The 8th international research symposium on the Marfan Syndrome and related conditions

Authors

  • Reed E. Pyeritz,

    Corresponding author
    1. Departments of Medicine and Genetics, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, Pennsylvania
    • Center for the Integration of Genetic Healthcare Technologies, Perelman School of Medicine at the University of Pennsylvania, Penn Tower 1115, 3400 Spruce St., Philadelphia, PA 19104.
    Search for more papers by this author
  • Bart Loeys

    1. Center for Medical Genetics, University of Antwerp, Antwerp University Hospital, Antwerp, Belgium
    Search for more papers by this author

  • How to Cite this Article: Pyeritz RE, Loeys B. 2012. The 8th International Research Symposium on the Marfan Syndrome and related conditions. Am J Med Genet Part A 158A:42–49.

Abstract

In the nearly quarter-century since the first international symposium on Marfan syndrome, enormous progress has been achieved in clinical, translational, and basic research. The 8th symposium, at the end of 2010, provides a useful summary of the current status of investigations, reveals why life-expectancy has improved so markedly during the past 30 years, and lays out a clear path for future endeavors, not only in Marfan syndrome, but in the expanding array of conditions related either through phenotype or pathogenesis. © 2011 Wiley Periodicals, Inc.

INTRODUCTION

In late summer, 2010, over 120 basic, clinical, and translational scientists gathered at Airlie Conference Center, in Virginia near Washington, DC, to discuss the latest research on Marfan syndrome and related conditions. The first such conference was held in Baltimore in 1988 [Pyeritz, 1989], and six similar meetings have occurred during the intervening years. In this article we review the major scientific advances presented at the 8th International Research Symposium on the Marfan Syndrome and related disorders and make suggestions for focused attention in the near future.

At the initial symposium in 1988, a number of clinical findings were presented for the first time. Long-term success of composite graft repair of the dilated aortic root was reported by several groups and provided stimulus for widespread adoption of prophylactic surgery to prevent aortic dissection. Follow-up of patients with Marfan syndrome (MFS) who had repair of severe pectus excavatum demonstrated that repair should be delayed, if possible, until the skeleton was nearing maturity. The preliminary results of the use of β-adrenergic blockade to delay aortic root dilatation and prevent dissection provided encouragement that non-surgical approaches held promise. Several “new” phenotypic features of MFS were described including dural ectasia and protrusio acetabulae, and attention deficit hyperactivity disorder was suggested to have increased frequency. Several series reported for the first time some of the psychological implications of MFS. Finally, a related condition, MASS phenotype, was first described. Much of the basic science focused on the search for a cause of MFS. A number of loci for fibrillar procollagens were excluded by linkage analysis using restriction fragment length polymorphisms. A large family from Kentucky was described; this pedigree soon was to be crucial in the correct linkage assignment. Immunohistopathologic studies of the extracellular microfibrils documented abnormalities in both skin biopsies and cultured dermal fibroblasts from people with MFS compared to unaffected relatives.

Since that first symposium, remarkable progress has been made in all aspects of MFS, and a wide range of related conditions has been separated from MFS, most notably the Loeys–Dietz syndrome. Advances in medical and surgical therapies have resulted in a life-expectancy in classic MFS approaching that of the general population. Mutations in FBN1, the gene that encodes fibrillin-1, the principal component of extracellular microfibrils, were shown to cause MFS in 1991, and more than one thousand pathogenic changes have been documented. However, the diagnosis of MFS in a proband remains largely based on the clinical phenotype. A set of diagnostic criteria was proposed in 1996 [De Paepe et al., 1996] and found widespread utility in both research and the clinic. Several years ago, a committee convened to consider revisions to the so-called Ghent criteria, and the results were published shortly before the 8th International Symposium in Virginia convened [Loeys et al., 2010a]. Presentation and discussion of the revised Ghent criteria provided a natural place to begin the Symposium.

