Muscarinic acetylcholine receptor M1 mutations causing neurodevelopmental disorder and epilepsy

De novo rare damaging variants in genes involved in critical developmental pathways, notably regulation of synaptic transmission, have emerged as a frequent cause of neurodevelopmental disorders (NDD). NDD show great locus heterogeneity and for many of the associated genes, there is substantial phenotypic diversity, including epilepsy, intellectual disability, autism spectrum disorder, movement disorders, and combinations thereof. We report two unrelated patients, a young girl with early‐onset refractory epilepsy, severe disability, and progressive cerebral and cerebellar atrophy, and a second girl with mild dysmorphism, global developmental delay, and moderate intellectual disability in whom trio‐based whole‐exome sequencing analysis uncovered de novo missense variants in CHRM1. Biochemical analyses of one of the NDD‐associated variants proved that it caused a reduction in protein levels and impaired cellular trafficking. In addition, the mutated receptor showed defective activation of intracellular signaling pathways. Our data strengthen the concept that brain‐reduced muscarinic signaling lowers the seizure threshold and severely impairs neurodevelopment.

diseases with onset within the first year of life where developmental impairment is recognized to occur as a direct consequence of the genetic mutation, in addition to the potential deleterious effect of epileptic activity on brain development (Scheffer et al., 2017). With the transition of next-generation sequencing to genetic diagnostics, the pace of DEE gene discovery has accelerated with over 100 currently known genetic etiologies (Brunklaus et al., 2020). Importantly, the genes underlying DEE are known to often cause a range of neurodevelopmental disorders (NDD) with or without epilepsy. Most of these genes encode proteins related to one of six fundamental processes: Ion transport, cell growth and differentiation, regulation of synaptic processes, transport and metabolism of small molecules within and between cells, and regulation of gene transcription and translation (Symonds & McTague, 2020).
Cholinergic signaling has been associated with the maintenance of cortical network excitability balance (Drever et al., 2011). To date, only mutations in the nicotinic acetylcholine receptor have been reported as disease-causing in genetic epilepsies (Steinlein et al., 1995). However, mutations affecting PLCB1, encoding phospholipase C isoform β1, which is placed downstream on the muscarinic cholinergic signaling pathway, have been associated with a DEE phenotype (Kurian et al., 2010) and so are mutations in the subunits of the CHRM1-regulated Kv7 channels (Jentsch, 2000).
In the present study, we identify two patients with point mutations in CHRM1, providing novel insight into the molecular mechanisms underlying DEE and neurodevelopmental impairment and further implicate defects in the cholinergic pathway in severe infantile epilepsies. Methods are described in the Supplementary section. The study was approved by the local Ethics Committee (PR(AG) 223/2017) and informed consent was obtained from the patient's parents according to the Helsinki declaration.
Case 1. The proband is a 10-year-old girl with epilepsy and severe encephalopathy. She was born at term to healthy, unrelated parents after uneventful pregnancy and delivery. After birth, she was noted to be hypotonic, irritable, and difficult to feed. At age 1.5 months, she developed tonic seizures featuring right arm extension and head version.
The reported physical exam showed no dysmorphic features or organomegaly, absent visual fixation, and global hypotonia. Electroencephalography (EEG) revealed multifocal epileptiform discharges and diffuse background slowing. Brain MRI (magnetic resonance imaging) (Figure 1a), array CGH, serum lactate, amino acids, sialotransferrin pattern, very-long-chain fatty acids, and biotinidase, urine organic acids, CSF folate, purines, neurotransmitter metabolites, and CSF/plasma glucose ratio were normal. Seizures persisted daily despite treatment with pyridoxine, phenobarbital, sodium valproate, or carbamazepine, and by age 4 months, manifested as a combination of tonic spasms, myoclonic jerks, and dysautonomic signs. At age 2 years 11 months, she displayed microcephaly with HC: 44.5 cm (−3 SD), right occipital plagiocephaly, and poor contact despite brief visual fixation and smiling in reaction to voice; other findings were horizonto-rotatory nystagmus, spastic quadriparesis, hyperreflexia, extensor plantar responses, and occasional axial or segmental myoclonia. The patient suffered on average two to three seizures per day, mostly generalized tonic, heralded by a flexor spasm and accompanied by oral automatisms and eyelid myoclonia, each lasting 1-5 min. EEG revealed a high-voltage background with multifocal spikes of occipitotemporal predominance.
