Shp1 function in myeloid cells

Review on the role of Shp1, and how myeloid-specific dysregulation may contribute to disease.

Abstract

The motheaten mouse was first described in 1975 as a model of systemic inflammation and autoimmunity, as a result of immune system dysregulation. The phenotype was later ascribed to mutations in the cytoplasmic tyrosine phosphatase Shp1. This phosphatase is expressed widely throughout the hematopoietic system and has been shown to impact a multitude of cell signaling pathways. The determination of which cell types contribute to the different aspects of the phenotype caused by global Shp1 loss or mutation and which pathways within these cell types are regulated by Shp1 is important to further our understanding of immune system regulation. In this review, we focus on the role of Shp1 in myeloid cells and how its dysregulation affects immune function, which can impact human disease.

Introduction

The cytoplasmic protein tyrosine phosphatase Shp1, encoded by the Ptpn6 gene, is a key regulator of immune cell function. This is particularly apparent in motheaten (me) mice that have a spontaneous mutation in Ptpn6, resulting in loss of Shp1, and develop a broad array of inflammatory and autoimmune symptoms. Given the important regulatory function of Shp1 in immune cells, there is growing interest in targeting this phosphatase in autoimmune disease and cancer [1, 2]. Indeed, compounds that could enhance Shp1 function may be useful for suppression of immune function [3]. Compounds that block Shp1 function and stimulate anti-tumor immune responses have been described [4]. However, the profound effect that global reduction in Shp1 activity produces, as evidenced by the mouse mutant models discussed below, warns of the risks of systemic approaches. Given that the pathology in this mouse model is driven largely by innate cells [5], a better understanding of the regulatory functions of Shp1 in these cell types is warranted.

Myeloid cells, including neutrophils, DCs, monocytes, macrophages, mast cells, eosinophils, and basophils, are activated by many different ligands. The subsequent activation of multiple downstream signaling cascades leads to diverse functional outcomes (such as proliferation, survival, adhesion, chemotaxis, phagocytosis, and degranulation). Shp1 impacts many different signaling pathways in myeloid cells in various different ways; although there are common themes among different myeloid cells, it is also clear that Shp1 regulates distinct pathways in each of these cell types. Here, we review the data concerning Shp1 function in myeloid cells, how Shp1 contributes to immune regulation, and its impact on human disease.

MOUSE MUTATIONS INVOLVING Shp1

The effect of Shp1 loss on immune system function was apparent long before the gene encoding it was discovered, as a result of the description of several spontaneous mouse mutations; these are summarized in Table 1 and Fig. 1. The first of these, known as me, was reported in 1975; the homozygous me/me mice are runted and develop skin lesions. These mice develop multiple inflammatory and autoimmune symptoms, resulting in lethality within 6 wk [6]. In 1984, a new mutant mouse was described with a similar but less severe phenotype, named motheaten viable (mev), which survives up to 12 wk, shown in Fig. 2 [7]. The positional cloning and sequencing of the gene came several years later; both of these strains of mice were found to have mutations in the gene Ptpn6, detailed in Fig. 1 [8]. Two subsequent mouse models—spin and meB2—involving mutations in Ptpn6 have been described with similar but milder symptoms [9, 10]. With the multitude of different symptoms and the broad hematopoietic cell expression of Shp1, researchers began to try and narrow down the roles of different cell types that generate the complex me phenotype. Many insights came from crossing Shp1 mutant mice with strains of mice deficient in a variety of proteins (summarized in Table 2). Some of the earliest of these experiments highlight the importance of Shp1 regulation in myeloid cells. Yu et al. [5] crossed mev mice with RAG-1-deficient animals; they found that the exaggerated myelopoiesis and inflammation were still present in the absence of B and T cells. Radiation chimeras reconstituted with BM from mev mice phenocopy mev mice and this phenotype are prevented by treatment with an anti-CD11b antibody that targets predominantly myeloid cells [11]. Even in zebrafish, knockdown of Shp1 early in development, when macrophages and neutrophils are present, but the adaptive immune system has not yet formed, leads to an inflammatory phenotype involving skin lesions, with an enhanced response to bacterial challenge but an inability to control infections [12].

