|Home | About | Journals | Submit | Contact Us | Français|
In this review, we focus on elucidating the cardiac function of germline mutations in the PTPN11 gene, encoding the SH2 domain-containing protein tyrosine phosphatase SHP2. PTPN11 mutations cause LEOPARD Syndrome (LS) and Noonan Syndrome (NS), two disorders that are part of a newly classified family of autosomal dominant syndromes termed “RASopathies,” which are caused by germline mutations in components of the RAS/RAF/MEK/ERK MAPK pathway. LS and NS mutants have opposing biochemical properties, and yet, in patients, these mutations produce similar cardiac abnormalities. Precisely how LS and NS mutations lead to such similar disease etiology remains largely unknown. Recent complementary in-vitro, ex-vivo and in-vivo analyses reveal new insights into the functions of SHP2 in normal and pathological cardiac development. These findings also reveal the need for individualized therapeutic approaches in the treatment of patients with LS and NS, and more broadly, patients with the other “RASopathy” gene mutations as well.
Congenital heart disorders (CHDs) are the most common type of birth defect (~1/100 live births) and the major cause of birth-related deaths (Weismann and Gelb 2007). Abnormalities in signaling molecules and/or pathways are implicated in CHD pathogenesis; however, underlying mechanisms remain poorly understood and/or unknown. Recently, a new family of autosomal dominant syndromes was recognized, termed “RASopathies” (Figure 1). These disorders, which include LEOPARD Syndrome (LS) (OMIM: 151100) and Noonan Syndrome (NS) (OMIM: 163950), are caused by germline mutations in components of the RAS/RAF/MEK/ERK mitogen activating protein kinase (MAPK) pathway (Tidyman and Rauen 2009), which is required for normal cell growth, differentiation, and survival. Aberrant regulation of this pathway has profound effects, particularly on cardiac development, resulting in various abnormalities, including valvuloseptal defects and/or hypertrophic cardiomyopathy (HCM). With perturbations of the MAPK signaling pathway established as central to RASopathy disorders, several candidate genes along this canonical pathway have been identified in humans with RASopathy disease phenotypes, including mutations in KRAS, NRAS, SOS1, RAF1, BRAF, MEK1, MEK2, SHOC2, and CBL (Carta et al. 2006; Cirstea et al. 2010; Cordeddu et al. 2009; Dentici et al. 2009; Martinelli et al. 2010; Niihori et al. 2006; Pandit et al. 2007; Razzaque et al. 2007; Roberts et al. 2007; Schubbert et al. 2006; Tartaglia et al. 2007) (Figure 1). The gene most commonly mutated in NS and LS is PTPN11 (Figure 1) (Tartaglia et al. 2001).
PTPN11, encoding the Src homology-2 (SH2) domain-containing non-transmembrane protein tyrosine phosphatase (PTP) SHP2. SHP2 is a ubiquitously expressed protein that contains two SH2 domains, a central PTP catalytic domain and a C-terminal tail with two tyrosine phosphorylation sites and a proline-rich motif. Resolution of the crystal structure, along with biochemical validation, has elucidated its mechanism of regulation, whereby in the inactive state, the backside loop of the N-SH2 domain folds into the PTP catalytic pocket rendering SHP2 structurally and biochemically inactive (Hof et al. 1998). Binding of phosphotyrosyl (pY) peptides to the SH2 domains induces a conformational change, unfolding the structure and thereby allowing substrates access to the catalytic pocket (Neel et al. 2003). The interaction of SHP2 with pY proteins, such as receptor tyrosine kinases (RTK), cytokine receptors and scaffolding adaptors (e.g. insulin receptor substrate 1 (IRS1), GRB2-associated binder 1 (GAB1)), mediates signaling events that control proliferation, differentiation and apoptosis within the cell.
