Whole genome genotyping and sequencing of MYLK

DNA was extracted from peripheral blood leukocytes of two patients, VII-3 and VI-26 (Fig. 1), using Qiagen’s QIAamp 96 DNA QIAcube HT Kit (#51331, Qiagen, Venlo, Netherlands) for whole genome genotyping to identify chromosomal regions with extended haplotype homozygosity. After purifying the DNA, whole genome genotyping was done using Illumina’s HumanCytoSNP-12v2.1 DNA Analysis BeadChip Kit (#WG320–2102; Illumina, CA, USA). According to the kit’s manufacturer, this array is a powerful whole-genome scanning panel that is designed for efficient high-throughput analysis of genetic and structural variations in the entire genome that are the most relevant to human disease, including the loci and genes which are known to be involved in TAA and TAAD [9–13]. Disease genes were also detected in the DNA samples using Superlink-Online SNP (http://cbl-hap.cs.technion.ac.il/superlink-snp/), an online system for exact and approximate genetic linkage of single nucleotide polymorphism (SNP) data in large pedigrees [17]. The exon sequences and the flanking intron sequences of MYLK were amplified by polymerase chain reaction (PCR), and the PCR products were sequenced by Sanger sequencing using the standard protocols. Finally, a segregation analysis to test for segregation of the rare variant in MYLK with disease in the family was done using DNA that were extracted from peripheral blood leukocytes of four more patients and 36 apparently healthy family members, and the paraffin-embedded aortic artery samples of two patients (Fig. 1, Table 1).

Myosin light chain kinase activity assays

The plasmid of human full-length short form of myosin light chain kinase was obtained (Oragene) and inserted into a pCMV6-XL4 vector with additional amino acids (Ala-Gly-Ala-Gly-Ser-Thr-Val) followed by 3 FLAG tags at the amino-terminus. Oligonucleotide-directed mutagenesis was introduced to produce the mutation with A1491S (G > T). The construct was confirmed by restriction enzyme digestion and DNA sequencing. HeLa cells were transfected with the FLAG-MYLK constructs using lipofectamine 2000 transfection reagent (Invitrogen) and cell lysates harvested 48 h later. Wild type and mutant MLCK assessed by immunoblotting with anti-FLAG monoclonal antibody (M2, Sigma). The proteins were purified using ANTI-FLAG M2 Affinity Gel beads (Sigma). Kinase activity was determined based on P³² incorporation into the MYLK substrate MYL (Biomol). K_m and V_max values were obtained from non-linear regression program (Prism 5). Calmodulin activation was assessed by P32 incorporation into MYL in the presence of saturating concentration of CaCl₂ and varying amounts of calmodulin. K_CaM values were assessed using non-linear regression program.

Group I patients

This group comprised 6 patients, V-2, VI-23, VI-26, VI-27, VII-1, and VII-3, with extensive arterial dissections involving many arteries. On presentation, these family members complained of dyspnea and sudden onset of severe pain in the chest, abdomen, or upper and lower extremities (Table 1). Diagnosis was made in 4 patients, within three hours after their admission to the hospital; for the remaining two patients the diagnosis was made two and eight days after their admission. All were diagnosed with aortic dissections, 5 involving the ascending aorta extending to the descending aorta (Stanford classification type A dissections) and one involving just the descending aorta (type B dissection) (Table 2). None of the patients had unequivocal evidence of an aortic aneurysm at the time of the dissection. An aneurysm of subclavian artery was identified in three patients, VII-26 and VII-27, who were sisters, and VII-3, and an aneurysm in the innominate and carotid arteries was found in two patients, VII-27 and VII-3. Patient VI-26 had a dilatation of the aortic arch between the left subclavian and carotid arteries (Table 2, Fig. 2).

