Friend or Foe? Unraveling the complex roles of protein tyrosine phosphatases in cardiac disease and development

Abstract

Regulation of protein tyrosine phosphorylation is critical for most, if not all, fundamental cellular processes. However, we still do not fully understand the complex and tissue-specific roles of protein tyrosine phosphatases in the normal heart or in cardiac pathology. This review compares and contrasts the various roles of protein tyrosine phosphatases known to date in the context of cardiac disease and development. In particular, it will be considered how specific protein tyrosine phosphatases control cardiac hypertrophy and cardiomyocyte contractility, how protein tyrosine phosphatases contribute to or ameliorate injury induced by ischaemia / reperfusion or hypoxia / reoxygenation, and how protein tyrosine phosphatases are involved in normal heart development and congenital heart disease. This review delves into the newest developments and current challenges in the field, and highlights knowledge gaps and emerging opportunities for future research.

1. Introduction

Protein tyrosine phosphorylation plays a pivotal role in the regulation of most, if not all, fundamental cellular processes such as growth, proliferation, differentiation, or survival. Rapid changes in protein tyrosine phosphorylation levels enable signalling networks to respond to various stimuli thus allowing for coordinated communication within and between cells. Importantly, protein tyrosine phosphorylation is not an isolated event, but needs to be evaluated in the context of other post-translational protein modifications [1]. Phospho-proteomic analyses have demonstrated that serine phosphorylation is the predominant post-translational modification in most tissues. For example, serine accounts for 88%, threonine for 11%, and tyrosine for less than 2% of all phosphorylation sites across various rat tissues [2]. However, these numbers are unlikely to reflect the relative importance of these different modifications. Protein tyrosine kinases (PTKs) and protein tyrosine phosphatases (PTPs) are found in all tissues, with most cell types expressing 30% – 60% of the entire complement of PTKs and PTPs [3]. Further underlining the importance of tyrosine phosphorylation, PTP mutations have been identified in various human diseases such as immunodeficiency [4] or diabetes mellitus [5, 6], and PTK dysregulation plays a prominent role in many forms of cancer [7]. Notably, treatment with tyrosine kinase inhibitors is associated with cardiotoxic side effects [8], indicating that appropriate protein tyrosine phosphorylation levels may be particularly important in the heart.

Despite substantial efforts to unravel cellular signalling cascades over the past decades, the complex and tissue-specific roles of PTPs in the heart are not yet fully understood. Importantly, individual PTPs play unique roles and may have numerous different as well as overlapping interaction partners [9]. PTPs function as independent and specific signalling regulators in coordination with, rather than opposition to, PTKs [10], and can exert opposite effects on downstream signalling depending on context [11]. Undoubtedly, PTPs play fundamental roles in numerous tissues including the heart, and many have been shown to be indispensable. As a systematic approach and to reveal the importance of structural similarities and differences, PTPs have been classified into four families based on the amino acid sequences of their catalytic domains [3]. Yet, as more data continue to emerge on the organ- or tissue-specific roles of individual PTPs, we recognize that further context in addition to the structural determinants is needed to thoroughly understand PTP physiology. We unfortunately still lack comprehensive data sets to fully answer whether PTPs primarily support or hamper cardiac physiology. However, recent detailed gain- and loss-of-function experiments in various disease models together with data from human samples have provided critical new insights (summarized in Table 1). As our ultimate goal is to gain new knowledge directly applicable to translation into the clinic, this review takes a disease-centric approach. The following sections will focus on the roles of PTPs in (i) cardiomyocyte hypertrophy and contractility, (ii) hypoxia / reoxygenation (H/R), oxidative stress, ischaemia / reperfusion (I/R) injury, and myocardial infarction (MI), and (iii) heart development and congenital heart disease. This provides a foundation to understand unique versus potentially overlapping or synergistic functions of various PTPs and can direct us towards the most pressing key questions to examine in the near future.

Table 1:

