Article Text

Download PDFPDF

The fibrillin-1 gene: unlocking new therapeutic pathways in cardiovascular disease
  1. Paddy M Barrett1,
  2. Eric J Topol1,2
  1. 1Scripps Translational Science Institute, La Jolla, California, USA
  2. 2Scripps Health, La Jolla, California, USA
  1. Correspondence to Eric J Topol, Scripps Translational Science Institute, 3344 N Torrey Pines Court Suite 300, La Jolla, CA 92037, USA; etopol{at}scripps.edu

Abstract

The dramatic reductions in DNA sequencing costs allow us to delve deeper into the genomic alterations that increase susceptibility to many polygenic cardiovascular diseases. One such condition is an abnormal proximal aorta. Until recently, many believed that dilated, distorted or dissected proximal aortas might represent a forme fruste of Marfan syndrome or a continuum of aortopathy. Although an FBN-1 mutation does not guarantee the diagnosis of Marfan syndrome it is clear however that FBN-1 mutations independently confer additional risk for many of the cardiovascular complications classically associated with the disease. Furthermore, treatment with an angiotensin receptor blocker has proven effective in reducing rates of thoracic aortic root dilatation in preliminary studies of Marfan syndrome patients. Awareness of an FBN-1 mutation then highlights the need for increased vigilance for the associated cardiovascular phenotypes. Knowledge of an FBN-1 gene mutation may allow actionable interventions earlier in the natural history of the condition.

  • Coronary artery disease
  • acute myocardial infarction
  • restenosis
  • myocardial infarction
  • interventional cardiology
View Full Text

Statistics from Altmetric.com

Introduction

FBN-1 gene mutations result in a constellation of phenotypes. Marfan syndrome is, however, the most common manifestation with significant associated cardiovascular pathologies.1 Isolated ascending aortic aneurysm,2 mitral valve prolapse3 ,4 and bicuspid aortic valve (BAV)5 ,6 have all been demonstrated in those with FBN-1 mutations. Beyond the classical cardiovascular phenotypes, FBN-1 mutations have been identified in a range of conditions including the mitral, aortic, skin and skeletal phenotype,7 familial ectopia lentis,1 ,8 isolated skeletal features9 ,10 and Weill–Marchesani syndrome.11 This paper explores the cardiovascular implications of an FBN-1 gene mutation and the possibility of a continuum of phenotype between those with thoracic aortic aneurysm and FBN-1 mutation and Marfan syndrome. Furthermore, with data suggesting a benefit of angiotensin receptor blocker (ARB) use in those with Marfan syndrome, the possibility exists as to whether this approach could be extrapolated for use in those with thoracic aortic aneurysm and an FBN-1 mutation.

The FBN-1 gene

The FBN-1 gene is approximately 200 kb12 in size and composed of a relatively large coding sequence divided into 65 exons, located on chromosome 15.13 ,14 It encodes for the protein fibrillin-1, which is a 350-KDa, 2871-amino acid cysteine-rich glycoprotein. Fibrillin-1 consists mainly of epidermal growth factor domains and a small number of transforming growth factor ß1-binding protein (TGFB1)-like domains15 (figure 1). Fibrillin-1 acts as a major structural component of extracellular matrix microfibrils, located primarily in the periphery of elastic fibres.17 Originally, disruption of the elastic network of the media by means of disruption of microfibrillary assembly was seen as the primary reason for vascular pathology. However, these pathologies are most likely to be as a result of the adventitia's microfibril inability to sustain physiological stress.18 Strategies to strengthen the vessel wall might then reduce the rate of progression of any associated vascular diseases.

Figure 1

Fibrillin-1 cbEGF12-13 pair of Ca2+ binding epidermal growth factor-like domains. Molecular graphics and analyses were performed with the UCSF Chimera package. Chimera is developed by the Resource for Biocomputing, Visualisation, and Informatics at the University of California, San Francisco, funded by grants from the National Institutes of Health National Center for Research Resources (2P41RR001081) and the National Institute of General Medical Sciences (9P41GM103311).16

Genomics of thoracic aortic disease

Most of aneurysmal thoracic aortic disease is sporadic. The genetic disorders associated with thoracic aortic disease such as Marfan syndrome, Ehlers–Danlos syndrome, Turner syndrome, Noonan syndrome, Alagille syndrome, osteogenesis imperfecta and adult polycystic kidney disease account for little of the overall disease prevalence.17 ,19–24 However, family studies have shown that up to 19% of those with a thoracic aortic dissection have an affected first-degree family member suggesting a familial clustering effect.17 ,25 ,26 Of particular interest is that this population tends to present at a much younger age than those without an affected relative.

