Abstract
Variants in RYR1 are associated with the majority of cases of malignant hyperthermia (MH), a form of heat illness pharmacogenetically triggered by general anesthetics, and they have also been associated with exertional heat illness (EHI). CACNA1S has also been implicated in MH. The authors applied a targeted next-generation sequencing approach to identify variants in RYR1 and CACNA1S in a cohort of unrelated patients diagnosed with MH susceptibility. They also provide the first comprehensive report of sequencing of these two genes in a cohort of survivors of EHI.
DNA extracted from blood was genotyped using a “long” polymerase chain reaction technique, with sequencing on the Illumina GAII® or MiSeq® platforms (Illumina Inc., USA). Variants were assessed for pathogenicity using bioinformatic approaches. For further follow-up, DNA from additional family members and up to 211 MH normal and 556 MH-susceptible unrelated individuals was tested.
In 29 MH patients, the authors identified three pathogenic and four novel RYR1 variants, with a further five RYR1 variants previously reported in association with MH. Three novel RYR1 variants were found in the EHI cohort (n = 28) along with two more previously reported in association with MH. Two other variants were reported previously associated with centronuclear myopathy. The authors found one and three rare variants of unknown significance in CACNA1S in the MH and EHI cohorts, respectively.
Targeted next-generation sequencing proved efficient at identifying diagnostically useful and potentially implicated variants in RYR1 and CACNA1S in MH and EHI.
Abstract
Variants in the ryanodine receptor gene were identified in 13 of 29 malignant hyperthermia patients, with one variant in Cav1.1. Targeted DNA sequencing is a potentially useful diagnostic approach to identifying genetic variants associated with malignant hyperthermia and exertional heat illness.
Variants in the genes for ryanodine receptor 1 and the Cav1.1 calcium channel account for most cases of malignant hyperthermia susceptibility
Rapid targeted DNA sequencing was used to identify variants in the coding sequence of these two genes in a cohort of patients with malignant hyperthermia susceptibility or exertional heat illness
Variants in the ryanodine receptor gene were identified in 13 of 29 malignant hyperthermia patients, with one variant in Cav1.1
Targeted DNA sequencing is a potentially useful diagnostic approach to identifying genetic variants associated with malignant hyperthermia and exertional heat illness
MALIGNANT hyperthermia (MH) is a potentially fatal adverse drug reaction, triggered during general anesthesia by volatile halogenated anesthetics and succinylcholine. The prevalence of genetic variants predisposing to MH has been estimated to be as high as 1:2,000.1,2 In most families, MH shows an autosomal dominant pattern of inheritance. Following a suspected MH episode, the diagnosis can be confirmed using an in vitro contracture test (IVCT), where skeletal muscle, obtained through biopsy of the vastus muscle, is exposed to incremental concentrations of caffeine or halothane and responses recorded.3
Physiological and biochemical studies show a severe defect in MH skeletal muscle calcium regulation.4 The ryanodine receptor (RYR1) forms a calcium channel in the sarcoplasmic reticulum and in its active state releases Ca2+ from the sarcoplasmic reticulum leading to muscle contraction.5 Cav1.1 is structurally a calcium channel located within the t-tubule membrane, but functionally, it is the voltage sensor that activates RYR1. During an MH episode, myoplasmic calcium concentration rises rapidly resulting in a hypermetabolic state characterized systemically by hyperthermia, acidosis, tachycardia, cardiac arrhythmias, skeletal muscle rigidity, and rhabdomyolysis. Similar features are observed in exertional heat illness (EHI), suggesting possible links between MH and EHI.6
Variants in the RYR1 gene (Mendelian inheritance in man, #180901) are thought to be responsible for about 75% of cases susceptible to MH. Less frequently MH is associated with variants in CACNA1S (Mendelian inheritance in man, #114208), encoding the α1 subunit of Cav1.1. Mutations in either of these genes may disrupt signaling between Cav1.1 and RYR1 and thus confer MH susceptibility.7–9 A significant number of families do not have a variant in a major locus identified to date.