TABLE 1.

Features of Shp1 mutant mice

Mouse model	Abbreviation used in text	Result of mutation	Survival	Severity of disease*	Reference
Ptpn6me/me	me	Loss of Shp1 protein	2–3 wk	++++	[6]
Ptpn6me-v/me-v	mev	Expression of Shp1 with ∼10–20% of wild-type phosphatase activity	9–10 wk	+++	[7]
Ptpn6me-B2/me-B2	meB2	Reduced Shp1 protein	>1 yr	++	[9]
Ptpn6m1Btlr/m1Btlr	spin	Disrupted Shp1 signaling	>1 yr	+	[10]

Four spontaneous mouse models identified with mutations in the Ptpn6 gene. All strains develop myelopoiesis, splenomegaly, inflammatory disease involving the skin, paws, and lungs, increased serum proinflammatory cytokines, and defects in B and T cells that lead to systemic autoimmunity involving anti-nuclear antibody production and immune-complex glomerulonephritis. The details of the mutations are shown in Fig. 1. *Severity of disease refers to the extent of these symptoms and ranges from less severe (+) to most severe (++++).

Figure 1. The structure of the Ptpn6 gene encoding the Shp1 protein, showing positions of mutations and key regulatory sites.

The Ptpn6 gene is found on mouse chromosome 6 and on human chromosome 12p13. The numbering shown is based on the mouse protein produced from the hematopoietic-specific promoter 2. The mutations that give rise to the four spontaneous mouse models detailed in Table 1 are indicated by boxes. The position of the loxP sites in the Shp1 floxed mice are shown and result in deletion of exons 1–9 in the presence of Cre protein. The amino acid changes shown in red lead to reduced phosphatase function; the C453S amino acid change creates a phosphatase-dead Shp1, whereas the other three mutations are spontaneously occurring (Y208N in mice; N225K and A550V in humans). When phosphorylated, the tyrosine and serine residues shown in black have been shown to be involved in increased or decreased phosphatase function, respectively. N-SH2, N-terminal SH2; C-SH2, C-terminal SH2.

Figure 2. The phenotype of the mev mouse.

A mev mouse and wild-type littermate at 6 wk of age, showing patchy fur and inflammation of paws and ears.

TABLE 2.

