Aortic gene dictionary in the precision medicine era—update from the Aortic Institute at Yale New Haven

Abstract

This paper reviews the current understanding of the inherited, genetic nature of thoracic aortic aneurysm and dissection (TAAD), as well as the practice of genetic testing for thoracic aortic disease at the Aortic Institute at Yale-New Haven Hospital.

Introduction

Thoracic aortic aneurysm (TAA) is often described as a “silent killer” [1], because of its virulent nature, as it is completely asymptomatic in the majority of patients [2]. Following an indolent pathologic process, TAAs typically grow over the course of many years. The disease frequently presents either with death or catastrophes such as dissection or rupture, which can rapidly lead to death unless treated urgently with a surgical operation [3].

In the United States, aortic aneurysms are responsible for approximately 10,000 deaths every year [4], making aortic aneurysm the 20th leading cause of death overall, and the 17th leading cause of death in individuals over 65 years [4]. However, even these statistics likely represent an underestimation of the true incidence of aortic death, due to the sudden occurrence of many aneurysm-related deaths. Without a postmortem examination, these deaths are likely to be falsely attributed to a myocardial infarction [5]. Large-scale autopsy studies of sudden death patients also support this statement, identifying aortic rupture and dissection as the true cause for an astonishing 8% of all out of hospital sudden cardiac deaths [6–8]. This indicates that the incidence of aortic disease is, in reality, much greater than previously suspected.

Despite its virulence, TAA can be treated safely by means of preemptive surgical intervention. When detected in asymptomatic patients, TAAs can be monitored through non-invasive imaging. The threshold for elective surgical repair has been well-studied. Surgery is recommended by international guidelines at an absolute diameter of ≥ 5.5 cm for ascending and ≥ 5.5 to 6.0 cm for descending TAAs [9, 10]. In our experience, mortality in all categories of elective ascending aortic and arch operations has been reduced to between 0 and 1.9% in the present day [11–15].

In the 18 years since Craig Venter first mapped the human genome (his own), technology has advanced remarkably, so that what cost one billion dollars to map the first human genome has now become affordable and nearly commonplace. Whole genome sequencing (WGS) maps each of the approximately 3 billion “letters” of the human genetic code. Whole exome sequencing (WES), on the other hand, maps only the 2.5% of the human genome known to be involved in directing manufacture of proteins. We routinely perform WES on all patients presenting to the Aortic Institute at Yale New Haven Hospital, when covered by health insurance.

In 2012, we published in Connecticut Medicine the Yale milestones in reading the playbook of thoracic aortic aneurysm, including its recognition as a familial, inherited disease [16]. To date, more than 30 genes are directly associated with thoracic aortic aneurysm and dissection, providing the basis for genetic screening in affected patients and their families. Here, we will focus on the genetic nature of thoracic aortic disease and emphasize the importance of genetic testing when caring for such patients. We hope that the following information, supplemented by key graphs and figures, will be of interest and benefit to clinicians.

Familial patterns of TAAD

The heritable nature of aortic aneurysm disease is currently well established. Although the role of genetics and family history in abdominal aortic aneurysm was known since the 1980s due to landmark studies by Tilson and colleagues [17, 18], it was not until the late-1990s when the familial nature of nonsyndromic thoracic aortic disease was described [19, 20]. For many years, Marfan syndrome was the only recognized genetic defect linked to TAA. Marfan syndrome, which was first described in 1896, is a connective tissue disorder characterized by long, thin extremities; skeletal abnormalities; ligamentous laxity; ectopia lentis; ascending aortic dilation; and aortic or mitral valve insufficiency [21]. Nowadays, other connective tissue disorders apart from Marfan syndrome are recognized to cause syndromic thoracic aortic aneurysm and dissection (TAAD), such as Ehlers-Danlos syndrome, Loeys-Dietz syndrome, arterial tortuosity syndrome, aneurysm-osteoarthritis syndrome, and cutis laxa syndrome [22]. Syndromic TAAD, which is associated with extra-aortic manifestations, such as in the aforementioned syndromes, accounts for less than 5% of all cases of TAAD, leaving many other aneurysm families unexplained [23, 24].