Seizures proved refractory to treatment with different combinations of topiramate, lamotrigine, gabapentin, clobazam, levetiracetam, lacosamide, and eslicarbacepine acetate. A repeat MRI at age 4 years showed mild enlargement of subarachnoidal spaces, a relatively thick corpus callosum, and marked cerebellar atrophy with vermian predominance (Figure 1a).
Magnetic resonance spectroscopy showed the presence of a creatine peak. She currently remains wheelchair-bound, with little awareness of her surroundings (Figure 1c). Seizure frequency has been reduced to a few per week while on perampanel monotherapy. Case 2. This is a 14-year-old girl born at 28 weeks gestation to healthy, unrelated parents. Birth weight was 860 g, length 34 cm, and head circumference 24 cm. Despite prematurity, the neonatal period was relatively uneventful, did not require ventilation and serial cerebral ultrasounds were all normal. She showed some dysmorphic features including right hemifacial microsomia, congenital torticollis with plagiocephaly, high-arched palate, hypertelorism, and bilateral fifth finger clinodactyly. She was hypotonic early on and displayed feeding difficulties and axial titubation. Psychomotor development was globally delayed. She sat unassisted at age 2 years and walked after age 3 years.
On follow-up, she displayed general motor clumsiness, crouch gait, right esotropia, very limited language with right-sided sensorineural hypoacusia, and poor social skills. Chromosomal analysis excluded 22q.11 deletion syndrome and the metabolic screen was normal. Brain Muscarinic acetylcholine receptors are involved in signaling pathways related to adenosine 3ʹ,5ʹ-cyclic monophosphate (cAMP) and calcium intracellular release. CHRM1 is mainly known to signal through Gq/11-activating phospholipase C that increases IP 3 calcium-related pathways. However, cAMP signaling has also been associated with CHRM1. Reporter assays were performed in HEK293T cells to detect cAMP- (Figure 2c) or IP 3 /Ca 2+ -associated transcription (Figure 2d). Both reporters displayed a reduced signaling activation for the mutant compared with WT.
In our first patient, seizure onset was in early infancy, yet she dis- CHRM1 is a protein whose molecular weight is 52 kDa, but the mature protein increases its molecular weight due to glycosylation. Different protein band pattern was obtained corresponding to (1) mature glycosylated protein form, (2) and (3) immature forms. Cells transfected with the mutant showed an increase in forms 2 and 3 and a decrease in mature glycosylated CHRM1 compared to cells transfected with WT. β-Actin was used to normalize protein levels. Statistics were performed from five independent experiments. (b) HeLa cells cotransfected with WT or mutant CHRM1 and PH-GFP, were studied by immunofluorescence. Co-localization between CHRM1 (red label) and PH-GFP (green label) was analyzed (Merge) by Pearson´s correlation analysis. Statistics were performed from 3 independent experiments. (c, d) Signaling pathways associated to CHRM1 were studied by a luciferase signal assay in HEK293T cells transfected with WT or mutant CHRM1, after treatment with Carbachol, a CHRM1 agonist. (c) cAMP signaling pathway activation was studied through a CRE gene reporter. Negative controls and forskolintreated cells (positive control) were used to normalize results. (d) IP 3 -Ca 2+ signaling pathway activation was studied through an NFAT gene reporter. Negative controls and Ca+ PMA (phorbol myristate acetate)-treated cells (as a positive control) were used to normalize results. Statistics were performed from three independent experiments. cAMP, adenosine 3ʹ,5ʹ-cyclic monophosphate; GFP, green fluorescent protein; WT, wild-type epileptic and developmental disorder. Our second patient had earlyonset cognitive and motor impairment as well, but her phenotype is better covered under the more unspecific label of NDD. De novo mutations in variance-intolerant genes are a well-known cause of NDD, including DEE. In fact, patients with developmental disorders resulting from mutations in novel genes, as opposed to many classical conditions, are often phenotypically dissimilar and phenotype-driven recognition of these genetic defects is revealing rather unlikely (Kaplanis et al., 2020). In keeping, we have previously reported wide phenotype variability, from epileptic encephalopathy to NDD and autism spectrum disorder, in association with variants in other synaptic genes such as VAMP2 or GRIA2 (Salpietro, Dixon, et al., 2019;Salpietro, Malintan, et al., 2019).