CLINICAL RESEARCH ON THE MARFAN SYNDROME

The revised Ghent criteria introduce several important modifications [Loeys et al., 2010a]. First, there is re-emphasis on the importance of ectopia lentis and dilatation of the aortic root; anyone with both warrants a diagnosis of MFS. Second, dural ectasia, previously a major criterion, is relegated to one of 13 phenotypic features, each of which is assigned a score, from 1 to 3. A total score of 7 out of a possible 22 constitutes a major criterion. Third, the importance of a mutation in FBN1, which is known to be pathogenic for MFS, is emphasized. Finally, the age-dependency of many features is overtly recognized; a child or adolescent who does not reach criteria can be labeled as “potential MFS,” with the expectation that regular follow-up will occur. Two important issues persist. First, the phenotypic scoring system has not been validated, so it is not clear if having a score of 7 or greater warrants status as a major criterion. Second, the definition of aortic root dilatation focuses on a Z-score criterion, with a measurement greater than a Z of +2 indicative of enlargement. The use of Z-scores is routine in pediatrics, but not in adult cardiology. In children, the aorta grows along with body size so the aortic diameter is indexed to some measure, typically body surface area (BSA). Such indexing can create confusion, for example if two children of the same stature, one lean and the other obese, have the same aortic root dimension, the lean child may be defined as having a dilated aorta while the obese one will not. This same issue applies to adults, but there is a deeper concern. Do very tall adults deserve larger aortic root dimensions? In other words, do the graphs plotting aortic dimension versus BSA in adults (those of Roman et al. [1989] being the most commonly used) truly extend linearly upward with BSA? The few data that exist suggest not [Reed et al., 1993]. Thus, a tall, non-obese adult (2.5 m2) with an aortic root dimension of 44 mm would have a Z score <2 when indexed for BSA, when 44 mm exceeds what most authorities would consider the criterion for dilated. This individual might not then be labeled as MFS. Richard Devereux (New York) discussed new results from a large population of “normal” adolescents and adults, of whom none had an aortic root diameter greater than 43 mm. His multivariate models found important relationships between aortic root diameter and both age and gender (larger in males from adulthood on) and either BSA or height alone (BSA or height yielded similar results in adults, in children BSA was the only parameter) (Table I). Importantly, this new analysis holds promise for removing the current discontinuity between existing relationships for adolescents and adults. Subsequent to the Symposium, a European consortium compared how a group of probands with pathogenic mutations in FBN1 would be diagnosed under the old and new criteria in a retrospective study [Faivre et al., 2011].

Table I. Proposed New Formulas for Calculating Z-Scores (Based on Devereux)
For BSA:
Predicted diameter at sinuses of Valsalva in cm:
=2.423 + (age [years] × 0.009) + (BSA [m2] × 0.461) − (sex [1 = M, 2 = F] × 0.267)
SEE = 0.261 cm
Z-score = (measured [cm] − predicted diameter [cm])/0.261
For height:
Predicted diameter at sinuses of Valsalve in cm:
=1.519 + (age [years] × 0.010) + (ht [cm] × 0.010) − (sex [1 = M, 2 = F] × 0.247)
SEE = 0.215 cm
Z-score = (measured [cm] − predicted diameter [cm])/0.215