At Yale, the quest for a possible heritable pattern of thoracic aortic aneurysm and dissection started after identifying a family in which type A aortic dissection had occurred in three consecutive generations, including a 12-year-old girl, who was the youngest person to die from an aortic dissection at Yale-New Haven Hospital [23]. At the Aortic Institute, we immediately sought to study in detail many families of patients with thoracic aortic disease in order to determine whether other family members (apart from our patient) were affected. We were able to show that in 21% of patients TAAD was familial, meaning that at least one other family member was known to be affected by aortic disease, and that the most frequent mode of inheritance was autosomal dominant [19] (Fig. 1). It is important to note that for a family to be classified as “positive” for a familial pattern, aneurysm in the family members had to have been demonstrated by imaging or surgery. The true rate is likely underestimated in this way, as no routine clinical screening of family members was executed as part of the study. We also found that, compared to the non-familial cases of TAAD, patients with familial TAAD presented with aortic aneurysm or dissection at an earlier age (58.2 vs. 65.7 years) and manifested a higher aortic growth rate (0.21 vs. 0.16 cm/year), highlighting the virulence of familial thoracic aortic disease [19].

Fig. 1.

Positive nonsyndromic thoracic aortic aneurysm family pedigrees among 100 family trees constructed. Squares represent the men and circles, the women. An arrow indicates the proband with a thoracic aortic aneurysm. Blackened squares or circles represent affected patients with aortic aneurysms. ([19] Reproduced with permission from Coady MA, Davies RR, Roberts M, et al. Familial patterns of thoracic aortic aneurysms. Arch Surg. 1999;134:361–7)

Importantly, a very similar study conducted independently by Dr. Diana Milewicz had reported the exact same finding—21% of aneurysms were familial [20]. Mirroring these studies on Mendelian genetics of TAA conducted at Yale [19, 25], as well as the ground-breaking Milewicz work [20], the current Guidelines for the Diagnosis and Management of Patients with Thoracic Aortic Disease [26] recommend the following: (1) aortic imaging for first-degree relatives of patients with TAA and/or dissection to identify those with asymptomatic disease (Class 1, Level of Evidence B recommendation) and (2) aortic imaging for second-degree relatives if one or more first-degree relatives are found to have thoracic aortic dilatation, aneurysm, or dissection (Class 2, Level of Evidence B recommendation) (Fig. 2).

Fig. 2.

Proposed flow chart for a dedicated screening program for relatives of patients affected by nonsyndromic diseases of the thoracic aorta. BAV, bicuspid aortic valve; CT, computed tomography; FDRs, first-degree relatives; MRI, magnetic resonance imaging; NS-TAD, nonsyndromic thoracic aortic disease; SDRs, second-degree relatives; TTE, transthoracic echocardiogram. ([43] Reproduced with permission from Mariscalco G, Debiec R, Elefteriades JA, Samani NJ, Murphy GJ. Systematic review of studies that have evaluated screening tests in relatives of patients affected by nonsyndromic thoracic aortic disease. J Am Heart Assoc. 2018;7:e009302)

Guilt by association

The “Guilt by association” concept has been proposed by our group to describe a number of seemingly unrelated conditions that are associated with TAAD, or are, in other words, its “guilty associates” [27]. These conditions include intracranial aneurysms, aortic arch abnormalities, abdominal aortic aneurysms, bicuspid aortic valve, renal cysts, a positive thumb-palm test, family history of aortic disease, and giant cell arteritis (Fig. 3). The above have been shown to have a higher prevalence among individuals with TAAD compared with the general population, and presence of one or more of these clinical markers should trigger suspicion for TAAD. Many of these conditions are thought to share similar genetics and pathophysiology to thoracic aortic aneurysm [27].

Fig. 3.

Paradigm of “Guilt by association” for detection of silent thoracic aortic aneurysms. ([27] Reproduced with permission from Elefteriades JA, Sang A, Kuzmik G, Hornick M. Guilt by association: paradigm for detecting a silent killer (thoracic aortic aneurysm). Open Heart. 2015;2:e000169)

Genes associated with TAAD

It is widely recognized by the medical community that, while descending thoracic aortic disease is degenerative in nature and associated with atherosclerosis, hyperlipidemia, and hypertension, genetic predisposition is the most crucial risk factor for ascending thoracic aortic aneurysm [28, 29]. To date, more than 30 genes have been linked to TAAD, and new candidate genes are being discovered regularly. More than 30% of familial nonsyndromic TAAD cases have been documented to carry a genetic variant in one or more of these genes [30]. Although most of these genes are associated with syndromic TAAD, pathogenic variants in 14 genes among them cause familial nonsyndromic TAAD [31].

