Congenital myasthenic syndrome in Golden Retrievers is associated with a novel COLQ mutation

Abstract Background Congenital myasthenic syndromes (CMSs) are a group of inherited disorders of neuromuscular transmission that may be presynaptic, synaptic, or postsynaptic. Causative mutations have been identified in 4 breeds including the Labrador Retriever, Jack Russell Terrier, Heideterrier, and Danish Pointing Dog. Hypothesis/Objective Clinical and genetic characterization of a neuromuscular disorder in Golden Retriever (GR) puppies. Animals Four GR puppies from California were evaluated for generalized muscle weakness beginning at weaning. Biological specimens were collected from the affected puppies, and familial information was obtained. Blood or buccal swabs were obtained from 63 unaffected GRs. Methods Complete physical, neurological, electrodiagnostic, and histological evaluations and biochemical quantification of muscle acetylcholine receptors were performed. Polymerase chain reaction was used to amplify the 17 exons of COLQ, and sequences were obtained by Sanger sequencing. Variant frequency was assessed in unrelated GRs and a public database. Results Clinical, neurological, and electrodiagnostic evaluations confirmed a disorder of neuromuscular transmission in a GR family. Sequencing of all exons and splice sites of a primary candidate gene, COLQ, identified a point mutation that predicts an amino acid substitution (G294R). The primary COLQ transcript was absent from affected muscle samples. All affected puppies were homozygous for the mutation, which was not detected outside this GR family or in other breeds. Conclusions and Clinical Importance We confirmed the diagnosis of a CMS in GR puppies and identified a novel COLQ mutation. The COLQ gene encodes the collagenous tail of acetylcholinesterase, the enzyme responsible for termination of skeletal muscle contraction by clearing acetylcholine at the neuromuscular junction. Clinicians and breeders should be aware of this CMS in GR puppies with an early onset of weakness.

Conclusions and Clinical Importance: We confirmed the diagnosis of a CMS in GR puppies and identified a novel COLQ mutation. The COLQ gene encodes the collagenous tail of acetylcholinesterase, the enzyme responsible for termination of skeletal muscle contraction by clearing acetylcholine at the neuromuscular junction. Clinicians and breeders should be aware of this CMS in GR puppies with an early onset of weakness. Fox terriers 8,9 and English Springer Spaniels, 10

| Animals
Four related GR puppies, 2 males and 2 females, ranging from 3 to 5 months of age, were evaluated. All puppies were from the same breeder in Southern California. Parentage information and pedigrees could be obtained for only 2 of the affected puppies. Unaffected GRs (n = 63) had no known relationships with this family.

| Electrodiagnostic testing
One affected pup was anesthetized, and complete neuromuscular examinations were performed including electromyography (EMG), measurement of motor and sensory nerve conduction velocities, and measurement of the compound muscle action potential (CMAP) after repetitive nerve stimulation at 1, 3, 10, and 50 Hz using a Nicolet Viking Select EMG/evoked potential system (Nicolet, Biomedical Inc, Madison, WI). Insulated stainless steel needle electrodes were used for both nerve stimulation and recording from muscle, and a platinum subdermal electrode (Grass-Telefactor) was employed as a ground. Motor nerve conduction velocity of the peroneal and ulnar nerves was determined by dividing the distance between proximal and distal stimulation sites by the difference in latency of the corresponding CMAP recorded from the extensor digitorum brevis 11 and palmar interosseous 12 muscles, respectively, after supramaximal stimulation (2 Hz stimulus rate, 0.2 millisecond stimulus duration). Sensory nerve conduction studies were performed on the peroneal, ulnar, and radial nerves using previously described techniques. 13,14 Amplitude (peak-to-peak) was measured from CMAPs, and percentages of decrement were calculated for each repetition rate. Muscle (vastus lateralis, triceps brachii, and cranial tibial) and nerve (peroneal nerve) biopsy specimens were collected from the side opposite that used for electrodiagnostic testing.

