Heterozygous TGFBR2 mutations in Marfan syndrome

Abstract

Marfan syndrome is an extracellular matrix disorder with cardinal manifestations in the eye, skeleton and cardiovascular systems associated with defects in the gene encoding fibrillin (FBN1) at 15q21.1 (ref. 1). A second type of the disorder (Marfan syndrome type 2; OMIM 154705) is associated with a second locus, MFS2, at 3p25-p24.2 in a large French family (family MS1). Identification of a 3p24.1 chromosomal breakpoint disrupting the gene encoding TGF-beta receptor 2 (TGFBR2) in a Japanese individual with Marfan syndrome led us to consider TGFBR2 as the gene underlying association with Marfan syndrome at the MSF2 locus. The mutation 1524G-->A in TGFBR2 (causing the synonymous amino acid substitution Q508Q) resulted in abnormal splicing and segregated with MFS2 in family MS1. We identified three other missense mutations in four unrelated probands, which led to loss of function of TGF-beta signaling activity on extracellular matrix formation. These results show that heterozygous mutations in TGFBR2, a putative tumor-suppressor gene implicated in several malignancies, are also associated with inherited connective-tissue disorders.

Figure 1. TGFBR2 was isolated from the 3p24.1 breakpoint of a Japanese patient with complex chromosomal abnormalities

The 3p24.1 breakpoint was analyzed because the MFS2 locus was also mapped to 3p24.2-p25. In the patient, a chromosomal segment 3p24.1-p 14.2 was inserted into 3q11.2. FISH analysis confirmed that RP11-775G14 spans the 3p24.1 breakpoint. A summarized physical map covering the 3p24.1 breakpoint is indicated. Horizontal bars show BAC and cosmid clones: green depicts clone spanning the breakpoint, and arrow shows gene. TGFBR2 exons are indicated as purple squares. Cosmid subclones, c1C and c3G spanned the breakpoint. c1C contains 35.2-kb genomic region of TGFBR2 (through intron 3 to the end of 3′ UTR) out of the 39.4 kb insert-DNA according to its end sequence information. The gene must have been disrupted at the breakpoint because signals on the 3q would have not been seen if the breakpoint existed outside TGFBR2.

Figure 2. Haplotype analysis of a large French family (MS1) and a mutation causing abnormal splicing

(a) Pedigree of family MS1 and segregation of 3p24.1 markers in members affected with MFS (filled symbol), having an ambiguous status (gray symbol) or unaffected (open symbol). Orange symbol indicates member with a mutation, 1524G>A (Q508Q), and green, member without the mutation. The haplotype co-segregating with the disease is shown in red. (b) Normal (green) and abnormal splicing (orange) caused by 1524G>A (Q508Q). The consensus value on splicing was modified from 95.6 to 82.3 by 1524G>A and another good consensus sequence (AGgt xxg, value 90.5) was located 23 bp away in the intronic sequence. In the mutated allele, 23 nucleotides of intron 6 were newly added after exon 6 and connected to regular exon 7 as predicted, resulting in a premature stop codon at amino acid 525.

Figure 3. Genomic structure of TGFBR2 and mutations found in MFS families

TGFBR2 consists of seven exons. Square indicates exon. A transmembrane domain, a kinase domain, and UTRs are shown in light blue, pink, and gray, respectively. (A) 10 and (GT)3 are genomic instability sites. Three other missense mutations, 923T>C (L308P), 1346C>T (S449), and 1690C>T (R537C) are found in families MS57, MS382, and MS587, respectively. Symbols (filled, open, green, and orange) are depicted in the same manner as in Fig. 2. Each mutation occurred at an amino acid that is evolutionally conserved or chemically similar (red square) in a kinase domain among mouse and rat Tgfbr2, mouse and fly Acvr2, and nematode daf genes. Multiple sequence alignment was done using the web-based software, CLUSTALW (http://clustalw.genome.ad.jp/).

Figure 4. Impairment of TGF-β signaling activity by TGFBR2 missense mutations found in patients with MFS

(a) Relative luciferase activity (RLA) of pTARE-Luc cis-reporter in HEK 293 cells after transient transfection with various TGFBR2 constructs was calculated by normalization using the activity of a co-transfected control vector, pRL-TK, containing distinguishable R.. reniformis luciferase. The pCIS-CK negative control plasmid containing no inducible cis-enhancer element was also used to evaluate whether effects are TGF-β signaling-specific. Data represent mean + SD. Basal activity only by pTARE-Luc cis-reporter is probably due to endogenous TGF-β1 in HEK 293. Transfection with wild-type (WT) TGFBR2 cDNA showed a significant high value (about 15) of RLA, while that with a truncated mutant, δ cyt, lacking the kinase domain, showed significantly low RLA. Transfection with other missense-type mutants, L308P, S449F, and R537C, also showed significantly low RLA, similar to that with δ cyt. *: Statistically different (P<0.0001) by Post-hoc test. (b) RLA of pT3P-Lux reporter in DR-26 (a Tgfbr2 negative cell line) after transient transfection. Mv1Lu (a Tgfbr2 positive cell line) was also used to show normal response of the pT3P-Lux reporter to TGF-J31. Transfection with wild-type (WT) TGFBR2 cDNA showed a significant high value (about 20) of RLA, while that with an empty vector [pcDNA3.1(−)] and the truncated mutant, δ cyt, showed significantly low RLA. Transfection with other missense-type mutants, L308P, S449F, and R537C, also showed significantly low RLA, similar to that with δ cyt. Co-transfection in DR-26 of WT and R537C with different ratios showed serial increase of RLA in parallel with WT cDNA quantity, suggesting that the mutation was less likely to cause a dominant negative effect. *: Statistically different (P<0.0001) by Post-hoc test.