A canine BCAN microdeletion associated with episodic falling syndrome

Episodic falling syndrome (EFS) is a canine paroxysmal hypertonicity disorder found in Cavalier King Charles spaniels. Episodes are triggered by exercise, stress or excitement and characterized by progressive hypertonicity throughout the thoracic and pelvic limbs, resulting in a characteristic 'deer-stalking' position and/or collapse. We used a genome-wide association strategy to map the EFS locus to a 3.48 Mb critical interval on canine chromosome 7. By prioritizing candidate genes on the basis of biological plausibility, we found that a 15.7 kb deletion in BCAN, encoding the brain-specific extracellular matrix proteoglycan brevican, is associated with EFS. This represents a compelling causal mutation for EFS, since brevican has an essential role in the formation of perineuronal nets governing synapse stability and nerve conduction velocity. Mapping of the deletion breakpoint enabled the development of Multiplex PCR and Multiplex Ligation-dependent Probe Amplification (MLPA) genotyping tests that can accurately distinguish normal, carrier and affected animals. Wider testing of a larger population of CKCS dogs without a history of EFS from the USA revealed that carriers are extremely common (12.9%). The development of molecular genetic tests for the EFS microdeletion will allow the implementation of directed breeding programs aimed at minimizing the number of animals with EFS and enable confirmatory diagnosis and pharmacotherapy of affected dogs.


Introduction
Episodic falling syndrome (EFS) is a well-recognized paroxysmal disorder found in Cavalier King Charles spaniels (CKCS). Episodes begin between fourteen weeks and four years of age and are triggered by exercise, stress, apprehension or excitement (Herrtage and Palmer, 1983). Episodes are of variable frequency and severity but are characterized by progressive hypertonicity involving thoracic and pelvic limbs (Fig. 1a) until the dogs are ultimately immobilized in a characteristic 'deer-stalking' or 'praying' position ( Fig. 1b). Stiffening of all four limbs during exercise can cause falling (Fig. 1c), although there is no loss of consciousness or cyanosis. Other clinical signs may include facial muscle stiffness, stumbling, a 'bunny-hopping' gait, arching of the back or vocalization. Curiously, between episodes, dogs appear to be completely normal neurologically. Spontaneous activity was not observed in muscle electrodiagnostic testing, ruling out myotonia congenita (Wright et al., 1986(Wright et al., , 1987. Muscle biopsies are typically normal at the light microscopic level, excluding many congenital myopathies. However, EFS has been linked to ultrastructural defects in skeletal muscle including dilatation and proliferation of the sarcoplasmic reticulum, mitochondrial swelling and degeneration (Wright et al., 1986(Wright et al., , 1987. EFS has also been compared (Rusbridge, 2005) with startle disease/hyperekplexia, typically characterized by noise-or touch-evoked neonatal hypertonicity due to defects in inhibitory glycine receptor (GLRA1, GLRB; Shiang et al., 1993;Rees et al., 2002) or glycine transporter GlyT2 (SLC6A5) genes (Rees et al., 2006;Harvey et al., 2008). However, a microdeletion in the GlyT2 gene in Irish Wolfhounds results in severe neonatal muscle stiffness and tremor in response to handling (Gill et al., 2011), which is inconsistent with the observed clinical signs of EFS. Comparisons with startle disease may have been made because affected dogs often respond well to the benzodiazepine clonazepam (Garosi et al., 2002), an effective anticonvulsant, anxiolytic and muscle relaxant that is the most effective known treatment for human hyperekplexia (Thomas et al., 2010). However, the carbonic anhydrase inhibitor acetazolamide, used to treat certain types of human episodic ataxia (Tomlinson et al., 2009) and hyperkalemic periodic paralysis (Matthews and Hanna, 2010), also appears to have therapeutic value in the treatment of EFS (http://www.cavalierhealth.org/episodic_falling).
Since a ten-year breeder-led investigation into the inheritance of EFS suggested an autosomal recessive mode of inheritance (http:// cavalierepisodicfalling.com/), we used a genome-wide association strategy (Karlsson et al., 2007) to map the EFS locus to a defined region of canine chromosome 7. Candidate gene analysis enabled us to identify a microdeletion affecting the brevican gene (BCAN), confirm the deletion breakpoint and develop rapid genotyping tests for EFS.