SHP2’s regulatory function in intracellular signaling is complicated. Genetic and biochemical evidence well demonstrate that SHP2 is a positive regulator of intracellular signaling pathways and is required for ERK/MAPK pathway activation by most, if not all, RTKs, as well as by cytokine receptors and integrins (Feng 1999; Neel et al. 2003; Tonks and Neel 2001). SHP2 binds directly to some RTKs or, more often, to scaffolds, and becomes activated. Cells expressing dominant negative SHP2 (Noguchi et al. 1994) or Ptpn11 Exon 3 deleted (Ex3−/−) mouse embryonic fibroblasts (Shi et al. 2000) exhibit defective RAS activation, suggesting that SHP2 acts upstream of RAS. However, early in vitro experiments showed that a catalytically inactive mutant of SHP2 could perturb (some) components of downstream signaling, even in the presence of a constitutively active RAS, suggesting that SHP2 may also function either downstream and/or in parallel to RAS (Yamauchi et al. 1995). As such, despite extensive study, the precise mechanisms by which SHP2 mediates ERK/MAPK activation remain controversial. SHP2 has been implicated in RTK-evoked src family kinases (SFK) activation, with SFKs, in turn, required for sustained RAS activation (Zhang et al. 2004). Other suggested targets include the RASGap binding sites on some RTKs (Agazie and Hayman 2003; Klinghoffer and Kazlauskas 1995), the Shp Substrate 1 (SHPS-1) glycoprotein (Fujioka et al. 1996; Kontaridis et al. 2001) and the regulatory pY sites on Sprouty proteins, a family of RAS inhibitors (Hanafusa et al. 2004).
Actions of SHP2 on other pathways have also been demonstrated to play a key role in functional regulation. JAK/STAT (Neel et al. 2003), NF-kB (Feng 1999), and RHOA (Kontaridis et al. 2004) signaling have all been proposed. In addition, SHP2 can also regulate activation of the phosphoinositide 3-kinase (PI3K)/AKT pathway in a growth factor-specific manner. In response to epidermal growth factor (EGF), SHP2 negatively regulates PI3K/AKT activation, likely through binding and dephosphorylation of the GAB1 p85 binding sites (Zhang et al. 2002). In contrast, in response to other growth factors such as the platelet-derived growth factor (PDGF) and insulin-like growth factor (IGF), SHP2 evokes a positive regulation on PI3K/AKT activation (Zhang et al. 2002). The basis for this differential regulation on AKT activity by SHP2 is not yet clear.
Interestingly, and in addition to its role as a phosphatase, there is evidence to suggest that SHP2 can function as an adapter. For example, SHP2 acts as a scaffold for recruitment of the GRB2/SOS complex to the cell membrane through its Tyr542 site in response to PDGF, suggesting a potential alternate mechanism for SHP2 regulation of RAS (Bennett et al. 1994; Li et al. 1994). Similarly, Tyr580 on SHP2 binds GRB2 (Vogel and Ullrich 1996). While intriguing, Araki et. al. demonstrated that SHP2 can only become tyrosyl phosphorylated at these sites in response to some (fibroblast growth factor (FGF), PDGF), but not all (EGF, IGF1) agonists (Araki et al. 2003), suggesting that the tyrosyl phosphorylation sites on SHP2 function to enhance, but not replace, SHP2 catalytic function on downstream signaling in a growth factor-specific manner. The roles, if any, of other functional domains of SHP2 have not yet been determined.
SHP2 mutants have functional significance and, therefore, biological consequences. Heterozygous missense mutations in PTPN11 are observed in up to 90% of LS cases. LS, a rare autosomal dominant disorder, is an acronym for its presenting features of multiple Lentigines, ECG conduction abnormalities, Ocular hypertelorism, Pulmonic stenosis, Abnormalities of genitalia, Retardation of growth and sensorineural Deafness (Digilio et al. 2002; Legius et al. 2002) and is also referred to as Noonan Syndrome with Multiple Lentigines, Multiple Lentigines Syndrome, and Cardiomyopathic Lentiginosis. ~50% of NS cases are also caused by missense PTPN11 mutations (Tartaglia et al. 2001). In contrast to LS, NS is fairly common (~1:1000–1:2500 live births) but is also characterized by multiple, variably penetrant defects, including proportionate short stature, facial dysmorphia, and/or CHD (Noonan 1968; Nora et al. 1974).