There was general agreement that chronic treatment with a negative-inotropic dose of β-adrenergic blockade remains the standard treatment for people of all ages with MFS, as long as no contraindication exists [Keane and Pyeritz, 2008]. Only a few controlled studies have supported this approach [Shores et al., 1994; Ladouceur et al., 2007], but the only evidence against β-blocker therapy suffers from methodologic flaws [Selamet Tierney et al., 2007]. Nonetheless, many patients with MFS when treated with adequate beta-blockade eventually develop progressive aortic root dilatation, or even acute dissection. Thus, the finding by Dietz and colleagues, first in the murine model of MFS [Habashi et al., 2006] and in a case series of young MFS patients [Brooke et al., 2008], that treatment with angiotensin receptor blockade reduces the rate of aortic root dilatation (in humans) and even reverses dilatation (in mice) generated considerable excitement. In the Marfan mice, treatment with losartan, either from birth or beginning at two months of age, normalized both aortic root diameter and aortic wall histopathology, while atenolol treatment had minimal effects on either measure [Habashi et al., 2006]. These preliminary findings led to a large, randomized, blinded human trial of losartan versus atenolol, primarily in North America and Belgium [Lacro et al., 2007]. This trial is supported by the Pediatric Heart Network through the U.S. National Heart, Lung and Blood Institute (NHLBI) and has a desired N of 604; at the time of the Symposium, only a few dozen additional subjects needed to be enrolled, and indeed planned enrollment was achieved in early 2011. Some subjects who were randomized at the outset have already completed the three-year period of treatment. The Data Safety and Monitoring Board has reviewed the data on several occasions and recommended continuing the trial. Final results might not appear until 2014. Meanwhile, nearly a dozen additional trials with different designs and inclusion criteria have either been started or are being planned around the world [Gambarin et al., 2009; Detaint et al., 2010; Radonic et al., 2010; Moberg et al., 2011]. Details of trials centered in Belgium, France, Italy, The Netherlands, Taiwan, and the United Kingdom were presented (Tables II and III). This enthusiasm for attempting to validate a potentially life-saving new therapy in MFS was more than evident among participants of the Symposium. However, concerns were raised for several reasons: the trials used different sartans, different drugs for comparison, and the subjects were of different ages. The dosing of sartans in some trials might be too low. Some of the trials are likely underpowered. If results of the NHLBI trial are published in the near future, recruitment in other trials might be impacted negatively. Indeed, some patients and parents of patients when offered entry into the NHLBI trial declined and began taking losartan off-label. Some of these additional trials will be done and completed, and eventually a meta-analysis might be performed, as long as the trials are similar enough to be compared. Of course the risk exists that one trial, especially an underpowered one, might produce disparate results from the others and lead to confusion as to standard of care. There was sentiment, but no consensus, to have the leaders of the various trials communicate as to how to optimize the chances for both statistically significant and clinically important results.

Table II. Existing and Proposed Drug Trials in Marfan Syndrome
CountryMulti versus single centerTreatment strategyDosing
  1. Based on [Lacro et al., 2007; Gambarin et al., 2009; Detaint et al., 2010; Radonic et al., 2010; Moberg et al., 2011].

US (PHN)MAtenolol versus losartan (blinded, no placebo)Atenolol ≥4 mg/kg/day or max 250 mg
Losartan ≥1.4 mg/kg/day or max 100mg
FranceMBB + Losartan versus BB only (blinded, no placebo)<50 kg: losartan 50 mg
>50 kg: losartan 100 mg
BelgiumSBB + Losartan versus BB only (blinced, no placebo)<50 kg: losartan 50 mg
>50 kg: losartan 100 mg
ItalySNebivolol versus Losartan versus Nebivolol + Losartan (open label, not blinded)Nebivolol: 10 mg
Losartan: 100 mg
NetherlandsMBB + Losartan versus BB only (open label, not blinded)Losartan 100 mg
TaiwanSBB + Losartan versus BB only (open label, not blinded)Losartan 100 mg for adult, 50 mg for children
UKMPrior therapy + irbesartan versus Prior therapy + placeboIrbesartan: 300 mg for adults, 150 mg for <50 kg
Table III. Details of Existing and Proposed Clinical Trials in Marfan Syndrome
CountryTarget sample sizeAge rangeFollow-up (year)Ghent criteriaAortic Z-score
US (PHN)6046 months–25 years (25% adult)3YesZ > 3
France300>10 years3YesZ > 2 or absolute > 38 mm for females and 40 mm for male
Belgium174>10 years3Yes or FBN1Z > 2
Italy29112 months–55 years (50% adult)4Yes + FBN1Z > 2.5 (now Z > 2)
Netherlands330≥18 years3YesAll
Taiwan44>1 year3YesZ > 2
UK4906–40 years4Revised GhentZ > 0