To date, beyond the FBN-1 mutations seen in the Marfan syndrome population, much of the thoracic aortic aneurysm and dissection disease burden is accounted for by mutations at six chromosomal loci: TAAD1 at 5q13-14, FAA1 at 11q23-24, TAAD2 at 3p24- 25, TAAD3 at 15q24-26, TAAD4 at 10q23-24 and MYH11 at 16p12-13.17 ,27–30 Many of the genes identified at these loci are involved in vascular function. MYH11 encodes for smooth muscle myosin heavy chain and ACTA2 at the TAAD4 locus encodes for smooth muscle α-actin. Several of the gene mutations implicated in familial thoracic aortic aneurysm and dissection could interrupt mechanotransduction of smooth muscle cells. Many of the genes mutations outlined affect the proteins involved in smooth muscle cell biology by means of loss of decreased contractility. TGF-ß receptor 1 and 2 (TGFBR1 and TGFBR2) mutations, which are also implicated in thoracic aortic disease, however, affect TGF-ß signalling and smooth muscle cell differentiation.31–33

Thoracic aorta dilatation and dissection

Thoracic aortic dilatation is predominantly an asymptomatic process, the diagnosis often being achieved by radiological screening or discovery as an incidental finding.34 Thoracic aortic dissection however can present acutely with devastating haemodynamic consequence. Dissections can be classified as either Stanford type A or Stanford type B depending on whether the dissection involves the ascending aorta or not35–38 (figure 2).

Figure 2

Stanford and Debakey thoracic aorta dissection classification. Anatomical location and classification of aortic aneurysms and dissections. (A) Anatomical location of aortic aneurysms. Ascending thoracic aortic aneurysms are highlighted in yellow. (B) Classification of aortic dissections initiating in the thoracic aorta.17

With the substantial genetic heterogeneity there is substantial clinical heterogeneity.17 ,28 ,39 ,40 To examine this further, LeMaire et al6 performed a multistage genome-wide association study on a spectrum of sporadic thoracic aortic aneurysm and dissection (STAAD) phenotypes to identify the single nucleotide polymorphisms (SNPs) associated with thoracic aortic aneurysm and dissection. A three stage genome-wide association study analysis was performed on 327 628 SNPs in 765 patients with sporadic thoracic aortic aneurysm and dissection (STAAD) and compared with 1355 controls from the Wellcome Trust Case-Control Consortium, 1958 birth cohort (C58) and 874 controls from the US National Institute of Neurological Disorders and Stroke Repository's Neurologically Normal Control Collection.

Importantly, none of the 765 STAAD cases had either a family history of STAAD or a syndromic form of TAAD such as Marfan syndrome.6 Five SNPs were identified with genome-wide significance (p<5×10−8) with combined ORs ranging from 1.6 (p=1.3×10−8) to 1.8 (p=4.6×10−8). All five SNPs fall into an area of linkage disequilibrium approximately 305 kb in size, encompassing the entire FBN-1 gene, which encodes fibrillin-1 at chromosomal locus 15q21.1 (figure 3; table 1).6 rs2118181 Was the most highly associated stage 1 genotyped SNP with STAAD (ORmeta=1.8, Pmeta=5.9×10−12). This is the only gene in this area of linkage disequilibrium.