Approximately 200 missense variants in RYR1 have been described in association with MH,10 with 31 of them known to be functionally relevant and used diagnostically,11 while only two variants described in CACNA1S in association with MH have been functionally characterized.12,13 Variants in RYR1 are also known to cause distinct rare myopathies, such as multiminicore disease and central core disease, with some variants associated both with MH and central core disease.14
Because of the large size of RYR1, screening the whole coding sequence (~15,000 base pairs) using Sanger sequencing is expensive and time-consuming. In a clinical setting, often only the limited number of diagnostic variants is screened. With the advent of next-generation sequencing (NGS), limitations of current approaches can be overcome with a potential for increased variant detection rate compared with Sanger sequencing.15 We have assessed the feasibility of NGS using groups of MH-susceptible and EHI patients previously screened for RYR1 diagnostic variants and found to be negative, for analysis of the whole coding sequences of RYR1 and CACNA1S. We applied a long-range polymerase chain reaction (PCR) technique as an enrichment method, to amplify coding sequences, and sequenced using Illumina GAII® and MiSeq® technology (Illumina Inc., USA).
Materials and Methods
Samples
Blood samples were obtained from patients referred to the Malignant Hyperthermia Unit, University of Leeds, for genetic and functional (IVCT) tests used in the diagnosis of MH susceptibility. The research was approved by the Leeds East Local Research Ethics Committee (Reference 10/H1306/70; Leeds, United Kingdom), and written consent was obtained from each patient. DNA was extracted from peripheral blood lymphocytes using a salting-out method. Briefly, after erythrocyte lysis with a buffer containing 155 mM NH4Cl, 10 mM KHCO3, and 1.0 mM EDTA, the leukocytes were pelleted and subjected to lysis with 2% sodium dodecyl sulphate, 25 mM EDTA solution. Next, a protein precipitation with 10 M ammonium acetate was performed. Then, the proteins were pelleted, and supernatant containing DNA was transferred into a new tube. DNA precipitation was carried out using isopropanol.
Fifty-seven DNA samples were chosen for sequencing the entire coding sequence of RYR1 and CACNA1S, as the two main genes known to contribute to MH susceptibility. The majority of samples had already been routinely screened for the 16 most common diagnostic variants found in the United Kingdom population: DNA results from these patients have not been reported in previous publications. We sequenced DNA from 29 MH-susceptible individuals, with susceptibility confirmed by IVCT, and 28 EHI individuals. All of the EHI individuals were military personnel who had had a clinical episode of heat illness during military exercises that required hospital treatment. They were referred for testing for MH susceptibility having all subsequently failed to demonstrate normal thermoregulation during a heat tolerance test carried out at the Institute of Naval Medicine (see appendix 1) on at least two separate occasions. As the sensitivity of the IVCT to detect genetic predisposition to EHI, as opposed to MH susceptibility, is unknown, we included EHI individuals in this study irrespective of their IVCT result. In the group of EHI samples, 16 were from MH normal individuals (MHN, responding to neither triggering agent used in the IVCT) and 12 from MH-susceptible individuals, where MH susceptibility is defined by the IVCT.
Each of the 57 samples represented a single family. For follow-up of NGS, DNA from additional family members and at least 150 MH normal and up to 556 MH-susceptible genetically independent individuals, including cases of EHI, were tested. If a variant is not found in the 300 chromosomes of 150 individuals, it can be concluded with 95% confidence that the prevalence of the variant in the population is less than 1%, assuming the data are from a Binomial distribution. Such a variant, by definition, is not a polymorphism.
Long PCR
The samples were processed in four batches. The coding sequence of RYR1 and CACNA1S was amplified in 23 and 9 PCR products, respectively (tables 1 and 2). Amplification was carried out in 20 μl reaction volumes containing 20 ng genomic DNA template, using SequalPrep Long PCR Kits with deoxyribonucleotide triphosphates (Invitrogen, USA). Reactions were optimized according to the manufacturer’s recommendations and contained 1× reaction buffer, 2% dimethyl sulfoxide, 0.5× SequalPrep EnhancerA, 1.8 U SequalPrep Long Polymerase, and 250 nM each of the forward and reverse primers. The thermal cycling program was 2 min at 94°C for denaturation, followed by 35 cycles of 10 s at 94°C, 30 s at 60°C, and 6 min 30 s at 68°C. The final extension step was 5 min at 72°C.