The outcome of this group was poor. Five of the 6 patients died shortly after presentation and diagnosis. Patients V-2, VII-1, and VI-27 died within three weeks after their admission to hospital. Emergent surgical repair of the ascending aorta was done on patients, VI-23, VI-27, and VII-3. The repairs were successful in patients VII-3 and VI-23 and they were discharged after four and six weeks, respectively, to a rehabilitation center. These two patients were re-admitted after four months (patient VII-3) and six months (patient VI-23) because of extensive vascular dissection of the abdominal aorta, and renal and superior mesenteric arteries and multi-organ failure, and they died within 48 h of their admission. TAAD and AAAD were diagnosed in patient V-2 on post-mortem examination. Patient VII-1 was initially diagnosed as having a mesenteric vascular embolus and underwent surgery for a segmental bowel resection. During the surgery, a abdominal aortic dissection was found, leading to bowel infarction. The attending surgeons were not able to complete an endovascular graft repair of the aorta and the patient died due rapid hemodynamic deterioration. Patient VII-26 is the only surviving patient of Group 1. She is currently 52 years of age. 8 Years after presenting with the type 3B aortic dissection, she had an extension involving the left renal artery, which arose from a false aortic lumen and resulted in a dysfunctional left kidney (Fig. 2d). Since she was hemodynamically stable at the time of diagnosis, she has been monitored by regular echocardiography and CTA and treated with an angiotensin converting enzyme inhibitor and furosemide and has been stable ten years. The siblings and close relatives of patients VII-26 and VII-27 were interviewed and their medical files reviewed. Our investigation revealed that both patients (VII-26 and VII-27) had complained of bilateral arm weakness, effort-induced shoulder pain, and general fatigue. The shoulder pain was induced by minor physical activity, such as raising their hands above the head, as done in hair dressing, and while doing housekeeping chores.

In the absence of a definitive diagnosis on post-mortem examination, we hypothesize that the death of individual VI-28, a 35-year-old brother of patients VI-26 and VI-27 and the son of the patient V-5, was due to rupture of an aortic dissection that either followed or caused a minor road accident.

Group II patients

This group comprised three patients, V-15, V-22, and VI-24, who presented with either TAA and aortic dissection or both confirmed by echocardiography and CTA. These individuals presented later in life than Group 1, in the sixth or seventh decade. In these cases, the diagnosis was made when the individual presented with acute chest pain, effort-induced breathlessness, and weakness. Patients V-22 and VI-24 had type-2 diabetes mellitus and TAA and TAAD were diagnosed after they were hospitalized with acute ischemic heart disease. Review of the medical history and the absence of a definitive diagnosis in the post-mortem reports, we suspect that three family members, IV-5, IV-7 and IV-8, who were either the parents or grandparents of the group I and group II patients, died because of complications of dissections. A few days before their deaths, these individuals all complained of chest and abdominal pain.

Whole-genome genotyping and sequencing of the MYLK gene

Since the mode of inheritance of familial TAAD is usually autosomal dominant [12], we posited that the six group I patients are homozygous for the familial mutation, and that this mutation was the underlying cause for the phenotype of these patients. Accordingly, we used the DNA of patients VII-3 and VI-26 to perform single nucleotide polymorphisms (SNPs) arrays and regions of shared homozygosity in their genomes were identified. We found an 18 Mb homozygous region on chromosome 3 (3q21.1), which harbors MYLK, a gene known to predispose to aortic dissection without aneurysm [9, 12]. Consequently, we sequenced the coding exons and their flanking regions of MYLK using the DNA of patient VI-26 (Fig. 3). We identified a missense mutation, c.4471G > T, Ala1491Ser, in exon 27 MYLK. This change was not found in the ClinVar, HGMD, Exome Aggregation Consortium (ExAC) and the 1000 genomes (http://www.1000genomes.org) databases. The mutation was not detected in the DNA of 100 healthy individuals of the same ethnic background. A BLAST comparison of the amino acid sequence of orthologous MLCK proteins from 35 different species revealed that Ala1491 is conserved (data not shown).

Moreover, we performed the segregation and LOD score calculation studies and for the disease model we used was additive with a penetrance of 0.3 for one copy of the mutation, a penetrance of 0.9 of two copies of the mutation, and using a disease prevalence of 0.001. The data consists of measuring the MYLK gene, and marking the results by Homozygous, Hetrozygous, and Wild. Exact two-point LOD score of 3.02 was computed using the Superlink-Online system [17].

