CASQ1 Gene Is an Unlikely Candidate for Malignant Hyperthermia Susceptibility in the North American Population

MALIGNANT hyperthermia (MH, MIM# 145600) is a subclinical complex pharmacogenetic disorder often inherited as an autosomal dominant trait. MH is triggered by exposure to inhalational anesthetics and depolarizing muscle relaxants,^1,2  and manifests as a hypermetabolic crisis characterized by a rapid and uncontrolled increase in myoplasmic Ca²⁺ concentration in skeletal muscle myofibers.³ 

Preoperative identification of susceptible individuals is crucial for the prevention of MH-related mortality and morbidity. The North American caffeine-halothane contracture test (CHCT) or its European equivalent, the in vitro contracture test are the gold standards for the diagnosis of MH susceptibility (MHS). The CHCT, with a high sensitivity of 97% but a relatively low specificity of 78%⁴  is invasive, requiring a muscle biopsy, and postoperative rehabilitation. Other limitations include its high cost and the existence of only a few testing centers in North America, entailing long and expensive travel.

The recognition that specific mutations in the RYR1 gene (MIM# 180901), encoding the Ca²⁺ release channel of the sarcoplasmic reticulum, are causal of human MH⁵  has made genetic testing for these mutations a viable diagnostic alternative to the CHCT in many MH families. In cases where a familial MH causative mutation is known, genetic testing is used efficiently for family counseling and significantly reduces the overall cost of MH diagnosis.

MH is a genetically heterogeneous disorder, and MH-associated mutations have been identified in two genetic loci so far, besides RYR1 also in CACNA1S encoding the α-subunit of the skeletal muscle voltage-dependent L-type Ca²⁺ channel.⁶  However, up to 40% patients do not carry mutations in either of these two genes, which results in a relatively low sensitivity of current MH genetic diagnostic tests.⁷  To fully comprehend the complex genetic nature of MH and to improve the sensitivity of the genetic testing it is critical to continue the search for new candidate genes that contribute to MH susceptibility.

Recent studies of a Casq1-null mouse line, which exhibited many of the symptoms of human MH susceptibility, including halothane-induced MH-like episodes prevented by dantrolene,^8,9  has raised the question of whether the CASQ1 gene (MIM#114250), encoding calsequestrin 1, the major luminal Ca²⁺ binding/buffering protein of the sarcoplasmic reticulum, might be a novel candidate gene for human MH. In an attempt to explain why mutations in RYR genes and in CASQ genes might have similar pathophysiological effects, a unifying theory^3,10  proposed a common mechanism for triggering of both the muscle contracture caused by RYR1 and CASQ1 MH mutations and the cardiac arrhythmias caused by RYR2 and CASQ2 catecholaminergic polymorphic ventricular tachycardia mutations. To date, there have been no reports of an association between human CASQ1 mutations and MHS in humans.

In an effort to associate mutations in the skeletal muscle calsequestrin 1 gene with MHS in humans, we have used our extensive depository of RNA and genomic DNA samples of MHS patients to identify genetic variants in CASQ1, which might be associated with MH susceptibility.

After institutional research ethics board approval, 497 individuals, referred to the MH Investigation Unit at the Department of Anesthesia, Toronto General Hospital (Toronto, Ontario, Canada) consented to the molecular genetics study. For this retrospective-observational study, we chose 75 consecutive unrelated MHS patients diagnosed by CHCT according to the North American protocol¹¹  between 2000 and 2011. This group included 50 probands who survived an MH reaction with clinical grading scale scores¹²  greater than 35 and 25 unrelated individuals from MH families in which probands were not available for the study.

In addition, three groups of unrelated Caucasian individuals (130 MHS, 100 MH negative, and 192 normal controls) were included in the study to be genotyped for the CASQ1 variant identified in the initial study group.

Nucleic acids were isolated according to published procedures^13,14  and complementary DNA synthesis was performed as described previously.¹⁵  All 75 MHS patients were screened for 30 RYR1 causative mutations by using either Sequenom platform or direct Sanger sequencing, and for the presence of 2 CACNA1S causative mutations, p.Arg1086His, and p.Arg1086Ser, by restriction fragment length polymorphism analysis.^16,17  In addition, using Sanger sequencing, 21 patients were screened for the entire RYR1 coding region, and 16 patients were screened for three MH hotspots regions as described.¹⁵ 

Characteristics of the patients and the results of the RYR1 screening are summarized in table 1.