Compound crosses of Shp1 mutant mice

Shp1 mutant strain	Crossed to…	Summary of phenotype	Reference
Ptpn6me-v/me-v	Rag-1−/−	Inflammatory disease; no mature B or T cells, no autoimmunity	[5]
Ptpn6me-v/me-v	Prkdcscid (scid)	No mature B or T cells, no rescue of me disease	[175]
Ptpn6me/me	Btkxid (xid)	Block in B cell development, reduced autoantibody production but no rescue of me disease	[176]
Ptpn6me-v/me-v	Igh-6−/−	Block in B cell development, reduced autoantibody production, but no rescue of me disease	[177]
Ptpn6me-v/me-v	Foxn1nu (nude)	Reduced autoantibody production but no rescue of me disease	[178]
Ptpn6me-v/me-v	Lystbg (beige)	Granule defect in NK cells, CTLs, neutrophils, no rescue of me disease	[175]
Ptpn6spin	G-CSF−/−	Neutrophil number reduced by 50%, prevented spontaneous inflammation	[67]
Ptpn6spin	Prtn3−/−Ela2−/−	Loss of proteinase 3 and neutrophil elastase does not prevent spontaneous inflammation	[67]
Ptpn6spin	Ncf−/−	No superoxide production, spontaneous paw inflammation suppressed	[67]
Ptpn6me/me	KitWv/Wv	Reduced inflammatory disease, increased survival	[131, 132]
Ptpn6me-v/me-v	KitW-Sh/W-Sh	Reduced pulmonary inflammatory disease	[133]
Ptpn6me-v/me-v	Hck−/−Fgr−/−Lyn−/−	Reduced inflammatory disease, increased survival	[64]
Ptpn6me-v/+	Lyn+/−	Autoimmune disease, no inflammation	[179]
Ptpn6me-v/me-v	CD45−/−	No rescue of me disease but normal development of B cells, reduced autoantibodies	[180]
Ptpn6me-v/me-v	IFN-γ−/−	Airway inflammation, increased Th2 skewing, no suppression of spontaneous paw inflammation	[135] and [unpublished results]
Ptpn6me-v/me-v	IL-4−/−	Slightly decreased lung inflammation, significantly decreased nasal inflammation	[134, 135]
Ptpn6me-v/me-v	IL-13−/−	Significantly decreased lung inflammation, slightly decreased nasal inflammation	[134, 135]
Ptpn6spin	Stat1m1Bltr/m1Bltr	No suppression of spontaneous paw inflammation	[10]
Ptpn6me-v/me-v	Stat6−/−	Significantly decreased lung inflammation	[134]
Ptpn6spin	Tnf−/−	No suppression of spontaneous paw inflammation	[10]
Ptpn6me/me	IL-1R−/−	Reduced skin and lung pathology, increased survival up to 12 wk	[181]
Ptpn6spin	IL-1R−/−	Spontaneous paw inflammation suppressed	[10]
Ptpn6me-v/me-v	IL-1α−/−	No suppression of spontaneous paw inflammation	[unpublished results]
Ptpn6spin	IL-1α−/−	Spontaneous paw inflammation suppressed	[78]
Ptpn6spin	IL-1β−/−	No suppression of spontaneous paw inflammation	[78]
Ptpn6spin	TLR4−/−	No suppression of spontaneous paw inflammation	[78]
Ptpn6me-v/me-v	MyD88−/−	No live pups born	[64]
Ptpn6spin	MyD88poc/poc	Spontaneous paw inflammation suppressed, reduced anti-nuclear antibodies	[10]
Ptpn6spin	TicamLps2/Lps2	No suppression of spontaneous paw inflammation	[10]
Ptpn6spin	IRAK4otiose/otiose	Spontaneous paw inflammation suppressed	[10]
Ptpn6spin	RIP1−/−	(Fetal liver transfer) Spontaneous paw inflammation suppressed	[78]
Ptpn6spin	RIP3−/−	No suppression of spontaneous paw inflammation	[78]
Ptpn6spin	Nlrp3−/−	No suppression of spontaneous paw inflammation	[78]
Ptpn6spin	Caspase1−/−	No suppression of spontaneous paw inflammation	[67, 78]
Ptpn6me-v/me-v	CD5−/−	Reduced inflammatory disease, increased survival	[182]
Ptpn6me-v/me-v	Ets2tmA72Osh (T72A)	Reduced inflammatory disease, increased survival	[183]

Many different strains have been crossed with Shp1 mutant mice to assess the role of specific genes in the me phenotype. The results are summarized in this table.

It is often difficult to tease apart the role that Shp1 plays in different cell types using these models, especially as it is clear that myeloid cell development is also abnormal in me mice [13]. Shp1 interacts with the Hox domain protein, HoxA10, which can impact gene transcription in myeloid precursors [14]. Reduced Shp1 activity causes constitutive activation of Stat5, which increases the growth of hematopoietic stem cells, leading to the development of a myelodysplastic/myeloproliferative phenotype [15]. Consequently, mice containing a floxed allele of Shp1 (Ptpn6^fl/fl) have been an invaluable tool for dissecting the complex phenotype of the me mouse [16]; Table 3 summarizes results from crossing Shp1 floxed mice with different Cre-expressing strains.

TABLE 3.

