Thoracic aortic aneurysm or dissection (GMS)

Gene: SMAD6

Red List (low evidence)

The rating of this gene has been updated to red following NHS Genomic Medicine Service approval.

Created: 11 Dec 2025, 5:14 p.m. | Last Modified: 11 Dec 2025, 5:14 p.m.

Panel Version: 4.4

Red List (low evidence)

Based on the review from Ian Berry and further personal communication with him regarding publications, it would appear that SMAD6 variants are not relevant to this panel.

Created: 15 Apr 2025, 9:48 a.m. | Last Modified: 15 Apr 2025, 9:48 a.m.

Panel Version: 3.20

Publications

• 39069920
• 40133303

Red List (low evidence)

SMAD6 has no constraint in gnomAD, and missense + LOF variants are far more common in the gene than would be expected by chance (which may explain the enrichment of rare SMAD6 variants in various cohorts including aortic phenotypes).

Summary of papers reviewed:
o Tan 2012 (PubMed: 22275001). Two missenses in a cohort of 450 patients in the MH2 domain, affecting p.Cys484 (12 missenses at this residue in gnomAD v4) and p.Pro415Leu (18 hets in gnomAD v4; Pro415 has a total of 34 hets mutating this residue in gnomAD v4). There is no segregation nor functional evidence noted.

o Luyckx et al. 2019 ([PubMed: 30796334). 473 non-syndromic TAAD patients tested. Rare SMAD6 variants were detected in 6 families. There is a mix of missense and truncating variants indicated, majority of the missenses are in domain MH1 (in contrast to MH2 in the other paper above). The only family with any segregation data is Family 1 – however, their phenotypes are not unifying in any obvious way (Note that only 3 of the 6 have any obvious aortic abnormality, and the phenotypes described are quite variable), they also have a deletion encompassing SMAD6 and part of SMAD3 (which IS an established functional disease gene) so any association may be due to joint LOF of both genes in some way. Some families listed appear to have dominant and some have recessive genotypes.
The paper includes no functional data.


o Gillis et al. 2017 (PMID: 28659821). 441 individuals in a BAV/TAA cohort. 2 nonsense, 2 frameshift, 7 in-frame/missense variants identified. Both this and the Lucyckx 2019 paper have Bart Loeys and a strong Dutch component, so some cases/families may be duplicated. The rare variants identified again are found across the gene (MH1, MH2 and neither domain). In two families, some relative testing was done (5 individuals in total), none of the 5 variant carriers had BAV/TAA (3 of whom were in later middle-age) suggesting either significant reduced penetrance or that the SMAD6 findings were due to random chance. There was no functional evidence.

o Strikingly, since 2019 there are no other papers linking SMAD6 to dominant FTAA in HGMDpro. Note that the only two variants associated with aortopathy/BAV and reported in more than one case in HGMD (c.42G>A and c.465_471del) have been reported across a range of unrelated phenotypes (incl. radioulnar stenosis and ID/autism). Both are also seen in gnomAD v4 (albeit at low levels). Lack of further convincing cases or association from 100K data etc suggests that the association is tenuous/unproven, or very low penetrance/risk factor in dominant form.


o A more recent review from the 2022 from the 2019 authors (again including Bart Loeys; Luyckx et al. 2022, PMID: 36414630) states that “Intriguingly, a similar spectrum of heterozygous loss-of-function variants has been reported to cause these clinically distinct disorders without a genotype-phenotype correlation. Even identical nucleotide changes have been described in patients with either a cardiovascular phenotype, craniosynostosis or radioulnar synostosis. These findings suggest that the primary pathogenic variant alone cannot explain the resultant patient phenotype.” - this casts further doubt on the dominant association, given the breadth of unrelated and non-overlapping phenotypes reported.

• There is no convincing animal model of BAV/TAA that I can see from the papers or from OMIM. The 2022 Luyckx paper summarises the three published mouse models, but these look at complete knockout (e.g. lacking in SMAD6 entirely – recessive) animals, these show early lethality but not clearly recapitulate the aortic or skeletal phenotypes theorised in relation to this gene in humans. One paper describes a range of severe cardiac defects in these mice but the phenotype and genotype are not a clear or useable analog for this association.

• Outside of OMIM, it is not recognised by any other gene-disease curation resource on DECIPHER. Gen2Phen lists it as a craniosynostosis gene with “limited” evidence. It would not satisfy the ClinGen criteria at any level for a dominant association with aortic phenotypes.

• It is green on the R125 panel on PanelApp, but the only green review (Kate Thomson) states: “This gene did not achieve a consensus Green rating; however, the group agreed that the existing evidence (published and in-house data) was sufficient to support inclusion in this panel.”

My impression is that no variant in this gene can be clearly classified as pathogenic or likely pathogenic for BAV/aortopathy as the association and penetrance remain unclear/unproven. PVS1 cannot be used for LOF variants due to the high proportion of LOF variants in gnomAD, lack of segregation in any cases, and lack of an animal model. None of the reported missense variants are highly recurrent in aortopathy cohorts nor have compelling segregation evidence.

I would propose demotion of this gene to Amber for these reasons.

Created: 24 Jan 2025, 4:58 p.m. | Last Modified: 24 Jan 2025, 4:58 p.m.

Panel Version: 3.19

Mode of inheritance
MONOALLELIC, autosomal or pseudoautosomal, NOT imprinted

Publications

• PMID:22275001
• 30796334
• 28659821
• 36414630)

Green List (high evidence)

Submitted on behalf of the GMS Cardiology specialist group. This gene did not achieve a consensus Green rating; however, the group agreed that the existing evidence (published and in-house data) was sufficient to support inclusion in this panel.

Created: 9 Dec 2019, 1:19 p.m. | Last Modified: 9 Dec 2019, 1:19 p.m.

Panel Version: 0.55

I don't know

Submitted on behalf of the GMS Cardiology specialist group. The group has agreed that this gene should be Amber on this panel. Therefore this gene has been promoted from Red to Amber.

Created: 18 Nov 2019, 4:33 p.m. | Last Modified: 18 Nov 2019, 4:33 p.m.

Panel Version: 0.35

I don't know

Literature associates SMAD6 with aortic valve disease and aneurysm.

Created: 2 Oct 2019, 3:17 p.m. | Last Modified: 2 Oct 2019, 3:17 p.m.

Panel Version: 0.32

Mode of inheritance
MONOALLELIC, autosomal or pseudoautosomal, imprinted status unknown

Phenotypes
Aortic valve disease 2 614823

Publications

• 30796334
• 28659821
• 30963242

I don't know

This gene was part of an initial gene list collated by Matthew Edwards Royal Brompton Hospital sent 16th Jan 2019 on behalf of the London South GLH for review by the GMS Cardiology Specialist Group. Only gene symbol from the Royal Brompton gene panel was provided - suggested initial gene rating and evidence for inclusion not provided with the list.

Created: 20 Feb 2019, 2:17 p.m.