Complex Roles of PTPN11/SHP2 in Carcinogenesis and Prospect of Targeting SHP2 in Cancer Therapy

Abstract

The non-receptor tyrosine phosphatase SHP2 has been at the center of cell signaling research for three decades. SHP2 is required to fully activate the RTK-RAS-ERK cascade, although the underlying mechanisms are not completely understood. PTPN11, coding for SHP2, is the first identified proto-oncogene that encodes a tyrosine phosphatase, with dominantly activating mutations detected in leukemias and solid tumors. However, SHP2 has been shown to have pro- and anti-oncogenic effects, and the most recent data reveal opposite activities of SHP2 in tumor cells and microenvironment cells. Allosteric SHP2 inhibitors show promising anti-tumor effects and overcome resistance to inhibitors of RAS-ERK signaling in animal models. Many clinical trials with orally bioactive SHP2 inhibitors, alone or combined with other regimens, are ongoing for a variety of cancers worldwide, with therapeutic outcomes yet unknown. This review discusses the multi-faceted SHP2 functions in oncogenesis, preclinical studies and clinical trials with SHP2 inhibitors in oncological treatment.

INTRODUCTION

SHP2, a protein tyrosine phosphatase (PTPase) that contains two Src homology 2 (SH2) domains, was discovered by several groups independently in the early 1990s (Feng & Pawson, 1994). The SH2 domain is a structural motif that mediates protein-protein interaction by recognizing phosphorylated tyrosyl residue (pTyr) on receptor tyrosine kinases (RTKs) and other membrane or cytoplasmic proteins (Koch et al, 1991). SHP2 has been shown to act downstream of various RTKs and cytokine receptors to regulate multiple signaling pathways. In particular, SHP2 plays a positive role in signal relay through the RTK-RAS-ERK pathway (Feng, 1999; Neel et al, 2003), and SHP2 is identified as the first oncogenic PTPase, with inherited or somatic gain of function (GOF) mutations detected in PTPN11 (coding for SHP2) in genetic diseases, leukemias and solid tumors (Chan et al, 2008; Chan & Feng, 2007; Tartaglia & Gelb, 2005). However, more recent studies revealed complex roles of SHP2 in tumor cells and microenvironment cells, especially in the liver (Chen et al, 2021). SHP2 has been at the center of research on cell signaling in health and disease over three decades. Currently, SHP2 is a prominent anti-tumor drug target in pharmaceutical industry. With many excellent reviews of SHP2 in the literature, this article mainly focuses on the newly deciphered SHP2 functions in carcinogenesis and on targeting SHP2 in cancer therapy. We provide candid views on a few critical, unsolved, or controversial issues, to facilitate healthy debate.

SHP2 STRUCTURE AND REGULATION

SHP2 is a broadly expressed cytoplasmic enzyme in various cell types, as opposed to the close relative SHP1 (gene PTPN6) whose expression is restricted to hematopoietic and epithelial cells. The orthologues of SHP2 are corkscrew (Csw) in Drosophila (Perkins et al, 1992), and Ptp-2 in C. elegans (Gutch et al, 1998). SHP2 is a protein of approximately 65 kDa that contains two tandem SH2 domains (N-SH2 and C-SH2), a typical PTPase domain and a proline rich C-terminal tail region with putative regulatory tyrosine residues (Figure 1). Like many other enzymes operating in intracellular signaling, SHP2 adopts an auto-inhibitory mechanism, which is relieved upon activation (Feng, 1999; Neel et al., 2003). This was first illustrated by biochemical data that a short peptide containing a SHP2-binding pTyr site (modeled on PDGFR^Y1009) drastically stimulated the phosphatase activity in vitro (Lechleider et al, 1993). A tethered peptide with two IRS-1-derived phosphorylation sites that bind the N-SH2 and C-SH2 domains simultaneously showed even more potent stimulation of the enzyme (Pluskey et al, 1995).

Figure 1. The structural domains and regulatory mechanism of Shp2.

A) SHP2 possess two SH2 domains and a PTPase catalytic domain and a C-terminal tail; it interacts with dimerized (and activated) RTKs with a pTyr site for association with the N-SH2 and C-SH2 domains or cytoplasmic scaffolding molecules with PH domain and two pTyr sites for simultaneous binding to the two SH2 domains. B) At the basal status, SHP2 assumes an auto-inhibitory conformation, which is disrupted by occupation of the two SH2 domains by pTyr-containing partners. Likewise, this enzyme is constitutively activated by mutations that disrupt the intramolecular interactions. C. Strategies of SHP2 inhibition: inactivation of the enzyme by orthosteric or allosteric inhibitors; induction of SHP2 degradation by PROTAC; disruption of SHP2 interaction with partners in cell signaling.

The crystal structure of SHP2 (containing amino acids 1–527, lacking 66 AAs at the C-terminal tail) reveals the molecular basis for the auto-regulatory mechanism (Hof et al, 1998). In its basal status, without a pTyr-containing partner, the N-SH2 domain physically interacts with and blocks the active site, and meanwhile the PTPase domain triggers a contortion of the N-SH2 domain, disrupting its phosphopeptide-binding cleft. Binding to a bisphosphorylated ligand relieves the intra-molecular interactions and activates the enzyme. Thus, the N-SH2 domain serves as a conformational switch, while the C-SH2 domain does not have a direct role in activation but contributes binding specificity and energy, such that dual steric inhibition of SHP2 exists in its closed or “off” state (Hof et al., 1998). This dualism was recently amended by a third conformational semiactive state, which occurs upon binding to monophosphorylated, instead of bisphosphorylated SHP2-binding partners (Tao et al, 2021). The biochemical and structural data on SHP2 regulation was convicted by identified GOF mutations of PTPN11 in developmental and malignant diseases, resulting in constitutively activated SHP2 by disruption of the auto-inhibitory mechanism (Chan & Feng, 2007; Tartaglia & Gelb, 2005). Finally, this “turn on and off” mechanism of SHP2 was vividly illustrated by allosteric inhibitors that bind the interface of N-SH2, C-SH2, and PTPase domains, thereby locking the enzyme in its inactive conformation (Chen et al, 2016).