Light and electron microscopy
For light microscopy, unfixed biopsies from the biceps femoris, vastus lateralis and triceps brachii muscles were collected from five affected CKCS dogs under general anesthesia and frozen in isopentane pre-cooled in liquid nitrogen. Cryosections were cut (8 μm) and the following histochemical stains and reactions performed: hematoxylin and eosin, modified Gomori trichrome, periodic acid Schiff, phosphorylase, esterase, ATPase reactions at pH of 9.8 and 4.3, nicotinamide adenine dinucleotide-tetrazolium reductase, succinic dehydrogenase, acid phosphatase, alkaline phosphatase and oil red O. For electron microscopy, glutaraldehyde-fixed muscle specimens were post-fixed in osmium tetroxide, and dehydrated in serial alcohol solutions and propylene oxide prior to embedding in araldite resin. Thick sections (1 μm) were stained with toludine blue for light microscopy and ultrathin sections (60-90 nm) were stained with uranyl acetate and lead citrate for electron microscopy.

Study cohort and DNA preparation
Our study cohort comprised: EFS affected-10 animals (6 from the USA, 2 from New Zealand and 2 from the UK); Obligate EFS carriers-8 animals (2 from the USA, 6 from New Zealand); Animals related to carriers or affected dogs-21 animals (7 from the USA, 14 from New Zealand); Controls-CKCS with no EFS history-14 animals (all from the USA). Genomic DNA was isolated from whole blood or buccal cells using the Gentra Puregene Blood Kit (QIAGEN, Valencia, USA). Additional DNA samples from 155 CKCS with no clinical history of EFS and other pure bred-dogs were available from unrelated studies and other sources (e.g. Cornell Medical Genetic Archive: http://www. vet.cornell.edu/research/dnabank/).

Genome-wide association mapping
Thirteen CKCS genomic DNA samples isolated from blood (five cases, one obligate carrier and seven controls from the USA) were genotyped for 127,000 SNPs on the Affymetrix Canine SNP Array version 2 (http:// www.broadinstitute.org/mammals/dog/caninearrayfaq.html). The two main drivers for sample selection were: i) lack of relatedness-i.e., that the animals used for case-control analysis should not share a common ancestor within at least three generations and ii) the quality and quantity of genomic DNA available. Arrays were processed at the Fig. 1. Clinical signs of episodic falling syndrome and muscle pathology. A 5-month-old female Cavalier King Charles spaniel presented with typical episodes of excitement or exercise-induced muscle stiffness (a, hypertonicity) that would involve all four limbs and progress to an usual 'deer-stalking' or 'praying' posture (b), eventually resulting in falling (c). While EFS muscle was normal histologically by light microscopy, electron microscopy (d) revealed that the sarcoplasmic reticulum (sr) appeared dilated and contained finely granular material compared to control muscle (e). Mitochondria (m) and myofibrils were normal in appearance in both tissues. Scale bars = 0.31 μm.
Clemson University Genomics Institute (http://www.genome.clemson. edu/) using the GeneChip human mapping 250 K Sty assay kit (Affymetrix, Santa Clara, USA). The GeneChip human mapping 500 K assay protocol was followed, but with a hybridization volume of 125 μl (Karlsson et al., 2007). Raw CEL files were genotyped using Affymetrix Power Tools software. SNPs having N10% missing data and ≥60% heterozygosity were removed. Data for 58,873 SNPs were formatted for PLINK (Purcell et al., 2007) and case/control analyses with 100,000 permutations were performed for five cases and seven controls (the obligate carrier was excluded from analysis).

PCR and DNA sequencing
PCR primers were designed to amplify exons and flanking splice donor, acceptor and branch-point sites, from gene structures derived in silico using the UCSC Genome Browser. For exon-specific primers for BCAN and HAPLN2 exon amplification see Table S1. PCR was performed using 50 ng genomic DNA as template and AccuPrime Pfx SuperMix supplemented with betaine for 40 cycles of 94°C for 1 min, 60°C for 1 min, 68°C for 1 min. PCR products were gel purified using a QiaQuick gel extraction kit (QIAGEN, Crawley, UK) for TOPO cloning (pCR4Blunt-TOPO; Invitrogen, Renfrew, UK). Sanger DNA sequencing was performed by DNA Sequencing & Services (MRCPPU, College of Life Sciences, University of Dundee, Scotland) using Applied Biosystems Big-Dye version 3.1 chemistry on an ABI 3730 automated capillary DNA sequencer. DNA sequences were analyzed using Sequencher 4.10 (Gene Codes Corporation, Ann Arbor, USA). For multiplex PCRs, the diagnostic primer set is: EFS1 5′-aaggtcttacacctgcaatgaatag-3′, EFS2 5′-agcaaatgtaaagtcctgtgaccat-3′ and EFS3 5′agttcacattgtgctctctctactg-3′.