The most common cardiac manifestation in NS is pulmonic stenosis (PS) resulting from dysplastic valve leaflets, but stenosis of other valves (mitral valve), atrial (ASD), ventricular (VSD) and atrioventricular (AVSD) septal defects or rarely, double outlet right ventricle (DORV), also are seen (Marino et al. 1999) (Yoshida et al. 2004). HCM has also been reported in a few NS patients without SHP2 mutations (Tartaglia et al. 2002), but genotype-phenotype correlation studies show that only 8% of SHP2 associated-NS patients present with HCM (Sznajer et al. 2007) (Table 1). In contrast, most cases of LS present with HCM, but similar valve anomalies to NS are also observed (Limongelli et al. 2007) (Table 1).
Why “hot spot” mutations within SHP2 lead to a disease phenotype is likely due to the fact that the mutations evoke a loss of negative regulation by SHP2. Whether PTPN11 itself serves as a “hot spot” for such mutations is not known; but speculatively, mutations in PTPN11 may be favored because, like other genes already identified along the RAS signaling pathway, 1) missense defects can give rise to gain of function capabilitites, 2) functional missense changes are mild enough as to not be embryonic lethal, 3) mutations may give spermatogonia a competitive advantage (Wilkie 1997), and/or 4) cells carrying these mutations may be favored during germ-line mitotic divisions (Wilkie 1997).
Because NS and LS share such similar phenotypic characteristics in patients, they were considered to have similar disease pathogenesis. Interestingly, the point mutations identified in PTPN11 that were associated with NS were distinct from those associated with LS (Table 2). Indeed, the biochemical properties between the PTPN11 NS- and LS-specific point mutations are quite different (Hanna et al. 2006; Kontaridis et al. 2006; Tartaglia et al. 2006). Most NS mutations reside within the N-SH2 domain interface that allows for the intermolecular interactions with the PTP domain. Therefore, NS mutations disrupt the ability of SHP2 to retain the closed, inactive conformation without affecting PTP enzymatic activity. As a result, NS mutations behave as gain-of-function (GOF) alleles with increased basal phosphatase activity (Keilhack et al. 2005).
In contrast, all LS mutations tested were catalytically inactive, with decreased and/or absent PTP catalytic activity that resulted in a loss-of-function (LOF) of the phosphatase. Indeed, LS mutants affect conserved residues important for PTP catalysis (e.g. Y279, which sets the depth of the catalytic cleft; T468, which lies within the area of the “signature motif.”) (Table 2) (Hanna et al. 2006; Kontaridis et al. 2006; Tartaglia et al. 2006). However, like the mutants associated with NS, LS mutations also perturb N-SH2/PTP domain interactions, suggesting that both NS and LS mutants can out-compete WT SHP2 for binding to RTKs/scaffolds (Kontaridis et al. 2006; Marin et al. 2011)
Understanding precisely how it is that GOF and LOF mutations lead to such similar clinical phenotypes is of great clinical importance. Attention has naturally focused on the ERK/MAPK pathway in explaining the pathogenesis of RASopathies. However, different RASopathies can have complex and distinct effects on multiple cellular signaling pathways. Knowledge of the detailed effects of these mutations on downstream molecular and cellular signaling events is essential for a more rational and personalized approach to therapy for these disorders.
Interestingly, initial in vivo studies reported conflicting results on the PTP function of the LS mutations in PTPN11. In contrast to the in vitro studies that provocatively suggested that LS mutations were LOF (Hanna et al. 2006; Kontaridis et al. 2006; Tartaglia et al. 2006), expression of the LS mutants Y279C or T468M in Drosophila resulted in ectopic wing veins and a rough eye phenotype, characteristics of increased ERK/MAPK activity in these tissues. In addition, genetic analysis of enhancers and suppressers were consistent with hypermorphism of the ERK/MAPK pathway in these mutant flies (Oishi et al. 2009). However, LS mutants of shp2 in zebrafish were found to have dominant negative effects (Jopling et al. 2007; Stewart et al. 2010), consistent with the in vitro studies. Recent experiments using induced pluripotent stem cells (iPSCs) from LS patient-derived fibroblasts showed both possibilities, with higher basal, but abrogated agonist-evoked, p-ERK levels when compared with control iPSCs (Carvajal-Vergara et al. 2011). The precise reasons for these signaling discrepancies are not yet entirely clear, but might relate to species differences, cell-type specificity and/or the effects of over-expression.