SHP2 IN CELL SIGNALING

Immediately after the initial discovery, SHP2 was found to bind ligand-activated RTKs, such as receptors for platelet derived growth factor (PDGF) and epidermal growth factor (EGF), through pTyr-SH2 interactions (Feng & Pawson, 1994). Cytokine receptors, SHPS1 or SIRPα (Fujioka et al, 1996; Kharitonenkov et al, 1997), and PZR (Zhao & Zhao, 1998) represent other transmembrane proteins that bind and regulate SHP2 activities. In addition, SHP2 is shown to associate with cytoplasmic adaptor/scaffolding molecules, such as Grb2-associated binders (GABs, acting downstream of EGFR and other RTKs), insulin receptor substrates 1–4 (IRS1–4), and FGF receptor substrates (FRS), in various signaling pathways (Lai et al, 2004; Neel et al., 2003; Qu, 2000). These cytoplasmic signaling molecules share two pTyr motifs for association with SHP2’s two SH2 domains (Figure 1A).

The findings of SHP2 docking on activated RTKs initially suggested a putative role of SHP2 in downregulating the activated RTK signals, in accordance with a general view on phosphatases counteracting with kinases. In contrast, both genetic and biochemical experiments demonstrated surprisingly that SHP2 acts to amplify, rather than inhibit, RTK signals in stimulating the RAS-ERK pathway (Feng, 1999; Neel et al., 2003). Detailed kinetics studies suggest that SHP2 deficiency leads to incomplete activation of RAS, RAF, MEK and ERK (Shi et al, 1998; Shi et al, 2000). Several mechanisms have been proposed to explain a SHP2 function in enhancing RTK-RAS-ERK signaling, through dephosphorylating and inactivating negative regulators of the pathway (Figure 2). SHP2 may dephosphorylate the RasGAP-binding site on RTK, thereby inhibiting RasGAP recruitment to the plasma membrane for RAS downregulation (Agazie & Hayman, 2003; Cleghon et al, 1998; Ekman et al, 2002). The membrane protein PAG/Cbp recruits Csk kinase that phosphorylates a Tyr residue at the C-terminal tail of c-Src for auto-inhibition, and dephosphorylation of PAG/Cbp by SHP2 could alleviate the inhibitory mechanism for Src kinase that promotes RAS signaling (Zhang et al, 2004). DOK1 is a protein at 62 kDa with pleckstrin homology (PH) and phosphotyrosine binding (PTB) domains, which binds RasGAP via pTyr362 and pTyr398. Dephosphorylation of these two sites by SHP2 reduced recruitment of RasGAP for RAS inhibition (Zhu et al, 2022). SHP2 may dephosphorylate RAS at Tyr³² directly, which promotes RAS-RAF association but impairs RasGAP binding and GTP hydrolysis (Bunda et al, 2015). The Sprouty (Spry) proteins negatively regulate RAS-ERK signaling, by interrupting the GRB2-SOS complex formation (Mason et al, 2006), and SHP2 may dephosphorylate Sprouty, potentiating RAS activity (Hanafusa et al, 2004). These logically proposed mechanisms are not mutually exclusive; they may operate in a cell type-dependent manner. More rigorous experimental validation is required for the proposed mechanisms (Figure 2).

Figure 2. Proposed mechanisms of SHP2 function in promoting RTK-RAS-ERK signaling.

SHP2 may dephosphorylate and inactivate these negative regulators of the pathway. a) The RasGAP-binding pTyr site on RTKs is responsible for its recruitment to the RTK complex and membrane relocation and access to RAS. b) The transmembrane protein PAG/Cbp acts to recruit of CSK, a negative regulator of SRC family kinases. c) Two sites, pTyr³⁶² and pTyr³⁹⁸, on DOK1 is involved in association with RasGAP. d) pTyr³² on RAS inhibits RAS-RAF interaction but enhance RasGAP binding. e) pTyr-containing Sprouty (Spry) impairs GRB2-SOS complex formation.

Large-scale phosphoproteomics has been performed to systematically search for pTyr sites and proteins that are potential SHP2 substrates in cell lines following stimulation with EGF, PDGF, FGF or IGF1, using chemical SHP2 inhibitor, SHP2 knockout cells or expressing SHP2 mutants (Batth et al, 2018; Vemulapalli et al, 2021; Zhu et al., 2022). Collectively, these searches validated previously recognized SHP2 substrates downstream of EGFR or PDGFR. One group identified PLCγ2, occludin, and ARHGAP35 as direct SHP2 substrates in breast cancer cells MDA-MB-468 following EGF stimulation (Vemulapalli et al., 2021). Another study with a SHP2-trapping mutant yielded 18 direct SHP2 substrates in EGF-treated HEK293 cells, including known substrates, such as GAB1–2, SIRPα and DOK1, as well as proteins involved in focal adhesion, cell communication and mRNA processing, splicing and translation (Zhu et al., 2022). SHP2 apparently has a multitude of substrates, and these large phosphoproteomics datasets will serve as valuable resources for further dissection of SHP2 functions and mechanisms. Some of the pTyr sites caught in large-scale screening may not be direct SHP2 targets, but instead represent indirect changes of pTyr profiles secondary to SHP2 inhibition or deletion. One challenging issue is to connect the biochemical activities to the biological functions.

