Analysis of TGFBR1*6A variant in individuals evaluated for Marfan syndrome
Abstract
Marfan syndrome (MFS) and Loeys–Dietz syndrome (LDS) are genetic disorders that affect connective tissue as a result of dysregulated TGF-β signaling. MFS is most frequently caused by mutations in FBN1 whereas Loeys–Dietz syndrome results from mutations in TGFBR1 or TGFBR2. There is substantial inter- and intra-familial phenotypic variability among these disorders, suggesting the presence of genetic modifiers. Previously, a polymorphism in the TGFβR1 protein termed the TGFBR1*6A allele was found to be overrepresented in patients with MFS and was identified as a low penetrance allele with suggestion as a possible modifier. To further investigate the importance of this variant, a retrospective review of genetic and phenotypic findings was conducted for 335 patients evaluated for suspicion of MFS or related disorders. In patients with a diagnosis of MFS, the presence of the TGFBR1*6A allele was not associated with phenotypic differences. Similarly, careful phenotyping of patients who carried the TGFBR1*6A allele but did not have MFS did not identify an altered frequency of specific connective tissue features. In this small cohort, the results did not reach significance to identify the TGFBR1*6A allele as a major modifier for aortic dilation, ectopia lentis, or systemic features associated with MFS or other connective tissue disorders. © 2016 Wiley Periodicals, Inc.
INTRODUCTION
Marfan syndrome (MFS) is a genetic disorder that affects connective tissue and is most frequently caused by mutations in FBN1. The FBN1 gene produces the protein fibrillin-1, an elastic fiber component that contributes to the structure and function of the extracellular matrix in skeletal, cardiovascular, and ocular tissues, among others [Dietz and Pyeritz, 1995; Cook and Ramirez, 2014]. Fibrillin-1 is a large glycoprotein and the primary component of microfibrils, which when mutated form defective connective tissues [Dietz and Pyeritz, 1995]. Clinical manifestations of MFS include cardiac, eye, and systemic features, including skeletal findings [Jondeau and Boileau, 2012; Cook and Ramirez, 2014; Romaniello et al., 2014]. While the causes of MFS have been identified, the sources of inter-and intra-familial variability in clinical manifestations are unclear, suggesting the presence of major genetic modifiers. This phenotypic heterogeneity complicates diagnosis and management.
Transforming growth factor-β (TGF-β) is a soluble cytokine that binds TGFβR1/TGFβR2 receptors, ultimately leading to initiation or repression of expression of target genes regulating cell growth, development, extracellular matrix (ECM) production, and many other cellular functions [Davis and Summers, 2012; Constam, 2014; Wheeler et al., 2014]. The bioavailability of TGF-β proteins is controlled by fibrillin-1 through sequestration. In MFS, mutations in FBN1 lead to impaired sequestration of TGF-β, increasing TGF-β signaling [Habashi et al., 2006; Halper and Kjaer, 2014]. Mutations in TGFBR1 and TGFBR2 also lead to dysregulated TGF-β signaling [Loeys et al., 2005]. Pathogenic mutations that modify TGF-β signaling lead to degenerative changes of ECM that contribute to clinical manifestations of disease. In a mouse model of MFS, excess TGF-β was associated with aortic aneurysms, mitral valve prolapse, and lung emphysema [Neptune et al., 2003; Ng et al., 2004].
MFS is caused by a largely heterogeneous group of mutations in FBN1, with few evident genotype–phenotype correlations. Mutations in exons 24–32 of FBN1 are associated with neonatal MFS and more severe MFS, although mutations causing classic MFS can also be found in this region [Tiecke et al., 2001; Ng et al., 2004]. Additionally, individuals with a cysteine substitution in FBN1 are more likely to have ectopia lentis [Biggin et al., 2004; Rommel et al., 2005; Faivre et al., 2007]. Significant variability is identified even between individuals that possess the same mutation, suggesting that additional genes contribute to modify phenotype. There is a limited understanding of how dysregulation of the TGF-β pathway leads to different phenotypes. An improved understanding of the genes that alter TGF-β signaling may clarify the mechanisms underlying this phenotypic variability and provide insight into the development of connective tissue feature in individuals who do not have MFS.