When the aortic root reaches a diameter of 45–50 mm in an adult, there is general agreement that surgical repair should be undertaken to prevent acute dissection. The “gold standard” surgical technique has been the Bentall composite graft, first described in 1968 [Bentall and De Bono, 1968], considerably modified over the years, and first applied routinely in MFS beginning in the mid-1970's. All of the case series from centers that specialize in MFS around the world have concurred that the composite graft is an excellent operation that leads to prolonged survival [LeMaire and Coselli, 1997; Gott et al., 1999], a point reiterated by surgeons at the Symposium from Hamburg, Madrid, and Tokyo. The principal detriment of the composite graft is the need for life-long anticoagulation because of the prosthetic aortic valve. Beginning with the pioneering work of Sir Magdi Yacoub in London and Tirone David in Toronto, various approaches to repairing the dilated aortic root while preserving the patient's native valve have been attempted. Several cardiothoracic surgeons presented highly encouraging results of employing the “David V” technique in MFS patients who have been followed for more than a decade. An observational, multicenter study (AVOMP registry) of patients who have had either composite graft or valve sparing repair is being led by Joseph Coselli (Houston) and recruitment was completed at the end of 2010 [Volguina et al., 2009]. A long period of follow-up would be ideal, and support is being sought to achieve this objective. At the meeting, a meta-analysis of studies comparing aortic root replacement and valve replacement with two approaches to valve-sparing root replacement was reviewed. The valve sparing procedures performed better with regards to thrombo-embolic events; the remodeling technique had a higher rate of re-intervention compared to the reimplantation technique [Benedetto et al., 2011].

Finally, attendees emphasized how multidisciplinary teams in organized centers and centralized databases, working in close collaboration with the patient organizations, benefit the care of people with MFS and related disorders (Y. von Kodolitsch, Hamburg; G. Jondeau, Paris).

MOLECULAR GENETICS OF THE MFS

Since mutations in FBN1 were identified as the cause of MFS 20 years ago [Dietz et al., 1991], more than a thousand different alterations have been identified in people who meet diagnostic criteria for MFS [Van Kien et al., 2010; Baetens et al., 2011; Faivre et al., 2011]. Few mutations have occurred with any frequency among unrelated patients. Unfortunately for those who seek specificity in molecular diagnosis, mutations in FBN1 have been found in individuals with a variety of conditions that are not MFS, but share some features (e.g., Shprintzen-Goldberg, MASS), and most interestingly, in conditions that are the antithesis of MFS (Table IV), such as stiff skin syndrome [Loeys et al., 2010b] and Weill–Marchesani syndrome [Faivre et al., 2003]. For people with classic MFS, the likelihood of finding a mutation in FBN1 approaches 95% if traditional sequencing is combined with methods for searching for large rearrangements and deletions [Baetens et al., 2011]. Mutations in the promotor or 3'UTR seem extremely rare and do not account for the missing 5% (Loeys, personal observation). Methylation, deep intronic variants and regulatory elements have not yet been investigated. Finding the mutation can be useful in confirming the diagnosis of MFS, if someone in the family who meets Ghent criteria has a known mutation, or in the unlikely possibility that the same mutation has been reported in an unrelated individual with unequivocal MFS.

Table IV. Mutations in FBN1 Associated With Phenotypes Other Than Marfan Syndrome
Amino acidExonExon typePhenotypeOMIM no.
  • EGF, epidermal growth factor-like; ncb, non-calcium-binding; cb, calcium-binding; C-term, C-terminus; TB, transforming growth factor beta Binding protein; AA amino acid. Table adapted from Pyeritz [2012].

  • a

    This mutation also causes classic Marfan syndrome.

  • b

    Other cysteine-related mutations in this domain have been described in stiff skin syndrome patients.