Multiplex ligation-dependent probe amplification (MLPA) analysis
Five MLPA probe sets were designed corresponding to the promoter region (PR) and exons 1 (5′ UTR), 2, 3 and 4 of the canine brevican gene (BCAN; NC_006589 on chromosome 7; Table S2). Criteria for MLPA probe design were as previously described (Schouten et al., 2002). A control probe pair was designed to recognize an unrelated gene (CFTR: NC_006596 on chromosome 14). Probes generated amplification products ranging in size from 88 to 115 bp and had annealing temperatures higher than 70°C as recommended in RAW Probe software (MRC-Holland, Amsterdam, The Netherlands) using standard MLPA conditions (Schouten et al., 2002). PCR products were analyzed on an ABI 3130XL capillary electrophoresis apparatus (Applied Biosystems, Lennik, Belgium). Normalization of BCAN-specific probe signals was performed by dividing the values obtained by the combined signal of the control probes.

Light and electron microscopy
Unfixed cryosections of muscle biopsies were histologically normal at the light microscopic level with no abnormalities detected following any of the histochemical stains and enzyme reactions employed. Contrary to previous reports (Wright et al., 1986(Wright et al., , 1987 electron microscopy revealed normal myofibrillar and mitochondrial morphology, although swelling of the sarcoplasmic reticulum was confirmed (Figs. 1d,e).

Deletion breakpoint identification and development of diagnostic tests
Further primer walking experiments enabled us to clone and sequence a DNA fragment containing the deletion breakpoint and develop a multiplex PCR assay that distinguishes between affected, carrier and normal dogs (Fig. 3b). Sequence analysis of the breakpoint amplicon revealed a 15.7 kb microdeletion starting 1.56 kb downstream of HAPLN2, encompassing BCAN promoter elements and exons 1 (5′ untranslated), 2 and 3, finishing 85 bp downstream of BCAN exon 3 (Fig. 3c). The microdeletion amplicon also contained a 6 bp inserted sequence (GGCCTT; Fig. 3d) typical of deletions resulting from nonhomologous end joining (NHEJ) or microhomology-mediated end joining (MMEJ). Several regions of microhomology (1-7 bp) were identified in a 30 bp region encompassing the breakpoints (Fig. 3e). Interestingly, 5 of the 6 bp of the reverse complemented inserted sequence align to the largest region of microhomology. We also noted an abundance of short interspersed element (SINE) insertions at the 5' end of the deleted sequence, which could cause the formation of secondary structures that facilitate chromosomal rearrangement (Chuzhanova et al., 2003). We also detected the presence of the  BCAN microdeletion (Fig. 4) using MLPA (Schouten et al., 2002) and canine-specific probe sequences (Table S3). This quantitative analysis also confirmed that EFS is associated with a loss of BCAN promoter/regulatory elements and exons 1-3 in heterozygous carriers and homozygous affected animals (Fig. 5).

Rapid genotyping using multiplex genotyping in different dog populations
To assess the prevalence of the EFS microdeletion, we used our multiplex PCR assay to test several CKCS populations and other dog breeds (Table 1). All affected dogs in our study cohort-from the UK (n = 2), USA (n = 6) and New Zealand (n = 2), were homozygous for the BCAN microdeletion. In addition, all obligate carriers were heterozygous (n = 2 from the USA and n = 6 from New Zealand). In animals related to affected dogs, we found 9 normal animals and 10 carriers. Interestingly, in this group, two dogs without a classical clinical history of EFS were homozygous for the microdeletion. Lastly, in dogs with no known clinical history of EFS sourced from the USA, the carrier frequency was 12.9% (20/155) suggesting that the EFS microdeletion is present at a high frequency in this population. Notably, the mutation was not detected in multiplex PCRs conducted on control DNA samples from 54 other breeds of dog (Table 1).