Amplicon size was verified on 0.7% agarose gels, and concentrations were measured using Quant-IT PicoGreen BR Kits (Invitrogen). For each patient, the PCR products were pooled in an equimolar ratio. The pooled PCR products were sheared using an ultrasonicator (Covaris S2; KBioscience, United Kingdom) to produce fragments of approximately 200 base pairs. Subsequently, sheared samples were purified with MiniElute Purification Kits (Qiagen GmbH, Germany) and the quality checked on an Agilent 2100 Bioanalyser (Agilent Technologies, Germany). The SPRIworks Fragment Library System I (Beckman Coulter, USA) was used for tagged library preparation. Ten samples can be processed in parallel. During the process, unique indexed adaptors are ligated to each pooled sample. In the next step, individual libraries were subjected to enrichment. Phusion High-Fidelity Master Mix (ThermoFisherScientific, USA) and 3′PTO modified primers (Fisher Scientific UK Ltd., United Kingdom) were used to amplify the libraries:
PTO-F 5′-AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATC*T-3′
PTO-R 5′-CAAGCAGAAGACGGCATACGAGATCGGTCTCGGCATTCCTGCTGAACCGCTCTTCCGATC*T-3′
PCR was carried out in 50 μl volumes containing 5 μl of library DNA, 25 μl of Phusion HF Master Mix, and 500 nM of each primer. The thermal cycling program was 98°C for 30 s, then 12 cycles of 98°C for 10 s, 65°C for 30 s, and 72°C for 30 s, and the final extension was 72°C for 5 min. Enriched libraries were then purified with Agencourt® AMPure® XP (Beckman Coulter) according to the manufacturer’s protocol. Purified libraries are pooled in equimolar ratios and loaded on one lane of a flow cell for cluster generation. Libraries were subsequently sequenced on an Illumina GAII® or MiSeq®.
Output files from the sequencing platform were sorted according to barcode tags using Illuminator Data Extractor (preprocessing software), and sequence analysis was performed using Illuminator software,16 both of which are freely available on the University of Leeds Web site.* This software filters out base calls with a Phred (Q) score less than 25. Reads were aligned against GeneBank genomic and coding reference sequences. For RYR1, the genomic reference sequence was NC_000019.9 (38923137-39079419) and transcript NM_000540.2. For CACNA1S, the genomic reference sequence was NC_0000001.10 (201008331-201081807) and transcript BC133671.1.
Testing of Missing Fragments, Variants in Family Segregation Studies, and Independent Samples
Sanger sequencing was used routinely for RYR1 exon 91, since amplification of long PCR fragments failed, and for exon 102 in all but one case for the same reason, despite protracted attempts to optimize primers. The primers used for Sanger sequencing to provide full coverage of these exons are provided in appendix 2. The primers for Sanger sequencing of other exons are available from the authors on request.
When a variant was identified in an MH patient or an EHI patient who tested MH susceptible by IVCT, there was a possibility of testing further family members to look for cosegregation of variant and disease phenotype. Sanger sequencing was used to test RYR1 and CACNA1S variants, identified through NGS, on additional family members (primers not supplied).
PCR based restriction fragment length polymorphism and amplification-refractory mutation system assays were designed for specific variants and used to screen unrelated MH-susceptible and MH-normal DNA samples. c.641C>T was screened by PCR and restriction fragment length polymorphism (F5′-ACC CTT GGC CTG AAA ATA CC, R5′-TGA AGT CAA GGG TTC AGC TC and BccI). c.9676G>C was also screened by PCR and restriction fragment length polymorphism (F5′-GCA CTG CAG CCT GAG TAA CA, R5′-CCC CCG AAC CAT AAA CTC TG and BsrBI). c.11958C>G was tested using an assay combining PCR products from exon 87 (the mutation site) and exon 24 to differentiate products after digestion with EcoNI (exon 87: F5′-GTG ATC CCT GAT CCC TTC TC, R5′-GAA GCA GGT GGA TGG AGA C; exon 24: F5′-GAC AAG GGT CAG CAG TCA GG, R5′-GGG TCA GAG TTG GGG TAG GA). c.12028G>A in exon 88 was tested by an amplification-refractory mutation system assay (mutation-specific primers: F′ CTA GGA CTC AAG CCA GAT CA, R′ GAT GGG GTT GAG GAT TAG GG; and exon 44 control primers: F′ GGG AGG TCT CTG ATG GTG, R′ CGG GAG ACT CAC TGC TCG).
Prediction of Variant Pathogenicity
To predict the possible functional relevance of all new variants, the impact of the amino acid substitution on the structure and function of the protein was estimated using the PolyPhen-2 (Polymorphism Phenotyping v2)17 and Combined Annotation–Dependent Depletion18 bioinformatics tools. The PolyPhen-2 program assigns the variant a score between 0 and 1.0: the higher the score, the more likely is the variant to be pathogenic. It should be noted that no bioinformatics tool has been validated for RYR1 or CACNA1S. Our previous work suggests that a PolyPhen-2 score of more than 0.95 is sensitive but not specific for pathogenic RYR1 variants.15 The Combined Annotation–Dependent Depletion is a recently published method that combines information from multiple sources into a single C-score. The C-score provides a better correlation with pathogenicity than any other in silico prediction program or combinations of these18 but has not been validated for, or even applied in the literature to, RYR1 and CACNA1S variants.
