A mutation update on the LDS‐associated genes TGFB2/3 and SMAD2/3

Abstract The Loeys–Dietz syndrome (LDS) is a connective tissue disorder affecting the cardiovascular, skeletal, and ocular system. Most typically, LDS patients present with aortic aneurysms and arterial tortuosity, hypertelorism, and bifid/broad uvula or cleft palate. Initially, mutations in transforming growth factor‐β (TGF‐β) receptors (TGFBR1 and TGFBR2) were described to cause LDS, hereby leading to impaired TGF‐β signaling. More recently, TGF‐β ligands, TGFB2 and TGFB3, as well as intracellular downstream effectors of the TGF‐β pathway, SMAD2 and SMAD3, were shown to be involved in LDS. This emphasizes the role of disturbed TGF‐β signaling in LDS pathogenesis. Since most literature so far has focused on TGFBR1/2, we provide a comprehensive review on the known and some novel TGFB2/3 and SMAD2/3 mutations. For TGFB2 and SMAD3, the clinical manifestations, both of the patients previously described in the literature and our newly reported patients, are summarized in detail. This clearly indicates that LDS concerns a disorder with a broad phenotypical spectrum that is still emerging as more patients will be identified. All mutations described here are present in the corresponding Leiden Open Variant Database.


BACKGROUND
The Loeys-Dietz syndrome (LDS, MIM# 609192, 610168, 613795, 614816, 615582) is an autosomal dominant connective tissue disorder with widespread systemic involvement. In 2005, Loeys and Dietz were the first to describe this disorder, which is-in its most typical presentation-characterized by vascular tortuosity and aneurysm in association with craniofacial and skeletal manifestations (Loeys et al., 2005;Loeys et al., 2006). On the one hand, LDS shows significant clinical overlap with Marfan syndrome (MFS, MIM# 154700), with vascular and skeletal features including aortic root aneurysm, arachnodactyly, scoliosis, and pectus deformity. On the other hand, a clear distinction between LDS and MFS can be made based on typical LDS findings such as widespread aortic/arterial aneurysm and tortuosity, club foot, craniosynostosis, hypertelorism, and bifid/broad uvula or cleft palate. Other LDS-specific features can include cervical spine malformation and/or instability, translucent skin with easy bruising and dystrophic scars, severe allergic tendency, and bowel inflammation including eosinophilic esophagitis/gastritis and/or inflammatory bowel disease. Soon after the initial description, the phenotypic spectrum was expanded to less syndromic presentations, including those that overlap with vascular Ehlers-Danlos syndrome (Loeys et al., 2005;Loeys et al., 2006). Mutations in the genes encoding the transforming growth factor (TGF-) receptor I (TGFBR1) and TGFreceptor II (TGFBR2) subunits were the first reported genetic causes of LDS. More recently, mutations in four additional genes, namely the mothers against decapentaplegic homolog 2 and 3 (SMAD2/3) and the TGF-2 and 3 ligand (TGFB2/3), have been shown to cause an LDS-like phenotype (Bertoli-Avella et al., 2015;Boileau et al., 2012;Lindsay et al., 2012;Micha et al., 2015;van de Laar et al., 2011).
SMAD2 and SMAD3, located on the long arm of chromosome 18 and 15, span about 130 kb and consist of 11 and 9 exons, respectively. Both the SMAD2 and SMAD3 proteins belong to the receptor-activated (R)-SMAD family, intracellular effectors of the canonical TGF-signaling pathway. The activating ligands of this pathway include TGF-2 and TGF-3, which are encoded by the TGFB2 and TGFB3 genes. These genes include 8 and 7 exons and are positioned on the long arm of chromosome 1 and 14, respectively. Including TGFBR1/2, all genes related to LDS spectrum disorders are part of the TGF-signaling pathway.
Additionally, fibrillin-1, which is encoded by the FBN1 gene and deficient in MFS, binds to the large latent TGF-complex and contributes to TGF-bioavailability and activation (Chaudhry et al., 2007;Isogai et al., 2003;Neptune et al., 2003). This clearly highlights the key role of the TGF-pathway in the pathogenesis of aortic aneurysm development in LDS and LDS-like disorders. However, the exact pathogenic mechanisms remain controversial.
Initially, it was suggested that two types of LDS could be distinguished: patients with LDS type 1 (LDSI) displayed typical craniofacial features, where LDS type 2 (LDSII) patients had more pronounced cutaneous features. Recently, this classification was revised and these LDS types are now thought to be part of a phenotypic spectrum of disease. Therefore, LDS type 1 (LDS1) and LDS type 2 (LDS2) now designate to the disease-responsible genes; TGFBR1 and TGFBR2, respectively. Mutations in SMAD3 were initially described as the genetic cause of aneurysms-osteoarthritis syndrome, but because of the many overlapping clinical features with LDS, including hypertelorism, bifid uvula, arterial tortuosity, and widespread and aggressive aneurysms, it is now also classified as LDS type 3 (LDS3) (MacCarrick et al., 2014). Similarly, patients with mutations in TGFB2 share clinical manifestations with LDS, and are for this reason now diagnosed with LDS type 4 (LDS4). More recently, patients with mutations in SMAD2 and TGFB3 (LDS5) were also shown to present with LDS-like features (Bertoli-Avella et al., 2015;Micha et al., 2015). It is thus suggested that a mutation in any of these six genes in combination with the presence of arterial aneurysm or dissection should be sufficient for the diagnosis of LDS (MacCarrick et al., 2014). Since most literature so far has focused on LDS1 and LDS2, this review will highlight known and novel SMAD2/3 and TGFB2/3 mutations and will summarize and compare the clinical features of affected individuals with those reported in the literature.