Roles of specific PTPs in cardiac homeostasis, development, and disease

Gene	Experimental model	Effects observed	References
ACP1	Human tissue, heart failure	Increased LMPTP expression	[18]
ACP1	Mice, TAC	Increased LMPTP expression	[18]
ACP1	Mice, global deletion, TAC	Reduced hypertrophy, ventricular dilation, fibrosis, improved contractile function	[18]
EYA1–4	human genetic analyses	copy number variants found in patients with oculo-auriculo-vertebral spectrum and congenital heart defects, mediated by components of the PAX-SIX-Eyes absent-Dachshund network (PSEDN)	[100]
EYA1–4	human genetic analyses	rare EYA1 mutations in patients with conotruncal defects	[101]
EYA1–4	human genetic analyses	frame shift EYA1 mutation associated with cardiofacial syndrome with patent ductus arteriosus	[102]
EYA1–4	human genetic analyses	a subset of EYA4 gene mutations cause autosomal dominant syndromic hearing loss with dilated cardiomyopathy	[104, 105]
EYA1–4	mice, deletion of both Six1 and Eya1	craniofacial, cardiac outflow tract, aortic malformations, and ventricular septal defects	[103]
EYA1–4	mice, overexpression of mutant EYA4	dilated cardiomyopathy	[107]
EYA1–4	mice, EYA2 overexpression	EYA2 regulates physiological hypertrophy induced by swimming exercise and pathological hypertrophy due to TAC	[109–111]
PTPN1	human tissue, ventricular dysfunction due to aortic stenosis	increased PTP1B activity	[15]
PTPN1	H9c2 cells, H/R, RNAi	decreased markers of apoptosis, normalized mitochondrial membrane potential	[39, 40]
PTPN1	zebrafish, mice, resection or MI, trodusquemine	increased rate of regeneration after caudal fin amputation or ventricular resection, improved heart function and increased survival after MI, reduced infarct size	[61]
PTPN1	mice, endothelial-specific deletion, TAC	increased survival, improved systolic function, preserved capillary density, improved perfusion, decreased cardiomyocyte hypertrophy and ventricular dilation	[17]
PTPN1	mice, aging	increased PTP1B expression coinciding with contractile dysfunction	[21]
PTPN1	mice, global deletion, aging	decreased hypertrophy, decreased fibrosis, increased ventricular capillary density, improved systolic and diastolic function	[22]
PTPN1	mice, MI, global deletion	improved diastolic function, increased capillary density in the border zone	[41]
PTPN1	mice, trodusquemine or myeloid-specific deletion, atherogenic conditions	reduced atherosclerotic plaques in the aortic root, decreased adiposity, circulating total cholesterol and triglycerides, and improved glucose homeostasis	[54, 55]
PTPN1	rat, TAC	upregulation of PTP1B activity after TAC	[15]
PTPN1	rats, doxorubicin-induced heart failure	prediction algorithms suggesting PTP1B may be a key regulator	[23]
PTPN1	rats, I/R or MI, microRNA	reduced myocardial size or infarct size, improved contractile function, decreased apoptotic markers	[39, 40]
PTPN4	H9c2 cells, H/R, miRNA, siRNA	knockdown of PTPN4 aggravates H/R injury, upregulation protects against H/R	[50, 52]
PTPN4	rat neonatal cardiomyocytes, H2O2	increased PTPN4/PTPRG reduces apoptosis markers, whereas downregulation increases apoptosis and reactive oxygen species	[51]
PTPN4	rats, I/R	upregulation of a miRNA that targets Ptpn4	[50, 52]
PTPN4	rats, I/R	downregulation of a miRNA that targets PTPN4 and PTPRG	[51]
PTPN6	H9C2 cells, hypoxia	SHP1 expression increased	[46]
PTPN6	mice, I/R, siRNA	reduced infarct size and TUNEL labeling	[46]
PTPN6	rats, MI, siRNA	reduced infarct size	[47]
PTPN6	rats, I/R, siRNA, VEGF	reduced infarct size, increased capillary density, increased markers of angiogenesis	[49]
PTPN11	human genetic analyses	germline point mutations cause NS and NSML	[69–71]
PTPN11	myocardial samples from NSML patients	upregulation of MAPK/ERK1/2 activity, attenuated mTOR activation	[96]
PTPN11	human induced pluripotent stem cell cultures from NSML patient fibroblasts	hypertrophic phenotype in cell culture, increased MAPK kinase 1-ERK1/2 activation	[95]
PTPN11	fibroblasts from NSML patients, chicken embryo explants, expression of NSML mutants	NSML mutations facilitate phosphoinositide 3-kinase (PI3K)/AKT/glycogen synthase kinase-3β stimulation through impaired GRB2 Associated Binding protein (GAB1) dephosphorylation	[85]
PTPN11	chicken embryo explants, expression of NS mutant	increased outgrowth from cultured valve primordia explants, mediated by ERK1/2 hyperactivation downstream of SHP2	[86]
PTPN11	fruit flies, expression of NS/NSML mutants	NS/NSML mutants cause ectopic wing veins and eye phenotypes with increased RAS-MAPK activation; genetic interaction with Notch, decapentaplegic, and JAK/STAT signaling	[87, 88, 90]
PTPN11	zebrafish, expression of NS/NSML mutants	cell migration defects, craniofacial and cardiac defects	[72–74]
PTPN11	zebrafish, mice, expression of NS/NSML mutants	protein-zero related (PZR) is a downstream mediator of the NS/NSML phenotypes	[91, 92]
PTPN11	mice, muscle-specific deletion, TAC	dilated cardiomyopathy at baseline, reduced hypertrophic response after TAC	[19, 20]
PTPN11	mice, small molecule inhibitor	reduced atherosclerosis by blocking smooth muscle cell proliferation	[56]
PTPN11	mice, global deletion	early embryonic lethality	[66, 67]
PTPN11	mice, expression of NS/NSML mutants	non-compaction or hypertrophic cardiomyopathy, septal defects, outflow-tract and valve malformations; downstream mediators include ERK1/2, focal adhesion kinase, AKT, MTOR, PI3K; association with elevated Ca2+ transients	[32, 75–83, 89]
PTPN11	mice, isolated cardiomyocytes and cardiac fibroblasts, SHP2 deletion and gain of function mutants	Ca2+ oscillations require the PTP activity of SHP2; gain-of-function mutants of SHP2 enhance Ca2+ oscillations while reducing activity of NFAT	[93]
PTPN11	rats, neonatal cardiomyocytes, expression of NSML mutants	NSML mutants induce cardiomyocyte hypertrophy mediated by AKT and focal adhesion kinase hyperactivation	[94]
PTPN12	primary mouse cardiomyocytes, H9c2 cells, H/R, RNAi	reduced cell death under H/R stress	[38]
PTPN12	mice, I/R, auranofin	reduced infarct size, improved contractile function	[38]
PTPRG	human exome sequencing	PTPRG variants associated with increased risk of stroke, MI, reduced ejection fraction	[53]
PTPRG	mice, global deletion	PTPRG adjusts microvascular perfusion and blood pressure during increased tissue metabolism and acid-base disturbances	[53]
PTPRG	rat neonatal cardiomyocytes, H2O2	increased PTPN4/PTPRG reduces apoptosis markers, whereas downregulation increases apoptosis and reactive oxygen species	[51]
PTPRG	rats, I/R	downregulation of a miRNA that targets PTPN4 and PTPRG	[51]
PTPRS	mice, MI, deletion or pharmacologic inhibition	restored sympathetic innervation and reduced arrhythmia susceptibility after MI	[60]

2. Cardiomyocyte hypertrophy and contractility

The prevalence of unexplained left ventricular hypertrophy has been estimated at 0.2% and up to 1.4% in the community [12]. The cardiac hypertrophic response is governed by numerous cellular signalling pathways (recently reviewed in [13]). Extensive research has centered on the roles of MAPK driving cardiomyocyte hypertrophy, but it should be emphasized that the duration and level of MAPK signalling are subject to negative-feedback regulation by tyrosine-, serine/threonine-, or dual-specificity phosphatases (DUSP) [14]. Therefore, it is not surprising that a number of studies have shown that certain PTPs play equally important roles in governing the hypertrophic response as detailed below and summarized in Figure 1.