Results
Quality of Sequencing
For all samples, the sequence quality reached the quality threshold and read depth was above the commonly applied minimum for diagnostic use of 50 by a considerable margin in the great majority of fragments. The mean read depth by exon was 403 (see table 3 for read depth for each exon). Despite extensive efforts to optimize primers, sequence quality was insufficient to reach the base-calling threshold in all samples for RYR1 exon 91 and all but one sample for RYR1 exon 102. Sanger sequencing was used for these exons and other exons that failed the quality threshold in individual samples. From the first batch of samples, RYR1 fragments 4 and 15 (table 1) failed in three samples and two samples, respectively. For subsequent batches of samples, the primers for these fragments were redesigned with good results (table 3). Otherwise, RYR1 fragments 13 and 16 (table 1) and CACNA1S fragments 1 and 3 (table 2) failed in one patient each, presumably as a result of pipetting errors.
RYR1
Nineteen RYR1 variants, four of which are novel (p.2248R>C, p.3226E>Q, p.4010E>K, and p.4230M>R), were found in 20 MH patients (table 4) and 11 variants, three of which are novel (p.492R>H, p.4282L>V, and p.4331A>T), in 13 EHI patients (table 5). Three polymorphic changes were found in the RYR1 coding sequence, based on frequencies supplied for all populations on the Exome Variant Server† (table 6), and a further two variants have been reported not to segregate with the MH phenotype (p.1787P>L19 and p.3253I>T20 ). All variants were seen as heterozygotes. RYR1 variants described in this article are illustrated in figure 1, and their PolyPhen-2 scores and C-scores are given in table 6.
Previously Reported RYR1 Variants.
We found three variants (p.163L>R,14,21 p.552R>W,22 and p.614R>C23 ), previously reported in MH patients, that have functionally characterized effects consistent with a pathogenic role in MH. Four further variants previously described in the literature in MH patients (p.177R>C, p.1342S>G, p.1571I>V, and p.3986D>E)14,19,24–26 are yet to be functionally characterized. p.177R>C affects a well-conserved amino acid, segregates with the MH-susceptible phenotype, and was not found in 100 control chromosomes.14,19,24 We found RYR1 c.11958C>G p.3986D>E in a further 3 of 541 independent MH-susceptible individuals and none of 177 MH-negative controls. Of the family members tested in all four families, disease phenotype and variant status were concordant but only in a small total number of meioses. p.3986D>E has a PolyPhen-2 score of 0.999 and C-score of 12.60.
The variant c.641C>T; p.214T>M was found in a patient who had a clinically suspected MH reaction during anesthesia but who did not have an IVCT. Two brothers of the proband did have an IVCT, however, and the p.214T>M variant was found in one brother who tested MH susceptible by IVCT but not the other brother who tested MH normal, which is consistent with the variant being associated with MH susceptibility. The PolyPhen-2 score for p.214T>M is 1.000 and the C-score is 14.48. Furthermore, p.214T>M was found in 3 of 556 MH-susceptible samples, and all members of the three families show concordance of disease and/or IVCT phenotype with variant status. All of 177 MH-negative samples tested lacked the substitution.
Variant p.4295A>V was described by Jeong et al.27 as a compound mutation (with p.2435R>H) in an MH family with histological minicores and elevated serum creatine kinase: some individuals developed a late onset myopathy. In our study, this variant appears in an EHI patient who tested MH normal by IVCT.
RYR1 c.14168G>A; p.4723R>H was found in a patient who had a clinically suspected MH reaction confirmed by IVCT. Despite annotation through the 1000 Genomes Project,28 the variant did not appear on the Exome Variant Server (table 6). However, the variant was found in the proband’s son who was MH negative by IVCT. No further work was carried out on this variant.
RYR1 c.4178A>G p.1393K>R was found in a patient with a clinically suspected MH reaction and detected in two further MH families through screening 556 unrelated MH-susceptible samples. Subsequent testing of family members revealed individuals discordant for variant status and disease phenotype in both families, which is perhaps not surprising considering the minor allele frequency reported in the Exome Variant Server, and suggesting that this variant is unrelated to MH status.
Novel RYR1 Variants.