CASQ1 gene on chromosome 1 consists of 11 exons that are spliced into a 1,993 bp messenger RNA transcript (GeneBank NM_ 001231). Full-length sequencing of the CASQ1 gene was performed on complementary DNA transcripts for patients with available muscle samples (44 patients) or, in their absence, on genomic DNA preparations (31 patients) by sequencing all CASQ1 exons, exon–intron boundaries, and the proximal regulatory regions. Two-kilobase fragments encompassing the complete CASQ1 transcript were amplified by reverse transcriptase-assisted polymerase chain reaction and sequenced using gene-specific primers. Alternatively, six genomic DNA fragments covering all 11 CASQ1 exons were amplified by polymerase chain reaction from the patient’s genomic DNA samples and sequenced directly (fig. 1).

Primers used for polymerase chain reaction amplification and sequencing are available on request. Sequencing of CASQ1 coding regions was bidirectional in all cases, using Big Dye Terminator v.3.1 chemistry on the ABI Prism 3730XL capillary sequencer instrument (reagents and instrument from Life Technologies Corporation, Carlsbad, CA). Sequencing reactions were run at the DNA Sequencing and Synthesis Facility of The Centre for Applied Genomics, Toronto, Canada. Genotyping for the c.260T > C variant was done either by DNA sequencing or by a restriction fragment length polymorphism assay that was based on the gain of BsmA1 endonuclease site in the presence of the variant. All changes identified by the restriction fragment length polymorphism analysis were confirmed by sequencing. There was no discordance in the results obtained by these two methods.

Raw sequence data analysis, contig building, and sequence comparison to the reference CASQ1 sequences of GenBank accessions NM_001231.4 and NC_000001.10 were done using Sequencher 4.10 software (Gene Codes, Ann Arbor, MI). All heterozygous calls were manually inspected and confirmed by sequencing of independent DNA fragments encompassing c.260T>C variant. Three software tools: PolyPhen-2,¹⁸  SIFT (Sorting Intolerant From Tolerant),¹⁹  and PMut^20 †† were used for prediction of the functional impact of the amino acid substitution identified.

Gracilis muscle biopsy specimens were collected from two male and six female unrelated MHS patients who were between 21 and 65 yr old. Gracilis muscle biopsy material from the MHN group was from 10 male and 8 female unrelated adult individuals. Muscle specimens were rinsed in ice-cold Ringer’s solution, snap-frozen in liquid nitrogen, and stored in air-tight containers at −70°C. Frozen specimens (approximately 150 mg) were powdered under liquid nitrogen and homogenized in 0.6 ml ice-cold buffer, containing 50 mm Tris-HCl (pH 7.4); 10 mm EGTA; 2 mm EDTA; 5 mm dithiothreitol; 0.5 mm phenylmethylsulfonyl fluoride, and protease inhibitor cocktail (Complete mini, Roche Diagnostics Canada, Laval, Quebec, Canada) by repetitive passage through 18- and 20-gauge needles. Protein concentration was determined by a modified Bradford procedure²¹  using bovine serum albumin as a standard.

The proteins of whole-muscle homogenates (20–200 μg total protein) were separated on 10% SDS-PAGE gels (Bio-Rad Laboratories Ltd., Mississauga, Ontario, Canada) and transferred electrophoretically onto Immobilon P^SQ membranes (0.2 μm pore size, EMD Millipore, Billerica, MA).

The membranes were blocked in 5% nonfat milk and 1% bovine serum albumin in phosphate-buffered saline solution and incubated with monoclonal antihuman calsequestrin 1 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for 2 h at room temperature. After incubation with horseradish peroxidase-conjugated rabbit antimouse polyvalent secondary antibodies (Sigma-Aldrich Canada Ltd., Oakville, Ontario, Canada), the immune complexes were revealed using Luminata^TM Forte Western horseradish peroxidase substrate (EMD Millipore). For a loading control and for relative estimation of calsequestrin protein content, the blots were reprobed with antidesmin, anti-α-tubulin, and anti-α-actinin mouse monoclonal antibodies (Sigma-Aldrich Canada Ltd.). The images were generated using Fluo S^TM Max MultiImager and Quantity One software (Bio-Rad Laboratories Ltd.). Normalization of calsequestrin band intensities was performed using ImageJ program (National Institutes of Health, Bethesda, MD) against density values of reference proteins within the same lanes.

Allele and genotype frequencies of the c.260T>C variant in MHS and control groups were obtained by direct counting. The variant allele frequencies in different groups were compared by tests for association that included Pearson chi-square statistics. The tests were adapted from the work by Sasieni²²  in a program provided on the website of the Institute of Human Genetics, Munich, Germany.^‡‡ In Western blotting analysis, unpaired two-tailed t test as implemented in the GraphPad Prism v.5 software (GraphPad Software Inc., La Jolla, CA) was used to compare the relative density data sets from normal and p.Met87Thr variant containing calsequestin 1 protein bands. For statistical tests, we considered P value less than 0.05 to be evidence of significance.