R122C4EGF (ncb)Atyical skeletal, no cardiovascular134797
G1127S27EGF (cb)Familial aortic aneurysm134797
C1223Ya29EGF (cb)Shprintzen–Goldberg syndrome182212
W1570Cb37TB4 domainStiff skin syndrome184900
Frame shift41TB5 domainMASS604308
8AA-deletion41TB5Weill–Marchesani608328
T1696C41TB5 domainGeleophysic dysplasia231050
A1728T42TB5 domainAcromicric dysplasia102370
E2447K59EGF (cb)Familial ectopia lentis129600
R2726W64C-termFamilial tall stature134797
W2756X65C-termMASS604308

Interest persists in attempting to define genotype–phenotype correlations, but little emerged beyond the old impressions that mutations in the middle third of fibrillin-1 (exons 24–32) tend to result in a more severe condition, that cysteine-replacing missense mutations are more likely to cause ectopia lentis compared to other missense mutations, and that premature truncation mutations are more likely to be associated with a more severe skeletal and skin phenotype (L. Faivre, Dijon). FBN1 deletion patients were shown to have a wide range of severity in clinical phenotype (Y. Hilhorst, Leiden and L. Faivre, Dijon). Reinhardt and colleagues (Montreal) reported that mutations associated with a severe clinical phenotype were more likely to render fibrillin-1 susceptible to protease cleavage and to impair attachment of fragments of fibrillin-1 to cultured cells.

CLINICAL AND MOLECULAR RESEARCH ON RELATED CONDITIONS

Mutations in several of the receptors for transforming growth factor-beta (TGFβ) were first reported to cause MFS. Whether these patients actually meet current diagnostic criteria for MFS remains debated. There is general agreement that the vast majority of people (95%) who meet Ghent criteria for MFS have a mutation in FBN1. What became strikingly clear several years ago is that a group of patients, many previously diagnosed as MFS, is phenotypically distinct and have mutations in TGFβ receptors 1 or 2 (TGFBR1 or TGFBR2). This condition, the Loeys–Dietz syndrome (LDS) is autosomal dominant with considerable variability. Features that set it apart from MFS include an absence of ectopia lentis and presence of arterial tortuosity, a tendency for aneurysms beyond the aortic root, a tendency to dissect at smaller aortic diameter, hypertelorism, bifid uvula or cleft palate, various autoimmune disorders, and club foot [Loeys et al., 2005, 2006]. There was general agreement that people with LDS require close monitoring and prophylactic aortic root surgery at smaller aortic root dimensions (e.g., 40 mm in adults) than in MFS.

From presentations at the Symposium, it was not clear to all groups that every person with a mutation in TGFBR1 or TGFBR2 should be labeled LDS or need to be subjected to such aggressive prophylactic surgery. Some investigators, who usually ascertain patients in adulthood, believe that a non-syndromic form of thoracic aortic aneurysm and dissection (TAAD) can be due to mutations in a number of genes, including TGFBR1 and TGFBR2. Attias et al. [2009] published data, updated by G. Jondeau at the symposium, that showed little difference in either cardiovascular presentation or response to treatment between adults with a mutation in TGFBR2 or with MFS and a mutation in FBN1. The mean age of the TGFBR2 patient cohort was 28 years and patients belonged to 26 families. Tran-Fadulu et al. [2009] ascertained families with TAAD and compared those with mutations in TGFBR1 and TGFBR2. The results, updated by D. Milewicz at the Symposium, showed that men with a mutation in TGFBR1 were more likely to dissect at an aortic root dimension <50 mm and women with a mutation in TGFBR1 developed aneurysms and dissections in arteries other than the aorta. Morisaki (Osaka) and Arbustini (Pavia) reported on 30 Japanese and 80 Italian patients, respectively, with TGFBR1/2 mutations and presented similar findings to the initial reports [Loeys et al., 2005, 2006]. Finally, Loeys presented a previously unpublished cohort of 105 TGFBR1/2 mutation positive probands and confirmed aortic dissections (n = 23) with a median age of 20 years. The current view on the phenotypical spectrum of TGFBR1/2 mutations involves a continuum starting from isolated young probands with more severe outward features of LDS to older familial cases with less outward features more belonging to the TAAD spectrum. Most likely genetic modifiers determine the degree of expressivity as the same TGFBR1/2 mutations have been found in typical LDS and TAAD-like presentations.