The four new variants found in MH patients were located in exons 41, 65, 88, and 91. The c.9676G>C; p.3226E>Q RYR1 variant was found in a MH proband whose mother was tested MH susceptible and also carried the variant. The brother and maternal cousin of the proband did not carry the variant and were tested MH negative by the IVCT. The PolyPhen-2 score was 0.997 and the C-score 14.93. This substitution was not seen in 211 MH-negative or 530 independent MH-susceptible samples. There were no family members to conduct further studies on p.2248R>C (PolyPhen-2, 1.000; C-score, 15.92) or p.4230T>G (PolyPhen-2, 0.932; C-score, 12.83).
Variant p.4010E>K was found in a MH-susceptible woman who had four other RYR1 variants (table 4): p.1571I>V, p.2060G>C, p.3366R>H, and p.3933Y>C. Her son was the clinical proband and was too ill for an IVCT: he was negative for all variants except p.4010E>K. The PolyPhen-2 score for this variant is 0.07, which suggests that it is unlikely to be damaging, although the C-score is 15.14. Two hundred MH-negative controls and 422 MH-susceptible individuals tested were negative for the variant. The father of the proband was MH negative by IVCT and negative for all variants.
One of three new RYR1 variants found in EHI patients is c.1475G>A; p.492R>H, which occurs in an EHI patient found to be MH susceptible by IVCT. At the same amino acid position, there is another variant described in the 1000 Genomes Project,28 c.1474C>T; p.492R>C. There were no family members available for segregation analysis. The PolyPhen-2 result for the new variant is 0.986 while the C-score is only 7.26. The other new variants identified in EHI patients, p.4282L>V and p.4331A>T, occur in EHI patients found to be MHN by IVCT, and so no further family members have been tested by IVCT. The PolyPhen-2 scores and C-scores (table 6) suggest that they are benign.
CACNA1S
Eight of nine missense variants found in the CACNA1S coding sequence were previously described.7,8 All of these are listed in the 1000 Genomes Project (table 7). In our patients, four polymorphic changes were found in the CACNA1S coding sequence, based on frequencies supplied for all populations on the Exome Variant Server, and a single additional low-frequency variant was seen, which had been reported as polymorphic elsewhere (table 7). Consistent with polymorphic status, c.206C>G; p.69A>G and c.1373T>A; p.458L>H were homozygous in some of the sequenced samples. A variant with a low minor allele frequency (p.683R>C; minor allele frequency, 0.00976) was found in an EHI patient who had tested MH negative by IVCT (table 5). The relevance of this variant to EHI remains to be established.
In the MH group, we found CACNA1S variant c.3026C>A, ACG/AAG; p.1009T>K in one individual (tables 4 and 7): this variant was previously found by exome sequencing in another family from the United Kingdom.15 In the 1000 Genomes Project, there is an annotated variant changing threonine at position 1,009 to methionine (c.3026C>T; ACG/ATG). The Exome Variant Server provides a minor allele frequency of 0.000154 for the T>M substitution. Both substitutions are predicted to give probably damaging amino acid changes with PolyPhen-2 scores of 1.0 and high C-scores (table 7). Segregation studies showed the new variant not to segregate with the MH IVCT phenotype in the family from the current study. Although three individuals susceptible to MH carried the variant and six MHN individuals did not carry the variant, there were two MHN individuals who carried the variant. This is a similar situation to the family with the same variant reported by Kim et al.,15 where there was a logarithm of odds score of 2.12 but one MHN individual carried the p.1009T>K variant.
Discussion
We have used NGS technology to sequence the coding regions of the RYR1 and CACNA1S genes in 29 MH patients and 28 EHI patients. By limiting our analyses to two genes, we were able to simultaneously sequence samples from multiple patients while achieving very high read depths (table 3) in all CACNA1S exons and all but two RYR1 exons. Combined with the high base-calling quality of the sequencing platforms used, this provides a high level of accuracy of sequencing in an efficient manner (once the primers were optimized, preparation of the enriched libraries for sequencing took 2 weeks). We had to resort to Sanger sequencing for RYR1 exons 91 and 102, where the quality of NGS did not meet our quality thresholds, presumably because of the high proportion of guanine and cytosine bases in these regions.
