My Best Guess

The hypothesis begins with the Marfan Syndrome (MFS MIM #154700), a disease affecting the connective tissue including the skeleton, arterial wall tissue and other elastic tissues [].   The disease is well known among clinical geneticists and to the public. The appearance of these patients is often striking: they are unusually tall, have long arms and long fingers.  They often have narrow long faces, crowded teeth with narrow, high-arched palates and can be very myopic because of dislocated lenses in the eyes.  Those with MFS are at great risk for developing aortic root aneurysms and need to be followed by echocardiography closely.  Patients with MFS are treated with beta adrenergic antagonists (beta blockers) to reduce stress on the aortic; this has been shown to delay the onset of severe aortic disease and postpone the need for surgical treatment [].

The genetic cause of MFS was identified in 1991; mutations were found in the fibrillin 1 gene []. The fibrillin 1 protein complexes with itself and other proteins (e.g collagen VI, fibrillin 2) to form a protein structure called the extracellular microfibril that associates at the margins of maturing elastic fibers forming part of a matrix during development of the embryo.  Fibrillin 1 and presumably all other proteins in the complex are needed to one degree or another to ensure that following birth, elastic fibers and non-elastic tissues are properly maintained. 

One function of the extracellular microfibril is to serve as a reservoir for an important family of growth factors called the Transforming Growth Factor beta (TGFbeta) family that regulate the growth and differentiation of cells.  The current understanding is that at least 5 of the 33 members of this family of growth factors are held in an inactive or latent state bound to the microfibrillar matrix [].  These include TGFbeta 1, 2, 3, GDF8 and GDF11 [].  When appropriate, these hormones are liberated by specific proteases and bind to specific receptors.  In this manner these powerful growth factors can be held inactive locally and released near their target cell only when appropriate.

Some of the clinical features -- importantly the vascular disease -- of MFS can be ascribed to activation of the TGFb pathway.  There are two lines of evidence supporting this.  First, a mouse model of Marfan syndrome – a mouse carrying one mutant allele of the fibrillin 1 gene and a wild-type fibrillin 1 – has many of the skeletal and other physical features of the human disease.  The vascular disease can be mitigated by treatment with an anti-TGFbeta antibody [].  A second line of evidence is the recent description of a new syndrome – Loeys Dietz - that presents with many Marfanoid features (MIM  #609192); Loeys Dietz Syndrome (LDS) is associated with mutations in the TGFbeta receptors TGFBRI and TGFBRII [].  This biochemical and genetic data support the view that some of the important pathophysiology of MFS and LDS involves the inappropriate activation of the TGFbeta signaling pathway []. 

The important details are such: fibrillin 1 binds a protein called latent TGFbeta-binding protein 1 one of whose properties is to bind a latent form of TGFbeta. [].  When the microfibrillar matrix is ill composed or structurally abnormal, concentrations of active TGFbeta rise [].  Excess TGFbeta binds to and activates to a greater degree than normal the TGFbeta receptors and its downstream intracellular signaling molecules, most importantly SMAD2/3 [].  This pathway activation has many different effects depending on the cells that are activated; for example, TGFbeta induces immune tolerance by regulating T-lymphocyte proliferation, differentiation and survival while under some circumstances other cells are inhibited from proliferating [].        

The tissue distribution and specific protein composition of the microfibril and surrounding matrix probably account for the different phenotypes seen with changes in the TGFbeta receptor genes.  One can infer from MFS and LDS that there are severe disturbances in the vascular tree beginning with the aortic root.  In the case of Beals Syndrome, associated with mutations in fibrillin 2, the disturbance would appear generally confined to the skeleton and the musculature.

Though LDS shares many features with MFS, notably the Marfanoid habitus, LDS is distinguished in three ways from MFS.  Patients with LDS have more extensive and more severe vascular disease than those with MFS accounting for the average age of death around 27 compared to MFS where the average age of death has historically been later often in the 4^th or 5^th decade[].  In addition, LDS patients typically have hypertelorism and a bifid uvula, clinical findings rarely found in MFS[].  

The TGFBRI and II mutations reported in LDS alter in some unknown way the balance of receptor activation of downstream targets such that there is an apparent activation of the TGFbeta pathway similar with that observed in the MFS.  SMAD2and SMAD3 are two substrates of the TGFBRI kinase and when phosphorylated, serve as the “second” messenger translocating to the nucleus in a complex with SMAD4 [].  There is evidence that the concentration of phosphorylated SDAD2/3 is elevated in the aortic tissue of patients with LDS [].  Exactly how a defective receptor could result in an activated pathway is the subject of active investigations but a mutation in a protein caveolin 3 involved in regulating the turnover of a TGFbeta-like receptor ActIIR (a.k.a. ACVR), also results in the toxic activation of the TGFBRI pathway [].  In the context of LDS, one interpretation is that if the mutant TGFbeta receptor lingers in an activated state, signaling persists as well.  The dominant nature of the inheritance of LDS indicates a single mutant TGFbeta receptor I or II protein in the heterotetrameric complex is sufficient to cause the phenotype.  It seems most likely that the aberrant signaling is mediated by a TGFbeta receptor I or II complex containing one copy of the wild-type protein bound to a copy of the mutant protein.  Following this model, one of the receptors composing the homodimer can be activated by phosphorylation, e.g. in the case of the Type II mutant/wild-type homodimer by ligand binding.  In the case of a mutant Type I receptor complexed with its wild-type cognate, the wild-type version of the Type I would be expected to be phosphorylated by the Type II complex; this complex is seems capable of phosphorylating its substrate, SMAD2 and SMAD3 but, critically, it is not subject to normal receptor down-regulation.   In mice, a deletion of one copy of either TGFbeta receptor genes has no phenotype making haploinsufficiency in LDS a less likely explanation []. 