Screening of the entire coding sequence of the CASQ1 gene in 75 unrelated MHS patients, 50 of whom had survived an MH reaction (probands), revealed a single variant c.260T>C in exon 1 that was present in the heterozygous state in 6 individuals. In two of them, the variant was present together with a concurrent, causative mutation in RYR1 (table 1). Genotyping of further 130 unrelated MHS patients revealed 9 more individuals heterozygous for this variant, 2 of them also carrying a causative mutation in RYR1.

Bioinformatic analysis shows that c.260T>C would result in the substitution of Thr for p.Met87 (p.Met87Thr) in the calsequestrin 1 protein sequence, which has the potential to be pathogenic. p.Met87Thr is predicted to be “probably damaging” by PolyPhen2 software (with the position-specific independent counts score difference of 2.263) and “intolerant” by SIFT (Sorting Intolerant From Tolerant) software (with the score of 0.00). PMut software predicted the p.Met87Thr to be neutral but with a low prediction reliability of three. The calsequestrin 1 Met87 residue is highly conserved throughout mammalian kingdom (fig. 2). It is located within the α2 helix, which is involved in the dimer interface formation.²³  The p.Met87Thr substitution introduces a shift from a nonpolar to a weakly negative charge. Potentially, this could have a destabilizing effect on α-helix structure, calsequestrin 1 dimerization and consequently, interfere with high-capacity Ca²⁺ binding.

To test whether the variant occurs with higher frequency in the MHS population than in the general population we estimated the frequency of this variant in 205 MHS, 100 MH negative, and 192 control genomic DNA samples. The results showed that the allele frequency in the MHS group (0.04) was not statistically significantly different from that in the MH negative (0.02, odds ratio = 1.861; 95% CI = [0.609–5.681]; P = 0.268) or control groups (0.018, odds ratio = 2.045; 95% CI = [0.825–5.072]; P = 0.115).

To test the possibility that the c.260T>C variant could affect the level of calsequestrin 1 in skeletal muscle, we assessed the level of total calsequestrin 1 protein in gracilis muscle biopsies from 18 unrelated MHN individuals and from 8 unrelated MHS patients, identified as carriers of the c.260T>C CASQ1 variant, by Western blotting in whole-muscle homogenates. Pilot experiments on MHN muscle biopsy specimens showed that chemiluminescence of the calsequestrin 1 protein band exhibited a linear range of intensity when 50–200 μg of whole-muscle protein were loaded per lane (data not shown). Accordingly, samples of 50, 100, and 150 μg total protein were used for comparative estimates of calsequestrin 1 levels in MHS and MHN individuals. The biopsy samples from MHS individuals were matched with control samples from healthy individuals who were of the same sex and belonged to a similar age group. The normalized expression level of calsequestrin 1 in normal muscle (mean ± SEM = 0.99 ± 0.038, 95% CI = 0.92–1.1) was not significantly different (P = 0.73) from that in the muscle of patients with c.260T>C variant (mean ± SEM = 0.97 ± 0.054, 95% CI = 0.86–1.1) (fig. 3). Thus, the results of Western blotting analysis indicate that the c.260T>C CASQ1 variant does not significantly compromise calsequestrin 1 protein expression in human gracilis muscle.

CASQ1 encodes skeletal muscle calsequestrin 1 (MIM#114250), a low-affinity, high-capacity, Ca²⁺ binding protein that is localized in the lumen of the sarcoplasmic reticulum. In a unifying theory,¹⁰  spontaneous premature activation of Ca²⁺ release, characteristic of episodes of MH and of catecholaminergic polymorphic ventricular tachycardia, is promoted either by mutations in RYR1 and RYR2, which lower the threshold of their activation by luminal free Ca²⁺ concentration, or by reduction in luminal Ca²⁺ binding and buffering capacity either through null mutations in CASQ1 that lead to loss of calsequestrin protein or missense mutations that prevent the appropriate folding of calsequestrin 1 into the polymeric structures that are required for high-capacity Ca²⁺ binding. In these cases, free luminal Ca²⁺ levels are no longer adequately buffered and can rise to levels that exceed the threshold of activation of a normal RyR1 Ca²⁺ release channel. Indeed, several recessive mutations of the cardiac CASQ2 gene have been detected and implicated in cardiac arrhythmias.²⁴  By contrast, overexpression of Casq2 in mice causes cardiac hypertrophy and cardiomyopathy.²⁵  Furthermore, Casq1-null mice, lacking skeletal muscle calsequestrin exhibited enhanced sensitivity to halothane, resulting in hyperthermia and rhabdomyolysis; these MH-like episodes were prevented by dantrolene pretreatment.^8,9  Thus, these new developments in the MH research have brought the CASQ1 gene forward as a candidate gene for MH, certainly in mice and potentially in humans.