Figure 1: Cardiac hypertrophy and failure after TAC may be ameliorated by targeting LMPTP or endothelial PTP1B.

Studies have shown that deletion of Acp1, which encodes LMPTP, reduces hypertrophy, ventricular dilation, and fibrosis, and improves contractile function in mice [18]. Similarly, deletion of endothelial PTP1B increased survival, improved systolic function, decreased cardiomyocyte hypertrophy, and decreased ventricular dilation after TAC [17]. This was also accompanied by preserved capillary density and improved perfusion. These data suggest that LMPTP and PTP1B may be ideal targets for future therapeutic approaches.

2.1. Protection against pressure overload-induced hypertrophy and failure

One of these critical PTPs is protein tyrosine phosphatase 1B (PTP1B, encoded by PTPN1). Recently, Nguyen and colleagues found that PTP1B activity is upregulated in rat hearts under chronic pressure overload [15]. Increased PTP1B activation preceded mitochondrial and contractile dysfunction after transverse aortic constriction (TAC). Consistent with the findings in their rat model, the authors observed significantly increased PTP1B activity in left ventricular samples from patients with systolic dysfunction due to aortic valve stenosis who underwent valve replacement. These data suggest that PTP1B may be a potential target to modulate contractile function in heart failure induced by chronic pressure overload. Mechanistically, effects of PTP1B on insulin sensitivity are thought to play a role, which will be discussed in more detail below.

Nguyen and colleagues analyzed tissue samples encompassing the entire ventricular wall, therefore conclusions regarding cell type-specific roles of PTP1B cannot be drawn. However, most of the proteins in lysates from homogenized ventricular tissue come from cardiomyocytes, suggesting that the increase in PTP1B activity may primarily be a cardiomyocyte-specific effect. However, stimulation of cultured cardiac rat fibroblasts with angiotensin II results in increased tyrosine phosphorylation of numerous proteins including signalling mediators of hypertrophy [16], indicating that changes in tyrosine phosphorylation in response to growth stimuli is not limited to cardiomyocytes. Furthermore, endothelial-specific deletion of Ptpn1 has been shown to protect against chronic pressure overload, indicating that other cell types in addition to cardiomyocytes and fibroblasts are likely to be involved as well. Gogiraju and co-authors showed that survival up to 20 weeks after TAC was significantly improved in mice lacking endothelial PTP1B [17]. At the same time, systolic pump function, cardiac hypertrophy, and left ventricular dilation were significantly improved in animals without endothelial PTP1B compared to controls. Notably, cardiac vascular endothelial growth factor (VEGF) signalling and angiogenesis were upregulated in these hearts, indicating that the resulting improvement in cardiac muscle perfusion may be the main cause of the protective effects.

Consistent with the protective effects observed with PTP1B deletion, it has been recently shown that global deletion of the low molecular weight protein tyrosine phosphatase (LMPTP, encoded by Acp1) 1 also confers beneficial effects against pressure overload. Wade and colleagues found that LMPTP expression is increased both in human heart failure and in mouse hearts after TAC [18]. Importantly, Acp1−/− mice were resistant to pressure overload-induced hypertrophy, fibrosis, and failure [18]. The substantial protection against pressure overload-induced heart failure in both PTP1B and LMPTP deletion models suggest that these two PTPs may play similar roles in cardiac homeostasis. Whether this is a common theme and applies to more than these two PTPs is not yet clear.

A third PTP may also be involved in the cardiac hypertrophic response to pressure overload. As a muscle-specific approach, muscle creatine kinase (MCK) Cre has been employed to delete src homology 2 domain-containing phosphatase 2 (SHP2, encoded by Ptpn11) in cardiomyocytes. In two independent studies, this resulted in a baseline phenotype of dilated cardiomyopathy [19, 20], indicating that SHP2 is indispensable in cardiomyocytes. If mice underwent 7 days of banding of the ascending aorta at an early age before the dilated cardiomyopathy was fully established, the hypertrophic response was reduced, suggesting that SHP2 may be required for the hypertrophic response to short-term pressure overload. These data are difficult to interpret due to the underlying dilated cardiomyopathy phenotype and the short follow-up after TAC. However, it is likely that other PTPs in addition to PTP1B and LMPTP may also become future therapeutic targets to ameliorate heart failure.

2.2. Protection against aging and doxorubicin-induced hypertrophy and dysfunction

Interestingly, the protective effects of Ptpn1 deletion are not limited to pressure overload-induced heart failure. PTP1B expression increases in aged mice, coinciding with contractile dysfunction [21]. Recently, Besnier and coworkers examined aging-associated cardiac remodeling in mice with global deletion of Ptpn1 [22]. This study demonstrates that old mice lacking PTP1B exhibit less hypertrophy and fibrosis, higher ventricular capillary density, and better systolic and diastolic function compared to wild type controls. In addition, PTP1B inactivation reduced aging-associated endothelial dysfunction in mesenteric resistance arteries. Furthermore, proteomic and metabolomic data point at PTP1B also being involved in doxorubicin-induced heart failure in rats [23], suggesting that PTP1B is involved in multiple forms of heart failure.