VARIANTS
Each published mutation was checked for accuracy and compared to the respective wild-type reference sequence. When a different reference sequence was used, nucleotide and codon numbers were converted so their annotation matched with reference tran- To obtain an indication on the pathogenicity, all variants are classified according to the ACMG guidelines (Suppl . Table S1) (Richards et al., 2015).
For TGFB3, we identified four additional probands with a mutation that was previously published and four new mutations. skeletal muscle, and aortic pathology of fibrillin-1-deficient mice (Cohn et al., 2007;Habashi et al., 2006;Neptune et al., 2003;Ng et al., 2004), indicating that fibrillin-1 is not only a structural component of the ECM but also a key regulator of TGF-signaling (Dallas, Miyazono, Skerry, Mundy, & Bonewald, 1995;Isogai et al., 2003). This new hypothesis was confirmed by a key experiment in which the mutant phenotype in fibrillin-1-deficient mice could be attenuated by the administration of TGF--neutralizing antibodies (Cohn et al., 2007;Habashi et al., 2006;Neptune et al., 2003;Ng et al., 2004). Shortly after this, the central role of TGF-dysregulation was proven by the identification of mutations in the TGFBR1 and TGFBR2 genes as the cause of LDS.
The effect of these mutations on the TGF-pathway is, however, complex as heterozygous loss-of-function mutations associate with paradoxical activation of TGF-signaling, which was demonstrated by increased phosphorylation of Smad proteins and increased output of TGF--driven genes.

TGF-pathway
TGF-is the prototype of a family of secreted polypeptide growth factors essential in development, differentiation, cell growth, migration, apoptosis, and ECM production (Derynck & Akhurst, 2007;Massague, Blain, & Lo, 2000). In humans, three TGF-ligand isoforms exist, TGF-1, TGF-2, and TGF-3, encoded by the TGFB1,  origins respond in different ways to TGF-stimulation, it is likely that cells of one lineage are more prone to perturbed TGF-signaling compared to cells of another origin (Topouzis & Majesky, 1996). Cell types that are more sensitive towards heterozygous LDS mutations might attempt to compensate for the initial loss in TGF-signaling by secreting excessive amounts of TGF-ligand in their direct envi-ronment, leading to overdrive of TGF-signaling in neighboring cells which are intrinsically less vulnerable to LDS mutations (Gallo et al., 2014). Finally, since increased expression of pSMAD2 is observed in the aortic media of LDS patients, this might indicate that other TGFrelated pathways such as angiotensin II or activin signaling cascades are involved as well (Bernard, 2004;Rodriguez-Vita et al., 2005).

F I G U R E 2
Schematic representation of the TGFB2 and TGFB3 gene with their protein coding domains. Boxes represent exons 1-8 and 1-7, respectively. On the left side of the schematic are the previously reported mutations, whereas on the right side mutations identified in this study are described. Mutations are annotated at the protein level (reference transcript: NM_001135599.2 for TGFB2 and NM_003239.3 for TGFB3)