Most of these 30-plus genes encode regulatory molecules for the extracellular matrix (FBN1, FBN2, COL1A1, COL1A2, COL3A1, EFEMP2, ELN, EMILIN1, MFAP5, LOX, BGN), smooth muscle cell contractile units (ACTA2, MYH11, MYLK, PRKG1, FLNA, MAT2A, FOXE3), or the TGF-β signaling pathway (TGFβ2, TGFBR1, TGFBR2, SMAD3, SKI, SLC2A10) [23, 24, 29]. Since the tunica media of the aorta is comprised mainly of smooth muscle cells, mutations in genes that encode the contractile units of these cells can disrupt the medial layer, and therefore, the smooth muscle cells’ ability to effectively remodel and regulate the surrounding extracellular matrix [32].

The majority of the TAAD-related genes predispose to aneurysm formation; however, some are exclusively associated with aortic dissection, such as the MYLK gene [33], or cause aortic dissections at a small aortic diameter, such as the ACTA2 gene [34, 35]. Mutations in the ACTA2 gene are the most common genetic alteration leading to nonsyndromic TAAD, accounting for 12–21% of cases [31]. Identification of patients harboring such mutations is essential, as early prophylactic aortic surgery is necessary to protect them from a catastrophic aortic dissection, which frequently occurs below the dimension of 5 cm in these cases. Another gene linked to dissection, KIF6, has been proven by previous studies to increase the risk for cardiovascular disease and myocardial infarction [36]. The KIF6 gene encodes a protein in the kinesin superfamily, which is involved in intracellular transport. Our group has shown that, even after controlling for the potentially confounding presence of coronary artery disease, patients who carried the KIF6 719Arg variant presented with a twofold higher risk of aortic dissection compared with noncarriers [36].

Syndromic and nonsyndromic TAAD are considered two distinct entities; however, studies have proven that there is an overlap between genes causing syndromic and nonsyndromic cases. Mutations in the FBN1 gene, for example, characteristically cause Marfan syndrome, but some variants have been implicated in sporadic nonsyndromic TAAD [37], as well as dissection at small aortic sizes [38], requiring early surgical intervention.

The evidence that mutations in different genes have distinct clinical implications, and that they determine the natural history of TAAD, makes genetic screening a crucial tool for formulating a personalized treatment strategy for individual patients, including, but not limited to, the appropriate size threshold for prophylactic aortic repair [30] (Fig. 4).

Fig. 4.

Ascending aorta dimensions for prophylactic surgical intervention. Abbreviations: ECM, extracellular matrix; SMC, smooth muscle cell; TAAD, thoracic aortic aneurysm and/or dissection; TGF, transforming growth factor. ([38] Reproduced with permission from Brownstein AJ, Kostiuk V, Ziganshin BA, et al. Genes associated with thoracic aortic aneurysm and dissection: 2018 update and clinical implications. Aorta (Stamford). 2018;6:13–20)

Routine whole exome sequencing for TAAD

Since 2012, we have been routinely performing genetic screening by means of WES on most thoracic aortic aneurysm or dissection patients presenting to the Yale Aortic Institute [29]. WES is a technique that identifies the deoxyribonucleic acid (DNA) sequence of all protein-coding regions (exons) in the genome, including, but not limited to, all TAAD-associated genes. The entire exome is scanned for genetic variants that might cause TAAD, but a panel of genes recognized by previous studies to underlie the phenotype are subject to more focused analysis. The most recent panel of 44 genes analyzed with WES at our laboratory includes the following genes: ACTA2, ADAMTS10, BGN, CBS, COL1A1, COL1A2, COL3A1, COL4A5, COL5A1, COL5A2, EFEMP2, ELN, EMILIN1, FBN1, FBN2, FLNA, FOXE3, GATA4, GATA5, GATA6, HCN4, LOX, MAT2A, MED12, MFAP5, MIB1, MYH11, MYLK, NOTCH1, PKD1, PKD2, PRKG1, ROBO4, SKI, SLC2A10, SMAD2, SMAD3, SMAD4, SOX18, TGFB2, TGFB3, TGFBR1, TGFBR2, TGFBR3.

A major advantage of WES is that it is comprehensive, allowing sequencing of all TAAD-associated genes in a single test in one laboratory, instead of (as per prior methods) individually testing for each TAAD gene and syndrome separately. In this way, WES is usually less expensive and more efficient than syndrome-specific tests in different laboratories. WES also permits sample re-analysis retrospectively as new genes are discovered [29]. For example, when we first started performing WES, there were only 12 TAAD-related genes analyzed with the test. Currently, with 44 genes analyzed, efforts are being undertaken to reanalyze all previous patient samples which had been reported as negative at the time. Furthermore, since WES encompasses genome-wide data, genetic data from patients who tested negative for mutations in TAAD genes, especially ones with family members affected with TAAD, can be used to “mine” information for the discovery of completely novel genes [29].