Another interesting finding is ironically the tyrosine-phosphorylation of SHP2 by RTKs and other kinases (Feng, 1999; Feng et al, 1993; Neel et al., 2003; Vogel et al, 1993). Tyrosine phosphorylation may modulate the PTPase activity or turn it into a binding partner for SH2-containing signaling molecules, such as GRB2 downstream of PDGFR, which led to a proposal that SHP2 acts as an adaptor connecting RTKs to GRB2-SOS, in promoting RAS activation (Bennett et al, 1994; Li et al, 1994). A recent phosphoproteomics study confirmed that Y546 and Y584 at the C-terminal tail were phosphorylated in PDGFbb-treated cells (Batth et al., 2018). However, most experimental data indicated a requirement of SHP2’s catalytic activity in upregulating RAS-ERK signaling, with the adaptor function of SHP2 remaining to be defined unequivocally.

In addition to the central RAS-ERK pathway, SHP2 has been shown to modulate other intracellular signaling cascades, including the PI3K, JAK/STAT, JNK and NF-κB pathways (Lai et al., 2004). SHP2 plays a role in focal adhesion and cell motility, via FAK dephosphorylation (Manes et al, 1999; Tsutsumi et al, 2006; Yu et al, 1998). SHP2 may regulate these signaling pathways directly or indirectly.

SHP2 IN GENETIC DISEASES

The SHP2 function in upregulating RAS-ERK signaling is solidified by uncovering germline PTPN11 mutations in Noonan syndrome (NS, MIM: PS163950) (Tartaglia et al, 2001), and Noonan syndrome with multiple lentigines (NSML, MIM: PS151100, also known as LEOPARD syndrome) (Digilio et al, 2002; Legius et al, 2002). NS is an autosomal dominant genetic disorder (1 in 1,000–2,500 births) featured by facial dysmorphia, short height, impaired hemostasis, skeletal malformations and cardiopathies. NSML is rare (1 in 100,000 births) and defined by features comprising lentigines, hearing loss, ECG conduction abnormalities, ocular hypertelorism and pulmonic stenosis (Gelb & Tartaglia, 1993). NS-causative PTPN11 mutations enhance catalytic activity by abrogating the autoinhibitory interaction between SHP2’s N-SH2 and catalytic domains (Figure 1B) (Chan & Feng, 2007; Tartaglia & Gelb, 2005). Mutations in PTPN11 occur in 50 percent of NS, while large-scale screens have identified mutated genes involved in the RAS pathway, including KRAS, NRAS, SOS1, NF1, SPRED2, RAF1, BRAF and MAP2K1, in those without PTPN11 mutations as well as in other NS-like syndromes (Tartaglia et al, 2022; Tartaglia & Gelb, 2005). In aggregate, these genetic alterations result in RASopathies, with clinically overlapped developmental disorders driven by excessive RAS-ERK signaling.

Like NS mutations, NSML displays autosomal dominant features (Ogata & Yoshida, 2005; Sarkozy et al, 2008). Controversial data and views were reported on whether NSML-associated PTPN11 mutations represent GOF or LOF (loss-of-function) mutations (Kontaridis et al, 2006; Yu et al, 2013). A comprehensive analysis of NSML-related SHP2 mutants demonstrated that these mutations, while attenuating the phosphatase activity, also weakened the N-SH2 and PTPase domains’ interaction. The mutant molecules are more readily activated by pTyr-containing ligands, resulting in GOF effect with enhanced ERK activation (Yu et al, 2014). In agreement with this notion, targeted deletion of amino acids 46–110 in the N-SH2 domain generated a biologically inert (or LOF) molecule (even with an intact catalytic domain), leading to defective activation of RAS-ERK signaling (Qu et al, 1997; Saxton et al, 1997; Shi et al., 1998; Shi et al., 2000). Therefore, the SHP2 function in cell signaling requires proper targeting to its biological substrates via the N-SH2 domain. More recent data suggest that the disease-associated SHP2 mutants can recruit and activate the wild-type enzyme in liquid-liquid phase separation to enhance ERK activation (Zhu et al, 2020).

Animal models for NS- and NSML-associated SHP2 mutations were characterized in fruitfly, zebrafish, frog, and mouse (Solman et al, 2022b). In Drosophila, expressing Y279C and T468M (NSML) or N308D (NS) developed ectopic wing veins and rough eye phenotypes, reminiscent of elevated RAS signaling (Das et al, 2021). A knockin zebrafish line with a D61G mutation recapitulated many aspects of NS phenotypes, including expansion of myeloid cell lineages, due to hyperproliferation of hematopoietic progenitor cells (Solman et al, 2022a). Several knockin mouse lines have been characterized, especially the NS-related D61G and N308D and NSML-relevant Y279C and T468M mutants (Araki et al, 2004; Marin et al, 2011; Solman et al., 2022b; Tajan et al, 2014). These mutant lines recapitulate developmental, cardiac and hematopoietic disorders in NS and NSML, respectively, serving as excellent models for mechanistic and pathogenic dissection of RASopathies (Solman et al., 2022b).