Figure 3

A Manhattan plot of stage 1 genome-wide association results from a comparison of a spectrum of sporadic thoracic aortic aneurysm and dissection cases to National Institute of Neurological Disorders and Stroke (NINDS) controls. For each tested marker, the significance is displayed on the y-axis as the −log10 of the p value. The −log10 results are ordered along the x-axis by chromosome, with each coloured bar representing a different chromosome.6

Table 1

Genotyped SNPs associated with STAAD (p<5×10−8) with variants within and flanking the FBN1 gene

Irrespective of the type of dissection, a variety of the SNPs at locus 15q21.1 were associated with thoracic aortic dissection. LeMaire et al also examined dissection types separately; rs10519177 (ORAD, A, meta=1.8, Pad, A, Meta=1.2×10−8) was the most significantly associated SNP with type A dissections.6 This study importantly demonstrates the association of several polymorphic variants at 15q21.1, encompassing the FBN-1 gene with STAAD in the absence of a syndromic aetiology or family history of STAAD. These findings add to the growing literature on patients with an FBN-1 mutation and thoracic aortic aneurysm formation but without a diagnosis of Marfan syndrome.7 In what often presents in an asymptomatic fashion, knowledge of an FBN-1 mutation may be critical to detection and accurately determining ones risk of thoracic aortic aneurysm and dissection.

Bicuspid aortic valve

As there is a strong association between concurrent BAV disease and thoracic aortic aneurysm and dissection, LeMaire et al further validated their findings by examining the five stage 1 SNPs in those with thoracic aortic dissection with or without BAV and those with sporadic non-dissection aortic aneurysms, again, with and without BAV. All five stage 1 SNPs maintained a significant OR at genome-wide significance 1.5 (p=3.8×10−11) to 1.8 (p=1.8×10−12) thereby demonstrating their relationship to BAV with either thoracic aortic dissection or sporadic non-dissection aortic aneurysms6 ,41–43 (table 2).

Table 2

Aneurysm, dissection and valve characteristics of stages 1, 2 and 3 cohorts6

Although BAV disease is the most common congenital heart defect occurring in 1%–2% of the general population and is commonly associated with Marfan syndrome it has an unclear aetiology.44 Echocardiography studies have demonstrated familial clustering and therefore the possibility of heritability.45 A suspected autosomal dominant mode of inheritance is suspected with almost 10% of first-degree family members of those with BAV disease having the same valvular defect.46 Release of matrix metalloprotinase-2 from microfibrils due to deficient expression of fibrillin-1 is a likely candidate mechanism for the valve pathology.5 This defect in fibrillin-1 production then provides a possible mechanistic linkage explaining why those with BAV disease are at significantly increased risk of developing concurrent aortic dysfunction.46

Marfan syndrome

Marfan syndrome is an autosomal dominant, systemic disorder of connective tissue with a population prevalence of approximately 1 per 5000.47 ,48 Classic features include aortic root dilatation, ectopia lentis, increased arm span and generalised joint hypermobility49 (figure 4). It is, however, the cardiac features that are responsible for the substantial morbidity and mortality associated with the disease.1 ,52 By age 60 years, 96% of Marfan syndrome patients develop ascending aortic dilatation and 74% will either suffer a dissection or undergo prophylactic surgery.1 These complications are not limited to the older subgroups with 16% of those under 30 years either undergoing prophylactic surgery or presenting with a dissection.1 Besides the connective tissue manifestations of ectopia lentis, aortic root dilatation and MVP, Marfan syndrome is also characterised by skeletal manifestations such as bone overgrowth.

Figure 4

Marfan syndrome: (A) dilated thoracic aortic root and (B) ectopia lentis. The most common cardiovascular manifestation of Marfan syndrome is thoracic aortic root dilatation. A second major feature is ectopia lentis, which is defined as displacement or malposition of the crystalline lens in the eye.50 ,51

The FBN-1 gene mutations that cause Marfan syndrome are usually private mutations and are distributed throughout the gene. The most prevalent are missense mutations substituting or creating a cysteine in one of the calcium binding EGF domains.53 When mutations are clustered between exons 24 and 32 they are associated with neonatal Marfan syndrome or severe rapidly progressive MFS.54

Small case series have described Marfan phenotypes even in the setting of an entire FBN-1 deletion.53 ,55–58 Multiple studies have demonstrated a variety of mutations across the entire gene, often with varying degrees of phenotypic expression.59 ,60 Many of these mutations are novel and often unique to the affected pedigree.4 ,61 In all, 40% of mutations of the terminal seven exons (exons 59–65) are associated with a mild phenotype without aortic root pathology. However, only 7% of mutations in the remaining exons are characterised by this milder phenotype. Indeed, whether the mutation occurs nearer the N-terminus or C-terminus portion of the gene appears to have varying phenotypic consequence.62 ,63 When the two primary mutations of cysteine-involving mutations and premature termination mutations are compared, significant differences in phenotype are observed. Knowledge then of an FBN-1 mutations presence and its specific location can assist in risk stratification for the development of an associated cardiovascular disease such as thoracic aortic aneurysm.