| Sample collection and nucleic acid isolation
Muscle biopsy specimens were collected from an affected GR puppy, and whole blood was obtained from all 4 affected puppies. Blood or buccal swabs were obtained from unaffected GRs. Informed owner consent was obtained before collection of all biologic samples, in accordance with protocols approved by the Clemson University Institutional Review Board (IBC2008-17). Owner consent was obtained for necropsy and postmortem tissue collection from 1 affected puppy at University of California, Davis.
Genomic DNA was isolated from blood and buccal swabs using Gentra Puregene Blood Kit (Qiagen, Germantown, Maryland) and from muscle using the DNEasy Kit (Qiagen). Ribonucleic acid was extracted from snap frozen muscle tissue from an affected dog using the ToTALLY RNA kit (Ambion, Foster City, California). Canine testis RNA was obtained from Zyagen (San Diego, California). All RNAs were treated with the Turbo DNA-free kit (Invitrogen, Carlsbad, California), and cDNAs were synthesized using an oligo dT primer and the RevertAid First Strand cDNA Synthesis kit (Thermo Scientific, Pittsburg, Pennsylvania).

| Sequencing and genotyping
Polymerase chain reactions (PCRs) to amplify the 17 exons of genomic COLQ were performed using previously described primers and protocols. 5 Primers to amplify the last 7 exons of the COLQ transcript were designed with the forward primer (5 0 GGGCAGAAAGGTGAAATGGGT) spanning the exon 10 and 11 junction and the reverse primer located in exon 17 (5 0 ATGTGAAGTAGCGGCAGGAC). A constitutively expressed gene (PSMB7) was used to ensure that cDNA was present. Amplicons were reamplified using band-stab PCR. Products were purified using an domestic dogs, including 20 purebred GR, and 54 wild canids. 16 In silico programs, PolyPhen-2 17 and CPD prediction tool, 18 were used to assess the impact of the variant.

| AChR quantification and antibody-bound AChR
Acetylcholine receptors were extracted from an external intercostal muscle specimen collected at necropsy from an affected puppy using a modification of a previously described procedure. 19 The muscle specimens were stored at −70 C before homogenization and extraction of AChR in 2% Triton X-100. Solubilized AChRs were labeled by incubation with an excess of 125 I-α-bungarotoxin ( 125 I-αbgt) followed by sequential addition of high titer rat-anti-AChR antibody and precipitation with goat anti-rat immunoglobulin G (IgG). The precipitate was Quantitative serum AChR antibody concentrations were determined as previously described using an immunoprecipitation radioimmunoassay procedure. 19 3 | RESULTS Laryngeal examination confirmed laryngeal paralysis, and a left laryngeal tieback surgery was performed. Two days after the surgery, the dog developed aspiration pneumonia and died. Of the remaining puppies, 1 puppy was lost to follow-up at 7 months of age, and 2 puppies were euthanized at 6 months of age. Necropsy was performed on the latter 2 puppies, and no histological abnormalities were found in peripheral nerves, muscle, brain, spinal cord, or other organs.

| Electrodiagnostic testing
Abnormalities were restricted to a mild decrease in the amplitude of the M wave, and a decremental response of the CMAP after repetitive nerve stimulation. These findings were consistent with a disorder of neuromuscular transmission (Figure 1). Biopsy specimens from the vastus lateralis, triceps brachii, and cranial tibial muscles were collected under anesthesia for histological evaluation.

| Muscle histology, histochemistry, and fluorescence microscopy
No specific pathological changes were identified in the muscle or peripheral nerve biopsy specimens that would support an underlying myopathy or neuropathy. The esterase reaction identified several F I G U R E 1 A series of recordings obtained using various repetition rates (1, 3, 10, and 50 Hz) during a repetitive nerve stimulation study in one affected dog. A 1-minute rest period was utilized between each train of stimuli. All frequencies resulted in decremental responses identical to those associated with myasthenia gravis (a change in amplitude of 10% or greater between the first and subsequent responses). This was the case even with a tetanizing stimulus of 50 Hz when pseudofacilitation (an incremental response in amplitude with little or no change in area) is a typical finding in normal dogs abnormally large motor end-plates in the affected puppy compared to control end-plates ( Figure 2). Using fluorescent labeled α-bungarotoxin, staining of motor end-plates for AChRs was decreased but not absent in the affected puppy compared to control muscle (Figure 2).

| AChR quantification
The AChR concentration from 1 affected GR puppy was determined in external intercostal muscle samples collected origin to insertion after euthanasia. The AChR concentration was decreased at 0.11 pmol/g tissue (reference range, 0.2-0.4 pmol/g tissue). Antibodies were not detected against AChRs in muscle or in serum.