2.3. Potential mechanisms underlying the beneficial effects of targeting specific PTPs in pressure overload

In order to exploit any protective effects triggered by PTP manipulation, further mechanistic insight regarding the downstream effectors will be needed. Even though we still lack a comprehensive picture of all the signalling pathways that are regulated by particular PTPs, we already have evidence of likely downstream mediators. PTP1B is a negative regulator of the insulin signalling pathway by binding to and dephosphorylating the activated insulin receptor [24–28], therefore beneficial effects on glucose metabolism could be one likely mechanism. Nguyen and colleagues showed that under insulin stimulation, tyrosine phosphorylation levels of the insulin receptor β subunit are altered by chronic pressure overload [15]. Importantly, they found that cardiac insulin responsiveness was markedly perturbed at 2.5 – 5 months after TAC, whereas systemic insulin action remained normal. The impaired cardiac insulin action was related to a decrease in insulin-stimulated phosphorylation of insulin receptor β and co-incided with elevated PTP1B activity under pressure overload. Wade and colleagues reported that insulin receptor β subunit phosphorylation was increased in Acp1−/− hearts early after TAC [18]. Although this does not constitute causality, these finding are intriguing. Further detailed investigations are needed to elucidate to which extent the favorable effects of PTP1B inhibition against pressure overload-induced heart failure are due to enhanced cardiac insulin sensitivity.

In contrast, protection against aging-associated cardiac remodeling induced by PTP1B deletion does not closely correlate with changes in insulin sensitivity. Besnier and co-authors observed that Ptpn1−/− mice did not yet exhibit alterations in insulin sensitivity at an age when left ventricular functional benefits were already observed [22]. This speaks against changes in insulin sensitivity playing a major role in their model but does not definitively exclude this possibility.

On the cellular level, vascular effects appear to play a prominent role in the protective effects of PTP1B inhibition. As described above, endothelial deletion of Ptpn1 resulted in less ventricular remodeling and dysfunction after TAC [17]. Importantly, increased numbers of proliferating endothelial cells resulting in preserved cardiac capillary density and improved perfusion were observed in these mice lacking endothelial PTP1B. This was associated with increased expression of caveolin-1 as well as reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase-4 expression, reactive oxygen species (ROS) generation, and transforming growth factor β (TGFβ) signalling, which may have mediated the cardioprotective effects of endothelial PTP1B deletion.

In addition to increased phosphorylation of insulin receptor β, multiple other signalling mediators have been implicated in the protective effects that were observed after short-term pressure overload in Acp1−/− mice [18]. In particular, protein kinase A (PKA) and ephrin receptor were upregulated on the protein level in Acp1−/− hearts. In contrast, no changes in protein kinase B (Akt), extracellular signal-regulated kinase (ERK), or p38 MAPK activation were observed in this model. However, phospholipase C β (PLCβ) and calcium/calmodulin-dependent protein kinase II δ (CaMKIIδ) phosphorylation were reduced, indicating that CaMKIIδ inhibition may contribute to the protection in Acp1−/− hearts. Notably, phosphorylated phospholamban was higher in LMPTP-deficient hearts both at baseline and early after TAC, suggesting that alterations in Ca2+ cycling may also contribute to the protective effects seen in Acp1−/− mice.

These findings are intriguing, as CaMKIIδC overexpression has been shown to induce heart failure and to alter Ca2+ homeostasis in transgenic mice [29]. Similarly, the protective effects of PTP1B inhibition may also be connected to changes in Ca2+ handling. Hsu and coworkers have shown in AR42J and HEK293 cell cultures that PTP1B modulates store-operated calcium influx [30]. In cultured adult ventricular myocytes isolated from rabbit hearts, PTP1B has recently been shown to mediate cyclic ADP ribose-induced Ca2+ signalling [31]. Furthermore, we observed that cardiomyocytes isolated from transgenic mice with tissue-specific expression of a loss-of-function mutant of SHP2 exhibit elevated calcium transients [32]. However, these studies still only provide circumstantial evidence that altered calcium handling may be an important downstream effect caused by inhibition of certain PTPs. Extensive further investigation is warranted to elucidate whether modulation of Ca2+ handling could be a common end-effector.

3. Hypoxia, ischaemia, and MI

Approximately every 40 seconds, a resident of the United States will suffer a myocardial infarct [33]. Despite many recent advances, timely reperfusion remains the only established treatment to salvage viable tissue injured by MI and to reduce the associated mortality and morbidity. Experimental studies have demonstrated that interventions such as ischemic preconditioning can effectively reduce the extent of myocardial damage. Therefore, it is critically important that we fully understand the signalling mechanisms underlying such cardioprotective effects. A number of recent studies have demonstrated that PTP are likely to be critical players in I/R injury as discussed in the following and illustrated in figure 2.

Figure 2: Specific PTP play critical roles in cardiac I/R injury and MI.

Importantly, PTP can exert either beneficial or detrimental effects. PTPN4 and PTPRG are thought to have protective functions, as knockdown of these proteins led to increased injury [50–52]. In contrast, non-selective PTP inhibition, deletion, knockdown, or inhibition of PTP1B, PTP-PEST, or SHP1 resulted in protection against I/R injury or MI [38–41, 46–49]. These findings indicate that PTP play unique roles in the setting of cardiac I/R injury and MI.

3.1. Protection against H/R or I/R injury by targeting PTP

Both in ischemic and pharmacological preconditioning, activation of myocardial PTKs has been proposed as one possible reason for cardiac protection [34, 35]. Supporting this concept, PTK inhibitors can abrogate the beneficial effects induced by preconditioning [34, 35]. Furthermore, PTK-mediated pro-survival signals may also be activated by PTP inhibition. For example, vanadate-derived compounds suppress PTP activity by oxidation of the cysteine residue in the catalytic domain and have been shown to induce cardioprotection against I/R in mice [36, 37].

Building on these findings, Yang and co-workers recently explored the roles of PTPs in cardiac I/R injury [38]. In mice undergoing coronary artery occlusion followed by reperfusion, overall myocardial PTP activity was elevated by ischaemia and remained high during reperfusion. In line with these findings, protein tyrosine phosphorylation levels were decreased in the myocardium. Treatment with the pan-PTP inhibitor phenyl vinyl sulfone attenuated infarct size as assessed by triphenyl tetrazolium chloride (TTC) staining. Using RNAi in a H/R cell culture model, Yang and co-workers identified PTP-PEST (also known as PTPN12) as a candidate PTP. Importantly, the PTP-PEST inhibitor auranofin reduced infarct size and improved cardiac function in I/R mice. This suggests that PTP-PEST may be responsible for I/R-induced tissue damage and could become a novel therapeutic target.