Marfan syndrome diagnosis

With multiple gene mutations resulting in a range of phenotypes and the diagnosis often relying on manifestations of these phenotypes, the condition is occasionally missed. The diagnosis relies on applying the 2010 revised Ghent criteria.49 Manifestation of two of the cardinal features, (1) aortic root aneurysm or (2) ectopia lentis, even in the absence of family history, is sufficient for the unequivocal diagnosis of Marfan syndrome. If either of these two features is absent, the presence of one of the 600 previously identified FBN-1v mutations along with a variety of systemic manifestations are required to achieve a diagnosis.54 When a definitive diagnosis of the Marfan syndrome is made, the 2010 American College of Cardiology/American Heart Association/American Association for Thoracic Surgery (ACC/AHA/AATS) guidelines recommend annual cardiac imaging, certainly to increase the likelihood of identifying aortic dilatation requiring intervention.64 The revised Ghent criteria, although achieving a diagnosis in over 95% of cases, occasionally fail to achieve a diagnosis even in patients with an FBN-1 mutation and features suggestive, but not diagnostic, of Marfan syndrome such as aortic aneurysm.7 ,49 Certain FBN-1 missense mutations may result in aortic disease in the absence of ocular or skeletal features.7 Failure to identify patients with such FBN-1 mutations but without evident aortic disease reduces the probability of undergoing appropriate imaging screening but may also preclude the use of treatments, which are showing promise in reducing aortic dilatation in Marfan syndrome patients. Importantly, failure to diagnose Marfan syndrome may also result in family members not undergoing appropriate clinical or genetic screening.

Medical therapy in Marfan syndrome

To date, the ACC/AHA/AATS recommended medical therapy for those with Marfan syndrome involve the use of β-adrenergic blockade. This serves to reduce haemodynamic wall stress and (to an extent) the rate of change of aortic root diameter.65–67 However, the results of such studies are mixed and although β blockade is frequently administered, the data do not support clear evidence of efficacy in reducing progression or the incidence of aortic dissection.47 ,67–70

Transforming growth factor β

The spectrum of phenotypes seen in Marfan syndrome, such as the skeletal disturbances, becomes biologically difficult to explain in the setting of a purely connective tissue disorder. The answer to this may lay in the fact that fibrillin-1 also acts as a regulator of TGF-ß.71–74

TGF-ßs are pluripotential cytokines involved in tissue morphogenesis and homeostasis. This inactive precursor, a large latent complex, binds to extracellular microfibrils of the extracellular matrix and specifically fibrillin-147 ,75 ,76 (figure 5). This latent complex requires regulated activation to release free TGF-ß for biological activity.78 ,79 In the setting of an FBN-1 mutation and abnormal fibrillin-1 production, failure of this latent complex binding could result in excessive TGF-ß activation.80 Further evidence for TGF-ß's role in Marfan syndrome is highlighted by the fact that up to 10% of cases are caused by mutations in TGFBR1 and TGFBR2 genes without an identifiable FBN-1 mutation.31–33

Figure 5

Latent transforming growth factor β (TGF-β)-binding proteins (LTBPs). LTBPs are a class of proteins that have many similarities with the fibrillins. Evidence so far suggests that rather than having a primarily structural function, their main role is in the regulation, targeting and release of TGF-β. TGF-β is an extremely important factor, both in normal development and also in a wide variety of pathological conditions. LTBPs contain many domains in common with fibrillins as well as tandem arrays of cbEGFs; they also contain TGF-β-binding protein-like (TB) domains, one of which is responsible for binding TGF-β's pro-peptide through a disulphide exchange mechanism, and in so doing retains TGF-β in a latent state. ECM, extracellular matrix.77