Analysis of Shp1 deletion in specific cell types

Cre-expressing strain	Cell type specificity	Phenotype		Summary of phenotype	Reference
		Autoimmunity	Inflammation
Vav-cre	Hematopoietic cells	Yes	Yes	Phenocopies me and mev mice	[64]
CD19-cre	B cells	Yes	No	Autoimmune disease	[16]
MB1-cre	B cells	Yes	No	Autoimmune disease	[unpublished results]
Aicda-cre	Activated B cells	No	No	No phenotype, poor memory response upon challenge	[184]
CD4-cre	T cells	No	No	Th2 skewing of T cells	[185]
Lck-cre	Mature T cells	No	No	Enhances CD8+ T cells response to LCMV challenge	[186]
MRP8-cre	Neutrophils	No	Yes	Rescue by compound cross with Sykfl/fl but not MyD88fl/fl	[64]
CD11c-cre	DCs	Yes	No	Partial rescue by compound cross with MyD88fl/fl	[64, 89]
Mcpt5-cre	Mast cells	No	No	Enhances allergen-induced atopic dermatitis	[137]
α-Chymase-cre	Mast cells	No	No	No phenotype	[unpublished results]
PF4-cre	Megakaryocytes and platelets	No	No	Platelets less responsive to collagen-related peptide and fibrinogen	[187] and [unpublished results]
NKp46-cre	NK cells	No	No	Reduces activation threshold of NK cells, hyporesponsive to tumor cells	[188] and [unpublished results]
LysM-cre	Macrophages/monocytes	No	No	No phenotype	[unpublished results]
GE-cre	Monocytes/neutrophils	No	Yes	Skin and lung inflammation	[unpublished results]
F4/80-cre	Tissue macrophages	No	No	No phenotype	[unpublished results]
Basoph8-cre	Basophils	No	No	No phenotype	[unpublished results]
CX3cr1-cre	Macrophages/monocytes/mast cells/cDCs	(Yes)	No	Mild autoimmunity in aged mice	[unpublished results]

The Shp1 floxed mice, Ptpn6^{tm1Rsky/tm1Rsky}, generated by Pao et al. [16], have been crossed with many different hematopoietic Cre-expressing strains, and the results are summarized in this table. LCMV, lymphocytic choriomeningitis virus.

Shp1 STRUCTURE AND REGULATION

An outline of Shp1 protein is shown in Fig. 1. The 68 kDa protein consists of two SH2 domains, followed by a tyrosine phosphatase domain and a C-terminal tail containing multiple sites of phosphorylation. Structural analysis of Shp1 lacking the C-terminal tail determined that the N-terminal SH2 domain of Shp1 forms an intramolecular interaction with the phosphatase domain, thereby inactivating it under basal conditions [17]. A more recent crystal structure of full-length Shp1 suggests that upon ligand binding to the SH2 domains, these domains are held out of the way, leading to stabilization of the activated phosphatase [18]. This process is summarized in Fig. 3. Determination of the optimal pTyr-containing substrates that bind to the SH2 domains of Shp1, implicated many signaling molecules in the activation of Shp1 phosphatase activity [19], in particular, proteins that contain an ITIM (sequence I/V/LxYxxL/V), found within the cytoplasmic domain of many cell surface inhibitory receptors. In myeloid cells, the two inhibitory receptors that have been most implicated in Shp1 activation are Sirpα and Pirb [20, 21]. Both of these receptors contain cytoplasmic ITIM domains, which when phosphorylated, associate with Shp1, resulting in phosphatase activation. However, it is likely that other ITIM-containing receptors in myeloid cells, such as FcγRIIb, CEACAM-1 (CD66a), and PECAM-1 (CD31), also associate with Shp1 to regulate downstream signaling [22–24].

Figure 3. The regulation of Shp1 phosphatase activity.