Two recent studies testing the roles of two other micro-RNAs, miR-135a and miR-203, respectively, indicate that their shared target PTP1B also plays an important role [39, 40]. Overexpression of either miR-135a or miR-203 reduced injury in in vitro and in vivo models by reducing PTP1B levels. In particular, adenoviral delivery of miR-135a after I/R reduced infarct size measured by TTC staining in rats. This was accompanied by reduced creatine kinase and lactic acid dehydrogenase serum levels, supporting that targeting PTP1B with miR-135a reduces cardiomyocyte damage. Similarly, lentivirus-mediated overexpression of miR-203 reduced myocardial scar size and improved cardiac function in rats 2 weeks after coronary artery ligation without reperfusion [40]. Notably, Besnier and co-authors showed in 2014 that mice with homozygous deletion of Ptpn1 were protected against heart failure induced by large MI [41]. In their study, PTP1B deficiency enhanced angiogenesis, increased perfusion, and improved diastolic function early after MI. It is not clear whether miR-203 delivery also affected angiogenesis, but taken together these studies provide support for PTP1B exacerbating injury after MI. The downstream protective mechanisms of PTP1B reduction or deletion still need to be further elucidated including dissection of anti-necrotic versus anti-apoptotic effects.

The non-receptor PTP src homology region 2 domain-containing phosphatase 1 (SHP1, or PTPN6) has also been implicated as an adverse mediator in MI. SHP1 is thought to act as a major apoptosis-regulating factor as it binds to and promotes tyrosine dephosphorylation of death receptors such as tumor necrosis factor 1 (TNFR1) and FAS receptor (FASR) [42–44]. Under normal conditions, SHP1 is predominantly expressed in hematopoietic cells and only at low levels in epithelial cells [45]. However, two studies have shown that SHP1 expression is increased both in hypoxic H9c2 cells and in rat myocardial samples taken after I/R [46, 47]. Strikingly, SHP1 knockdown using siRNA with an amphipathic delivery system significantly reduced infarct size and number of terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) positive cells after I/R [46]. Similarly, Sugano and colleagues observed reduced infarct sizes in rat hearts after permanent coronary artery ligation and siRNA treatment to reduce SHP1 protein levels [47]. These protective effects could be cardiomyocyte-specific or may be due to SHP1’s negative effects on angiogenesis during ischaemia, possibly via regulation of hypoxia-inducible factor-1α (HIF-1α) and ROS formation [48]. In a follow-up study, Kim and co-workers combined administration of siRNA against SHP1 with plasmid DNA to increase VEGF expression during ischaemia [49]. These two treatments acted synergistically, and together not only significantly reduced infarct size but also increased the number of small vessels in the myocardium. Knockdown of SHP1 alone resulted in a smaller but significant increase in angiogenesis. These data would be consistent with the notion that SHP1 suppresses neovascularization, but whether SHP1 also has cardiomyocyte-specific harmful effects during ischaemia remains to be shown.

3.2. Detrimental effects of specific PTP inhibition or deletion in H/R or I/R models

However, not all PTPs may have detrimental effects in I/R-induced cardiac injury. In particular, protein tyrosine phosphatase non-receptor type 4 (PTPN4) has recently emerged as a possible mediator of protective effects. Three recent studies showed that overexpression of miR-181c-5p or miR-208a results in knockdown of PTPN4 and is associated with increased H/R-induced damage in cell culture models [50–52]. Vice versa, reduction of miR-181c-5p or miR-208a resulted in beneficial effects. However, miR-208a also targets protein tyrosine phosphatase receptor type G (PTPRG) [51], and either micro-RNA may also have other as yet unidentified targets. Therefore, it remains unclear to which extent PTPN4 alone is responsible for the observed protective effects.

PTPRG is thought to function as an HCO3⁻ sensor and is widely expressed in the vascular endothelium, with prominent levels in cerebral and coronary arteries [53]. Using mice with global deletion of Ptprg, Hansen and co-workers demonstrate that PTPRG facilitates endothelium-dependent relaxation of resistance arteries through mechanisms regulated by extracellular HCO3⁻. Furthermore, they show that PTPRG adjusts microvascular perfusion and blood pressure during increased tissue metabolism and acid-base disturbances. These data indicate that PTP-mediated effects on the regulation of regional blood flow and microcirculation are likely to be critical. In addition, human exome sequencing data indicate that predicted loss-of-function variants in PTPRG are associated with increased risk of stroke, MI, and reduced cardiac ejection fraction [53]. These are intriguing findings that may lead to the discovery of so far unexplored therapeutic opportunities.

3.3. PTPs in atherosclerosis, post-MI arrhythmogenesis, and cardiac regeneration

As MI remains a leading cause of death and disability, we not only need to develop better treatments to prevent or at least reduce myocardial damage, but also need to continue all efforts to prevent or slow down the progression of atherosclerosis. In two separate studies, Thompson and co-authors demonstrated that either myeloid-specific deletion or pharmacological inhibition of PTP1B protects against atherosclerotic plaque formation under atherogenic conditions [54, 55]. More specifically, administration of the PTP1B inhibitor trodusquemine (MSI-1436, a naturally occurring aminosterol) in mice lacking the low density lipoprotein receptor (LDLR) decreased adiposity, reduced circulating total cholesterol and triglycerides, improved glucose homeostasis, and diminished atherosclerotic plaque areas in aortic root sections [54]. Remarkably, a SHP2-specific small molecule inhibitor has also been shown to ameliorate atherosclerosis by blocking smooth muscle cell proliferation [56]. This suggests that in the future, both PTP1B and SHP2 inhibitors could be employed in the clinic to reduce atherosclerosis.