SHP2 IN CANCER

SHP2 in hematopoiesis and hematopoietic malignancies

Earlier clinical studies found association of NS with myeloproliferative disorders, especially juvenile myelomonocytic leukemia (JMML) (Bader-Meunier et al, 1997; Choong et al, 1999). Indeed, a subset of NS-related germline PTPN11 mutations have a predisposition to developing JMML, and somatic PTPN11 mutations were detected in childhood neoplasias, such as JMML, childhood myelodysplastic syndromes (MDS), acute myeloid leukemia (AML), and B-ALL (Kanumuri et al, 2022; Tartaglia et al, 2003). The PTPN11^D61G knock-in mouse recapitulated the NS phenotype and developed mild myelodysplastic disorders (Araki et al., 2004). In PTPN11^D61G knock-in mouse, hematopoietic stem cell cycling was accelerated as a consequence of aberrantly enhanced cytokine signaling (Xu et al, 2010). Of note, there is a different spectrum of germline mutations in NS patients with SHP2 activity mildly upregulated, and somatic leukemia-associated mutations resulting in stronger activation of the enzyme (Chan & Feng, 2007; Tartaglia et al., 2022; Tartaglia & Gelb, 2005).

The effect of SHP2 mutants in leukemogenesis is in agreement with prior studies that unveiled a positive role of SHP2 in hematopoietic progenitor cells and hematopoiesis in mouse models (Chan et al, 2011; Chan et al, 2003; Qu et al, 2001; Qu et al, 1998; Zhu et al, 2011). More recently, a requirement of SHP2 for BCR-ABL1 induced hematologic neoplasia has been purported (both in CML and B-ALL); Ptpn11 deletion suppressed initiation of CML-like myeloproliferative disease and of BCR-ABL1 positive B-ALL (Gu et al, 2018). A tumorigenic effect of dominant Ptpn11 mutations in the bone marrow microenvironment was reported (Dong et al, 2016). Specifically, GOF Ptpn11 mutations in mesenchymal progenitors and osteoprogenitors may cause excessive production of MIP-1α that recruit monocytes into the HSC niche, resulting in myeloproliferative neoplasm (MPN). Interestingly, deletion of SHP2 neutralized MPN driven by PTEN deficiency as expected, but removal of both SHP2 and PTEN also triggered lethal anemia in mice, indicating distinct signaling mechanisms in control of myelopoiesis and erythropoiesis (Zhu et al, 2015). The cell type-specific signal cross-talks in blood cell lineages raise caution on targeting SHP2 in treatment of hematologic diseases.

SHP2 in solid tumors

Despite the robust leukemogenic effect, GOF mutations of PTPN11 have been detected at very low frequencies in solid tumors (Bentires-Alj et al, 2004; Chan et al., 2008; Martinelli et al, 2006). NS patients with congenital heart disease may have higher risk for neuroblastoma (Cotton & Williams, 1995), and PTPN11 mutations are found in recurrent glioblastoma (GBM) (Wang et al, 2016a), suggesting a putative role of SHP2 in brain tumorigenesis. However, clinical data analyses show that elevated SHP2 expression may increase the risk of non-small-cell lung cancer (NSCLC), gastric and cervical cancer, while reducing the overall survival (OS) of NSCLC (Li et al, 2023). SHP2 over-expressions are steadily observed in several widely used breast cancer cell lines and in 72% of infiltrating ductal carcinoma of the breast correlated with tumor grades and metastasis (Zhou et al, 2008). A tissue microarray analysis showed SHP2 overexpression in 46% of 1401 breast cancer specimens with significantly worse OS (Muenst et al, 2013). Thus, in parallel to GOF mutations in leukemias, aberrant expression of SHP2 may also be implicated in pathogenesis of solid tumors.

Another carcinogenic mechanism is abnormally elevated expression of SHP2 partners, such as GAB2 in breast cancer, owing to focal amplification of chromosomal region 11q14.1, independent of 11q13.2 (CCND1) (Bocanegra et al, 2010). Overexpression of GAB2 enhanced proliferation in breast cancer cells associated with excessive ERK activation, and aggravated mammary tumorigenesis in HER2 transgenic mice (Bentires-Alj et al, 2006). Cytoxin-associated antigen A (CagA) is a virulence factor of Helicobacter pylori, a micro-aerophilic spiral-shaped bacterium that colonizes the stomach in half of the world’s human populations, causing chronic atrophic gastritis, peptic ulcer and gastric adenocarcinoma (Hatakeyama, 2004). CagA is a protein of 120–145 kDa that, upon entering gastric epithelial cells, gets tyrosine-phosphorylated on EPIYA motifs by SRC kinases, which induces association with and activation of SHP2 (Higashi et al, 2002). Thus, CagA can act as a scaffolder, like GAB2, to activate SHP2 excessively. This notion was supported by experimental data that transgenic expression of CagA drove gastrointestinal and hematopoietic malignancies (Hatakeyama, 2014; Ohnishi et al, 2008). However, NS patients do not display a higher risk of gastric cancer. Thus, it is unlikley the aberrant SHP2 activation by CagA but the CagA-induced cell reprogramming, epigenetic changes, and chronic gastric inflammation that contribute to gastric cancer in the CagA-positive H. pylori-infected population.