(A) Under basal conditions, the N-terminal SH2 domain of Shp1 forms an intramolecular interaction with the phosphatase domain, restricting access of substrates to the phosphatase active site. The C-terminal tail of Shp1 is disordered in published crystal structures and is shown here as a dotted line. Phosphorylation of serine residues (pS) in the C-terminal tail also negatively regulates Shp1. (B) Upon engagement of the SH2 domains of Shp1 by tyrosine phosphorylated (pY) ITIMs, the SH2 domains swing out of the way, allowing access to the phosphatase active site. The N-terminal SH2 domain is held out of the way by a further intramolecular interaction (red dashed lines) to stabilize the active phosphatase. Phosphorylation of tyrosine residues in the C-terminal tail contribute to phosphatase activation.

It is clear that Shp1 itself is also regulated by phosphorylation. Two tyrosines in the C-terminal tail, Y536 and Y564 (of mouse Shp1; note that there is 95% homology between murine Shp1 and human SHP1), have been identified as likely substrates of SFKs. Biochemical analysis by Zhang et al. [25] suggested that phosphorylation at Y536 engages the N-terminal SH2 domain to relieve basal inhibition, whereas phosphorylation of Y564 results in an additional, smaller increase in phosphatase activity, perhaps as a result of an indirect interaction with the C-terminal SH2 domain. However, more recent work by Xiao et al. [15], in which Shp1 mutants were expressed in mev hematopoietic stem cells, found that phosphorylation at Y564 was critical for the phosphatase activity of Shp1, whereas phosphorylation at Y536 was more important for binding signaling molecules, such as Grb2 or Stat5, and a Y536/564E double-mutant Shp1 shows enhanced phosphatase activity. Phosphorylation of S557 by PKD in T cells and phosphorylation of S591 by PKC have been proposed to negatively regulate Shp1 [26–29]. Other features of the C-terminal tail of Shp1 that could also impact regulation include a nuclear localization signal [30], a lipid raft localization signal, and protein–protein interaction domains, reviewed in Poole and Jones [31]. The subcellular localization of Shp1 will clearly impact its function. For example, following chemokine stimulation of neutrophils, dissociation of Shp1 from the membrane, which is associated with reduced phosphorylation of Pirb, temporarily releases a brake on signaling to maximize downstream MAPK signaling [32]. Relocalization of Shp1 to the uropod may also affect the migratory capacity of chemokine-stimulated neutrophils [33]. How all of these overlapping features of Shp1 regulation work together to coordinate phosphatase function is not fully understood.

Shp1 is expressed broadly throughout the hematopoietic system. Expression is also found in nonhematopoietic cells, although at lower levels [34]. In epithelial cells, transcription from an alternative promoter leads to a different isoform of Shp1 than is found in hematopoietic cells [35]. This could lead to regulatory differences; indeed, Craggs and Kellie [36] report that epithelial cell Shp1 is localized to the nucleus, whereas hematopoietic Shp1 is cytoplasmic.

Shp2 is a closely related family member, but its expression is more ubiquitous than Shp1, and it is generally thought to positively regulate cell signaling [37]. In hematopoietic cell lineages, where Shp1 expression is highest, such as neutrophils, inflammatory monocytes, and pDCs, Shp2 expression is much lower than Shp1 (Immunological Genome Project, www.immgen.org).