Testing for diagnostic variants by amplification-refractory mutation system and PCR-restriction fragment length polymorphism methods is time-consuming and usually a staged procedure, taking place over a considerable period of time. Necessarily, the most prevalent variants are likely to be assayed first. An NGS diagnostic approach is more efficient with individuals fully sequenced for all variants at one time, with both the diagnostic variants and the wealth of uncharacterized variants, many of which may still prove to have functional relevance, detected. An NGS approach using long-range PCR has been employed successfully for diagnostic detection of BRCA1 and BRCA2 mutations in cases of familial breast cancer.29,30 A real clinical and possible ethical problem with NGS approaches to diagnosis is deciding what information is given to referring clinicians and patients concerning variants of unknown or uncertain significance. Guidelines for reporting variants have been produced by professional bodies in Europe‡ (joint British–Dutch guidelines) and the United States.31 The joint British–Dutch guidelines are especially useful as they provide a consensus view on the evidence required for inclusion in each class of variant. This includes the necessity to demonstrate the functional effects of missense variants before the variant can be considered to be pathologic, reiterating the position of the European MH Group with respect to the use of DNA findings in MH diagnosis.32 The guidelines also caution against reliance on in silico predictions of pathogenicity in the diagnostic context and we would agree with this based on our previous analyses of RYR1 and CACNA1S in control populations.15 We propose to limit the use of PolyPhen-2 and C-scores as part of the process to prioritize variants for functional studies.
In the MH cohort, we detected three variants that meet the European MH Group criteria for diagnostic use in assigning high risk of MH susceptibility. DNA from the individuals carrying these variants had not been sequenced before this study, and screening for a full panel of diagnostic mutations was not complete in all cases. One of the variants detected was yet to be screened in the affected individual and another had been missed on routine screening, highlighting the sensitivity of NGS sequencing compared with conventional screening, including Sanger sequencing in this context.15,29 Of the remaining 26 MH patients, we detected variants of possible clinical relevance in 10 patients. In the 28 EHI patients, we detected variants of possible clinical relevance in seven individuals, two of whom had abnormal IVCT and five of whom had normal IVCT responses. In the MH cohort, there was only one CACNA1S variant of possible relevance, the remainder being in RYR1, while three of seven variants of possible relevance in the EHI cohort were in CACNA1S.
This study represents the first report of sequencing the complete coding sequence of RYR1 and CACNA1S in a cohort of survivors of EHI who have been physiologically characterized using heat tolerance testing and investigated by IVCT. Other than for several common variants, the functional and clinical relevance of the variants found in our EHI cohort is unclear. Uncommon variants were found in RYR1 and CACNA1S in EHI patients who tested either positive or negative in the IVCT. However, the prevalence of susceptibility to EHI is unknown, making interpretation based on variant frequency difficult. This is compounded by the absence of family segregation studies and indeed uncertainty regarding the genetic model operating in heritable predisposition to EHI. It is more than 20 yr since we first reported familial IVCT abnormalities in survivors of EHI,33 but the sensitivity and specificity of the IVCT for identifying skeletal muscle abnormalities associated with EHI is unknown. Rather than trying to define crude IVCT cutoffs for this purpose (which has been clinically useful in MH susceptibility), it may be more fruitful to utilize the extensive quantitative information that can be derived from these tests, as we have done when analyzing the effects of different RYR1 mutations in MH-susceptible patients.34 As all of the variants that were found in both MH and EHI cohorts were polymorphisms or of uncertain significance, our findings are not helpful in advancing understanding of the relationship between the two conditions.
Variants in this study were detected throughout RYR1. Due to the size of RYR1, attention historically focused on three hotspot regions where the first diagnostic variants were found.34 Sequencing of the entire gene is revealing variants throughout the gene. The RYR1 protein has many direct, diverse protein and subunit interactions. Although particular regions and residues will be conserved and functionally important, the expanding range of variants is not surprising. To date, the assumption has been made that primary mutations causing MH are likely to be missense, coding variants. However, the expression of messenger RNA in heterozygous MH-susceptible individuals has been explored by Grievink and Stowell,35 and their findings suggest that allele-specific differences in RYR1 expression may also contribute to phenotype. The possibility that noncoding variants are involved has not been explored.
Although MH shows an autosomal dominant pattern of inheritance, in some MH families discordant individuals have been identified, where IVCT-confirmed phenotype is discordant with genotype,36–38 which suggests complexities in the genetics of the disease.11 One explanation for discordancy in IVCT-tested MH-susceptible individuals is that there are some as yet undiscovered variants in RYR1 or CACNA1S or other genes, which contribute to MH phenotype, while modifier loci have been implicated39,40 : NGS may help to resolve discordancy in some families.
