An Myh11 single lysine deletion causes aortic dissection by reducing aortic structural integrity and contractility

Pathogenic variants in myosin heavy chain (Myh11) cause familial thoracic aortic aneurysms and dissections (FTAAD). However, the underlying pathological mechanisms remain unclear because of a lack of animal models. In this study, we established a mouse model with Myh11 K1256del, the pathogenic variant we found previously in two FTAAD families. The Myh11∆K/∆K aorta showed increased wall thickness and ultrastructural abnormalities, including weakened cell adhesion. Notably, the Myh11∆K/+ mice developed aortic dissections and intramural haematomas when stimulated with angiotensin II. Mechanistically, integrin subunit alpha2 (Itga2) was downregulated in the Myh11∆K/∆K aortas, and the smooth muscle cell lineage cells that differentiated from Myh11∆K/∆K induced pluripotent stem cells. The contractility of the Myh11∆K/∆K aortas in response to phenylephrine was also reduced. These results imply that the suboptimal cell adhesion indicated by Itga2 downregulation causes a defect in the contraction of the aorta. Consequently, the defective contraction may increase the haemodynamic stress underlying the aortic dissections.


Myh11 ∆K/∆K aorta showed thickened media and adventitia. To evaluate the structural features of
Myh11 ∆K/∆K aortas, we excised the aortas (ascending, descending and abdominal) from WT, Myh11 ∆K/+ and Myh11 ∆K/∆K mice at 12 weeks of age (n = 5) for histological observation. The Myh11 ∆K/∆K aortas showed increased medial thickness (WT vs Myh11 ∆K/∆K , p < 0.05) and thickened adventitia (p < 0.05) but no significantly expanded aortic lumen (p = 0.30). Conversely, the Myh11 ∆K/+ aortas did not show any statistically significant increase in the thickness of the media or adventitia compared to the WT aortas ( Fig. 2A, B and Supplementary Fig. 2). None of the aortic regions showed a higher rate of dissection compared to the other regions (Supplementary Table 1).
In all Myh11 ∆K/∆K mice dissected, a shunt blood vessel connecting the distal aortic arch and pulmonary artery was observed (Fig. 2C) and resembled PDA, which is frequently associated with FTAAD. Furthermore, an enlarged bladder, swollen kidneys indicating hydronephrosis and hypoplastic uteri were observed in the Myh11 ∆K/∆K mice ( Supplementary Fig. 3A). Non-uniform staining, such as punch-out and gaps between the smooth muscle layers, was commonly observed in Myh11 ∆K/∆K bladders ( Supplementary Fig. 3B). Quantifying the thickness of the smooth muscle layers (myometria), we observed that the myometrium of Myh11 ∆K/∆K uterus were thinner than those of the WT (Supplementary Fig. 3C and D).

Cell adhesion of SMCs and composition of ECM are decreased in Myh11 mutant aortas.
To analyse the aortic microstructures, we examined the descending aortas of WT and Myh11 ∆K/∆K mice at 18 weeks of age by electron microscopy assuming that each aortic region was similarly predisposed to dissection since occurrence of dissection was not biased towards a specific region (Supplementary Table 1). The results revealed a thinner nuclear envelope and numerous granular components in the mutant nucleus (Fig. 3A, upper row), which appeared in the nuclear morphology as prevalent euchromatin and less heterochromatin, indicating active transcriptional activities. These nuclear features suggested a latent shift from the contractile phenotype to the synthetic phenotype in a part of mutant SMCs. Myh11 ∆K/∆K SMCs also showed an attenuation of cell adhesion to other adjacent cells and were larger than WT SMCs (Fig. 3A, middle and bottom rows). Furthermore, we detected some of features often observed in dead cells in Myh11 ∆K/∆K SMCs. Increased organelles were often observed in Myh11 ∆K/∆K SMCs, and debris was occasionally detected (Fig. 3B). We also found concentrically layered osmiophilic material, known as myelin figures, inside or outside the mutant SMCs, which were quite rare in WT (Fig. 3C).