After the acute phase of MI, especially during the first 30 days, survivors with left ventricular dysfunction continue to be at high risk of sudden death [57]. One important contributing factor is that the infarcted area constitutes an anatomical substrate which promotes re-entrant ventricular tachycardia. Studies in patients with implanted cardioverter defibrillators indicate that the amount of ventricular sympathetic denervation after MI may predict the risk for ventricular arrhythmias [58, 59]. Interestingly, PTPs may also play important roles in this context. Gardner and co-authors showed that either genetic manipulation to delete or pharmacological targeting of the neuronal PTP receptor σ (PTPσ, encoded by PTPRS) in mice after MI, sympathetic innervation is restored, and arrhythmia susceptibility is reduced [60]. This raises the possibility that targeting PTPσ may improve post-infarction outcomes by preventing sudden death.

Envisioning further future advancements and in particular new approaches to repair ventricular tissue, the role of PTPs in cardiomyocyte regeneration is another avenue under active investigation. Intriguingly, zebrafish studies have shown that PTP1B inhibition with trodusquemine (MSI-1436) increases the rate of regeneration after caudal fin amputation or ventricular resection [61]. Furthermore, the same study showed that intraperitoneal administration of trodusquemine to adult mice for 4 weeks after induction of MI improved heart function, reduced infarct size, and increased survival. This suggests that PTP1B inhibitors may not only induce multiple beneficial effects in various tissue types, but also have the potential to stimulate tissue repair and regeneration.

4. Cardiac development and congenital heart disease

Congenital cardiovascular defects are due to abnormal development of the heart, valves, and/or blood vessels in early pregnancy. Defects range from minor abnormalities to complex malformations that are associated with high morbidity and mortality. In the United States, at least 40,000 infants are born with congenital cardiovascular defects each year. In about a quarter of these cases (2.4 per 1000 live births), invasive treatment is required in the first year of life [33]. Health outcomes, including survival, continue to improve for congenital cardiovascular defects, which has led to a population shift into adulthood for these patients. Approximately 1 in 150 adults are expected to have some form of congenital heart defect, including minor lesions such as bicuspid aortic valve as well as severe malformations, for instance hypoplastic left heart syndrome [62]. Figure 3 spotlights the functions of various PTP during cardiac development known to date, and the respective details are described in the next sections.

Figure 3: Cardiac development is governed by a distinct set of PTP.

Non-selective PTP inhibition disrupts myofibrillogenesis [63]. SHP2 has been shown to be indispensable in cardiac development, as tissue-specific deletion leads to outflow tract malformation and dilated cardiomyopathy [19, 84]. Importantly, mutations in SHP2 and EYA proteins cause various forms of congenital heart disease including ventricular non-compaction, hypertrophic and dilated cardiomyopathy, outflow tract and/or valve malformation, as well as septal defects. Schematic adapted from [125].

4.1. Roles of PTPs in early heart development

Undoubtedly, appropriate levels of protein tyrosine phosphorylation are essential during normal cardiovascular development. For example, delivery of a general anti-phosphotyrosine antibody into embryonic axolotl hearts resulted in disrupted myofibrillogenesis [63]. This supports the notion that many PTPs are likely to play critical roles in cardiovascular development, but we still lack a comprehensive picture of prenatal PTP function.

Although LMPTP and PTP1B play such central roles in heart failure due to chronic pressure overload, it appears unlikely that either are critical during heart development. In particular, absence of LMPTP in Acp1−/− mice did not lead to cardiovascular defects, which was surprising as LMPTP expression levels are high in the embryonic mouse heart [18]. Similarly, Ptpn1−/− mice do not exhibit major cardiovascular malformations [64]. However, Ptpn1−/− mice have been shown to develop high blood pressure as measured via telemetry in conscious animals [65]. If measured under invasively anesthesia, only a trend but no significant increase in blood pressure was found in the same mouse model [22]. Notably, deletion of Ptpn1 affects vascular function in obese, but not lean, mice [65]. It is not yet clear to which extent PTP1B’s effects on blood pressure are due to modulation of sympathetic tone versus peripheral effects on the vasculature.

In contrast, we already have strong evidence that the non-receptor PTP SHP2 plays a major role in cardiac development and congenital heart disease. SHP2 is essential for key functions in numerous physiological processes and in organism patterning during embryo development. Multiple animal models have demonstrated that SHP2 is indispensable during early development. For example, global SHP2 deletion in mice results in embryonic lethality as early as peri-implantation or gastrulation stages [66, 67]. Tissue-specific deletion or knockdown of Shp2 has demonstrated numerous roles of SHP2 in tissue/organ development and/or homeostasis, including bone, nervous system, intestine, spermatogonia, etc. (reviewed in [68]). Moreover, genetic analyses have revealed that germline point mutations in PTPN11 cause Noonan Syndrome (NS) [69, 70] and Noonan Syndrome with Multiple Lentigines (NSML) [71]. In NS, frequent cardiac abnormalities include valve stenosis, arterial and/or ventricular septal defects, and hypertrophic cardiomyopathy, whereas hypertrophic cardiomyopathy is the most prominent cardiac characteristic in NSML.

4.2. PTP mutations in congenital heart disease

Numerous animal models have allowed us to gain deeper insight into the role of SHP2 in the pathogenesis of NS and NSML. Studies in zebrafish demonstrated that expression of NS or NSML-associated shp2 mutants induces cell migration defects and results in craniofacial and cardiac defects that are reminiscent of human disease characteristics [72–74]. Similarly, multiple mouse models using different genetic approaches have been successfully employed to recapitulate NS and NSML phenotypes, in particular hypertrophic cardiomyopathy, septal defects, and outflow-tract and valve malformations [32, 75–84]. In addition, drosophila and chicken tissue explant models have been used to deepen mechanistic insights into signalling dysregulation in the presence of NS and NSML mutants [76, 85–88].