A close look at the location of mutations in the TGFBRI and II reported for LDS reveals that they are missense mutations found almost exclusively in the portion of the gene coding for kinase domains of the two receptors [].  Although it is early in the history of this disease with the full spectrum of mutations and the full spectrum of phenotypes yet to be fully described and correlated, it does suggest that the distinguishing aspects of LDS – diffuse, aggressive vascular disease, hypertelorism, and bifid uvula – may be owing solely to mutations in the kinase domains in otherwise intact receptor proteins that can assemble with a wild-type counterpart.  More importantly for this discussion, the dominant and activating nature of the kinase domain mutations on the pathway might serve as a general model for activation in other TGFbeta superfamily receptors.

It was obvious my daughter had many features that could be described as Marfanoid including the pectus deformity, pes planus, lax ligaments, and hyperextensibility of her joints but the additional findings of bifid uvula and hypertelorism strongly suggested the diagnosis of LDS.  But it was difficult to conclude she had LDS because no mutation was found in either TGBFR genes, she thankfully has no evidence of vascular disease in three successive echocardiograms and her chief clinical concern was diminished muscle mass with its attendant weakness, a feature not noted in any of the 80 reported LDS patients[].  This last point seemed worthwhile focusing on as it was the cause of her morbidity and patients with LDS or MFS are not noted to have significant delays in achievement of gross motor milestones as my daughter did.

TGFbeta pathway activation seemed a necessary condition to satisfy for any biochemical or genetic explanation of her problem; more specifically, the ideal hypothesis would involve activation of SMAD2/3 given the bifid uvula and hypertelorism.  The open question was: what could account for hypomyoplasia?  Suspicion fell on other members of the TGFbeta superfamily, principally myostatin.

Myostatin or growth/differentiation factor 8 (GDF8) is a member of the TGFbeta family of secreted proteins [].  Myostatin was named for its principle property of inhibiting muscle growth [].  The protein has been shown to play a key role in regulating muscle mass [].  Myostatin is highly conserved among vertebrates and mutations in the myostatin gene have been reported in a several species of domestic animals with muscular hypertrophy.  A child presenting with extraordinary musculature with two null copies of the myostatin gene established the role of myostatin in regulating muscle mass in humans [].  Myostatin gene expression is largely confined to cells of skeletal-muscle lineage inhibiting the activation of satellite cells, the stem cells of skeletal muscle.  Its expression begins in the early embryo and persists throughout life regulating both hypertrophy and hyperplasia of myocytes thereby influencing both muscle mass and strength []. 

Much interest has focused on the potential to therapeutically regulate muscle mass by targeting myostatin with the goal of modifying the clinical course of muscular dystrophies or acquired diseases with muscle wasting where excess myostatin is thought to play a role [].  This hope is founded on the observation that antibodies directed against myostatin significantly increase muscle mass in normal mice []. Likewise, chronic administration of myostatin to normal mice reduces muscle mass [].  There are no reports of excess myostatin or activation of myostatin signaling in heritable human disease.

Myostatin binds to the activin receptors (ActR or ACVR), receptors very closely related to the TGFbeta receptor.  These receptors, like the TGFBRs, are heteromeric complexes composed of two Type I homodimers and two Type II homodimers.  The Type II protein component of the complete (heterotetramer) receptor mediates extracellular ligand binding while the Type I component specifies the intracellular signaling mediator.  Ligand binds the Type II receptor protein inducing its association with the Type I receptor. The regulation of the receptor, the association of Type I and Type II proteins and the intracellular signaling is mediated by phosphorylation.  Each receptor protein has a serine/threonine kinase activity that is essential for proper signaling.  The control of ligand signaling is further complicated by the existence of several distinct Type I receptor proteins that can associate with Type II receptor proteins.  Indeed, the TGFBR and the activin receptors can share a common Type I receptor, TGFBRI, one of the two genes mutated in LDS.  Myostatin signaling occurs principally through the activin receptor IIB (ActRIIB) or ActRIIA using as a Type I component TGFBRI or Alk4, a Type I receptor closely related to TGFBRI.  There is additional overlap between TGFBR signaling and that of the activin receptor beyond use the same Type I receptor components.  Ligands signaling through either the TGFbeta or myostatin (activin) receptor complex use SMAD2/3 to mediate intracellular signaling.

The activin receptor genes, ActRII, ActRIIB, ActRIB were candidates for mutations in my daughter because all mediate myostatin signaling.  By analogy with LDS, a mutation in the kinase domain of an Activin receptor might be expected to have the same molecular phenotype, namely enhanced intracellular signaling through elevated phosphor-SMAD2/3 leading to inhibition of muscle development. Though there may be considerable overlap, any clinical difference between LDS and a defect in activin receptors might be accounted for on the basis of differences in the cellular and tissue distribution of the two groups of receptors.  The discovery that caveolin 3, the cause of autosomal dominant limb girdle muscular dystrophy, regulated myostatin receptor signaling confirmed indirectly the clinical importance of myostatin.

Thus, I designed the necessary oligonucleotides to sequences these three genes using as my guide the sequencing strategy for the TGFbeta receptor genes.