Although PTPN11 mutations are rare events, a functional role of SHP2 was demonstrated experimentally in various solid tumors. SHP2 was found essential for KRAS-mediated pulmonary and pancreatic carcinogenesis in murine models, where genetic deletion of PTPN11 profoundly inhibited tumor development (Ruess et al, 2018). Knocking down SHP2 in established breast tumor cells inhibited their growth and metastasis both in HER2-positive and triple-negative breast cancer xenografts via c-MYC and ZEB1 (Aceto et al, 2012). SHP2 may promote triple-negative breast cancer development through stabilizing β-catenin (Martin & Agazie, 2021). GAB2 deficiency ameliorated mammary tumors in mice driven by Neu oncogene (Bentires-Alj et al., 2006), and another study showed that deleting GAB2 suppressed Neu-induced mammary tumor metastasis, without a profound impact on tumor progression (Ke et al, 2007). SHP2 was shown to play a critical role in diffuse type of gastric carcinoma (Nagamura et al, 2021), ovarian cancer (Hu et al, 2017), and prostate cancer (Zhang et al, 2016). SHP2 inhibition was detrimental to glioma cells defined by INK4A/ARF deficiency and diminished PDGFRα-driven tumorigenesis dependent on the PI3K/AKT/mTOR pathway (Liu et al, 2011).

SHP2 in liver cancer and beyond

The complex roles of SHP2 in oncogenesis have been extensively investigated in liver cancer, especially hepatocellular carcinoma (HCC). Removal of SHP2 or its partner GAB1 indeed suppressed hepatocyte proliferation and ERK activation in regenerating mouse livers following partial hepatectomy (Bard-Chapeau et al, 2006). Consistently, SHP2 deletion in hepatocytes (SHP2^ΔHep, SHP2^fl/fl:Alb-Cre) abrogated HCC driven by RTK (MET) in combination with activated PIK3CA or β-catenin mutant (Liu et al, 2018), underlining an essential role of SHP2 in RTK-elicited oncogenic signaling. Wild-type SHP2, but not a catalytically inactive SHP2^C463S mutant, rescued HCC development in SHP2^ΔHep mice (Liu et al, 2023). Of note, dissecting a positive role of SHP2 in proliferation of hepatocytes and cancer cells has uncovered a novel cell-cell communication mechanism via CD133⁺ vesicles (named intercellsome) induced under intracellular signal deficit (Kaneko et al, 2022; Kaneko et al, 2023), which requires further characterization.

Surprisingly, SHP2^ΔHep mutant mice spontaneously developed hepatocellular adenoma with age and displayed higher sensitivity to chemical carcinogenesis, revealing a tumor suppressor role of SHP2 (Bard-Chapeau et al, 2011). Further mechanistic dissection revealed a SHP2 function in bile acid biosynthesis regulated by FXR and FGF15 (FGF19 in human), with cholestasis developed in SHP2^ΔHep mice (Li et al, 2014). Combined deletion of SHP2 and PTEN in hepatocytes aggravated hepatocarcinogenesis driven by non-alcoholic steatohepatitis (NASH) (Luo et al, 2016). Moreover, concurrent removal of SHP2 and IKKβ, two well-known oncoproteins, triggered spontaneous HCC development, owing to deregulated circadian rhythm and numerous hepatic disorders (Hanley et al, 2021). Consistent with the animal data, either elevated or impaired SHP2 expressions were detected in tumors, relative to surrounding tissues, in human HCC specimens, suggesting pro- and anti-tumoral functions (Bard-Chapeau et al., 2011; Han et al, 2015; Luo et al., 2016).

The mechanisms underlying paradoxical roles of SHP2 in liver tumorigenesis were deciphered at least in part in a MYC-induced HCC model by single cell RNA-sequencing (Chen et al., 2021). Although HCC was drastically exacerbated by SHP2 deficiency in hepatocytes, tumors were grown selectively from sporadic SHP2-positive hepatocytes transfected by MYC in SHP2^ΔHep livers, indicating a cell-intrinsic requirement of SHP2 for MYC-driven oncogenesis. Biochemically, SHP2 is necessary for an intact RAS-ERK pathway to sustain MYC stability, although SHP2 loss generated a tumor-conducive hepatic microenvironment. These findings reinforced an essential role of SHP2 in intracellular oncogenic signaling through the RTK-RAS-ERK-MYC pathway and also revealed a profound impact of the microenvironment on HCC development. Similar to the bi-directional SHP2 functions, several groups reported an anti-oncogenic effect of pro-oncogenic molecules in the liver, including MET, JNK, CTNNB1 and IKKβ (Feng, 2012). Despite frequent hyperactivation of AKT kinases in human cancers, deleting AKT1/2 induced spontaneous HCC development associated with FoxO-dependent hepatic injury and inflammation (Wang et al, 2016b).

In agreement with SHP2’s anti-tumor effect in the liver, mice with Ptpn11 ablated in cathepsin K-expressing cells (Ptpn11^fl/fl:Ctsk-Cre) displayed excessive proliferation of mesenchymal progenitors and developed features similar to metachondromatosis (Yang et al, 2013), a rare inherited bone disorder with heterozygous inactivating PTPN11 mutations (Bowen et al, 2011; Sobreira et al, 2010). Despite an inhibitory effect on invasive cell proliferation by SHP2 silencing in intestinal epithelial cells (IECs), SHP2 removal in IECs drove development of colitis-associated adenocarcinoma in aged mice associated with persistent activation of β-catenin, NF-κB and STAT3 (Gagne-Sansfacon et al, 2016). These findings document anti-oncogenic roles of SHP2 in the bone and gut.