This broad range of experimental models facilitated the identification of underlying signalling mechanisms. In NS models, gain-of-function mutations of Shp2 result in hyperactivation of RAS/MAPK and in particular ERK1/2 signalling [73–75, 77, 86, 89]. Genetic or pharmacological rescue strategies targeting the RAS/MAPK pathway provide strong support that signal hyperactivation through this pathway plays a causal role [73, 77, 80, 89]. However, other dysregulated mediators such as janus kinase / signal transducer and activator of transcription (JAK/STAT), histone deacetylase (HDAC), or protein zero-related (PZR) may also play important roles [87, 90–92]. Further downstream, it has also been shown that NS mutants of SHP2 disrupt calcium-dependent control of nuclear factor of activated T-cells (NFAT) activity [93], which is critical for cardiac morphogenesis.

In NSML models, there is strong agreement among multiple studies using different experimental approaches that hyperactivation of signalling through the mammalian target of rapamycin (MTOR) / AKT pathway is a hallmark in NSML-associated hypertrophic cardiomyopathy and valve malformation [76, 79, 81, 83, 92, 94]. To which extent RAS/MAPK hyperactivation also plays a role in NSML, is still under debate. ERK1/2 hyperactivation has been observed in fly, zebrafish, and mouse models of NSML [73, 82, 83, 87, 88]. However, other studies also using zebrafish or mice have shown downregulation of ERK signalling in the presence of NSML mutants of SHP2 [72, 79]. These differences may be due to differences in models, agonist doses, or SHP2 mutant expression levels, but this has not yet been conclusively resolved.

Most recently, the use of induced pluripotent stem cells (iPSCs) generated from patient fibroblasts has opened new ways for exploring signalling dysregulation downstream of mutant SHP2 [95]. Interestingly, iPSC studies indicated that NSML mutations in SHP2 primarily lead to increased MEK1-ERK1/2 activation, suggesting that this pathway plays a larger role in NSML-associated cardiomyocyte hypertrophy than previously indicated by animal models. Importantly, a significant increase in S6 protein phosphorylation could not be confirmed in iPSCs, arguing against upregulation of MTOR/AKT signalling in these patient-derived cell cultures. Furthermore, studies using human myocardial samples from NSML patients also did not confirm MTOR hyperactivation as previously observed in animal models [96]. Therefore, it still remains unclear which pathways may be the best therapeutic targets for future pharmacologic treatment of NSML-associated hypertrophic cardiomyopathy. To resolve this, high-throughput studies as recently conducted in drosophila to screen libraries of pharmacological compounds show great promise [90].

Intriguingly, a very different class of PTP has also been implicated in congenital cardiovascular disease. The Eyes Absent (EYA) proteins uniquely contain independent PTP and protein threonine phosphatase domains as well as a transcriptional activation domain [97]. As EYA proteins do not have the signature cysteine-containing motif of most other PTPs and instead contain an aspartic acid-based catalytic domain, they form their own group in the original PTP classification system [3]. To date, the only other tyrosine phosphatase shown to share this reaction mechanism is the TFIIF associating component of CTD phosphatase/small CTD phosphatase (FCP1/SCP) [98]. This suggests that this unusual reaction mechanism may serve a highly specific function, but this has not yet been explored in detail.

The 4 human EYA paralogs (EYA1–4) are part of a conserved cell-fate determination cascade originally described in drosophila eye development [99]. With a few exceptions, EYA1 and EYA2 expression is restricted to developmental stages, EYA3 is expressed during development and in most adult tissues, while a more restricted pattern of expression is present for EYA4 in both developing and adult tissue [97]. To date, EYA1 and EYA4 have been implicated in congenital cardiovascular defects. In particular, very recent genetic studies indicate that copy number variants found in patients with oculo-auriculo-vertebral spectrum and congenital heart defects involving components of the PAX-SIX-Eyes absent-Dachshund network (PSEDN) may contribute to this morphogenetic disorder [100]. Rare EYA1 mutations have been found in patients with conotruncal defects [101], suggesting that EYA1 may play a role in the second heart field. In one case, an EYA1 mutation was associated with cardiofacial syndrome with patent ductus arteriosus [102]. Further supporting a role of EYA1 in early cardiovascular development, genetic deletion of both Six1 and Eya1 in mice recapitulated most features of human del22q11 syndromes including craniofacial, cardiac outflow tract, aortic malformations, and ventricular septal defects [103].

In contrast, the majority of EYA4 gene mutations identified so far are associated with non-syndromic hearing loss and only a subset of EYA4 gene mutations has been found to cause autosomal dominant syndromic hearing loss with dilated cardiomyopathy [104, 105]. It has been proposed that early truncations of the N-terminal variable regions, but not of only the C-terminal EYA domain, lead to syndromic hearing loss with cardiomyopathy, but this does not appear to apply to all cases [106]. So far, EYA4’s function as a transcriptional co-activator appears to be the primary role that is disrupted in EYA4 mutation-associated cardiomyopathy. In mice, overexpression of mutant EYA4 protein resulted in dilated cardiomyopathy similar to the human phenotype [107]. These data suggest that the EYA4/SIX1 complex acts as a transcriptional repressor of p27, which results in the development of hypertrophy. It is thought that truncation of EYA4 disrupts the transcriptional inhibition of p27^kip1, leading to subsequent inhibition of HDAC2 phosphorylation. To which extent EYA4’s phosphatase function is involved in the maintenance of normal cardiac function has not yet been explored.

To date, neither EYA2 nor EYA3 mutations have been identified in patients with congenital cardiovascular defects. The role of EYA3 in the heart appears to be small. One mouse study showed that deletion of Eya3 leads to only minor changes such as electrocardiographic abnormalities, suggesting that Eya3 does not play a significant role during cardiovascular development [108]. In the adult mouse heart, EYA2 is thought to regulate physiological hypertrophy induced by swimming exercise [109] as well as pathological hypertrophy due to TAC [110, 111]. Importantly, an EYA2 mutation in which the phosphatase activity is thought to be disrupted, was still able to inhibit phenylephrine-induced cardiomyocyte hypertrophy, suggesting that phosphatase activity of Eya2 is not essential for its function in cardiac remodeling [110]. But, Eya PTP activity is clearly important in various cellular processes including developmental angiogenesis (reviewed in [112]). No doubt future investigations will shed further light on the function of Eya PTP activity in the heart.