MYELOID SUBSTRATES OF Shp1

Large-scale screening of combinatorial peptide libraries was carried out by Ren et al. [38] to determine the optimal substrates of Shp1. Both Shp1 and Shp2 show a preference for acidic and aromatic residues on either side of the target pTyr residue, with just subtle differences observed between the two phosphatases. More recent computational methods have confirmed these general principles for Shp1 and Shp2 substrate preferences [39]. These observations match well with known substrates of Shp1 found in myeloid cells, including p62Dok [40], Syk [41], Vav [42], Slp65 [43], Slp76 [44], and Sirpα [20], with the notable exception being Pirb [20], which surprisingly, does not contain any sequence corresponding to an optimal substrate. Shp1 has also been shown to dephosphorylate SFKs, including Lyn [45], and given that Src substrates tend also to be good Shp1 substrates [46], tight control of signaling pathways involving these proteins is obtained. An example of this balance is the coordinated regulation of Caspase 8 by Lyn and Shp1, which impacts neutrophil survival [47]. A further layer of complexity is added by the fact that Y536 of Shp1 itself is also a good substrate for Shp1 phosphatase activity [48]. Consequently, the specific substrates for Shp1 that are crucial for mediating key signaling pathways in different myeloid subsets in vivo are not well defined. Furthermore, Sirpα and Pirb are hyperphosphorylated in macrophages from mev mice but hypophosphorylated in me macrophages [20], perhaps highlighting phosphatase-dependent and independent functions of Shp1, which are important to consider in the context of inhibitors.

COMMON THEMES IN MYELOID SIGNALING PATHWAYS THAT USE Shp1

Shp1 has been implicated in many signaling pathways. The most well-defined signaling mechanism used by Shp1 in myeloid cells is mediated by SFKs and Syk kinase phosphorylation of ITIM domains on inhibitory receptors, leading to recruitment and activation of Shp1 [49] (Fig. 4). Shp1 (and other inhibitory phosphatases, such as Shp2 or Ship1) turn off activating signals initiated by ITAM-containing proteins, such as integrins, FcRs, and C-type lectin receptors, to maintain an appropriate balance of signaling. Well-studied examples of ITIM-containing inhibitory receptors in myeloid cells include Pirb [50], Sirpα [20], and the FcR, FcγRIIb [51]. How inhibitory ITIM signaling achieves specificity to target different activating ITAM signals is unclear. In some cases, regulation of a specific pathway may be achieved through coaggregation of activating (ITAM) and inhibitory (ITIM) receptors, such as the well-known example of coligation of the BCR with FcγRIIb [52]. Furthermore, there are examples of ITAM-containing receptors (most notably, FcγRIIa) that are able to mediate inhibitory signals through recruitment of Shp1, following low-affinity/avidity ligand interactions [53]. Under these conditions, the signaling is referred to as an ITAMi pathway.

Figure 4. Key signaling pathways regulated by Shp1 in myeloid cells.

Examples of signaling pathways impacted by Shp1 in (A) neutrophils, (B) DCs, (C) macrophages and monocytes, and (D) mast cells, as discussed in the text. Common themes in these signaling pathways include recruitment of Shp1 by ITIM-containing proteins to dampen down ITAM receptor-mediated signaling and negative regulation of signaling through TLRs and cytokine receptors by Shp1.

Shp1 has also been implicated downstream of receptor tyrosine kinases, such as c-Kit and CSF1R, where it is activated directly through binding to receptor pTyr residues via its SH2 domains, allowing it to dephosphorylate signaling proteins [54–56]. Much research has implicated roles for Shp1 downstream of cytokine receptors, such as GM-CSF1R, erythropoietin receptor, IL-3R, IL-4R, IL-13R, and IFNAR/IFNGR, the receptors for type 1 and 2 IFNs, respectively. In these pathways, Shp1 is also thought to bind receptor pTyr residues through its SH2 domains, leading to dephosphorylation of JAKs and/or STAT proteins downstream [57–60]. TLR signaling is also impacted by Shp1, potentially through interactions with downstream signaling proteins or recruitment of Sirpα or by regulation of PI3K pathways [61–63]. However, given that TLR signaling is primarily mediated through pathways involving serine/threonine phosphorylation, it remains very unclear how a tyrosine phosphatase would be involved.

Given the impact of Shp1 on all of these pathways, it was surprising that upon analysis of several myeloid cell-specific conditional knockouts of Shp1, all of these pathways were not activated. Instead, it seems that different subsets of myeloid cells use Shp1 to regulate different signaling pathways [64]. These pathways will be considered for each myeloid cell type below (summarized in Fig. 4).