Our findings hint at CHRM1 dysfunction as a possible novel cause of NDD. Additional evidence may come from the recent finding of a missense CHRM1 variant in a young boy included in an autism spectrum disorder cohort sequencing study (Satterstrom et al., 2020). Thus, a total of three putative pathogenic CHRM1 variants have been associated with neurodevelopmental phenotypes and all of them are assessed as diseasecausing by main in silico predictors (Table S1). It is also the first instance where the metabotropic muscarinic receptors are linked to an epileptic disorder. Admittedly, this is based only on the fact that we have identified one de novo mutation in a single patient. However, we believe that this mutation is pathogenic as it affects a conserved residue in all CHRM proteins that is present in an important structural element of the transmembrane segment 6 (TM6). The TM6 suffers a small rotation and an outward displacement during receptor activation that allows the G protein to engage the receptor core (Maeda et al., 2019). Mutation p.(Phe425Ser) may also partially destabilize the binding of the G11 protein, as it is very close to the residues cysteine 421 and asparagine 422, which have been involved in the binding of the G protein. Thus, it could be that both mutations impair receptor activation. In addition, expression of the Pro380Leu mutant protein in transfected cells at the membrane is reduced possibly due to a folding defect. We reasoned that, based on the fact that this mutant is expressed at low levels, it is very difficult to consider that the mutant protein may exert a dominantnegative effect. Rather, we support the hypothesis that the patient may suffer a reduction of cholinergic activity due to haploinsufficiency. This hypothesis is in agreement with the fact that a minor reduction of KCNQ activity, a known target of muscarinic regulation, is enough to cause an epileptic phenotype (Jentsch, 2000).
How a reduction in cholinergic activity in humans may lead to an early epileptic or developmental phenotype? One of the bestknown targets of muscarinic regulation is the M current formed by KCNQ channels through G q/11 mediated protein signals that increase phospholipase C-beta activity, which results in consumption of phosphatidylinositol 4,5-bisphosphate (PIP2) resulting in KCNQ inhibition. In this pathway, loss of function mutations in KCNQ channels or PLCB1 cause epileptic encephalopathy (Schoonjans et al., 2016). Thus, considering this well-known pathway, one simple hypothesis is that muscarinic inhibition might result in increased KCNQ activity, which may affect neuronal excitability properties. However, gain-offunction variants in KCNQ2 do not show epileptic seizures (Miceli et al., 2015). Thus, modulation of M current by CHRM1-PLCB1 might not be the only acting pathogenic mechanism.
CHRM1 might regulate many different targets including other ion channels depending on the developmental stage and in a cell-specific manner. It is noteworthy that, although CHRM1 is nearly not expressed in the cerebellum (Bakker et al., 2015), our patient 1 developed prominent cerebellar atrophy suggesting that a defective CHRM1-mediated cholinergic activity may have resulted particularly damaging for Purkinje cells. Patient 2 had also cerebellar vermian and pontine atrophy, while her moderate ventricular enlargement could conceivably relate to white matter injury in association with encephalopathy of prematurity.
In summary, our work further suggests that muscarinic activity in the brain might affect multiple processes regulating seizure susceptibility and neuronal development and therefore, CHRM1 can be proposed as a novel gene associated with DEE and NDD.