5. Conclusions and outlook

At first glance, it could appear that PTPs may be our friends during cardiac development, but become enemies during disease. But upon closer inspection, none of the PTPs investigated to date would fit that description. Under chronic pressure overload, reduced PTP1B or LMPTP activity results in cardioprotective effects, indicating that these two PTPs play detrimental roles in this context. It is possible that inhibition of other PTP such as SHP2 may also protect against hypertrophy and contractile dysfunction after TAC, but this has not yet been thoroughly tested. Similarly, non-selective PTP inhibition as well as specific targeting of PTP-PEST, PTP1B, SHP1, or neuronal PTPRσ confers protective effects against I/R-induced injury in mice. However, we cannot infer from these data that all or most PTPs play detrimental roles in this context, as other PTPs such as PTPN4 and PTPRG protect against hypoxia-induced apoptosis / oxidative stress and support microvascular perfusion. In contrast, PTPs such as SHP2 and Eya1/4 are essential during cardiovascular development, whereas PTP1B or LMPTP appear to play no or only minor roles. Therefore, PTPs do not appear to serve general beneficial or detrimental roles, but should rather be seen as ‘frenemies’ since individual PTPs have highly context-specific and unique functions in the heart.

It still remains unclear whether these are the only PTPs that regulate essential processes in cardiac homeostasis and disease. Clearly, numerous other PTP have not yet been thoroughly investigated. As it is not clear how many and which PTPs have unique versus overlapping functions, studies have focused on one PTP at a time. In addition, examination of pathway activation / deactivation may not be sufficient to understand the roles of PTP as fine-tuners of signalling. We now appreciate that protein kinases may primarily control signal amplitude, whereas PTP may rather control rate and duration of the resulting responses [113, 114]. To fully understand how protein kinases and phosphatases act in coordination, new methodologies are needed to better detect PTP substrates and assess activity levels. For example, a novel fluorescent probe allows monitoring of PTP-PEST activity [115]. Not to forget, some PTP exhibit not only phosphatase-dependent, but also phosphatase-independent functions [72, 116], which require more complex experimental designs to assess. Another critical challenge in the field lies in the still only partially understood roles of PTP-associated adaptor and interacting proteins that further complicate the puzzle. For example, protein tyrosine phosphatase interacting protein 51 (PTPIP51) has been found to regulate ischaemia-reperfusion processes [117], which are difficult to dissect from PTP-specific effects. Similarly, scaffolding proteins such as Shc1 direct the temporal flow of signalling information by recruiting Ptpn12 [118]. Last but not least, protein tyrosine and serine / threonine phosphorylation are closely linked, and often depend on each other. For example, tyrosine phosphorylation of protein kinase C δ (PKCδ) by src leads to autophosphorylation of a threonine in its activation loop, thus enhancing PKCδ’s enzymatic activity [119]. Further complicating the picture, DUSP dephosphorylate both serine / threonine as well as tyrosine residues and thereby regulate targets such as various mitogen-activated protein kinases (MAPKs) [120–122]. DUSP play critical roles in cardiac physiology as well [123], therefore we need to develop better tools to investigate highly complex multidimensional signalling networks.

In summary, PTPs promise to have great potential as future therapeutic targets. Importantly, PTPs are powerful regulators of complex signalling networks and intervention at their level would leave essential signalling cascades intact while allowing modulation of pathway activation. Due to overlapping functions, multiple cardioprotective processes could possibly be controlled by a single inhibitor. Furthermore, substantial progress in recent years regarding the development of small molecule compounds to target for example the PTP SHP2 [124] is likely to accelerate progress towards fully translating our increasing insights into cardiac PTP function into the clinic.

Abbreviations

CaMKIIδ

calcium/calmodulin-dependent protein kinase II δ

DUSP

dual-specificity phosphatase

ERK

extracellular signal-regulated kinase

EYA

eyes absent

FASR

FAS receptor

FCP1/SCP

transcription factor IIF associating component of c-terminal domain phosphatase/small c-terminal phosphatase

HIF-1α

hypoxia-inducible factor-1α

HDAC

histone deacetylase

H/R

hypoxia / reoxygenation

iPSCs

induced pluripotent stem cells

I/R

ischaemia / reperfusion

JAK/STAT

janus kinase / signal transducer and activator of transcription

LDLR

low density lipoprotein receptor

LMPTP

low molecular weight protein tyrosine phosphatase

MAPK(s)

mitogen-activated protein kinase(s)

MCK

low molecular weight protein tyrosine phosphatase

MI

myocardial infarction, in experimental models coronary artery ligation without reperfusion

MTOR

mammalian target of rapamycin

NADPH

nicotinamide adenine dinucleotide phosphate

NFAT

nuclear factor of activated T-cells

NS

Noonan Syndrome

NSML

Noonan Syndrome with Multiple Lentigines

PI3K

phosphoinositide 3-kinase

PKA

protein kinase A

PKCδ

protein kinase C δ

PSEDN

PAX-SIX-Eyes absent-Dachshund network

PLCβ

phospholipase C β

PTK(s)

protein tyrosine kinase(s)

PTP(s)

protein tyrosine phosphatase(s)

PTP1B

protein tyrosine phosphatase 1B

PTPIP51

protein tyrosine phosphatase interacting protein 51

PTPN4

protein tyrosine phosphatase non-receptor type 4

PTPRG

protein tyrosine phosphatase receptor type G

PTPσ

PTP receptor σ

PZR

protein zero-related

ROS

reactive oxygen species

SHP1

src homology region 2 domain-containing phosphatase 1

SHP2

src homology 2 domain-containing phosphatase 2

TAC

transverse aortic constriction

TGFβ

transforming growth factor β

TNFR1

tumor necrosis factor 1

TTC

triphenyl tetrazolium chloride

TUNEL

terminal deoxynucleotidyl transferase dUTP nick end labeling

VEGF

vascular endothelial growth factor