The current strategy for MH is to regard individuals carrying a familial diagnostic variant as susceptible. Diagnosis is conservative, and these individuals would not require an IVCT. However, individuals who do not carry a familial diagnostic variant cannot be regarded as unaffected, due to our incomplete knowledge of the genetics of MH, and these individuals undergo the IVCT. Of course, this strategy is only appropriate for families where a proband has been screened for mutations and a familial diagnostic variant already identified. Incomplete knowledge concerning the genetics of MH means that a sensitive test based on identification of particular diagnostic variants is some way away. The last decade has seen many groups sequencing and cataloguing variants relevant to MH, some with information relating to diagnosis, focusing on RYR1 and CACNA1S. Recently, NGS methods have been employed to sequence genes involved in excitation–contraction coupling20 or the whole exome.15,41 To expand the range of diagnostic variants, parallel efforts are required to identify further variants in RYR1, CACNA1S, and other loci and to obtain functional information on the nature of the variants.
A number of samples sequenced for RYR1 and CACNA1S in the present study carried several variants, particularly those of higher frequency. Minor allele frequencies from public databases, or inferred from our own screening, were combined with other information for interpretation of the likely nature of variants (see appendix 3 for an example). The description of variants as “polymorphic” was not necessarily consistent across possible sources. Although the prevalence of MH is difficult to assess, primary causative missense mutations are highly unlikely to have frequencies greater than 0.001. However, the notion of common polymorphisms, or even rare variants, modifying the effects of mutations, such as the predicted damaging change CACNA1S c.3026C>A, cannot be ruled out in some instances.42,43 RYR1 variants are discussed here in the context of MH. However, RYR1 mutations are also known to be responsible for rare, congenital myopathies, which show both dominant and recessive modes of inheritance44 such as autosomal recessive central core disease.45,46 NGS will undoubtedly reveal variants of interest to a number of disorders.
In summary, we have reported NGS using long-range PCR of RYR1 and CACNA1S, the only two loci with variants with functional evidence supporting their role in MH. NGS is a fast and efficient diagnostic tool and a significant improvement on the current screening of diagnostic genetic variants. Additionally, complete coding sequence information will rapidly expand our knowledge of variants associated with MH and related disorders. However, functional studies will then be required to define the importance of these variants to the condition in question.
Acknowledgments
Support was received from the National Institutes of Health (Bethesda, Maryland): National Institute of Arthritis, Musculoskeletal, and Skin Diseases (2P01 AR05235 to Dr. Hopkins); the National Institute of General Medical Sciences (T32 GM086270 to Dr. Kim); and the National Center for Advancing Translational Sciences (KL2 TR000421 to Dr. Kim). Support was received from the Medical Research Council (London, United Kingdom), Medical Bioinformatics Centre (grant MR/L01629X/1) to the University of Leeds (Leeds, United Kingdom). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Competing Interests
The authors declare no competing interests.
Available at: http://dna.leeds.ac.uk/illuminator/. Accessed October 3, 2014.
Available at: http://evs.gs.washington.edu/EVS/. Accessed October 12, 2014.
Available at: http://www.acgs.uk.com/media/774853/evaluation_and_reporting_of_sequence_variants_bpgs_june_2013_-_finalpdf.pdf. Accessed October 12, 2014.
References
Appendix 1. Summary of the Heat Tolerance Assessment Undertaken at the Institute of Naval Medicine, Alverstoke, United Kingdom
For the British military, there are published guidelines,47 which require all military personnel who have suffered a significant episode of heat illness or required urgent admission to hospital with heat illness to be subsequently referred and assessed by the Institute of Naval Medicine. A function of the assessment is to establish if thermoregulation is affected by an underlying metabolic, biochemical, or physiological disorder.
Exertional heat illness encompasses a spectrum of disorders deriving from the combined stresses of exertion and thermoregulation.48 Such disorders may include heat-related dehydration, stroke, cramps, exhaustion, injury, stress, syncope, rhabdomyolysis, acute renal failure, and hyponatremia.49,50 Individuals across the whole spectrum may be assessed. The assessment usually takes place a minimum of six calendar weeks following the original episode or following full biochemical recovery. Individuals also undergo a limited guided and supervised return to physical activity before the assessment.
The heat tolerance assessment is conducted in an environmentally controlled climatic chamber (dry bulb temperature, 34°C (± 0.25°C); 40% (± 2%) relative humidity; wet-bulb globe temperature index, 27°C; and wind speed, 7 km/h). Patients first undergo a full medical examination, including cardiac screening and 12-lead electrocardiogram. Anthropometric measurements, height weight, estimation of body surface area, lean mass, and fat mass are taken. This is followed by a maximal oxygen uptake assessment (Vo2 max) with electrocardiogram telemetry. This exercise test is conducted wearing shorts and a T-shirt on a treadmill in the climatic conditions described. Following this assessment, patients are allowed an hour rest in a cool room to allow body temperature to return to normal. They are also allowed to drink ad libitum.