Angiotensin receptor blockade in Marfan syndrome

Knowledge of the role of TGF-ß in response to an FBN-1 mutation prompted researchers to investigate the possibility of ARBs as a therapy for those with Marfan syndrome.47 ,71 ARBs are classically used as an antihypertensive by selective inhibition of angiotension 2 type 1 (AT1) receptors within the renin-angiotensin-aldosterone system.81 ,82 Blockade of this pathway decreases TGF-ß signalling, reduces plasma levels of free TGF-ß, reduces tissue expression of TGF-ß responsive genes and reduces levels of the intracellular mediators, such as phosphorylated Smad2.81 ,83–85

The ARB Losartan is efficacious in reducing the rate of thoracic aortic aneurysm formation in murine models of Marfan syndrome.86 ,87 Furthermore, this intervention has also proven effective in a small study in the paediatric population with Marfan syndrome.81 The study demonstrates a significant decrease in the rate of aortic root enlargement in populations who have severe aortic disease at baseline and had failed conventional medical therapy. The rate of change of distal ascending aortic aneurysms which are not classically associated with Marfan syndrome are unaffected. A number of studies are currently underway to assess the use of ARBs in an adult setting.47 ,65 ,72 These trials will compare the efficacy of ARBs versus placebo, β blockade and best medical therapy. If the benefit observed in mouse and paediatric populations extends to the adult population it will be a major leap forward in the treatment of Marfan syndrome.

Targeting the disease not the defect

Surgical corrections, even prophylactically, of the cardiovascular manifestations of Marfan syndrome improve survival.88–91 Although the anatomical defect is corrected, the underlying pathology remains and continues to progress. Indeed, up to 53% of patients undergo a second surgery to repair subsequent aortic aneurysms or dissections.89 With β blockade offering potentially limited reduction of progression of aortic dilatation, a therapy that targets the underlying TGF-ß pathophysiology such as ARBs has considerable potential. In light of the above evidence, the 2010 ACC/AHA/AATS guidelines suggest the use of an ARB, such as Losartan, as a reasonable intervention to reduce the rate of thoracic aortic dilatation.64

ACE inhibitors versus ARBs

Much of the focus on reducing the rate of aortic root dilatation has been on ARBs rather than ACE inhibitors. Data from a small study of adult MFS patients receiving perindopril did result in a modest reduction in aortic root dilatation.92 The preferential use of ARBs rather than ACE inhibitors in this setting may be as a result of their differing mechanisms of action on the renin-angiotensin-aldosterone cascade. ARBs act by means of selective inhibition of the angiotensin 2 type 1 (AT1) receptor.93 AT-1 receptor inhibition can decrease production of TGF-ß ligands and receptors.81 ,94 In a somewhat opposing fashion, activation of the angiotensin type 2 (AT2) receptor results in an antiproliferative and anti-inflammatory effect. This would imply a beneficial effect of maintaining AT2 activation in aortic-wall homeostasis.81 ,95 As ACE inhibitors decrease the amount of available angiotensin 2, inhibition of both the detrimental AT1 receptor and the protective AT2 receptor occurs. Selective inhibition of AT1 induces a protective mechanism and results in overactivation of the protective AT2 receptor.71 ,81

Mitral valve prolapse

Although a non-specific feature of Marfan syndrome, mitral valve prolapse is present in up to 54% of diagnosed cases.3 ,4 When present with limited MFS systemic features (<5)49 it is referred to as mitral valve prolapse syndrome. This is likely an autosomal dominant condition95 with several candidate gene loci on chromosome 16.96 The mitral, aortic, skin and skeletal phenotype includes MVP as a cardinal feature and FBN-1 mutations have been described in this population.49 ,97 ,98 However, the likelihood of progression to aortic aneurysmal dilatation has not been defined in this subgroup.