In an effort to resolve this conundrum, we generated the first mammalian model of LS by inducibly knocking-in a PTPN11 allele harboring the LS Y279C mutation in mice (iLS/+) (Marin et al. 2011). When crossed to EIIA deleter-Cre mice, these (now termed) LS/+ mice recapitulated nearly all aspects of the human LS phenotype, including hypertrophic cardiomyopathy (HCM). LS/+ mice also had short stature, craniofacial dysmorphia, and skeletal abnormalities, consistent with LS in human patients (Gorlin et al. 1971; Tartaglia and Gelb 2005). Importantly, heart and/or cardiomyocyte lysates from LS/+ mice showed decreased SHP2 catalytic activity, suggesting that LS mutants are indeed catalytically inactive in vivo. Biochemical analyses revealed that the LS/+ mutant hearts and/or cardiomyocytes showed enhanced binding of SHP2 to the scaffolding adapter proteins Gab-1 and IRS-1 and had abrogated agonist-evoked ERK/MAPK activation (Figure 2), consistent with previous in vitro reports (Hanna et al. 2006; Kontaridis et al. 2006; Tartaglia et al. 2006). By contrast, and different from NS biochemistry, AKT/mTOR signaling was enhanced significantly in LS/+ hearts, as well as in cardiomyocytes, consistent with in vivo findings in fibroblasts isolated from LS patients (Edouard et al. 2010). In addition, we found increased phosphorylation of FAK, STAT3 and JNK1/2 (Figure 2), all signaling molecules that are also implicated in cardiomyopathy (Ng et al. 2003; Wang et al. 1998). Although there is a possibility that these other signaling pathways play a contributory role in the development of HCM in vivo, we found that inhibiting mTOR signaling with rapamycin, an immunosuppressant drug that specifically targets and suppresses TOR activity, completely normalized and reversed the LS cardiac defects. These data directly implicate excessive AKT/mTOR activity, and not ERK/MAPK, in LS pathogenesis. Further, they suggest that therapeutic efficacy of rapamycin and/or its derivatives should be evaluated in LS patients specifically.
More recently, Ishida et. al. analyzed the effects of yet another LS PTPN11 mutant, the Gln510Glu (Q510E) mutation, and also found that this LS mutation evokes a severe HCM in their cardiomyocyte differentiation model system (Ishida et al. 2011). In addition, the Q510E mutants differentiated into cardiac progenitors, but were unable to undergo terminal differentiation into mature cardiomyocytes, suggesting that LS mutations may have an effect on cardiac progenitor expansion. Here too, however, activity of AKT and its downstream effector, glycogen synthase kinase (GSK)-3β, were upregulated. In these experiments, upstream inhibition of the pathway with the PI3K/AKT inhibitor LY-294002 partially rescued the hypertrophic defect (Ishida et al. 2011). These findings were further validated in an in vivo characterization of the Q510E mutation using a transgenic overexpression mouse model showing that prenatal, but not postnatal, overexpression of the Q510E mutation causes enlargement of cardiomyocytes, an increase in heart-to-body weight ratio, and thicker interventricular septum (Schramm et al. 2012). Here again, the biochemical signaling pathway most robustly affected by the LS mutation in PTPN11 is the AKT/mTOR pathway (Schramm et al. 2012). Taken together, the combined in vitro, ex vivo, and in vivo data all confirm that 1) LS mutations primarily affect the AKT/mTOR signaling pathway, not the ERK/MAPK pathway and 2) the cardiac abnormalities associated with LS, namely HCM, can be treated and reversed with a TOR inhibitor such as rapamycin.