NEUTROPHILS

Neutrophils are an important and rapid first responder in the host response to infection, but their function must be tightly regulated to avoid tissue damage. Neutrophils express a broad array of cell surface receptors (reviewed in Futosi et al. [65]). Activation of signaling pathways downstream of these receptors leads to survival, migration, adhesion, phagocytosis, and production of superoxide. Shp1 has been reported to impact many of these functions. Early characterization of me and mev mice showed an increase in circulating neutrophils, as well as their accumulation at sites of lung and skin inflammation, suggesting that these cells play a role in some aspects of the me phenotype [7, 66].

Neutrophils from me and mev mice have been described as phenotypically and morphologically immature, implicating Shp1 involvement in their development. Likewise, meB2 mice have increased numbers of immature neutrophils, which is a result of their delayed apoptosis and not increased proliferation [9]. Myelopoiesis in mice with the milder spin mutation appears normal [67]. In neutrophil-specific, Shp1-deficient mice, increased numbers of neutrophils are found in the spleen and BM [64]. G-CSF and G-CSFR null mice are severely neutropenic, suggesting that G-CSF is the principal cytokine regulating granulopoiesis [68]. Neutrophil precursors from mev mice are more proliferative and less differentiated in response to G-CSF [69]. Shp1 expression is up-regulated during differentiation of the myeloid cell line 32D, and overexpression of Shp1 both reduces proliferation and enhances differentiation [70]. Although several tyrosine residues within the cytoplasmic domain of the G-CSFR are required for this effect, Shp1 does not appear to bind to the receptor directly [70]. However, Dong et al. [71] report a weak interaction between the C-terminal tail of the G-CSFR and Shp1 when overexpressed in 293T cells, which is required to limit the activation of Stat3 but not Akt or MAPK. Baumann et al. [72] suggest that G-CSF signaling activates MAPK during neutrophil maturation to promote proliferation, but up-regulation of Shp1 inhibits this pathway to enhance differentiation in mature neutrophils. More recently, CEACAM-1, which is an ITIM-containing receptor expressed highly on the surface of neutrophils and up-regulated upon neutrophil activation [73], has been suggested to mediate Shp1 inhibition of G-CSF signaling. In BM progenitors stimulated with G-CSF, Shp1 coassociates with G-CSFR only when CEACAM-1 is present, leading to dephosphorylation of G-CSFR. CEACAM-1-deficient mice develop neutrophilia as a result of hyperproliferation of progenitors, which is rescued only by neutrophil re-expression of CEACAM-1 containing the ITIM domains [74]. This phenotype is not seen in mice that lack other ITIM-containing receptors, indicating a level of specificity in recruitment of Shp1-mediated inhibition, perhaps regulated by localization, expression levels, or ligand binding.