Before undergoing the full heat tolerance assessment, patients are weighed nude, they self insert a rectal thermistor to a depth of 10 cm, are instrumented with further skin thermistors for the measurement of skin temperature,51 electrocardiogram telemetry, and heart rate monitors. They, then dress in combat trousers, T-shirt, and combat jacket. They reenter the environmental chamber and are asked to march on the treadmill, carrying a weighted rucksack at a speed and gradient that is adjusted until they are working at 60% of their Vo2 max. Steady-state Vo2 measurements are taken at regular intervals throughout the assessment. The duration of the assessment is a minimum of 60 min but may proceed to a maximum of 90 min. After 30 min of marching with the load, subjects remove the rucksack and jacket and then continue walking. Patients who are able to rapidly attenuate the rate of rise of their deep body temperatures and demonstrate the ability to achieve plateau in their core temperature are considered to be able to demonstrate physiological thermoregulation and the attainment of thermal equilibrium. Those who cannot do so, or those whose deep body temperatures rise rapidly toward the safety threshold of 40°C are considered heat intolerant. Other measures such as sweat rate, which can be shown for body surface area, Vo2 consumption, heart rate, and rating of perceived exertion are made, and occasionally blood sampling for biochemical analysis is considered if clinically indicated.
Patients who have an uneventful heat tolerance assessment are passed as fit to return to fitness and exercise training and are eventually returned to full military activities. Those who are thermally intolerant of exercise in the proscribed conditions remain in a protected medical category. They are retested at intervals of no less than 6 weeks. Persistent heat intolerance under these conditions, especially when this is combined with evidence of exertional rhabdomyolysis or elevations of muscle markers after the assessment and the absence of other medical conditions, may be referred to The Malignant Hyperthermia Investigation Unit at the University of Leeds, United Kingdom.
Appendix 2. Primers for Sanger Sequencing
RYR1 Exon 91
Fragment 1
F5′-GCT GAC GGC GCC CTA TCC TGT
R5′-GCG CCG CCG CAG GCT GCG GTA A
Fragment 2
F5′-GCC GGG CCC TGC GAG GCC TCA
R5′-GGT GGG GTC GGG CAT GCC TGC C
Fragment 3
F5′-GCT CTG GGC AGC AGT GAC
R5′-ATC CCC CAT CTT TCC AAA AC
RYR1 Exon 102
F5′-AAT GTC GAA TGA ATG CGT GA
R5′-CTG GGC CTG CAT TCT TAG C
Appendix 3. Interpretation of the Results of a Malignant Hyperthermia Sample in Which Five Missense Variants in RYR1 Were Found
In one MH sample, five missense variants in the gene coding for RYR1 were found (p.1571I>V, p.2060G>C, p.3366R>H, p.3933Y>C, and p.4010E>K). Histopathology showed nonspecific myopathic changes in the skeletal muscle of our patient, but serum creatine kinase concentration was normal. The variants p.1571I>V and p.3933Y>C in cis were first described in one family with four missense changes by Tammaro et al.,26 while we found p.1571I>V in one United Kingdom family as a single variant (unpublished sequence data, March 2013, Hopkins PM, Leeds, United Kingdom). The Italian family members carrying these two variants were MH negative, but the patient who inherited an additional variant from her father (p.3903R>Q) was diagnosed by IVCT as MH susceptible. The MH-susceptible phenotype seemed to segregate with the p.3903R>Q variant, but there were discordant members in this family who were MH negative while carrying the p.3903R>Q variant, which may suggest that more than one variant is necessary to develop MH susceptibility.26 Duarte et al.45 also reported p.3366R>H and p.3933Y>C inherited in cis in a patient with a family history of serious adverse reactions to anesthetics. These variants were considered pathogenic as they were not detected in 150 control samples and involve with phylogenetically conserved residues. Since the RYR1 variant p.2060G>C is polymorphic14,23 and the novel variant p.4010E>K shows a very low PolyPhen-2 score of 0.07, it could be suggested that a compound effect of the three other variants may contribute to the susceptibility of our patient. However, a limited family study does not support this suggestion since p.4010E>K was the only variant clearly segregating with disease in this particular family. It may well be that variants in other genes contribute to, or are wholly responsible for, MH susceptibility in these cases.