Discussion
The genomic architecture of pure-bred dog lines is ideal for the identification of loci responsible for autosomal recessive traits using genome-wide association mapping (Karlsson et al., 2007;Drögemüller et al., 2009). In this study, we demonstrate that this technique can be used successfully on minimal samples sets, since we located the EFS locus using DNA samples from only five affected and seven breedmatched control dogs. Since a homozygous haplotype spanning 6.35 Mb was identified in affected animals, it is questionable whether further SNP typing would have generated additional useful data. In fact, a single recombination event in an obligate carrier allowed us to narrow the critical interval to 3.48 Mb. This region contained N100 genes, including ligand-gated ion channels, K + channels, transporters, mitochondrial proteins and several genes known to be involved in neurological disorders in humans. For example, mutations in CHRNB2 are associated with nocturnal frontal lobe epilepsy (De Fusco et al., 2000) and TPM3 mutations are associated with nemaline myopathy (Laing et al., 1995). However, many of these genes were rapidly eliminated as candidates due to either: i) poor correlation of EFS clinical signs with the equivalent human disorders or ii) systematic resequencing of the genes. Consistent with the unique clinical signs observed in affected dogs, we discovered that a homozygous microdeletion affecting BCAN is associated with EFS in CKCS dogs, confirming that this disorder is inherited in an autosomal recessive manner. This mutation was not detected in control DNA samples from 54 other dog breeds, confirming the unique nature of this genomic rearrangement.
Brevican belongs to the lectican family of aggregating extracellular matrix (ECM) proteoglycans, which comprises aggrecan, brevican, neurocan and versican. Although mutations in the aggrecan and versican genes (ACAN: 15q26.1 and VCAN: 5q14.2-14.3) have been linked to different connective tissue disorders (Gleghorn et al., 2005;Stattin et al., 2010;Miyamoto et al., 2005;Fig. 6), no mutations in the brevican or neurocan genes (BCAN: 1q23.1 and NCAN: 19p13.11) have been identified to date. Brevican and neurocan are highly expressed in the central nervous system, where they are found in specialized extracellular matrix structures called perineuronal nets that play a role in cell adhesion, migration, axon guidance and neuronal plasticity (Brakebusch et al., 2002). Brevican, versican, HAPLN2/BRAL1, tenascin-R and phosphacan are also present at the nodes of Ranvier on large diameter myelinated axons (Bekku et al., 2009(Bekku et al., , 2010 where cations are accumulated and depleted in the local extracellular nodal region during action potential propagation. The ECM complex at nodes of Ranvier is thought to play a pivotal role in maintaining a local microenvironment, acting as a diffusion barrier for K + and Na + around the perinodal extracellular space (Oohashi et al., 2002;Bekku et al., 2010). Thus, disruption of ECM complexes governing synapse  Genotypes revealed by multiplex PCRs were determined as described in Materials and methods. Dogs were evaluated on the basis of available clinical data and placed into one of the phenotype categories above. Note that all clinically affected animals were homozygous for the BCAN deletion, whilst obligate carriers were heterozygous. As well as wild-type animals and carriers of the BCAN deletion, two dogs homozygous for the BCAN deletion, which were not reported to have classical clinical signs of EFS, were detected in a cohort of animals related to known EFS dogs. Carriers were also detected in CKCS with no history of EFS, but not in control DNA samples from 54 other dog stability and nerve conduction velocity is likely to underlie the EFS phenotype. Certainly, since EFS appears to result from a central nervous system rather than a muscle defect, the associated sarcoplasmic reticulum pathology is likely to be a secondary manifestation of muscle overstimulation (Engel and Banker, 2004). Interestingly, BCAN knockout mice were not reported (Brakebusch et al., 2002) to have a phenotype similar to EFS -although an increased grip-strength was noted, which could be indicative of increased muscle tone. However, it is questionable whether current mouse phenotyping tests will reveal neurological disorders evoked by strenuous exercise, stress, apprehension or excitement, since these conditions are generally avoided in mouse care. It is also noteworthy that EFS clearly shows variable age of onset and penetrance. For instance, we identified two dogs that were homozygous for the BCAN microdeletion that had not shown classical clinical signs of EFS in the presence of their owners. Interestingly, similar findings were reported for the dynamin (DNM1) mutation (p.R256L) underlying exercise-induced collapse in Labrador retrievers (Patterson et al., 2008). It is plausible that these dogs have not been exposed to sufficient exercise or excitement to trigger an episode, or that genetic or environmental modifying factors (e.g. diet) affect the onset of these disorders. However, it is notable that one of the two asymptomatic dogs in our study was described as 'exercise resistant' by their owner, suggesting that adaptive behavior may occur in dogs that are homozygous for the BCAN microdeletion. It must also be emphasized that such non-classical cases are rare-we did not find a single dog that was homozygous for the BCAN microdeletion in 155 CKCS with no family history of EFS (Table 1).
In summary, we have shown that an inactivating microdeletion affecting BCAN is associated with EFS in CKCS dogs. Identification of the deletion breakpoint has allowed the development of diagnostic tests that have revealed a high prevalence of carriers (12.9%) in clinically unaffected dogs from the USA. These genetic tests (available via Laboklin: http://www.laboklin.co.uk/) will enable future identification of heterozygous animals (which have no discernable phenotypic difference to wild-type animals) allowing directed breeding programs to be implemented, and confirmatory diagnosis and appropriate pharmacotherapy of affected animals. Since this also represents the first report of a genetic disorder involving a neuronalspecific ECM proteoglycan, we suggest that BCAN and NCAN should be considered as candidate genes for genetic analysis in unresolved cases of human disorders with similar clinical presentations to EFS, such as paroxysmal exercise-induced dyskinesias (Weber and Lerche, 2009) or episodic ataxias (Tomlinson et al., 2009).

Role of the funding source
The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. None of the authors declare a conflict of interest.
Supplementary materials related to this article can be found online at 10.1016/j.nbd.2011.07.014.