SHP2 in tumor microenvironment

How SHP2 deficiency induces tumor-prone microenvironment is complex and apparently organ-specific, which remains to be fully elucidated. In MYC-transfected SHP2^ΔHep liver (Chen et al., 2021), an immunosuppressive tumor microenvironment is likely triggered by altered expression of cytokines and chemokines, such as Lect2, CCL2, CCL9, CCL17 and MIF, which influenced recruitment of innate and adaptive immune cells to the tumor site. The tumor-initiating stage is featured by changes of classical dendritic cells (cDCs), Kupffer cells, macrophages, neutrophils and B cells in SHP2-deficient livers, accompanied by defective expression of M1 macrophage markers and phagocytosis-related genes, leading to defective clearance of MYC-transformed tumorigenic cells. SHP2 loss also triggered local increase of basal Wnt/β-catenin signaling, which is augmented by MYC transfection, constituting a cooperative tumor cell and environmental interaction in accelerating tumor progression (Chen et al., 2021). Cell type-specific SHP2 deletion also unveiled functional heterogeneity of Kupffer cells and monocyte-derived macrophages (MDM) in liver tumorigenesis (Du et al, 2023). SHP2 ablation in Kupffer cells induced cell apoptosis, which triggered hepatic recruitment of MDMs with pro-tumoral properties.

With regard to potential roles in T lymphocytes, SHP2 was first found to interact with CTLA-4 (Marengere et al, 1996), which might be involved in negative regulation of TCRζ signaling (Lee et al, 1998). Another study suggests that PD-1-associated SHP2 preferentially dephosphorylates the co-receptor CD28, rather than TCR (Hui et al, 2017). The interaction of SHP2 and PD-1 is mediated by cooperative interaction of SHP2’s N- and C-SH2 domains with the immune receptor tyrosine–based inhibitory motifs (ITIMs) and immune receptor tyrosine–based switch motifs (ITSMs) on PD-1 (Marasco et al, 2020). However, the ITIM and ITSM motifs bind both SHP1 and SHP2, which may compound the biological significance of these biochemical data. Targeted SHP2 ablation (mediated by Lck-Cre) unveiled an overall positive effect of SHP2 in pre-TCR and TCR signaling and T lymphopoiesis, especially CD4 T cells (Nguyen et al, 2006). Therefore, the widely expressed enzyme SHP2 does have complex roles in immune cells, which raises caution on SHP2 as a drug target in oncological treatment. Although deletion of SHP2 neutralized excessive myeloid cell proliferation driven by PTEN loss, the mutant mice with combined SHP2 and PTEN deficiencies displayed lethal anemia (Zhu et al., 2015). However, these genetic data, while deciphering positive and negative effects of SHP2 in various cell types, may have limited value in predicting outcomes of pharmaceutically targeting SHP2. The therapeutic efficacy and benefits may largely depend on the overall responses of tumor cells and immune cells in the microenvironment (Fedele et al, 2021; Liu et al., 2023).

DEVELOPMENT AND APPLICATION OF PHARMACEUTICAL SHP2 INHIBITORS

SHP2-inhibitory compounds and strategies

Earlier efforts were mainly devoted to orthosteric inhibitors that target the catalytic center of SHP2 (Figure 1C); these compounds were found to have limited value, due to lack of specificity and poor membrane permeability. A turning point was the discovery of an allosteric inhibitor SHP099, with high selectivity for SHP2 over other PTPases (Chen et al., 2016). The compound, acting as “molecular glue”, binds and stabilizes SHP2 in the auto-inhibited closed conformation, preventing its interaction with pTyr-containing partners (Figure 1C). SHP099 robustly suppressed proliferation of RTK-driven tumor cells in vitro and tumor growth in animals (Chen et al., 2016). Due to the intriguing efficacy and specificity of SHP099, a few more have been developed in this class, which target either SHP2^WT, SHP2^E76A or SHP2^F285S by binding to the “tunnel” allosteric site 1 at the C-SH2/PTP domain interface, or the “latch” allosteric site 2 at the N-SH2/PTP domain interface (Kerr et al, 2021; Song et al, 2022; Wu et al, 2021; Yuan et al, 2020).

Another new strategy is the proteolysis-targeting chimera (PROTAC) approach by decreasing protein levels instead of transiently blocking activity (Figure 1C). Specific PROTAC degraders of SHP2 have been designed, including SHP2-D26 made by tethering compound 5 to the VHL-1 ligand (Song et al., 2022; Wang et al, 2020). As a proof of concept, SHP2-D26 was shown to reduce SHP2 protein levels in cancer cells by 95 percent and exhibited robust inhibition of ERK phosphorylation and cell proliferation in esophageal cancer and AML cell lines. Given the necessity of SHP2 association with partners in cell signaling, one enticing opportunity represents the design of peptides or compounds that disrupt protein-protein interaction (Figure 1C) (Bobone et al, 2021; Fan et al, 2020; Sha et al, 2013; Song et al., 2022). Methylene blue (an FDA approved chemical) was shown to block the interaction between SHP2 and a Y²⁴⁸-phosphorylated ITSM of human PD-1, resulting in shrinkage of PD-L1 expressing tumor allografts and autochthonous lung cancers in a transgenic mouse model (Fan et al., 2020).

SHP2 inhibitors in pre-clinical studies of cancer

The development of allosteric SHP2 inhibitors has facilitated pre-clinical assessments on targeting SHP2, with encouraging results in animal tumor models. Oral gavage of SHP099 (100 mg/kg) for 24 days into nude mice markedly inhibited tumor growth in subcutaneous xenograft model using KYSE520 esophageal squamous cell carcinoma, and a daily dose at 75 mg/kg almost eradicated circulating leukemic cells in an AML model driven by FLT3-ITD (Chen et al., 2016). Likewise, oral administration of RMC-4550 robustly suppressed tumor growth of NCI-H355 and MIA PaCa-2 xenograft models in immune-deficient mice, as well as in patient-derived xenograft (PDX) models for NSCLC harboring mutations in BRAF, NF1 or KRAS (Nichols et al, 2018). Notably, these compounds displayed tolerance and safety in the animal models tested, as manifested by no weight loss.