To evaluate the structural integrity of the aortas, we further analysed the cell adhesion and elastin lamella by using comprehensive images of cross-sectional aortas, which were obtained from large-scale electron microscopic images. We measured the length of the intercellular adhesions of longitudinally sectioned SMCs (Supplementary Fig. 4A; a detailed protocol is provided in the supplementary material online), and the intercellular adhesions were significantly shortened in Myh11 ∆K/∆K aortas (WT vs Myh11 ∆K/∆K , p < 0.01, Fig. 3D). The area of elastic lamellas, identified as an area of plain texture (Supplementary Fig. 4B; a detailed protocol is provided in the supplementary material online) was also decreased in Myh11 ∆K/∆K aortas (p < 0.01, Fig. 3E).
Contractile function was attenuated in Myh11 ∆K/∆K SMCs. Considering these morphological features in the Myh11 ∆K/∆K aortas, we hypothesised that Myh11 K1256del negatively affects SMC function. To evaluate SMC contractility in the aortas, we measured the isometric force of aortic rings from WT, Myh11 ∆K/+ and Myh11 ∆K/∆K mice at 12 weeks of age in response to contractile agonists and vasodilators. The force developed in response to phenylephrine was significantly decreased in the thoracic aorta from Myh11 ∆K/∆K mice compared to that in WT mice (p < 0.01, Fig. 4A). The maximum force developed in response to phenylephrine or potas- www.nature.com/scientificreports/ sium chloride treatment was significantly reduced in the Myh11 ∆K/∆K aorta compared to the WT aorta (Fig. 4B). This reduced contractility in Myh11 ∆K/∆K SMCs may contribute to the decreasing mechanoadaptation of the aortic wall. By contrast, the vasodilation functions in response to acetylcholine (for endothelium-dependent vasodilation) or nitroprusside (for endothelium-independent vasodilation) were comparable between the WT, Myh11 ∆K/+ and Myh11 ∆K/∆K aortas (Fig. 4C).

Ang II induced aortic dissections in Myh11
K1256del mutant mice. The attenuation of the structural intensity and maladaptation against mechanical stress in the mutant aortas may predispose people to developing thoracic aortic dissection. To investigate the pathological mechanisms of aortic dissection associated with Myh11 K1256del, we administered Ang II (1000 ng/kg/min) to WT (n = 16) and Myh11 ∆K/+ (n = 15) males at eight weeks of age with osmotic pumps. Systolic blood pressure was measured before infusion pump implantation and after two weeks of Ang II treatment. Ang II-treated mice showed an increase in systolic blood pressure, but there was no significant difference between WT and Myh11 ∆K/+ mice (before implantations: p = 0.97,  5C). In contrast, intramural haematoma or aortic dissections were not induced in Ang II-treated WT mice. We also attempted to generate a sufficient number of Ang II-treated Myh11 ∆K/∆K mice for analysis, but it was difficult to obtain aortic samples because most mice died from aneurysmal ruptures. In a preliminary experiment, three of four Ang II-treated Myh11 ∆K/∆K mice suddenly died in from two to four days after the start of the infusion. We dissected these mice and found a rupture in the ascending aorta or the aortic arch (data not shown). Histologically, Ang II-treated www.nature.com/scientificreports/ Myh11 ∆K/+ aortas showed a fragmentation of the elastic lamellae and fibrotic tissue deposition and displayed luminal expansion (Fig. 5A, B). Significant difference was not observed in the expression of the Ang II type 1 receptor (AGTR1) in the WT and Myh11 ∆K/∆K aortas ( Supplementary Fig. 5E).