One polymorphism that is a candidate for modifying this signaling pathway is the TGFBR1* 6A variant. This variant in the TGFBR1 gene has a deletion of three alanines within a nine alanine tract in exon 1 [Schirmer et al., 2009]. It is hypothesized that the TGFBR1*6A allele is associated with a greater number and severity of the signs and symptoms of MFS. Lucarini et al. [2009] reported a significant difference in TGFBR1*6A allele frequency between Italian individuals with MFS (n = 160; 0.13) and controls (n = 237; 0.08). The authors suggested the TGFBR1*6A allele may act as a low penetrance allele in MFS. The primary objective of this study was to identify any clinical manifestations associated with the TGFBR1*6A allele, as we hypothesized that this allele serves as a modifier associated with a more severe phenotype. Identifying clinical manifestations associated with the TGFBR1*6A may help in risk assessment and management of individuals with MFS or individuals with signs and symptoms of connective tissue disorders.
METHODS
Study Population
This was a retrospective cohort study. The study population included patients of any age who were evaluated by Cardiovascular Genetics at Cincinnati Children's Hospital Medical Center (CCHMC) and underwent clinical genetic testing for FBN1, TGFBR1, and TGFBR2 (Marfan Syndrome and Marfan Related Syndromes Disorders Gene Panel), Thoracic Aortic Aneurysm (TAA) Panel (including minimally FBN1, TGFBR1/2, MYH11, and ACTA2; newer version includes 14 genes), or TGFBR1 and FBN1 full mutation analysis through the Heart Institute Diagnostic Laboratory (HIDL) at CCHMC from August 1, 2010 to February 26, 2014. HIDL is a clinical laboratory certified by the College of American Pathologists (CAP) for molecular genetic testing. Patients from outside institutions who were not evaluated at CCHMC but underwent genetic testing through HIDL were excluded. Patients were separated into groups based on diagnosis and presence or absence of 6A allele: MFS and TGFBR1*6A (+MFS/+6ala), MFS without TGFBR1*6A (+MFS/−6ala), TGFBR1*6A without MFS (−MFS/+6ala), absence of TGFBR1*6A and MFS (−MFS/−6ala). The study was approved by the Institutional Review Board at CCHMC, which waived patient consent.
Instrumentation
A search in CCHMC's i2b2 Research Data Warehouse was performed to identify patients who met the inclusion criteria. Demographic data were collected, including age, gender, and race/ethnicity. Clinical data were recorded, including diagnosis based on the revised Ghent criteria for MFS [Loeys et al., 2010], Beighton criteria for hypermobility, and additional clinical features associated with MFS and related disorders. Clinical data were classified by affected body systems including cardiovascular, musculoskeletal, craniofacial, ocular, gastrointestinal, and pulmonary. The clinical data included the presence or absence of cardiovascular features (ascending aorta dilation, aortic root dilation, aortic root dissection, aortic root surgery, aortic valve dysfunction, mitral valve prolapse (MVP), mitral valve dysfunction, valve surgery, pulmonary artery dilation, patent ductus arteriosus (PDA), congenital heart disease, arterial tortuosity, easy bruising, aneurysm), musculoskeletal (wrist and thumb sign, limitation of elbow extension, pectus carinatum/excavatum, hindfoot deformity, pes planus, dural ectasia, protusio acetabuli, reduced upper segment to lower body segment with arm to height ratio >1.05, clubfoot, cervical spine instability, scoliosis, arachnodactyly), craniofacial (presence of three or more of features including dolicocephaly, malar hypoplasia, enophthalmos, retrognathia, and down-slanting palpebral fissures), ocular (myopia, ectopia lentis), gastrointestinal (eosinophilic esophagitis and hernia), skin (striae, abnormal scarring) and pulmonary (pneumothorax). Patients included in the study underwent both clinical evaluation and genetic testing at CCHMC, so that all clinical data listed above were available for each patient. All genetic testing results, including TGFBR1*6A status, were recorded.
Data Analysis
Demographics, genetic testing results, and clinical features were examined in a descriptive manner. Data was exported to JMP 12.0.1 (SAS, Cary, NC) for statistical analysis [JMP®, ]. To test our hypothesis that the TGFBR1*6A allele is associated with a more severe phenotype, we compared clinical features of patients with the TGFBR1*6A allele versus those without the allele stratified by MFS presence. Chi-square test or Fisher's exact test and Kruskal–Wallis test were used. Clinical features were combined into body systems for additional analysis. Systemic features were weighted according to revised Ghent criteria for analysis of features combined into body systems (e.g., pectus carinatum = 2, pectus excavatum = 1). Due to the limited sample size, power analysis indicated the data were underpowered to detect anything other than full penetrance.
