).

Stereotypy

Self-grooming behaviour was significantly higher in Fmr1 KO2 mice compared to wild-type animals and was reduced in Fmr1 KO2 mice treated with VSN16R to a similar level to that found in wild-type animals (Fig. 1F).

Aggression

The time before attack was significantly shorter in Fmr1 KO2 mice compared to wild-type animals, which indicated increased aggression and was increased in Fmr1 KO2 mice treated with VSN16R to a similar level to that found in wild-type animals (Fig. 1G).

BK channel and FMRP modelling and polymorphism mapping

The location of polymorphisms in the BK channel^11,19 and FMRP²⁰ that give rise to a clinical phenotype are shown in Fig. 2.

Figure 2.

Cryo-EM structure of KCNMA1/KCNMB4 complex. Cryo-EM structure of KCNMA1/KCNMB4 complex showing gain-of-function and loss-of-function variants/polymorphisms. (A) Side on view and (B) top view looking from outside of plasma membrane, of channel with main chain only shown as a ribbon. The recently disclosed cryo-EM structure of the BK channel complexed with the β4 subunit allowed us to map the BK channel variants (Supplementary Table 1) onto the channel structure of the KCNMA1/KCNMB4 complex (hsSlo1-beta4, pdb 6v22).¹⁹ This revealed that most variants were in the calcium-sensing cytoplasmic domains of the tetramer though some loss-of-function variants were present close to the K⁺ channel itself. Most loss-of-function polymorphisms and variants (red) were linked to intellectual disability while gain-of-function changes (green) were more often linked in epilepsy and other movement disorders. Some of the reported changes have not been analysed using electrophysiology and these are shown in black. The structure of the N-terminal domain of FMRP was also reported, together with the R138Q variant linked to a FXS-like disorder. Mapping this variant onto the FMRP protein colours KCNMA1 blue to cyan; KCBMB4 black. Variants shown as space-filling spheres, red = loss-of-function; green = gain-of-function; black = not determined. Co-ordinates taken from pdb 6v22. (C) FMRP protein showing the position of R138Q, coordinates from pdb 4qw2. Mapping this variant onto the FMRP protein (Fig. 1C) revealed it to be in a highly accessible position of the molecule.²⁰ KCNMA1 variants were searched using the term ‘KCNMA1’ and a citation search on the early paper.¹¹ A similar process was followed for the FMRP structure (pdb 4qw2). Only clinically relevant variants were included. The consensus sequence Q12791-1 was used. Many variants of the BK channel sequence appear in the literature and sometimes a given variant may have a different sequence number. The sequence context is shown in Supplementary Table 1 to avoid ambiguity. Modelled from Tao and MacKinnon.¹⁹

Discussion

The behavioural findings described in this study provide direct evidence that chronic treatment with VSN16R, a selective activator of BK channels, can rectify the hyperactivity, short-term and long-term memory deficits and reduce stereotypy and aggression that occur in the Fmr1 KO2 mouse model of FXS. This effect is likely to be due to activation of BK receptors on neurons in brain regions involved in these behaviours (as opposed to smooth muscle) and the in silico analysis of published BK receptor expression studies supported this view (Supplementary Figs 2 and 3). Consequently, the drug may also be of use in patients with FXS as there is much overlap with expression of the BK receptor between mice and humans.

A range of tests were used to explore the effect of VSN16R on FXS-like behaviours in the mouse model of the disorder. Marble burying behaviour in mice is analogous to activities of daily living in humans. These impairments are frequently more of a problem to the patient than the loss of more complex cognitive abilities.²¹ VSN16R treatment restored marble burying to control levels in Fmr1 KO2 mice. There is debate on whether this test measures anxiety or obsessive-compulsive disorder-like behaviours and hence, which symptoms it relates to in patients with FXS. However, it does show sensitivity to anxiolytic compounds and drugs that affect the serotonergic system and since Fmr1 KO2 mice do less marble burying, indicating less repetitive (obsessive-compulsive like) actions in this test, it is likely that marble burying behaviour is more a measure of anxiety than obsessive-compulsive behaviour.²² However, Bhattacharya et al.²³ reported increased marble burying in Fmr1 KO mice, which suggested an opposite action. Marble burying is a species-typical behaviour that has been shown to be sensitive to animal species, strain and hippocampal lesions; thus, differences between Fmr1 KO and Fmr1 KO2 models or subtle differences in the methods employed (e.g. marble burying depth) could affect this test outcome.