In a search for genetic variants predisposing to the development of sporadic thoracic aortic aneurysm and dissection (STAAD), Milewicz and colleagues [Kuang et al., 2011] identified a highly significant association between TAAD and a duplication on chromsome 16p13 (encompassing the MYH11 gene, a gene previously identified as causal for familial thoracic aortic aneurysm with patent ductus ateriosus). This suggests that copy number variation might affect smooth muscle contraction, but further validation of this hypothesis is awaited.

In line with previous observations, mutations in familial TAAD patients were reported in genes encoding components of the smooth muscle contractile apparatus, including smooth muscle specific alpha actin (ACTA2), myosin heavy chain (MYH11), and myosin light chain kinase (MYLK). Decreased SMC contractile function leads to the activation of different SMC cellular pathways causing proteoglycan accumulation, elastin degradation, and SMC proliferation [Milewicz et al., 2010].

Several clinical-molecular study results on other MFS-related conditions were presented. Defects in filamin A were reported in four patients with non-syndromic tetralogy of Fallot with ascending aortic aneurysm. It was hypothesized that at least a subset of these mutations affect the calpain cleavage site of filamin A. Decreased filamin A correlated with increased ERK signaling (D. Kim, Baltimore). LTBP4 mutations cause the recently described Urban–Rifkin–Davis syndrome. In this condition the aortic elastic fibers seem largely preserved but the pulmonary (emphysema), skin (cutis laxa), bladder (diverticula), and intestinal (dilatations) complications are severe (Z. Urban, Pittsburgh). Null mutations in LTBP2 cause a wide range of ocular phenotypes (primary congenital glaucoma, ectopia lentis, microspherophakia, megalocornea, high myopia). The absence of cardiovascular and skeletal abnormalities suggest a key role for LTBP2 in maintaining ciliary muscle tone and support of the lens (A. Manir, Leeds). Ectopia lentis is also the main presentation of autosomal recessive ADAMTSL4 mutations (N. Hanna, Paris and A. Child, London). Interestingly, mutations in either ADAMTS10 or ADAMTSL2 cause acromelic dysplasia, a group a skeletal dysplasias with short stature, brachydactyly and joint stiffness. ADAMTS10 mutations are found in the autosomal recessive form of Weill–Marchesani syndrome [Dagoneau et al., 2004], whereas ADAMTSL2 mutations can cause geleophysic dysplasia [Le Goff et al., 2011]. Recently mutations in the 5th LTBP domain of FBN1 were identified in patients with geleophysic and acromicric dysplasia. Fibrillin-1 and LTBP1 were identified as binding partners of ADAMTSL2 and increased levels of TGFβ signaling were found in fibroblasts cultured from patients with geleophysic dysplasia. Sengle and colleagues (Portland) reported a deletion of exons 9–11 in FBN1 leading to Weill–Marchesani phenotype in a single family. A mouse model lacking the same exons showed only thick skin. Further biochemical studies suggested a binding site for ADAMTSL2, 3 and 6 in this region of fibrillin-1. Moreover, interaction between ADAMTSL3, ADAMTS10 and fibrillin-1 was demonstrated. It is striking that disruption of the function of specific binding partners of fibrillin-1, such as ADAMTS10 and ADAMTSL2 leads to short stature and skin phenotypes without obvious aortic root dilatation.

NEW INSIGHTS INTO THE PATHOGENESIS OF MARFAN SYNDROME AND RELATED CONDITIONS

Over the past decade, it has become increasingly clear that fibrillin and microfibrils act to localize, concentrate and stabilize the latent TGFβ binding complex. Although until recently most studies have focused on the canonical TGFβ signaling, there is emerging evidence that non-canonical pathways, such as mitogen-activated protein kinase (MAPKs), may have a role in aneurysm development [Habashi et al., 2011; Holm et al., 2011]. It was demonstrated in the aorta of fibrillin-1 deficient mice that TGFβ- and angiotensin II type 1 receptor (AT1R)-dependent activation of the extracellular-signal regulated kinases (ERK1 and ERK2) is involved in the pathogenesis of aneurysm, and further evidence for their importance was obtained from the abrogation of pathological aortic root growth after treatment with a specific ERK inhibitor. Similar lessons were derived from the study of a fibulin-4 (Fbln4) smooth muscle knock-out model [Huang et al., 2010].