RESULTS
Study Population
The i2b2 Research Data Warehouse identified 373 patients who underwent genetic testing for MFS or TAA at CCHMC. Four patients whose testing was ordered from an outside institution and nine patients who were not evaluated in genetics clinic were excluded due to lack of clinical data. Sixteen patients were excluded for variants of uncertain significance, leaving 344 patients who underwent both a clinical evaluation and genetic testing at CCHMC.
Genetic testing included MFS and related disorders panel (n = 284), TGFBR1 sequencing (n = 31), or TAA panel (n = 29). Based on revised Ghent criteria, 28 individuals were diagnosed with MFS (Table I). One patient was diagnosed with MFS based on follow-up deletion/duplication testing of FBN1. Two patients had a clinical diagnosis with no positive molecular testing results and were excluded from the study. Five individuals with MFS had the TGFBR1*6A allele (19.2%) and 40 (11.9%) of those who did not meet diagnostic criteria for MFS had the allele. Six patients were diagnosed with LDS and one was diagnosed with Sphrintzen–Goldberg syndrome, none of whom also had the TGFBR1*6A allele. Due to the absence of the TGFBR1*6A allele and limited sample size in patients with LDS and Sphrintze–Goldberg syndrome, these patients were excluded from further analysis. The final sample size was 335.
| Gender | |
| Male | 229 (68.4%) |
| Mean Age | 19.1 (range 1–68) |
| Race | |
| Caucasian | 303 (90.4%) |
| African American | 22 (6.56%) |
| Bi-racial | 6 (1.8%) |
| Other | 2 (0.6%) |
| Asian | 2 (0.6%) |
| Syndromes | Mutation in FBN1 (unless otherwise noted) |
| +MFS/−6alaa (n = 21) | c.8176C>T (p.Arg2726Trp) |
| c.6890C>T (p.Thr2297Met) | |
| c.1609C>T in TGFBR2 (p.Arg537Cys) | |
| c.1426T>G (p.Cys476Gly) | |
| c.1426T>G (p.Cys476Gly)b | |
| c.7240C>T (p.Arg2414*) | |
| c.4079delAA (p.Lys1360Serfs*53) | |
| c.6658C>T (p.Arg2220) | |
| c.874_875insC (p.Cys292Serfs*9) | |
| c.2273C>A (p.Ser758*) | |
| c.3650G>A (p.Gly1217Asp) | |
| c.3026C>G (p.Pro1009Arg) | |
| c.1787G>A (p.Cys596Tyr) | |
| c.4043G>A (p.Cys1348Tyr) | |
| c.8176C>T (p.Arg2726Trp) | |
| c.3413G>A (p.Cys1138Tyr) | |
| c.2728G>T (p.Asp910Tyr) | |
| Heterozygous deletion exons 48–49c | |
| c.897_898delTG (p.Cys299*) | |
| c.4282C>T (p.Arg1428Cys) | |
| c.8176C>T (p.Arg2726Trp) | |
| +MFS/+6alad (n = 5) | c.1052delA (p.Gln351Argfs*3) |
| c.7039_7040delAT (p/Met2347Valfs*9) | |
| c.3596A>G (p.Asp1199Gly) | |
| c.5885A>G (p.Tyr1962Cys) | |
| c.3130 T>G (p.Cys1044Gly) | |
- a Participants with MFS without the TGFBR1*6A allele.
- b Unrelated to the other proband with this variant to the best of our knowledge.
- c Reference sequence: NM_000138.4 and genomic coordinates: chr15:48,736,819–48,737,653 [hg19/GRCh37].
- d Participants with Marfan syndrome (MFS) and the TGFBR1*6A allele.