In this study aimed at the identification of MH-associated mutations in the human CASQ1 gene, we have performed sequence analysis of the CASQ1 gene in 75 unrelated individuals who survived an MH crisis and/or tested positive by CHCT. On the basis of the assumption that genetic variants in CASQ1 might either be MH causative or contribute to the MH phenotype caused by defects in RYR1, this group included patients carrying MH-associated RYR1 mutations. Sequence analysis revealed a very low level of variation in the CASQ1 coding regions. The only variant detected was c.260T>C/p.Met87Thr. This variant was also detected independently in two large genotyping projects involving more than 2,000 individuals (National Center for Biotechnology Information dbSNP rs150330307). The low level of CASQ1 coding sequence variation observed in our study is consistent with findings of the previous investigations where the CASQ1 gene was systematically screened for variants associated with type 2 diabetes and where only two rare coding variants (minor allele frequency of 0.005) were detected.^26,27 

The c.260T>C variant identified in this study was present in both MHS and MH negative samples, thereby demonstrating that it is not directly associated with MH. Moreover, Western blotting analysis showed that the presence of the variant did not significantly compromise the level of calsequestrin 1 protein (fig. 3). However, bioinformatic analysis, as well as the location of amino acid residue p.Met87 in a region of the three-dimensional structure that is critical to calsequestrin 1 polymerization suggests that the p.Met87Thr has the potential to alter the function of calsequestrin 1 as the sarcoplasmic reticulum Ca²⁺ buffer through interference with polymerization. Multiple sequence alignments place p.Met87 within the sequence motif FEMEEL (PheGluMetGluGluLeu) in a position equivalent to the middle of helix α2 of domain I in the rabbit calsequestrin 1 structure. It lies close to the point of contact between two calsequestrin monomers within a dimer in the three-dimensional structure.²⁸ 

Multiple alignment of calsequestrin 1 and calsequestrin 2 amino acid sequences from various species²⁹  also shows that helix α2 is evolutionarily conserved and that p.Met87 in calsequestrin 1 is positionally equivalent to p.Leu72 in calsequestrin 2. It has been shown that in calsequestrin 2, p.Leu72 together with p.Ile75 and p.Leu76 contribute a hydrophobic interaction to the front-to-front dimer interface. Furthermore, the calsequestrin 2 polymorphism p.Leu76Met, although not clearly associated with catecholaminergic polymorphic ventricular tachycardia, was shown to confer a slightly diminished Ca²⁺ binding capacity probably due to its effect on both local conformation of the monomer and its Ca²⁺ dependent dimer and polymer formation.²⁹  Thus, p.Met87Thr, located within a conserved secondary/tertiary structure implicated in calsequestrin 2 multimer formation, may have an impact on calsequestrin 1 polymer formation and hence on its Ca²⁺ binding capacity. Although this study did not establish a direct causal relationship between p.Met87Thr and MH susceptibility trait, this variant has the potential to act as an enhancer of the disease phenotype. Clearly, in vitro functional study of the mutated calsequestrin 1 protein is needed to assess the effect of the p.Met87Thr substitution on Ca²⁺ binding function.

Because our analysis was limited to the coding regions of the CASQ1 gene, it is possible that some MHS patients may carry mutations in the regulatory regions or within introns that were not identified in this study, but which might modify calsequestrin 1 stability or level of its expression, possibly reducing luminal Ca²⁺ binding/buffering capacity. Besides, 75% of individuals included in our study group were of west European origin. It is possible that genetic analysis of MHS populations from different ethnic groups will reveal CASQ1 variants with direct or modifying effect on MH phenotype. The results obtained in the mouse Casq1 knockout imply that an MH-like phenotype would be generated either by a coding sequence mutation that impairs calsequestrin 1 function, or by a noncoding sequence mutation inducing a considerably lower level of expression of the protein. Although we have not observed either of these events in our cohort of MH patients, it is conceivable that epigenetic inhibition in CASQ1 expression, mediated by genomic DNA methylation, could be brought about by random events, such as persistent bacterial infection or adverse conditions in the environment.^30,31  It is also possible that transient nonhereditary impairment of calsequestrin 1 function might occur as a result of posttranslational modifications, such as phosphorylation and/or glycosylation.^32–34 

In conclusion, this study revealed a low level of variability within the CASQ1 gene indicating that CASQ1 is not a major MHS locus in the North American population. Further analysis in other populations and functional studies of potentially pathogenic variants will be necessary to elucidate the role of CASQ1 in MH.