Although ineffective in KRAS-mutant tumor cells if given alone, combination of SHP099 with MEK inhibitor AZD6244 (Selumetinib) showed a synergistic effect in subcutaneously implanted PDX model for NSCLC and pancreatic ductal adenocarcinoma (Mainardi et al, 2018; Ruess et al., 2018). Due to rapid development of adaptive resistance, MEK inhibitors showed rather limited efficacy; SHP2 inhibitors represent a means to overcome resistance to MEK-inhibition in multiple tumor models expressing wild-type or mutant KRAS (Fedele et al, 2018). Likewise, SHP2 inhibition overcomes adaptive resistance to ALK (anaplastic lymphoma kinase) inhibitors in ALK-rearranged cancers (Dardaei et al, 2018). Combinations of PI3K and SHP2 inhibitors overcame resistance to PI3K inhibition in breast cancer cells mediated by activated RTKs (Heynen et al, 2022). Combined transcriptomic, proteomic, and phosphoproteomic analyses of 43 KRAS mutant cancer cell lines identified subsets with differential drug sensitivities (Liu et al, 2021). A compensatory phosphoprotein biomarker analysis (CPBA) suggested new drug combinations with improved therapeutic potentials. For example, one combination of SHP099 and SGC0946 (inhibitor of histone H3 lysine methyltransferase, DOT1L) was particularly active in a subset of KRAS mutant cancers, in SU8686 subcutaneous xenograft model and the PANC-02–03 model. These studies are focused on disrupting alternative or compensatory activation of intracellular signaling pathways to overcome drug resistance, an area extensively explored in cancer therapy. However, usage of appropriate animal tumor models is crucial for evaluation of therapeutic means; the subcutaneous tumor model in mice, although most frequently used, may provide misleading data (Xin et al, 2022).

The therapeutic efficacy in the clinic is largely influenced by pharmaceuticals’ effects on tumor cells as well as the niche cells. In a model of CT26 colon cancer, the anti-tumor effect of SHP2 inhibitor RMC-4550 was abolished in immunocompromised RAG2-deficient mice and likewise in mice depleted of CD4 and CD8 T cells (Quintana et al, 2020). Treatment with RMC-4550 induced shifts in polarized macrophage populations, with increased M1 and decreased M2, in favor of anti-tumor immunity, and exhibited additive effect with PD-L1 blockage. In PDAC model, combined treatment of SHP099 and ARS1620 (RAS^G12C mutant inhibitor) triggered decrease of CD11b⁺ myeloid cells, especially granulocytic myeloid-derived suppressor cells and preferentially increased CD8 T cells (Fedele et al., 2021). Reprogramming the innate immune system appears to be crucial for the anti-tumor effect of pharmacological SHP2 inhibition potentially mediated by IFNβ secreted by macrophages, and suppression of the CCR5 signaling axis, in a liver cancer model (Liu et al., 2023). Combination of SHP099 and PD-L1 inhibition could overcome resistance to PD-1 directed therapies in a model of NSCLC (subcutaneously injected anti–PD-1–resistant 344SQ NSCLC adenocarcinoma cell line). A triple-combination of radiotherapy with SHP099 and anti-PD-L1 increased the M1/M2 ratio and the number of CD8 T cells with reduced Treg cells, resulting in diminished tumor metastasis and improved survival in NSCLC model resistant to anti-PD-L1 therapy (Chen et al, 2020). Therefore, despite the opposing roles of SHP2 in tumor cells and microenvironment revealed in genetic dissection, these preclinical data encourage the application of SHP2 inhibitors in clinical studies for cancers.

SHP2 inhibitors in clinical trials

As summarized in Table 1, multiple allosteric SHP2 inhibitors in the pyrazine pharmacophore class, including TNO155, RMC-4630, BBP-398, JAB-3068, JAB-3312, RLY-1971, are already in phase I, I/II or II clinical trials (Kerr et al., 2021; Song et al., 2022; Yuan et al., 2020). These orally bioactive SHP2 inhibitors are being evaluated for efficacy and safety alone or combined with EGFR inhibitor (Nazartinib or Osimertinib), KRAS^G12C inhibitor (MRTX849 or AMG-510), BRAF^V600E inhibitor (Dabrafinib), MEK inhibitor (Cobimetinib), ERK1/2 inhibitor (LTT462), Cyclin D1/CDK4/6 inhibitor (Ribociclib) or anti-PD-1 (Spartalizumab), in solid tumors under various clinical conditions or disease stages. For example, TNO155 (Novartis) is in phase I trial for advanced solid tumors (NCT03114319), and another multicenter and open-label clinical trial of RLY-1971 (Relay Therapeutic) was initiated in 2020 as a dose escalation phase I for advanced and metastatic solid tumors (NCT04252339). A combination regimen of sotorasib and SHP2 inhibitor RMC-4630 (Revolution Medicines) is being evaluated in phase 1b multicenter, open-label trial NCT04185883, for patients with locally advanced or metastatic KRAS^G12C positive solid tumors. This study achieved responses in both KRAS^G12C inhibitor-pretreated NSCLC (ORR 27%, 3/11) and KRAS^G12C-inhibitor-naïve NSCLC (ORR 50%, 3/6), while also reporting a partial response in a patient with ovarian cancer and several stable diseases in other tumor types including colorectal cancer and pancreatic adenocarcinoma. More examples of combination trials include TNO155 with ribociclib or spartalizumab (NCT04000529), RMC-4630 and cobimetinib in relapsed/refractory solid tumors and RMC-4630 and osimertinib in EGFR-mutation positive locally advanced/metastatic NSCLC (NCT03989115) and JAB-3312 together with either a PD-1 antibody or a MEK inhibitor in advanced solid tumors (NCT04720976).