Mouse models
For SMAD3, TGFB2, and TGFB3, multiple mouse models have been developed through the years confirming the hypotheses about the molecular and cellular mechanisms underlying LDS pathogenesis.
However, initially, these mouse models were not reported to develop a vascular phenotype (Bonniaud et al., 2004;Li et al., 2009). After identification of SMAD3 as an aortic aneurysm disease causing gene, the vascular phenotype of the Smad3 knock-out mice was studied in more detail (Ye et al., 2013). Necropsy of the Smad3 −/− mice revealed that the majority died from a ruptured aneurysm. Additionally, ultrasound imaging of these mutant mice revealed progressive aortic root and ascending aortic dilation as early as 2 months.
Sanford et al reported that Tgfb2-null mice died shortly after birth and had small, thin walled ascending aortas in addition to other developmental defects involving different organ systems (Sanford et al., 1997). In a later study, Bartram et al. (2001) also observed that Tgfb2-null mice die during gestation due to congenital heart disease and display aortic anomalies. Therefore, Lindsay et al. studied the Tgfb2 heterozygous knockout mouse model to study aneurysm formation in more detail. Tgfb2 +/− mice showed dilatations of the aortic annulus and aortic root, a pattern similar to that of LDS and MFS patients (Loeys et al., 2005;Mc Kusick, 1955), confirming that loss-of-function of one Tgfb2 allele is sufficient to cause aortic root aneurysm. As Western blot of Tgfb2 +/− mouse aortas showed a similar increase in phosphorylation of Smad2/3 and Erk1/2, recapitulating the observations in the human aorta, it can be concluded that increased activation of TGFsignaling underlies the observed phenotype (Lindsay et al., 2012).
Besides Tgfb2 +/mice, the aortas of double-heterozygous knock-out mice, Tgfb2 +/− ; Fbn1 C+/C1039G , were examined as well. The aortic root dimensions of Tgfb2 +/− ; Fbn1 C+/C1039G were significantly increased compared with the wild-type mice at 2 and 4 months of age. Also elastic fiber fragmentation, higher collagen deposition and increased nuclear accumulation of pSmad2 in the aortic media were observed.
Tgfb1 mRNA levels in the proximal aorta of Tgfb2 +/− ; Fbn1 +/C1039G mice at 2 months of age were elevated, which is similar to the observations in human. This is again underlining the paradoxical TGF-pathway activation and stressing the complex pathogenesis that goes together with TGF-dysregulation.

CLINICAL AND DIAGNOSTIC RELEVANCE
The most important clinical finding in LDS patients is dilatation of the aortic root at the level of the sinuses of Valsalva, a feature that nearly all LDS patients will develop ultimately. Aneurysms of the ascending or descending aorta are less frequently observed. Dissection and rupture of these aortic aneurysms tend to occur at a younger age and at smaller diameters in LDS patients compared to MFS patients . The aortic phenotype in patients with SMAD3 mutations is very similar to TGFBR1/2 patients, whereas TGFB2 and TGFB3 cardiovascular features tend to be milder, although severe aortic presentation at young age has also been observed. Non-penetrance seems more common in TGFB2/3 families. Using CT or MRI, 3D reconstruction of images from the head to pelvis is needed to identify arterial tortuosity, present in most individuals with a TGFBR1/2, TGFB2, or SMAD3 mutation, and aneurysms in the rest of the arterial tree. This is

FUTURE PROSPECTS
The recent identification of SMAD2,3 and TGFB2,3 as disease causing genes responsible for LDS phenotypes further pinpoints altered TGFsignaling as the culprit in aortic aneurysm pathology. As we anticipate that more genes will cause similar clinical LDS phenotypes, there is a clear rationale for the development of gene panels. This makes it possible to screen multiple candidate genes at one go, taking advantage of the next-generation sequencing techniques, which are more time and cost efficient compared with the elaborate Sanger sequencing technique. As shown by recent studies, this will facilitate the assessment of an accurate genetic diagnosis for LDS patients hereby surely benefiting patient management (Campens et al., 2015;Proost et al., 2015).
Cardiovascular manifestations of LDS can be managed in a medical and/or surgical treatment strategy. Since vascular disease is more aggressive in LDS compared with MFS, prophylactic aortic surgery is already being recommended at smaller aortic root dimensions of 4.0-4.5 cm (MacCarrick et al., 2014). In order to reduce aortic wall-shear stress and eventually aortic dilatation, beta-blockers, such as atenolol, have been the first-line treatment (Shores, Berger, Murphy, & Pyeritz, 1994). However, since aortic-root tissue of LDS patients shows excessive TGF-activation and signaling, therapies that reduce TGF-pathway activation offer attractive therapeutic targets. Indeed, angiotensin II type 1 receptor blockers, such as losartan, reduce the rate of aortic root growth by lowering the expression of TGF ligands, receptors, and activators (Everett, Tufro-McReddie, Fisher, & Gomez, 1994;Fukuda et al., 2000;Habashi et al., 2006;Naito et al., 2004). Although a randomized clinical trial comparing losartan with atenolol did not show a significant difference in the rate of aortic-root dilatation between these two treatment groups, the observation that aortic-root z-scores decrease especially in younger patients advocates to start therapy earlier in the disease course (Lacro et al., 2014). Alternatively, the use of ERK inhibitors (RDEA119) and angiotensin II type 2 receptor agonists offer interesting alternative treatment options (Loeys, 2015).
The full spectrum of phenotypes associated with some genes of