Diagnoses and Clinical Features
Table II shows the cardiac findings and systemic scores for the cohort. Individuals with MFS, with or without the TGFBR1*6A allele were compared. Likewise, individuals with connective tissue features in the absence of MFS were compared. For each of the cardiac features, there is no statistically significant difference between the +MFS/−6ala and +MFS/+6ala or −MFS/+6ala and −MFS/−6ala groups. In this small sample, there was a marked increase of mitral valve prolapse in +MFS/+6ala over +MFS/−6ala (80.0% vs. 42.9%, respectively), but statistical significance was not reached (P = 0.32). Similarly, for each of the systemic features, there is no statistically significant difference between the +MFS/−6ala and +MFS/+6ala or −MFS/+6ala and −MFS/−6ala groups.
| Group | ||||
|---|---|---|---|---|
| +MFS/+ala | +MFS/−ala | −MFS/+ala | −MFS/−ala | |
| MVP | 83.3% (5/6) | 40.9% (9/22) | 5.1% (2/39) | 7.1% (19/268) |
| Dilated aortic root | 50% (3/6) | 77.3% (17/22) | 25.6% (10/39) | 26.0% (66/269) |
| Dilated ascending aorta | 0% (0/6) | 13.6% (3/22) | 10.3% (4/39) | 13.5% (37/268) |
| PDA | 0% (0/6) | 0% (0/22) | 5.1% (2/39) | 2.60% (7/268) |
| CHD | 0% (0/6) | 27.3% (6/22) | 0.0% (0/39) | 17.9% (48/268) |
| Facial features | 83.3% (5/6) | 36.3% (8/22) | 10.3% (4/39) | 14.6% (39/268) |
| Hindfoot deformity | 66.7% (4/6) | 36.3% (8/22) | 11.4% (5/39) | 17.5% (47/268) |
| Pectus carinatum | 50% (3/6) | 31.8% (7/22) | 7.7% (3/39) | 9.0% (24/268) |
| Arachnodactyly | 50% (3/6) | 72.7% (16/22) | 30.83% (12/39) | 25.0% (67/268) |
| Scoliosis | 33.3% (2/6) | 27.3% (6/22) | 5.1% (2/39) | 11.2% (30/268) |
| Reduced elbow extension | 33.3% (2/6) | 27.3% (6/22) | 2.6% (1/39) | 7.1% (19/268) |
| Skin striae | 33.3% (2/6) | 18.2% (4/22) | 23.1% (9/39) | 23.5% (63/268) |
| Ectopia lentis | 16.7% (1/6) | 31.8% (7/22) | 2.6% (1/39) | 0.7% (2/268) |
| Myopia | 16.7% (1/6) | 45.5% (10/22) | 20.5% (8/39) | 23.5% (63/268) |
| Pectus excavatum | 16.7% (1/6) | 22.7% (5/22) | 35.9% (14/39) | 27.6% (74/268) |
| Eosinophilic esophagitis | 0% (0/6) | 9.1% (2/22) | 7.5% (3/40) | 11.9% (32/269) |
- CHD includes bicuspid aortic valve, atrial septal defect, ventricular septal defect, patent foramen ovale, tetralogy of Fallot. Only features which were documented as present or absent in the medical record were recorded, therefore the denominators are not identical for each feature.
The mean MFS diagnostic criteria systemic scores are summarized in Table III. In one-way analysis of variance, there was no difference between +MFS/−6ala and +MFS/+6ala or −MFS/+6ala and −MFS/−6ala groups. Features were also combined into body systems for additional analyses including eye, cardiac, and musculoskeletal systems. No differences were identified between +MFS/−6ala and +MFS/+6ala or −MFS/+6ala and −MFS/−6ala.
| Group | Mean systemic score (SD) | |
|---|---|---|
| +MFS/+6alaa | 7.40 (2.30) | |
| +MFS/−6alab | 5.67 (2.74) | |
| −MFS/+6alac | 2.88 (2.11) | |
| −MFS/−6alad | 2.87 (2.31) | |
- Revised Ghent nosology systemic features scoring includes wrist and thumb sign − 3 (wrist or thumb sign – 1), pectus carinatum – 2 (pectus excavatum − 1), hindfoot deformity − 2 (pes planus − 1), pneumothorax − 2, dural ectasia − 2, protrusion acetabula − 2, reduced upper segment/lower segment and increased arm/height and no severe scoliosis − 1, scoliosis or thoracolumbar kyphosis − 1, reduced elbow extension − 1, facial features (3/5) − 1 (dolichocephaly, enophthalmos, downslanting palpebral fissures, malar hypoplasia, retrognathia), skin striae − 1, myopia − 1, mitral valve prolapse − 1. Maximum total is 20 points, score ≥7 indicates systemic involvement.