Biochemical, cell biological and genetic evidence indicate that PTPN11 mutations associated with NS are hypermorphs that can enhance ERK/MAPK pathway activation (Keilhack et al. 2005). However, differing NS PTPN11 mutations appear to result in uniquely differing cardiac phenotypes (Araki et al. 2009; Araki et al. 2004; Krenz et al. 2008). For example, ~50% of the knock-in mice for the Ptpn11 D61G NS exhibit valvuloseptal defects similar to those observed in NS patients, but with no evidence of HCM (Araki et al. 2004). Overexpression of PTPN11 Q79R in the developing endocardial cushions show significantly enlarged endocardial cushions in the atrioventricular canal and in the outflow tract (Krenz et al. 2008), whereas overexpression in the developing myocardium causes ventricular noncompaction, myocardial thinning and ventricular septal defects without evidence of HCM (Nakamura et al. 2007). Nevertheless, breeding the Ptpn11 Q79R mice into ERK1/2-null backgrounds confirmed that the MAPK/ERK pathway is indeed causative to NS. Moreover, constitutive activation of ERK1/2 through MEK1 overexpression during valve development results in an NS-like phenotype (Krenz et al. 2008), whereas transgenic overexpression in adult hearts leads to HCM (Bueno et al. 2000). Interestingly, absence of Ptpn11 in cardiomyocytes also causes a pathological cardiac phenotype; hearts from mice with SHP2-cardiomyocyte specific deletion (using the cardiomyocyte specific promoter alpha myosin heavy chain (αMHC)-Cre recombinase) develop severe dilatation (Kontaridis et al. 2008), suggesting a causal role for PTPN11 in cardiomyocyte regulation and hypertrophy.
The differential expression of the individual mutants, and/or their levels of activity, localization, and feedback regulation in different cell/tissue types, may account somewhat for the variable severity and/or phenotype in each RASopathy. NS-associated HCM does occur, however, usually in other RASopathy gene mutations. Knock-in mice expressing the NS-associated Sos1-E846K mice develop HCM with incompletely penetrant aortic stenosis (Chen et al. 2010) Mice expressing the NS-associated Raf1L613V allele develop HCM, but with normal valvuloseptal development (Wu et al. 2011). Assessment of ERK/MAPK signaling in vitro and in vivo in hearts from these mice revealed increased activation of Mek1/2 and ERK1/2, suggesting that blockade of the ERK/MAPK signaling pathway in these mice could be beneficial. Indeed, treatment with the MEK inhibitor, PD0325901, blocked and reversed HCM in vivo (Wu et al. 2011).
Dhandapany, et. al., further explored the ability of the NS/LS Raf1 mutants S257L, L613V, and D486N mutants to regulate hypertrophy in cardiomyocytes (Dhandapany et al. 2011). Interestingly, they found that the two mutants associated with HCM in NS/LS patients, the L613V and S257L mutations, also promoted NFAT signaling and interacted with calcineurin in co-immunoprecipitation studies. Furthermore, treatment with the calcineurin inhibitor cyclosporine-A prevented RAF1-induced hypertrophy in neonatal cardiomyocytes (Dhandapany et al. 2011). Taken together, these studies suggest the intriguing possibility that canonical MAPK signaling may not be the sole driver for NS-associated HCM.
Collectively, the results of these studies raise awareness for the need for treatment of diseases and disorders based on biochemical, rather than phenotypic presentation. This provides further impetus to proceed with efforts to identify the other disease genes underlying these disorders and to generate animal models as well as human cell model systems. For example, the continued development of iPSC technology might enable a deeper understanding of the molecular mechanisms underlying human disease, as well as improve our ability to achieve successful individualized therapies for treating patients. Ultimately, it will be the combination of understanding the pathogenesis of each disorder and the identification of small molecules that will enable us to treat patients.
