Here we report a twelve-fold overrepresentation of chromosome 16p13.1 duplications in patients with thoracic aortic disease (1.06% versus 0.09% in controls), indicating greater enrichment of this duplication in thoracic aortic disease than the three-fold overrepresentation previously identified in schizophrenic patients and the five-fold increase in attention-deficit hyperactivity disorder 
. Note that the frequency of the 16p13.1 duplication in European controls or controls of European descent of 0.09% was identical between our study and both the schizophrenia and attention-deficit hyperactivity disorder studies 
. Although the 16p13.1 duplications vary in size, the duplicated regions associated with these three disorders overlap. Therefore our data support a stronger predisposition for thoracic aortic disease than for neuropsychiatric disease with 16p13.1 duplications. In contrast with other contiguous gene defects, such as the 1q21.1 deletion syndrome, we found that 16p13.1 duplications are associated with an adult-onset cardiovascular disorder in the absence of significant neuropsychiatric abnormalities 
. These data indicate that CNVs may be associated with non-overlapping phenotypes that affect more than one organ system, an observation critical to our understanding of the role of recurrent CNVs in human disease and a finding that may be common to other recurrent CNVs involving multiple genes.
Significant associations were recently reported between neuropsychiatric disorders and rare but recurrent deletions and duplications involving the short arm of chromosome 16. These studies used cohorts of patients with and without a family history of the disease, similar to the study reported here 
. Although our data on segregation of the 16p13.1 duplications were limited in this study, both inherited and de novo
duplications were identified in FTAAD patients. As in prior studies of duplications or deletions involving this region, we found that the association between 16p13.1 duplications and STAAD is moderated by decreased penetrance, as illustrated by the identification of these CNVs in controls and unaffected family members. Among 16p13.1 duplication carriers, the risk for thoracic aortic disease is relatively higher than the risk for schizophrenia, but less than the risk associated with single gene mutations that cause familial TAAD. We hypothesize that the duplications may act to modify the age of onset and dissection risk in families with single gene mutations. At the same time, the effect of 16p13.1 CNVs is substantially greater than the typical effect size of common variants identified in genome-wide association studies. Although we cannot exclude the possibility that the 16p13.1 leads to another, unidentified trait that increases the risk for TAAD, our findings suggest that rare CNVs with moderate effects are an important part of the allelic spectrum that contributes to the risk for TAAD.
This is the first report of an association between a recurrent CNV and an adult-onset vascular disease, although the assessment of CNVs in vascular diseases has been limited 
. In addition to the increased frequency of 16p13.1 duplications in TAAD in this study, we also identified increased numbers of rare CNVs in TAAD patients compared to controls, as well as a greater prevalence of additional rare CNVs in TAAD patients with 16p13.1 duplications 
. In addition, ontology analysis of the rare CNVs identified in sporadic and familial TAAD patients were more significantly more likely to include genes encoding proteins involved in contraction and adhesion of cells when compared to rare CNVs found in controls 
Nine genes are commonly duplicated amongst all TAAD patients harboring 16p13.1 duplications, and of these genes, MYH11
is the most likely candidate for the predisposition to thoracic aortic disease. MYH11
encodes the smooth muscle cell (SMC)-specific β-myosin heavy chain isoform. The monomeric unit of myosin is a multimeric complex consisting of two heavy chains associate with two pairs of light chains. These units then assemble into thick filaments that slide along adjacent α-actin-containing thin filaments to contract SMCs using the force generated by the myosin heavy chain. Prior studies on MYH11
mutations that cause familial TAAD suggested that the mutant myosin molecules have a dominant negative effect on filament formation, supporting the hypothesis that MYH11
mutations will disrupt SMC contractile function 
. We demonstrated that the 16p13.1 duplications are associated with increased MYH11
mRNA levels in aortic tissue. Data from a transgenic mouse model overexpressing an isoform of Myh11
, SM1, similarly showed increased SM1 mRNA levels, but no increase in SM1 protein levels 
. Studies in C. elegans
have shown that a precise ratio of β-myosin to its cellular chaperone, UNC45, is required for proper folding of myosin and assembly into thick filaments, and an imbalance in this ratio causes the degradation of myosin heavy chain protein and dysfunction of the contractile complex 
. Therefore, we hypothesize that overexpression of MYH11
does not lead to increased β-myosin protein levels, possibly due to imbalance of β-myosin to its chaperone, leading to degradation of β-myosin and dysfunction of the SMC contractile unit.
One important limitation of our study is potential bias due to the differential sensitivity of SNP array platforms and quantitative PCR to detect 16p13.1 duplications, which could potentially lead to a spurious association between this CNV and TAAD. We found that the prevalence of 16p13.1 duplications was similar in discovery and replication cases, which were screened using these two different methods. In addition, the frequency of duplications in controls matched the population frequency reported in multiple previous publications. Therefore, our conclusions are unlikely to be altered significantly by the bias related to using different CNV discovery methods.
In summary, although the presence of the 16p13.1 duplication confers a risk for thoracic aortic disease, the decreased penetrance of TAAD associated with the duplication suggests that other risk factors are required for expression of the clinical phenotype 
. The risk factors can be a genetic variant, such as another CNV, or possibly the presence of a BAV or a single gene mutation. Alternatively, other known risk factors for TAAD, such as poorly controlled hypertension, could contribute to aortic dissections in 16p13.1 duplication carriers by increasing the hemodynamic forces on the ascending aorta. With these risk factors, the presence of the 16p13.1 duplication predicts development of an acute aortic dissection at an aortic diameter less that 5.0 cm rather than a stable aneurysm when compared with thoracic aortic disease patients without the duplication. At the same time, the lack of schizophrenia and ADHD in these patients also implies that a second, and most likely different event, is required for development of neuropsychiatric disease. Further studies in larger cohorts with complete phenotypic data are needed to further define the additional genetic and environmental risk factors leading to aortic dissection or schizophrenia in patients with 16p13.1 duplications.