- a Participants with Marfan syndrome (MFS) and the TGFBR1*6A allele.
- b Participants with MFS without the TGFBR1*6A allele.
- c Participants without MFS with the TGFBR1*6A allele.
- d Participants without MFS or the TGFBR1*6A allele.
DISCUSSION
Greater than 1,000 mutations have been identified in FBN1 and TGFBR1/2 in association with MFS, with few evident genotype–phenotype correlations [Dietz and Pyeritz, 1995; Biggin et al., 2004; Rommel et al., 2005; Faivre et al., 2007; Cook and Ramirez, 2014]. Additionally, there is clinical variability among individuals with identical mutations, implicating modifying effects of additional genes. The TGFBR1*6A variant has been studied extensively for an association with tumorigenesis, with evidence as a low penetrance tumor susceptibility allele in some cancers. The TGFBR1*6A variant was previously identified as over-represented in patients with MFS and related disorders and was suggested as a low penetrance allele in MFS with suggestion of further research to evaluate the allele as a modifier [Lucarini et al., 2009]. In functional studies, the presence of the 6A allele was associated with limited downregulation of SMAD3 in the presence of TGFβ1 [Schirmer et al., 2009]. SMAD3, a secondary messenger of the canonical TGF-β signaling pathway, serves a stimulatory function, but is repressed by high concentrations of TGFβ1 [Wheeler et al., 2014]. This results in increased levels of TGF-β, a known disease mechanism in MFS and related disorders [Neptune et al., 2003; Wheeler et al., 2014].
Lucarini et al. reported a higher TGFBR1*6A allele frequency in Italian individuals with MFS (0.13) versus controls (0.08). The clinical significance of this finding is unclear, posing challenges to both diagnostic laboratories and clinicians. Reference SNP data reports a frequency of the TGFBR1*6A allele (rs11466455) of 0.087 for African Americans (n = 46) and 0.125 for Europeans (n = 40). Unfortunately, due to the repetitive nature of the polyalanine tract, reference data from exome sequencing is not available for this region. A larger MFS and control cohort is necessary to establish a more definitive TGFBR1*6A allele frequency, and to determine if there is a true difference in the allele frequency between the two populations.
This analysis was undertaken to define the spectrum of clinical features in patients carrying the TGFBR1*6A allele who were evaluated in a cardiovascular genetics clinic. The cohort was carefully phenotyped for findings associated with a connective tissue disorder (Table II). In this small sample, no clinical feature reached a level of significance between the +MFS/+6ala and +MFS/−6ala or −MFS/+6ala and −MFS/−6ala groups. In addition, the presence of the allele alone does not explain manifestations of features associated with MFS in individuals who do not meet diagnostic criteria. Limitations of this study include its descriptive nature and the small sample size, thus the results of this study should be interpreted with caution. Continued research is necessary as the sample population increases to gain further evidence of the role of the TGFBR1*6A allele in features of MFS. A recent study using mice on a mixed, rather than inbred, background, showed strain-dependent phenotypic variability that was useful for mapping five modifier loci [Fernandes et al., 2016]. Interestingly, each affected organ system was found to have its own set of modifier genes, illustrating the different biology of the affected tissues and the importance of careful phenotyping for human genetic modifier studies in MFS.
Recommendations for management of TAA vary depending on the underlying genetic basis. Patients with MFS or LDS are managed more aggressively [Hiratzka et al., 2010]. Recommendations for aortic root replacement occur at the smallest aortic root diameters for LDS patients, followed by MFS, and then TAA patients with other underlying causes. Evidence for genetic modifiers that could serve to further refine risk stratification in these cohorts would be valuable. With further research identifying the impact of the TGFBR1*6A allele on features associated with MFS, management recommendations may be made. Currently, patients with MFS carrying this allele can continue to be managed according to ACC/AHA guidelines for patients with MFS related thoracic aortic disease [Hiratzka et al., 2010].
ACKNOWLEDGMENTS
This study was supported by the American College of Medical Genetics (ACMG) Summer Scholars Program grant (AES) and the Indiana University Health—Indiana University School of Medicine Strategic Research Initiative and Physician Scientist Initiative (SMW).