Myh11 del1256K pathogenic variation does not alter the expression of smooth muscle myosin heavy chain isoforms and the expression or phosphorylation of proteins involved in smooth muscle contraction. We investigated the gene and protein expressions of aortas from WT and Myh11 ∆K/∆K males at 12 weeks of age. Because the morphological features in mutant SMCs on electron micrograph suggests transformation from a contractile phenotype to a synthetic phenotype in SMCs, we analysed the expression of SM isoforms (SM1 and SM2) in the aortas, which indicate SMC differentiation [14][15][16] . However, we found that aortic smooth muscles does not undergo phenotypic modulation overall as seen in acute vascular injuries There is no significant difference in the endothelium-dependent or endothelium-independent vasodilation between the WT and Myh11 ∆K/+ aortas. www.nature.com/scientificreports/ because the SM1 and SM2 expression in the mutant aorta did not show any abnormality compared to WT mice at 12 weeks of age (Fig. 6A, B and Supplementary Fig. 11). To further investigate whether the expression or phosphorylation of proteins involved in smooth muscle contraction is altered in Myh11 ∆K/∆K SMCs, we measured the expression of these genes by RT-qPCR and protein levels by immunoblotting. However, no significant differences were noted between the WT and Myh11 ∆K/∆K aortas in the expression of α-smooth muscle actin (Acta2), SM-MHC (Myh11) and calponin (Cnn1) or in the level of myosin regulatory light chain (RLC) phosphorylation ( Fig. 6C-F and Supplementary Figs. 6A, 12). The phosphorylation level of focal adhesion kinase (FAK), a key molecule in the signalling cascades of focal adhesion, was also comparable between the WT and Myh11 ∆K/∆K aortas ( Fig. 6G and Supplementary Figs. 6A, 13, 14). The expression of a proliferation marker, proliferating cell nuclear antigen and Cyclin D1 were also comparable between the WT and Myh11 ∆K/∆K aortas ( Supplementary  Fig. 6B). We also investigated the gene expression of TGF-ß (Tgfb1) and the transcriptional factors of its downstream cascade [connective tissue growth factor (Ctgf), MMP2 (Mmp2) and MMP9 (Mmp9)], which are associated with the pathogenesis in TAD, but no significant differences in their expression were observed between the WT and Myh11 ∆K/∆K aortas (Fig. 6C).   Fig. 7A) 17 . The number of alkaline phosphatase positive colonies did not differ significantly ( Supplementary Fig. 7B, C). We then evaluated the pluripotency by generating embryoid bodies (EBs) by the hanging drop method and determined the tri-lineage differentiation of each genotype. As shown in the supplementary movie, the WT EBs produced beating cells. However, most of the cells that grew out of the Myh11 ΔK/ΔK EBs were non-beating and non-adherent, suggesting that the Myh11 K1256del pathogenic variant impairs the maintenance of pluripotency. Each genotype had a similar morphology for nine days after retroviral transduction, but the Myh11 ΔK/ΔK colonies did not maintain their morphology, and granular cells started to appear at passage 3 ( Supplementary Fig. 8A). We then transduced Nanog into the MEFs in addition to the Yamanaka factors. Surprisingly, the forced expression of Nanog along with the Yamanaka factors during somatic cell reprogramming allowed the Myh11 ΔK/ΔK iPSCs to maintain the ESC-like morphology ( Supplementary Fig. 8B).

Genes related to cell adhesion are downregulated in Myh11 ΔK/ΔK aortas and in SMC-lineage cells differentiated from iPSCs.