Since the initiation of our WES program, more than 500 patients have been screened, with a single saliva or blood sample. If mutations are found in one or more of the TAAD-associated genes in a patient’s sample, they are classified according to the criteria proposed by the American College of Medical Genetics and Genomics (ACMG) in one of the following five categories: pathogenic, likely pathogenic, benign, likely benign, or variant of uncertain significance (VUS) [39]. Pathogenic variants are definitively disease-causing, and benign variants do not cause TAAD. Likely pathogenic or likely benign variants have at least a 95% probability of being disease-causing or benign, respectively. If a variant does not fulfill criteria either for the likely pathogenic/pathogenic, or for the likely benign/benign categories, or if the evidence for benign and pathogenic is conflicting, it is classified as a VUS [39]. It is important to recognize that, in the present era, until multiple reports of positive family pedigrees have been published, most newly discovered pathogenic genes will be mis-classified initially as VUS. So, the categorization as “VUS” should not be viewed negatively, as exonerating the variant.

For patients with findings in TAAD-related genes, especially pathogenic or likely pathogenic variants, we recommend echocardiographic screening of family members to exclude the possibility of a thoracic aortic aneurysm. In certain cases, family members are also genetically tested with the simpler, less expensive technique of single-site sequencing (Sanger) to confirm or exclude the presence of the variant previously identified in the proband [29]. Sanger sequencing focuses on sequencing only the specific gene in which a mutation was found in the proband, instead of scanning the entire genome. Such Sanger sequencing can spare non-mutation carrying family members from decades of worry and from repeated imaging studies (and the imposed radiation burden).

Patients who undergo WES are thoroughly interviewed using the Aortic Institute Family History Questionnaire, which inquires in detail about history of aneurysm disease, dissection, or sudden death for all their blood relatives. The combination of the family history and clinical and genetic data facilitates the construction of a complete pedigree for the family, which is particularly useful in interpreting incoming genetic data and determining if the specific variant co-segregates with disease in the family; that is, in simple terms, if the individuals affected by TAAD in the family all harbor the variant, while the healthy individuals in the family do not [29].

In 2015, we published our initial findings performing WES in 102 patients [29]. Twenty-six genetic variants in the 21 TAAD-related genes analyzed were identified in 23 patients (22.5%); 3 patients were carriers of more than one mutation. Four of these variants were classified as deleterious, in the COL5A1, FBN1, MYLK, and FLNA genes, and the remaining variants were deemed to be of uncertain significance. Two additional patients harbored probable disease-causing mutations in genes not belonging to the panel of TAAD genes [29] (Fig. 5). More recently, our group published the largest (to our knowledge) genetic analysis of nonsyndromic TAAD (over 1000 patients), in collaboration with colleagues from the United Kingdom [37]. Polymerase chain reaction and next-generation sequencing of a gene panel were used on blood and saliva samples from United Kingdom (UK) patients, as well as surgical aortic specimens from Yale Aortic Institute patients. After classifying the variants identified into the tiers proposed by the ACMG guidelines, the study concluded that pathogenic or likely pathogenic variants were more likely to be found in genes implicated in syndromic TAAD (FBN1, TGFBR1, TGFBR2, COL1A1, COL1A2, COL3A1, COL5A1, COL5A2), and in patients with one or more of the following characteristics: a syndromic phenotypic element, ascending aortic aneurysm, positive family history of aortopathy, and early age at presentation [37].

Fig. 5.

Frequency distribution of genetic defects in TAAD-related genes based on the Yale Aortic Institute routine genetic testing program. ([29] Reproduced with permission from Ziganshin BA, Bailey AE, Coons C, et al. Routine genetic testing for thoracic aortic aneurysm and dissection in a clinical setting. Ann Thorac Surg. 2015;100:1604–11)

Variants of uncertain significance: a challenge in clinical practice

Apart from the well-characterized disease-causing mutations in TAAD-related genes, there is an abundance of genetic variants identified in our patients which are currently classified as variants of uncertain significance (VUS). In simple terms, that means that there is not currently enough clinical or structural information on these mutations to categorize them with certainty as disease-causing or benign. Individual variant characteristics, such as the frequency with which the variant is found across large population databases, can be indicative of their real significance [39]. A variant found in a population at a much higher frequency than the prevalence of thoracic aortic disease cannot cause TAAD with a high penetrance; conversely, a very rare variant found in an individual with TAAD is highly suspicious of being pathogenic. Other notable variant features include, among others, the degree of phylogenetic conservation across species of the specific amino acid residue affected by the mutation, as well as whether in silico (computational) algorithms consistently predict a deleterious effect of the particular mutation on the product of the gene [39].