Effect of MYLK p. Ala1491Ser on myosin light chain kinase activity

The MYLK variant, Ala1491Ser, falls in the kinase domain of the protein. Kinase activity of the Ala1491Ser protein showed a significant reduction in kinase activity compared to wildtype but no further decrease based on calcium/calmodulin activation (Fig. 5). Thus, the MYLK pathogenic variant does disrupt kinase activity but to a lesser degree than previously reported mutations causing autosomal dominant thoracic aortic disease [12].

Histologic studies

Medial necrosis was seen in all of the aortic samples from the living and deceased patients in whom an emergency endovascular repair was performed as seen in elastica – van Gieson aorta staining (Fig. 4).

Immunohistochemical staining for MLCK: Fig. 4-II, displays representative images of aortic specimens for patients VI-27, VII-1 and V-22 that were immunostained for MLCK. MLCK was barely not detected (score 0–1 for both intensity and extent) in the medial layer of the aorta from patients VI-27, VII-1 and was weak but diffuse (score 3) in patient V-22, compared with normal staining (score 6) as seen in normal aortic specimen.

Phenotype-genotype correlations

We then investigated whether a genotype-phenotype correlation exists in the five living patients and all living family members, and found that the c.4471G > T transversion segregated and adequately matched the clinical phenotypes of the 45 tested individuals. Specifically, three of the group I patients, VII-3, VI-26, and VI-27, were homozygous for the c.4471G > T mutation and their parents were heterozygous. The DNA that was extracted from the paraffin-embedded specimen of patient VII-1 had degraded and did not meet the quality requirements for Sanger sequencing analysis. We could not test the DNA of deceased patient VI-23 because there were no specimens for DNA extraction. When we tested the parents, V-20 and V-21, of patient VI-23, we found that both were heterozygous for the mutated allele. Accordingly, we inferred that the patient VI-23 was homozygous for the mutation because the clinical phenotype was well correlated with that of the group I patients (Table 1). We found 22 individuals were heterozygous for the mutant allele. Two of these individuals, V-22 and VI-24, were group II patients, and had successfully undergone reconstructive surgery for TAA and TAAD. One young asymptomatic carrier, VI-1, was of special interest because he was the son of the first documented patient in the family, V-2, who was diagnosed as having familial TAAD after analyzing specimens that were collected from the post-mortem examination. The carrier status of invidual VI-1 provides a strong evidence that patient V-2 was probably homozygous for the family mutation and his wife, V-3, most likely has the wildtype alleles. Twenty heterozygous carriers of the MYLK mutant allele (average age 45-years-old) were investigated for aortic dilatation using echocardiography. The diameters of the aortic root and aortic artery of 18 carriers were within the normal range (less than 4 cm, and 3.5 cm, respectively). In individuals V-7 and V-9 (brothers of patient V-2), the diameters of the aortic root (3.9 and 3.8 cm) and ascending aorta (3.4 cm/m² for both). Diameters of the the descending aorta of the two were outside the normal range (4.5 cm and 4.7 cm, respectively).

Four family members, IV-5, IV-7, IV-8, and V-15 (Fig. 1), died suddenly when they were in the six or seventh decade of their life. Aortic dissection was documented only in patient V-15, ten days after admission to hospital and before any surgical intervention could be done. We surmise that this patient was heterozygous for the mutated allele because three of his offsprings, VI-9, VI-12, and VI-13, are heterozygous while the other tested offsprings VI-5, VI-10, and VI-11 were homozygous for the wild allele. We also surmise that the family member IV-5, whose wife, IV-6, who is unrelated, is heterozygous for the mutated allele because his son, V-15, died of TAAD and the daughter, IV-16, and the granddaughter, VI-14, are both heterozygous for the mutated allele. The mother, VI-14, of the two deceased homozygous patients, VII-1 and VII-3, is also heterozygous for the mutated allele.