Several other new clues regarding the assembly and homeostasis of microfibrils were derived from the study of various mouse models. Masahiro et al. (Tokyo) demonstrated that ADAMTSL6 expression rescues microfibril formation in a periodontal ligament injury model in mgR/mgR mouse model. Sakai and colleagues (Portland) created a mouse model with a conditional truncated fibrillin-1 tagged with eGFP (called GT8/+). These heterozygotes show progressive fragmentation of elastic fibers from two months of age. The latter was absent in the endothelial cell specific (under control of Tie2) mutants and happened slower in vascular smooth muscle cell (VSMC) specific mutants, suggesting that VSMC contribute to fibrillin-1 secretion but are not the sole contributor. Pereira and colleagues (Sao Paulo) found tissue specific differences in the expression levels of deficient fibrillin-1 molecules in different genetic backgrounds (129/Sv versus C57Bl/6). This may suggest that genetic background and epigenetic factors also contribute to the clinical variability in MFS. Further evidence for epigenetic factors was provided in the study by Gomez et al. [2011]; they showed that SMAD2 overexpression is dependent of H3 histone acetyl transferases. The study of different mouse models for Loeys–Dietz syndrome suggests a gain-of-function mechanism for the TGFBR1/2 mutations. This is evidenced by the absence of vascular problems in haploinsufficient tgfbr1 or 2 mice and the presence of severe vascular disease in two knock-in strains with increased canonical and non-canonical TGFβ signaling in aortic tissues.

These studies emphasize the strength of using mouse models for the study of human disease, such as through examining the effect of different mutations in different genes under specific temporal, cellular, and tissue specific conditions. Despite these advantages, one should always be aware that mice are not humans and not all mice are created equal.

At the symposium a lot of attention was also given to the developmental aspect of the aorta. Truly, the aorta is a developmental mosaic and the effects of TGFβ signaling on different tissues from different embryonic origin seem to be different and extremely relevant in the development of aortic aneurysms. Vascular smooth muscle cells (VSMCs) at the root of the aorta and pulmonary artery are derived from the second heart field (cardiogenic mesoderm), whereas, the ascending aorta is a chimera of cells derived from the second heart field and the ectodermal cardiac neural crest (CNC). In the more distal ascending aorta and the transverse arch all VSMCs are derived from the CNC. There is an abrupt transition to somatic mesoderm-derived cells in the proximal descending thoracic (juxtaductal) aorta and a contribution of splanchnic mesoderm to the descending aorta beginning just below the diaphragm. Remarkably, although there is not a common origin for VSMCs, the sites of interaction between cells of divergent origin seem the most vulnerable location for the development of aortic aneurysms. Lineage tracing experiments in Marfan mouse models (C1039G) suggest a proliferation of the secondary heart field derived cells (M. Lindsay, Baltimore). It is also hypothesized that cells of one lineage, such as mesodermal cells in the aortic root and descending thoracic aorta, would be more sensitive to a perturbation of TGFβ signaling.

In addition to the TGFβ pathway, many other pathways have been invoked in the pathogenesis of aortic aneurysm. The angiotensin pathway cross-talks with the TGFβ signaling pathway. Knocking-out the angiotensin II receptor type 2 exacerbates the aneurysm phenotype in C1039G mouse models. In an AngII-induced aortic aneurysm model in LDL receptor-deficient mice, the deficiency of angiotensin II receptor 1a did ablate the development of aortic aneurysms. After excluding a central role of ATIIR1a in leucocytes and vascular smooth muscle cells, endothelial-specific ATIIR1a plays a crucial role in the development of aortic aneurysms [Rateri et al., 2011].

Finally, some evidence was presented that inflammatory processes might also play a role in the progression of aortic dilatation. Smooth muscle specific over-expression of S100A12 leads to increased IL-6 production, activation of TGFβ pathways and increased metabolic activity with enhanced metabolic stress [Hofmann Bowman et al., 2010]. Radonic and colleagues (Amsterdam) also found evidence for increased response of Th1 cells based on gene expression profiling study of skin fibroblasts and serum cytokine analysis.