Aside from enabling genetic testing for these RASopathy disorders, genotype/phenotype associations have been established for many, but much work is left to be done. These types of observations allow us to raise important biological and clinical questions. How are tissues/organs differentially affected by specific RASopathy mutations? If proteins are ubiquitously expressed, why are some organs/tissues spared while others are severely impacted? What is the basis of the specificity between altered gene, or even specific mutation, and the phenotype? How do we determine which pathways are affected by each of the mutations? Can inhibition of these ameliorate features of the disorders, and without side affects?
For treatment of NS-associated hypertrophy where the cause is associated with increased activation of MEK1/2 and/or ERK1/2, blockade of MAPK signaling seems efficacious. Indeed, in the Raf1L613V model, treatment of the mice with the MEK inhibitor, PD0325901, blocked emergence of HCM as well as reversed it. However, clinical trials with PD325901 for the treatment of several types of cancers were terminated early due to ophthalmologic and neurologic toxicity (Haura et al. 2010). Therefore, since RAS/MAPK signaling is present in a wide array of normal cells and RAS proteins control multiple cellular processes and target several downstream effectors, titrating down to the right level of inhibition to provide therapeutic efficacy without incurring intolerable side effects will be critical to maintain therapeutic advantage. Alternatively, more specifically targeted (and therefore more tolerable) drugs for this pathway are currently in the pipeline.
LS models, on the other hand, support the conclusion that excessive AKT/mTOR activity, not MAPK, is essential for the development and maintenance of HCM, and that a potential effective treatment for the reversal of HCM in these patients is rapamycin and/or its analogs. Rapamycin is commonly used as an immunosuppressive drug (Alexandre et al. 1999). Nevertheless, it also has been used for other purposes, both in animal models and humans. Rapamycin has proven safe and beneficial in several therapeutic protocols; for instance, rapamycin is used for treatment of polycystic kidney disease in clinical trials (Serra et al. 2007) and a rapamycin analog is now approved for the treatment of renal cancer (Dancey et al. 2009).
Previous pharmacological studies have shown that treatment of cardiomyocytes with rapamycin could inhibit the hypertrophic response evoked by stimulation with various agonists (Boluyt et al. 1997; Sadoshima and Izumo 1995; Wang and Proud 2002; Wang et al. 2000). Rapamycin has also been shown to attenuate or reverse pressure overload- or constitutively active AKT-induced cardiac hypertrophy (Soesanto et al. 2009) (Gao et al. 2006; Shioi et al. 2003), supporting a potential regulatory role for mTOR in HCM. However, recent evidence argues that mTOR expression either has no effect (Shen et al. 2008) or may even be protective against stressed-induced pathological hypertrophy (Song et al. 2010). Indeed, two independent studies indicate that complete ablation of mTOR (Zhang et al. 2010) or the regulator of mTORC1, RAPTOR (Shende et al. 2011), leads to DCM. These findings imply that rapamycin could have detrimental effects, and as a result, we cannot exclude the possibility that chronic inhibition and/or inhibition below physiological levels of mTOR could be deleterious in vivo. Consequently, chronic and/or high-dosed inhibition of mTOR in patients with severe HCM and/or compromised contractile function may not be advisable. However, in the animal models of LS described herein, acute and/or low-dose administration of rapamycin was indeed able to rescue the hypertrophic phenotype without adverse effects on cardiac function (Marin et al. 2011; Schramm et al. 2012). Therefore, a potential short term and/or low-dose administration of rapamycin could have advantageous and beneficial effects in LS patients. Taken together, clinical trials of rapamycin and/or analogs should be considered for the treatment of LS-associated HCM.
We have made significant progress in understanding the differential signaling mechanisms affected by specific mutations in RASopathy disorders. We still have a long way to go. But given what we have learned so far, it is clear that the strategy for therapy must be based on the elucidation of the biological mechanism evoked by each specific mutation, rather than the physiological characteristics of the disorder itself.
This work was supported by National Institutes of Health Grants HL088514, the Milton Fund and the Beth Israel Deaconess Medical Center Division of Cardiology (to M.I.K.).
The authors wish to disclose that there are no conflicts of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.