Nanog binds to an enhancer region of the human α-catenin gene (Ctnna2) (GeneCards website: http:// www. genec ards. org) 18 . The observation that iPSC stemness was improved by forced Nanog expression prompted us to examine the downregulated genes in the Myh11 ΔK/ΔK aortas we identified by RNA sequencing analysis for involvement in cell adhesion using the Gene Ontology (GO) database 19,20 . An examination of those genes for interactions with each other in the STRING database 21 then revealed that Ctnna2 had the highest number of interactions with the molecules related to intercellular adhesion (Supplementary Fig. 9). We further investigated the functional difference between WT and Myh11 ΔK/ΔK SMCs by inducing the differentiation of iPSCs into the SMC lineage by culturing the cells in media containing retinoic acid without the leukaemia inhibitory factor. At day 5 of differentiation, the cells had not assumed an SMC-like morphology (Fig. 7A). Myh11 was upregulated regardless of the genotype at day 3 of differentiation ( Fig. 7B), indicating the cells had committed to the SMC lineage. Integrin subunit α2 (Itga2) was downregulated in the aneurysmal aortas of SMC-specific Smad4 knockout mice 22 , and a STRING protein-protein interaction analysis showed that Itga2 had the second highest number of interactions with the focal-adhesion-related molecules that were downregulated in the Myh11 ΔK/ΔK aorta ( Supplementary Fig. 9). Measurements of the mRNA expression of Itga2 in the SMC-lineage cells also showed a significant reduction in Itga2 expression in the Myh11 ΔK/ΔK cells (Fig. 7C).

Discussion
We generated a novel mouse model of FTAAD with high reproducibility of aortic dissection and intramural haematoma by Ang II-treatment. We found that Myh11 K1256del leads to structural fragility and maladaptation against mechanical stress in aortas by decreasing the composition of elastin lamellae and SMC contractility. This altered property of Myh11 K1256del aortas increases association with the onset of aortic dissection. We predict that myosin filament assembly may be interrupted in the presence of K1256del based on molecular structure modelling 23 . Myosin is typically divided into two fragments: N-terminal heavy-meromyosin (HMM) and C-terminal light-meromyosin (LMM) 24 . HMM consists of the globular motor domain (S1) and the coiledcoil region connecting S1 and LMM (S2). LMM forms a thick filament by intertwining with the LMMs of other myosin molecules. K1256 is located on the N-terminal side of LMM, and the amino-acid sequence around K1256 adopts a typical pattern of a heptad repeat composed of hydrophobic residues at the first (a) and fourth (d) positions ( Supplementary Fig. 10A). These hydrophobic residues are required to form a coiled-coil structure by providing hydrophobic contacts between two heavy chains 25 . Supplementary Fig. 10B shows a structural model of the K1256-containing region (1226-1288 residues) of Myh11. The hydrophobic residues of WT, except for V1242, are at the interface between two heavy chains and contact each other, suggesting that this region can form a coiled-coil structure. In contrast, the hydrophobic residues located on the C-terminal side of K1256 are exposed to the outside by the K1256 deletion because the residues rotate 100° on the α-helix structure. This change in side-chain orientation likely disturbs the coiled-coil structure stably assembled by the hydrophobic contacts, which may partially affect HMM functions, such as thick filament formation.
Large-scale electron microscopic imaging is useful for comprehensively observing tissue sections; this is the first example of applying this method to analyse a luminal organ. We found a significant decrease in the area of elastic lamellae, suggesting a thinning of the elastic lamellae. As a previous study of FTAAD has shown, the thinning of elastic lamellae seems to be a common structural modification of aortas with the pathogenic variant that causes FTAAD 6 . Furthermore, by conventional transmission electron microscopy, we observed pathological features associated with cellular stress responses, such as debris and myelin figures. The expression and distribution of SM isoforms (SM1 and SM2), which are molecular markers of SMC differentiation [14][15][16] , showed no obvious changes in the Myh11 ∆K/∆K mutant. This suggests that smooth muscle in Myh11 K1256del does not undergo extensive phenotypic modulation.