Table 1.

Ongoing Clinical Trials of SHP2 Inhibitors.

Agent	Phase and additional interventions	Cancer type	Sponsor	NCT number
TNO155	Phase I, with or without EGF816 (EGFR inhibiutor)	Solid tumors	Novartis	NCT03114319
	Phase I, with spartalizumab (anti-PD-1) or ribociclib (CDK4/6 inhibitor)	Solid tumors		NCT04000529
	Phase I/II, with MRTX849 (KRAS G12C inhibitor)	Solid tumors with KRAS G12C mutation		NCT04330664
	Phase I/II, with JDQ443 (KRAS G12C inhibitor) and tislelizumab (anti-PD-1)	Solid tumors		NCT04699188
RLY-1971	Phase I	Solid tumors	Roche	NCT04252339
RMC-4630	Phase I	Solid tumors	Revolution Medicines/Sanofi	NCT03634982
	Phase I/II, with cobimetinib (MEK inhibitor) or osimertinib (EGFR inhibitor)	Solid tumors, NSCLC with EGFR mutation		NCT03989115
	Phase II, with sotorasib (KRAS G12C inhibitor)	NSCLC with KRAS G12C mutation	Revolution Medicines/Sanofi/Amgen	NCT05054725
	Phase I, with LY3214996 (ERK1/2 inhibitor)	Solid tumors with KRAS mutation	The Netherlands Cancer Institute	NCT04916236
	Phase I/II, with sotorasib (KRAS G12C inhibitor)	Solid tumors with KRAS G12C mutation	Amgen	NCT04185883
JAB-3312	Phase I	Solid tumors	Jacobio	NCT04045496
				NCT04121286
	Phase I/II, with JAB-21822 (KRAS G12C inhibitor)	Solid tumors		NCT05288205
	Phase I/II, with other agents	Solid tumors		NCT04720976
JAB-3068	Phase I/II	Solid tumors	Jacobio	NCT03565003
				NCT03518554
BBP-398	Phase I, with sotorasib (KRAS G12C inhibitor)	Solid tumors with KRAS G12C mutation	Navire	NCT05480865
	Phase I, with nivolumab (anti-PD-1)	NSCLC with KRAS mutation		NCT05375084
	Phase I	Solid tumors		NCT04528836
				NCT05621525
PF-07284892	Phase I, with or without other agents	Solid tumors	Pfizer	NCT04800822
ET0038	Phase I	Solid tumors	Etern BioPharma	NCT05525559
				NCT05354843
HBI-2376	Phase I	Solid tumors with KRAS or EGFR mutations	HUYABIO International	NCT05163028
ERAS-601	Phase I/II, with or without other agents	Solid tumors	Erasca	NCT04670679
				NCT04866134
HS-10381	Phase I	Solid tumors	Jiangsu Hansoh	NCT05378178
BPI-442096	Phase I	Solid tumors	Betta	NCT05369312
SH3809	Phase I	Solid tumors	Nanjing Sanhome	NCT04843033

Given the ubiquitous expression pattern and involvement in RTK-proximal signaling events, it is hard to predict the outcomes of targeting SHP2 in the ongoing clinical trials around the globe. Based on the high potency and tolerance in animal models with PDX and other tumor settings, it is reasonable to be optimistic for the clinic trials. However, if the SHP2 inhibitors eventually fail in phase III, it is not owing to lack of potency and selectivity, but is more likely due to treatment-associated adverse effects, which are yet to be determined. The therapeutic effects, especially survival benefit of cancer patients, are largely influenced by the management of adverse effects related to ages and overall health conditions. These important aspects are not considered sufficiently in experimental and preclinical studies during drug development.

CONCLUDING REMARKS

Given the numerous SH2-containing cytoplasmic molecules acting downstream of RTKs, it is astonishing that genetic and biochemical analyses unveil stringent requirement for Shp2 in RTK-driven cell proliferation and oncogenesis. Despite identification of multiple candidate substrates, it is yet to be definitively elucidated how this PTPase promotes the RAS-ERK pathway. Following detection of PTPN11 mutations in leukemias and less frequently in solid tumors, several other mechanisms, such as overexpression of itself or partners, are revealed for SHP2 function in carcinogenesis. In contrast to its essential role in cell-autonomous oncogenic signaling, anti-oncogenic effects of SHP2 have been found in the liver, bone and gut. The conflicting roles of SHP2 in oncogenesis unearth complexity in tumor cell communications with the microenvironment, which remain poorly understood. Isolation of orally active inhibitors has made SHP2 a prominent pharmaceutical target, with encouraging results from preclinical studies in animal tumor models. Positive outcomes are to be hoped for in the numerous ongoing clinical trials with SHP2 inhibitors worldwide. However, promising results in preclinical studies may not predict successful clinical treatment, particularly for patients at the advanced cancer stages, which need to be recapitulated by proper design of animal models. A failure of these SHP2 inhibitors in the clinic, which have passed rigorous preclinical tests, may not indicate flawed design in medicinal chemistry, but rather calls for rethinking of the classical drug development procedure from bench to bedside.