Our most recent data indicates that approximately a quarter of our patients who underwent WES carry VUS in one or more TAAD-related genes [39]. While efforts are being undertaken to reclassify these variants to pathogenic or benign by combining the clinical and genetic information we obtain from our patients’ families with the already established structural variant characteristics, many of these variants remain characterized as VUS until further in vitro or in vivo studies are conducted, or until more families harboring the same variant are reported in the literature. These variants represent a challenge in the clinical setting, as their presence should not influence clinical decision-making, according to the ACMG guidelines [39]. However, we feel that many of the rare VUS that we find in TAAD patients, especially the ones with a positive family history, will prove to be pathogenic as more evidence accumulates—after all, in the present era of clinical genetics, many pathogenic variants first emerge as VUS. Also, the viewpoint of the geneticist (who fears mis-classifying a gene as disease-causing) differs from that of the physician or surgeon (who wants to keep the patient alive and not be swayed by a technicality of classification).

The ribonucleic acid (RNA) signature

Apart from compiling our patients’ genetic profiles (DNA), we have been investigating specific RNA expression patterns of patients with thoracic aortic disease. If we consider DNA as the blueprint for the structure of the human body, then the RNA can be considered the “worker” molecules which execute the blueprint [23]. Upregulation or downregulation of various RNAs has been documented in different genetic diseases [23]. We sought to study whether up- or downregulation of certain RNA molecules could be useful as a diagnostic tool for TAAD. The concept of this study originated from the necessity for clinical biomarkers [40, 41] that could potentially detect silent thoracic aortic aneurysm before the occurrence of rupture or dissection. We analyzed 33,000 RNAs in patients with TAAD and their spousal controls, and identified 41 RNA molecules (highly up- or downregulated, relative to controls) that were quite accurate in predicting presence of thoracic aortic disease. This test was able to predict TAA with an 80% overall accuracy [42] (Fig. 6). We call this our “RNA signature” test. Interestingly, the RNA signature was also able to differentiate between ascending and descending aneurysm, and also between familial and non-familial aneurysm disease [42]. The validity of the RNA signature test was recently verified in a separate independent patient population, as well as in patients with acute aortic dissection and patients with Marfan syndrome (unpublished data). As additional replicatory studies from our group are nearing completion, we feel that the RNA signature test has the potential to become a simple and accurate blood test that could serve as a screening mechanism for detection of silent thoracic aortic aneurysm [23].

Fig. 6.

The RNA signature test for thoracic aortic aneurysm. In the hierarchical cluster diagram on the left (A), each vertical line represents a patient, and each horizontal line represents an RNA. In the grid (A), the green indicates underexpression and red indications overexpression. Note in the diagram on the left (A) how the overexpression and underexpression cluster, depending on phenotype. In the figure on the right (B), note that if all the greens were together and all the reds were together, the test would have been 100% accurate. As it turns out the overall accuracy was > 82%. ([5] Reproduced with permission from Elefteriades JA, Farkas EA. Thoracic aortic aneurysm clinically pertinent controversies and uncertainties. J Am Coll Cardiol. 2010;55:841–57)

Conclusion

The current understanding of the genetic nature of TAAD stems from the immense progress made in this field over the past two decades, including the discovery of the familial clustering of TAAD cases, as well as the development of an expanding genetic dictionary, encompassing all the genes associated with this virulent disease to date. Routine genetic screening for TAAD by means of WES opens up the possibility of personalized aortic care, not only by assisting in the prevention of catastrophic aortic events, such as dissection or rupture, but also by unmasking silent disease in undiagnosed family members. The future in this field lies in accruing more evidence to elucidate the implications of mutations now classified as variants of uncertain significance, as well as developing a simple blood test to diagnose TAA. Further studies will hopefully enhance our knowledge on TAAD genetics by discovering new genes and clarifying the natural history and prognosis of the disease associated with specific mutations in each TAAD-related gene, permitting even earlier diagnosis of asymptomatic TAAD cases and halting the progression of this silent killer.

Funding

None.

Declarations

Research involving human participants and/or animals

Not applicable for this review article.

Informed consent

Not applicable for this review article.

Conflict of interest

The authors declare no competing interests. Dr. Elefteriades is a Principal of CoolSpine, serves on the Data and Safety Monitoring Board of Terumo, and consults for CryoLife.