| Sequencing and genotyping
Because 1 affected puppy had a decreased (but not absent) concentration of AChRs at the motor end-plates and did not improve after IV administration of edrophonium chloride, we hypothesized that CMS could be the result of an end-plate AChE deficiency and prioritized COLQ as a candidate causal gene. Evaluation of pedigree information from 2 affected puppies identified extensive inbreeding (Figure 3), F I G U R E 2 Esterase reaction for localization of motor end-plates and fluorescent labeled α-bungarotoxin for localization of acetylcholine receptors (AChRs) in cryosections of muscle from a GR with CMS and archived control muscle. Esterase staining showed enlarged motor endplates (arrows with long tail) compared to control muscle (arrows with short tail). AChRs were decreased (red fluorescence) but not absent in the affected puppy compared to control muscle, consistent with the biochemical findings. Dapi staining highlights muscle nuclei. Bar in lower right figure = 50 μm for all images F I G U R E 3 Pedigree for two affected puppies (shaded) shows extensive inbreeding. The affected puppies share a sire that is also the grandsire on the maternal side for both puppies suggesting homozygosity for a recessive allele inherited identically by descent. Sequences from 1 affected GR indicated that 16 of 17 COLQ  Figure 4A) in a position conserved across placental mammals ( Figure 4B). PolyPhen-2 predicts the variant to be benign likely because arginine is the wild-type residue in some organisms (eg, wallaby, chicken). A second in silico prediction program indicates that the arginine may be pathogenic in dogs but compensated by additional COLQ amino acid changes in other organisms (8.1% likelihood). 18 The G>A transition changes a GG to AG within exon 13 that could F I G U R E 4 A, Chromatogram from a small section of COLQ exon 16 from a control GR (top) and an affected GR (bottom). The G>A mutation predicting a Gly>Arg amino acid change is indicated with a red asterisk (*). B, A partial amino acid sequence of COLQ surrounding the predicted substitution is shown from several species. The Glycine (G) that is changed to an Arginine in the affected GR puppies is shown in bold and is conserved across mammals but not chicken or wallaby F I G U R E 5 Agarose gel electrophoresis of cDNA amplicons from COLQ and a housekeeping control. The muscle of an affected GR (M) lacks the primary 748 bp COLQ amplicon present in testis from a control dog (T). N denotes a cDNA negative control. Selected fragment sizes (in bp) for the molecular ladder are marked to the left Cholinergic agonists inhibit AChE and prolong the action of available acetylcholine, thereby improving the safety margin of neuromuscular transmission. 1 The AChR content is only mildly decreased in COLQ variants, and no further benefit can be derived from increasing available acetylcholine. The mechanism by which adrenergic agonists improve neuromuscular transmission is not understood. 1 Esterase histochemistry similarly localizes motor end-plates with variants in both genes. Staining for AChRs with fluorescein labeled α-bgt, however, was not detectable in dogs with the CHRNE mutation, 5 and was decreased but detectable compared to control muscle in the puppies with the COLQ mutation. Consistent with this finding, the decremental responses of the CMAP to repetitive nerve stimulation were more marked in the CHRNE mutation (decrements of the CMAP of up to 66%) 5 compared to mild decreases in the dogs with the COLQ mutation. The G>A transition in the GR creates a novel AG sequence within exon 13 that we hypothesize could serve as a splice acceptor and cause aberrant splicing. Amplification of cDNAs from affected muscle did not identify abnormal transcripts but rather a complete lack of the primary COLQ isoform. A minor isoform that lacks exon 13 was present. This isoform, which creates a frameshift, also was present in the control tissue and has been previously described in humans. 24 The mechanism behind the absence of COLQ transcripts in the affected dog remains unclear. If the G>A transition does create an alternative splice acceptor, it would cause the loss of 140 nucleotides and a frameshift, likely triggering nonsense-mediated decay. Alternatively, our approach could have missed an intronic mutation that impacts the cDNA before exon 11 and triggers nonsense-mediated decay. Regardless of the mechanism, the lack of COLQ transcripts is consistent with the severe CMS phenotype observed in the GRs.