Response to “De novo mutation of the <i>TGFB3</i> latency‐associated peptide domain in a patient with overgrowth and Loeys–Dietz syndrome features”

The correspondence from Matyas et al. [2014] is intriguing. The authors have identified a mutation [c.899G>A, p.Arg300Gln] in the portion of the TGFB3 gene coding for the cleavage recognition site between TGFB3 propeptide or latency-associate peptide (LAP) and the mature TGFB3 ligand. This mutation alters the second arginine in the highly conserved cleavage site, RKKR. Though the authors do not provide evidence, in this instance, the LAP likely envelops the ligand normally, but free ligand cannot be liberated; hence, predicted production of mature ligand would be reduced by 50% and the mutation could be considered hypomorphic, though expressivity and penetrance might vary as a function of differing proteases among diverse tissues. The mutation was presumably identified by candidate gene analysis that included several of the more commonly interrogated genes mutated in patients with Marfanoid features. Though their genetic analysis does not capture all the variation in the exome, it is entirely believable that the phenotype is largely owing to the TGFB3 mutation. The authors make the point that this patient shared many clinical features with the first case report of a TGFB3 coding mutation [Rienhoff et al., 2013] but was also phenotypically distinct in some potentially clinically significant ways: the Maytas et al. patient showed skeletal overgrowth, her musculature was essentially normal, and all joints were hyperextensible with no evidence of contracture. The aortic root dimensions were toward the upper limit of normal, which must be taken at face value: there was no evidence of vascular disease. Could these differences simply be attributed to different TGFB3 concentrations with this new mutation reducing TGFB3 levels by a lesser degree than the dominant negative TGFB3 mutation reported by Rienhoff et al. [2013]? The Maytas et al. case is provocative because it begs a question beyond the scope of their report: Why are there similar clinical findings among these various allied syndromes—Camurati Engelman, Marfan, Loeys–Dietz, Beals-Hecht—when some mutations are thought to be hypermorphic and others hypomorphic? Is it possible that the newly reported mutation, as well as the many other mutations affecting proteins in the TGFB pathway, have both effects in a body, spatially distinct and determined at the level of each affected tissue? There are many examples in developmental biology of TGFB signaling having opposite effects depending on, for example, intracellular signaling pathways, competing ligands, local processing proteases, or extracellular matrix compositions. This case coupled with the previous TGFB3 report [Rienhoff et al., 2013] suggests that loss-of-function mutations reducing activity or production of TGFB3 ligand do not adversely affect the cardiovascular system at a young age, unlike Loeys–Dietz syndrome. But TGFB3 mutations also have clinically distinct effects for reasons unknown. That is a call to action among those who study TGFB. The point of reporting these individual cases is to present the conundrums these mutations pose—there is not a comprehensive molecular understanding of TGFB biology accounting for the contradictory phenotype–genotype correlations. The Maytas et al. case represents a wonderful example of how human mutations can show us what we don't know, when we thought we did.