In this study, aortic dissection and intramural haematoma developed in Myh11 ∆K/+ mice within two weeks of Ang II infusion. In our model, we observed a reduced contracting force of the aortic ring following stimulation with phenylephrine or KCl. This observation agrees with a previous report of attenuated SMC contractility induced by the deletion variant in the rod portion and in the motor domain of myosin 11 . Large-scale electron microscopy also indicated a reduction in the structural integrity of the aortic wall. Furthermore, the expression of the Ang II type 1 receptor in the Myh11 ∆K/∆K aorta was not significantly different from that in the WT, suggesting  26 . In the model, increased haemodynamic stress due to the stiffening of the aorta was proposed to cause dissection in aortas with reduced wall strength 26 . Thus, the amplified haemodynamic stress caused by attenuated contraction might have ultimately caused aortic dissection. The stabilisation of the submembranous cytoskeleton is viewed as important for the efficient force development of SMCs, and two pathways of smooth muscle contraction have been proposed 27 . One pathway involves myosin RLC phosphorylation 27 , but our RLC phosphorylation analysis by immunoblotting did not indicate any attenuation of this pathway in Myh11 ΔK/ΔK aortas. The other pathway involves mechanotransduction at focal adhesion junctions 27 . Our RNA sequencing of the Myh11 ΔK/ΔK aorta showed a downregulation of the Itga2 gene that encodes integrin subunit alpha 2 (Itga2), which is involved in focal adhesion 28 (Myh11 ΔK/ΔK /WT ratio (log2) = − 1.76). Regardless of their genotypes, the iPSCs upregulated Myh11 after culture with retinoic acid for three days, suggesting that K1256 deletion did not inhibit differentiation into SMCs. The expression of Myh11 was comparable between the WT and Myh11 ∆K/∆K cells, but the expression of Itga2 was lower in the Myh11 ∆K/∆K iPSCs than in the WT iPSCs. The newly differentiated cells showed a decrease in Itga2 expression, indicating that the downregulation of Itga2 detected in the Myh11 ΔK/ΔK aorta detected by RNA sequencing analysis is directly caused by the Myh11 K1256del pathogenic variant rather than a secondary response to defects caused by the pathogenic variant. This implies a possible role of Myh11 in the modulation of focal adhesion through the regulation of Itga2 expression. Previous reports have indicated that polymorphism in Itga2 is associated with ischemic stroke and coronary atherosclerosis 29 and that Itga2 is downregulated in the aortic aneurysm model www.nature.com/scientificreports/ induced by the inactivation of Smad4 22 . The present study is the first to report a downregulation of Itga2 in the aorta in an FTAAD model. Previous studies by our group as well as several other groups have shown that primary SMCs rapidly lose their original phenotypes when they are cultured in vitro 15,[30][31][32] . Thus, in vitro studies of primary SMCs are not likely to represent in vivo events. To obtain cultured SMCs that better resemble tissue SMCs, we established iPSCs and induced the differentiation of the iPSCs into SMCs. The iPSCs that we generated from Myh11 ΔK/ΔK MEFs with the forced expression of Yamanaka factors also lost pluripotency and the ability to self-renew and to remain undifferentiated within five passages. Interestingly, the forced stable expression of Nanog, along with the Yamanaka factors, improved the stability of Myh11 ΔK/ΔK iPSCs. Since Nanog is known to bind to one of the enhancer regions of human α-catenin 18 , the abilities of iPSCs to self-renew, remain undifferentiated and maintain pluripotency might have been rescued by the upregulation of α-catenin and the subsequent stabilisation of intercellular adhesion. Our RNA sequencing data of the aorta also showed a decrease in the expression of Ctnna2 (Myh11 ΔK/ΔK /WT ratio (log2) = − 10.8) as well as in a few other genes involved in intercellular adhesion. The STRING database showed that Ctnna2 encoding α-catenin interacted with the highest number of the intercellular adhesion-related molecules that were downregulated in the Myh11 ΔK/ΔK aortas. Previous work has shown that α-catenin connects cadherin and the actomyosin network and that the loss of α-catenin function disrupts intercellular adhesion 33 . Moreover, α-catenin is known to act as a mechanosensor 33 that enhances cell adhesion and promotes actin reorganisation at cell junctions 33 . In the future, a study using aortas either from Ctnna2 knockout or overexpression mice may reveal the involvement of α-catenin in the stabilisation of the submembranous cytoskeleton and its contribution to aortic contraction. Furthermore, increasing the expression of α-catenin could be an effective intervention for improving intercellular adhesion and may be a novel strategy for the treatment of FTAAD.