Future directions in clinical practice

Multiplex genotyping of known sequence variants of FBN-1 can be readily and inexpensively obtained, but are not yet commercially available. However, ‘full disclosure’ with complete FBN-1 gene sequencing would be expected to be more useful in defining structural variations, insertions and deletions, and detecting novel variants in either exons or regulatory elements. FBN-1, although a major factor in the development of aortopathy or Marfan syndrome, is not the only responsible gene. There are substantial knowledge gaps in how modifier genes affect FBN-1 or the role of gene–gene interactions in the development of the pathologies described. With the range of phenotypic differences displayed within FBN-1 gene mutations alone, the inclusion of modifier gene effects and so on will surely add an additional layer of complexity in determining the effect or significance of such gene mutations.

Actionable strategies

With an expanding knowledge base of how FBN-1 gene mutations intersect with a host of cardiovascular pathologies we must now develop plans to implement actionable clinical strategies. What is clear is that because of FBN-1's genetic heterogeneity, there is substantial phenotypic heterogeneity. As Marfan syndrome presents with a range of clinical phenotypes it occasionally goes undiagnosed. Although the current ACC/AHA/AATS recommendations place more emphasis on gene analysis, it is not unreasonable to suggest genetic screening in all those suspected to have Marfan syndrome. Although detection of an FBN-1 gene mutation does not guarantee a diagnosis of Marfan syndrome, its presence should certainly alert the physician as to the possibility of underlying cardiovascular pathology, such as thoracic aortopathy, BAV disease or MVP.

With the significant cardiovascular morbidities associated with Marfan syndrome, its early diagnosis and timely treatment is critical. Screening for the presence of FBN-1 mutations in all those diagnosed with Marfan syndrome may become standard and specific mutations may identify those most likely to benefit from such therapies as ARBs. Even screening at a population level for FBN-1 mutations could result in radiologically screening for aortopathy at an earlier time point than conventionally performed. Pharmacogenetically guided therapy could then be instituted either earlier in the disease or even prior to its manifestation to delay time of onset in those at the highest risk of aortopathy who harbour a relevant FBN-1 mutation.

Limitations

As FBN-1 mutations were sequentially discovered in Marfan syndrome it was apparent that different mutations resulted in different phenotypes. Although ARB therapy may play a significant role in the treatment of Marfan syndrome, it would be unwise to expect that all mutations would result in a disease state that would be responsive to such treatments. Extrapolating its use to all those with dysmorphic thoracic aortas, for example, would need to be done with the support of pharmacogentically guided studies. These studies would subselect those most likely to benefit from specific drug therapies, as has been demonstrated with the use of Ivacaftor in those with the G551D-CFTR mutation in cystic fibrosis.99

Currently, over 600 FBN-1 mutations have been identified as causing Marfan syndrome, but this list continues to grow. Whether testing for these known variants or sequencing the entire gene are questions that remain unanswered. Even with a whole gene sequencing strategy, de novo variants may be discovered, but their pharmacogenetic significance would potentially be as of then undetermined.

Available genomic information will be viewed not in isolation but rather as an arrow in the quiver of how we ultimately predict and manage disease. Awareness of an FBN-1 mutation certainly does not guarantee developing a cardiovascular disease phenotype but it certainly highlights the possibility for a pathophysiological process to be at work at a molecular, cellular or anatomical level.

Conclusions

The expanding knowledge of FBN-1 gene mutations illuminates much of the cardiovascular implications beyond that of Marfan syndrome. Studies have highlighted FBN-1's significance in those with dysmorphic thoracic aortas, MVP and BAV. More importantly, understanding the pathophysiology of the molecular defect has resulted in therapies which substantially reduce the rate of thoracic aortic dilatation in murine and paediatric Marfan syndrome patients. The extrapolation of the genomic insights of FBN-1 mutation screening is predicated on the information being actionable. If current trials with ARBs in adults do not show efficacy, this will be stuck in the realm of knowing sequence variants but not having a strategy to prevent Marfan's or sporadic aortic dysmorphic conditions. Should ARBs or other agents be demonstrated to have marked efficacy for prevention of complications, this could lead to a marked transformation of our approach to aortic diseases.

References

View Abstract

Footnotes

  • Funding Dr Barrett has received an unrestricted research grant from A Menarini Pharmaceuticals Inc. and Daiichi Sankyo Inc.

  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.