Loeys–Dietz syndrome is a specific phenotype and not a concomitant of any mutation in a gene involved in TGF-β signaling

To the Editor: The authors of the “narrative review” of Loeys–Dietz syndrome (LDS), “Loeys-Dietz Syndrome: A Primer for Diagnosis and Management,” published in this issue, propose a new nosology for LDS based on genotype and provide their perspectives on the pleiotropic phenotype and management of all affected organ systems.¹ TGFBR2 mutations were initially described in patients diagnosed with Marfan syndrome.² In 2005, Loeys and colleagues³ described severely affected patients with aggressive aortic disease with dissections with minimal enlargement of the aortic root and diffuse arterial tortuosity, coupled with other clinical features (bifid uvula, craniosynostosis, hypertelorism, and clubfoot) due to mutations in TGFBR1 and TGFBR2. This phenotype was labeled LDS. Subsequently, patients with thoracic aortic disease with minimal or no craniofacial dysmorphism were also found to have mutations in TGFBR1 and TGFBR2, as well as in SMAD3 and TGFB2, other genes encoding proteins in the TGF-β pathway. MacCarrick et al.¹ propose to define any patient with a pathogenic variation in any of these four TGF-β pathway genes as having LDS. Furthermore, they suggest that any patient with a mutation in any of these four genes should undergo valve-sparing aortic root surgery with minimal enlargement, criteria previously reserved for patients with the severe LDS phenotype with mutations in TGFBR1 and TGFBR2. The opinion that anyone with a pathogenic variation in any of these four genes qualifies for a diagnosis of LDS has been codified in Online Mendelian Inheritance in Man (http://omim.org/).

In our collective published and unpublished experience, diagnosing and managing all patients with mutations in these genes similarly is inappropriate. Many patients with mutations in TGFBR1, TGFBR2, SMAD3, and TGFB2 do not have a phenotype consistent with LDS nor do they have an aggressive arterial disease.^4,5 Although arterial involvement beyond the aortic root is established for TGFBR1, TGFBR2, and SMAD3, there is no evidence that TGFB2 mutations cause arterial disease to the same extent. Osteoarthritis and neurological features can be associated with mutations in SMAD3 but not with the other genes. The clinical picture also differs considerably between those with TGFBR1 and TGFBR2 mutations.⁵ Moreover, some mutations in genes involved in the TGF-β pathway do not lead to aortic disease (e.g., Camurati–Engelmann disease with mutations in TGFB1). Arterial tortuosity, which has been proposed to be diagnostic for LDS, is observed in patients with mutations in other genes, such as FBN1, FBNL4, PRKG1, and SLC2A10.

Therefore, labeling mutations in all these genes as diagnostic for LDS risks improper management (limitations of exercise, insurance and employment, unnecessary medication, and prophylactic surgery with only mild aortic dilatation), inaccurate reproductive counseling, and psychosocial distress. Furthermore, MacCarrick et al.¹ suggest high-dose angiotensin receptor blockade in all patients, evidence for which is entirely lacking in LDS and still emerging for Marfan syndrome. Finally, labeling a phenotype based solely on mutations in specific genes sets a precedent that conflicts with experience with other Mendelian conditions. To cite but two examples, mutations in FBN1 cause not only Marfan syndrome but at least eight other conditions, some completely distinct from Marfan syndrome (e.g., geleophysic dysplasia, Weill–Marchesani syndrome), and mutations in TGFBR1 cause multiple self-healing squamous epithelioma syndrome, a phenotype completely distinct from LDS. In addition to such intragenic variability, we are well aware of the possibility of intrafamilial variability, which emphasizes the utility of biomarkers that might help tailor therapy and indications for vascular surgery. Clinical features of LDS and arterial tortuosity may well be such markers.⁶

We recently proposed an alternative classification scheme for heritable thoracic aortic diseases.⁷ This nosology is based on the underlying gene but recognizes that diverse aortic phenotypes can arise from the same locus. In addition to classic heritable disorders of connective tissue (e.g., vascular Ehlers–Danlos syndrome, Marfan syndrome) and heritable thoracic aortic diseases due to mutations in genes for proteins involved in smooth muscle contraction (e.g., ACTA2, MYH11, MYLK, PRKG1), another category is genes specifying components of the TGF-β signaling pathway (e.g., TGFBR1, TGFBR 2, TGFB2, SMAD3). This latter category also encompasses a variety of heritable thoracic aortic diseases, including LDS.

As MacCarrick et al.¹ note, they have published a “narrative review,” which is “based on a systematic literature review, expert opinion, and standard-of-care practices from the center with the largest patient population of these disorders in the world.” If “expert opinion” is drawn principally from a center (albeit a great academic one) to which are referred the most severe and difficult cases, then the perspectives of those experts may be biased. Gene-specific clinical data from multiple institutions should be collected and analyzed to define vascular treatment guidelines for all heritable thoracic aortic disease genes. Based on our experience with diagnosing, counseling, and caring for patients and families with aortic diseases from general international populations, we respectfully disagree with lumping together as having LDS all people with mutations in TGFBR1, TGFBR2, SMAD3, and TGFB2, along with “other TGF-β signaling pathway genes yet to be associated with disease” for which the clinical characteristics have yet to be defined.

Disclosure

The authors declare no conflict of interest.