Hyperactivity is a confounding factor in the Fmr1 KO mouse models; nevertheless, fear conditioning has an amygdala and hippocampal element. VSN16R reduces Fmr1 KO2 locomotor activity but also improves hippocampal-dependent activities of daily living, indicating that it has a role in improving hippocampal functions, like memory, beyond simple improvement of locomotor tasks. Fmr1 KO2 mice show a significant deficit in contextual fear conditioning memory to an aversive stimulus compared with wild-type littermate mice and treatment with VSN16R before acquisition improved long-term memory retention 24 h after training. VSN16R could have improved memory retention by enhancing memory acquisition, memory consolidation or both processes. However, this effect could have been due to reductions in hyperactivity, as were seen in the open field tests, although it is important to take into consideration that VSN16R also improves hippocampal-dependent activities of daily living, indicating that it has a role in improving hippocampal functions, like memory, beyond simple improvement of locomotor tasks.

Fmr1 KO2 mice also showed increased self-grooming compared with wild-type littermates. One possible interpretation is that Fmr1 KO2 animals are more prone to repetitive action patterns (i.e. stereotypy). VSN16R decreased stereotyped self-grooming behaviour in Fmr1 KO2 mice suggesting that this task could be more sensitive to the drug. This finding seemingly contradicts the observation of reduced marble burying in Fmr1 KO2 described above. However, previous hippocampal lesion studies showed that lesioned mice displayed reduced marble burying behaviour with little change in self-grooming, which suggests that these behaviours are independently mediated.²²

A recent study by Wheeler et al.²⁴ reported most patients with FXS surveyed were aggressive in the previous 12 months and self-injury and impulsive behaviour are also more prevalent in individuals with FXS.²⁵ Aggressive behaviour was significantly increased in Fmr1 KO2 mice and was reversed by chronic VSN16R treatment.

Studies examining expression of BK channels in FmR1 KO mice are limited. RNA-sequencing has been performed on Fmr1 KO mice and differential Kcnmb4 neuronal expression was found in some neuronal populations, such as cortical neurons,^26,27 but not others such as in the CA1 region of the hippocampus.^28,29

We collated the published KCNMA variants and polymorphisms known to influence intellectual disability and mapped them onto the KCNMA/KCNMB4 cryo-EM structure revealing that most variants were in accessible cytoplasmic regions and therefore likely to adversely affect BK channel function (Fig. 2 and Supplementary Table 1). The sequence context of variants is informative from a structural perspective for design of novel drugs and understanding the functional significance of polymorphisms. For example, the A138V variant—when viewed in the context of the INNGSSQADG sequence—introduces a large (valine) amino acid change in an otherwise polar region. Dysregulated interaction with FMRP, as highlighted by the R138Q FMRP variant, may disrupt BK channel function and an effect on intellectual disability can result. Unfortunately, the details of the interaction between BK channel and FMRP are not known at the present time. The distribution of BK channels in the brain was consistent with VSN16R-inhibiting spasticity, which is associated with stress-induced hyperexcitation of glutamatergic nerves.¹⁵ This suggested the potential for control of autism spectral disorders and other conditions, that are likewise associated with an imbalanced extra- and intra-neuronal environment that favours excitation over inhibition and glutamate-mediated hyperexcitability and can be regulated by BK channels.^13,27,30,31

Importantly, VSN16R showed no evidence of tolerance after chronic treatment, which has been a problem with other drug classes used as putative treatments for FXS in patients and models of the disorder.³² This, combined with the effective reversal of the FXS behavioural phenotype in the Fmr1 KO2 mice means VSN16R may be of benefit in treating FXS and autism spectrum disorder-related behaviours and is therefore a strong candidate for future clinical trials.

Abbreviations

BK channel

large-conductance voltage and calcium-activated potassium channel;

FXS

fragile X syndrome