TREATMENT PROSPECTS BEYOND SARTANS

The enthusiasm for angiotensin receptor blocker treatment for MFS did not obscure the need to explore other treatment strategies. In the mgR/mgR mouse model combined treatment with doxycycline (a non-specific matrix metalloprotease inhibitor) and losartan was superior in reducing aortic root dilatation compared to each treatment alone (T. Baxter, Omaha). In another Marfan mouse model (C1039G), no beneficial effect of doxycycline alone could be shown, but pravastatin (an inhibitor of isoprenoids which are essential in the transportation and secretion of metalloproteinases) was equally effective as losartan in reducing aortic root dilatation (D. McLoughlin, Dublin). Caution emerged about the use of calcium channel blockers, as amlodipine worsened ascending aortic dilatation and led to increased mortality in a Marfan mouse model (C1038G) (J. Doyle, Baltimore). Finally, alendronate improved bone quality in the mgR/mgR mouse model, a finding not seen after losartan treatment [Nistala et al., 2010].

FUTURE PROSPECTS

At the end of the meeting, Reed Pyeritz defined some key challenges and questions for the future of clinical care and research in MFS and related disorders. Some questions will need answers in the short term: What is the best way to define aortic root dilatation? A critical re-evaluation of the use of Z-scores is warranted. What is the best and safest way to assess the arterial system, how early, and how often? How will the field proceed if the results of one or more of the angiotensin receptor blocker trials are disappointing? Potential alternatives such as pERK inhibitors and AT2 agonists were identified, but need further experimental validation in mouse models before translation to human care can be realized.

Many other questions remain grand challenges and will require well resourced, collaborative, international efforts to be solved. How can we better understand the variable clinical expression of FBN1 and TGFBR1/2? Can we identify better predictors of progression of aortic root dilatation? Although major progress has been made in understanding the pathogenesis of aortic aneurysmal disease, further dissection of the role of TGFβ, angiotensin and other signaling pathways will be essential for the development of new treatment strategies. The current clinical studies are mainly focused on disease progression, but through a better understanding of development, homeostasis and pathogenesis, we should aim to develop models for, and clinical trials of, tissue remodeling and reversal of existing disease. We still lack knowledge of the precise role of specific cell types from different embryonic origin, the precise mechanism by which TGFβ exerts its function in the cardiovascular system, the function of the AT1 receptor, and the roles of hemodynamic stress, inflammation and proteases.

One of the common characteristics of human Mendelian disease is variable expression. This is pervasive in MFS and some of the related disorders. Attempts have been made to explain variability, primarily by using linkage approaches with candidate genes in large pedigrees of MFS, to little effect. Studies using genome-wide association with SNPs are underway. Ultimately, whole exome or genome sequencing within and among families may be needed to identify modifying genes. Furthermore, epigenetic effects will need to be considered. Finally, we should marvel at, but attempt to explain, some apparent paradoxes: how can different mutations in FBN1 cause long and short bones, stiff and stretchy skin, big and small lenses, aortic aneurysm and no aortic dilatation, etc.?

Acknowledgements

The National Marfan Foundation (Jon Tullis, President, Carolyn Levering, Executive Director, Josephine Grima, PhD, Scientific Director) and the International Federation of Marfan Syndrome Organizations provided the stimulus for organizing this symposium. The Program Committee was chaired by Peter Byers, MD, Hal Dietz, MD, and Bart Loeys, MD, PhD. The following organizations provided financial support for the symposium: University of Washington Collagen Diagnostic Laboratory; Genzyme Corporation; the March of Dimes; and the Airlie Conference Center. We thank Ron Lacro, MD, and Richard Devereux, MD, for sharing their insights. Preparation of this report was supported by the NHLBI through a contract to RTI International (GenTAC Registry NIH contracts HHSN268200648199C and HHSN268201000048C). This manuscript was prepared while REP was in residence at the Brocher Foundation, Hermance, Switzerland.

Ancillary