Finally, mechanical signals are relayed by three mechanisms. Mechanical signals from (1) the focal adhesion are relayed to (2) the intercellular adhesion junction via (3) the actomyosin network 34 . From the intercellular adhesion junction, the signals are passed on to the neighbouring cells 34 . In this study, we found genetic and phenotypic change pointing towards defects in all three mechanisms in our FTAAD model. We showed the possibility of attenuated focal adhesion. Itga2, which is involved in focal adhesion, is downregulated in both Myh11 ΔK/ΔK aortas and in cells differentiating into the SMC lineage. Then, we obtained ultrastructural anomalies by large-scale electron microscopy suggestive of defective intercellular adhesion. In addition, the expression of genes, such as Ctnna2 encoding α-catenin, that regulate intercellular adhesion was reduced. We predict, based on structural analysis ( Supplementary Fig. 10), that the coiled-coil structure of the Myh11 K1256del myosin II molecule is not optimal; therefore, this structure may attenuate the formation of the actomyosin network. Hence, we suspect that Myh11 K1256del causes a defect in all three mechanisms mentioned above and that they all contribute to the reduced contractility of Myh11 ΔK/ΔK aortas. We point out the importance of testing for the integrity of all of the three mechanisms and monitoring the contractility of the SMC network rather than focusing on the dysfunction of individual smooth muscles. When we develop a new therapeutic strategy to improve the contractility of SMCs, it may be essential to target all of the three mechanisms.
Although the attenuation of SMC contractility did not lead to luminal expansion in the Myh11 ∆K/∆K aorta, the uterus showed marked dilatation with a thinner wall. Furthermore, stillbirth was significantly more frequent with Myh11 ∆K/∆K mothers compared to WT mothers, supporting that Myh11 plays a role in uterine contraction. We also observed PDA in all Myh11 ∆K/∆K mice, whereas few Myh11 ∆K/+ mice exhibited PDA. The inadequate contraction of the ductus arteriosus in response to oxygen may induce PDA in Myh11 ∆K/∆K mice. Compared with previously reported animal models with PDA which died within 24 h after birth [35][36][37][38] , most Myh11 ∆K/∆K mice survived for more than 18 months with PDA, suggesting that these mice are a unique animal model of long survival with PDA.
In summary, Myh11 K1256del induces pathogenic stress on the aortic wall through the induction of structural fragility and a disorder of mechanoadaptation in the aorta. These changes may be caused by defects in focal adhesion, intercellular adhesion and the actomyosin network. Further studies are necessary to understand the linkage between the abnormal LMM structure and the downregulation of genes that form focal adhesions and intercellular adhesions as this knowledge is critical for developing a preventive therapy for FTAAD.

Methods
A detailed description of the methodology is provided in the Supplementary data. Animals. C57BL/6 J mice with the wild type or pathogenic variant in Myh11 were kept under a 12-h light/ dark schedule (lights on from 07:00 h to 19:00 h). Before invasive procedures such as the implantation of an infusion pump, the mice were anaesthetised by a single intraperitoneal injection of a mixture of medetomidine chloride (0.3 mg/kg), midazolam (4 mg/kg) and butorphanol tartrate (5 mg/kg) 39 . The absence of pedal reflex was used as the indicator of deep anaesthesia. Before sampling tissues from the mice, they were euthanised by a single intraperitoneal injection of an overdose of sodium pentobarbital (100 mg/kg). All animal handling procedures in this study complied with the Jichi Medical University Guide for Laboratory Animals, the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (NIH Publication, eighth edition, 2011) and the ARRIVE guidelines 40 . The Institutional Animal Care and Concern Committee at Jichi Medical University approved all experimental protocols. With the exception of the histological analysis of the uterus, this study used male mice.