Early studies of neutrophils isolated from me and mev mice show that loss of Shp1 function enhances phosphorylation of several proteins in response to a variety of stimuli, including fMLF and opsonized zymosan. Neutrophils from these mice express higher levels of the β2 integrin CD11b and display increased functional properties, including hyperadhesion to fibrinogen and enhanced superoxide production in response to fMLF and C5a [75, 76]. The hyperadhesion likely leads to the defect observed in chemotaxis in vitro in response to several stimuli, including opsonized zymosan, fMLF, and keratinocyte-derived chemokine (KC) [75]. In vivo, neutrophil tethering, rolling, firm adhesion, and extravasation require the coordination of several signaling pathways involving chemokine receptors, selectins, and integrins. Shp1 is involved in several different steps, such that its loss impacts rolling and firm adhesion. Neutrophils deficient in Shp1 manifest increased neutrophil adhesion to vascular endothelium in vivo but also show reduced crawling, transmigration, and chemotaxis [77]. We also observed reduced PT-elicited recruitment of Shp1-deficient neutrophils [unpublished results]. Croker et al. [67] report that neutrophils from spin mice show enhanced LPS-stimulated NF-κB and MAPK signaling, leading to increased TNF-α and IL-1α and -β production. Stimulation with IL-1β itself also leads to increased IL-1β production. However, Lukens et al. [78] see an increase in TNF-α but no significant increase in IL-1β production from LPS-treated spin neutrophils. Spin mice are also more resistant to Listeria infection, likely as a result of the increased number of neutrophils [10]. The spontaneous skin inflammation observed in the spin mutant mouse is attributed to neutrophils, as inflammation is reduced in neutropenic spin mice that lack G-CSF [67]. Therefore, an increase in neutrophil numbers and their hyperactivity leads to enhanced responses to sterile wounding, leading to chronic tissue inflammation. Mice that specifically lack Shp1 in neutrophils also develop myelopoiesis and spontaneous paw inflammation at 8–12 wk of age. Neutrophils from these mice do not show enhanced responses to TLR ligands, fMLF, C5a, G-CSF, or GM-CSF, but they are hyper-reactive to integrin stimulation, resulting in elevated MAPK signaling and enhanced superoxide production [64]. In a subset of humans with neutrophilic dermatoses, including Sweet’s syndrome and Pyoderma gangrenosum, alterations in the PTPN6 gene have been described that reduce Shp1 expression, leading to increased neutrophil infiltration of the skin, implicating Shp1 in human disease [79].

What are the pathways impacted by Shp1 that lead to the inflammatory phenotype? Mature neutrophils express several ITIM-containing receptors that can recruit Shp1, and this seems to be the predominant mechanism for harnessing the negative regulatory role of Shp1 (Fig. 4A). The hyperactive integrin signaling that is observed in neutrophils lacking Shp1 is mediated through SFKs and Syk kinase, because genetic deletion of these kinases in the context of Shp1 deletion rescues the inflammatory phenotype [64]. It is likely that an ITIM-containing receptor is recruiting Shp1 to down-regulate integrin signaling, but which one is unknown.

Pirb is an important ITIM-containing inhibitory receptor in neutrophils. In wild-type mouse neutrophils in suspension, Pirb is constitutively phosphorylated [21]. This is mediated by the SFKs Hck and Fgr, leading to basal Shp1 association with Pirb [32]. In mev neutrophils, Pirb phosphorylation is increased. Following chemokine stimulation of neutrophils, a transient reduction in Pirb phosphorylation leads to reduced membrane-associated Shp1 and results in increased MAPK signaling [32]. In addition to chemokine signaling, Pirb recruits Shp1 to downmodulate integrin signaling; me and mev neutrophils are hyperadhesive to fibrinogen-coated plastic [75], and Pirb-deficient neutrophils have a similar phenotype [80]. It is interesting to note that loss of Pirb in mice leads to defects in B and T cell function but does not lead to inflammation [81]. Perhaps neutrophil-specific deletion using floxed Pirb mice [82] would uncover an inflammatory phenotype.

Several other ITIM-containing receptors can impact neutrophil function by recruiting Shp1. mSiglec E and its counterpart hSiglec 9 are constitutively expressed on neutrophils and associate with Shp1 via their ITIM domains. Siglec E can act as a negative regulator of integrin signaling. Neutrophils from Siglec E-deficient mice, plated onto fibrinogen, show higher levels of Syk and p38 phosphorylation than wild-type cells [83]. In vivo, the CD11b-dependent pulmonary neutrophil infiltration observed upon intranasal LPS or zymosan administration was increased in Siglec E-deficient mice [83]. The MHC class I receptor, Ly49Q, is another ITIM-containing receptor expressed on neutrophils that associates with Shp1 in the steady state, reducing phosphorylation of both SFKs and PI3K. In the absence of stimulation, neutrophils expressing Ly49Q with a mutated ITIM domain are hyperadhesive; however, no difference in fMLP-induced superoxide production is observed, which may be a result of recruitment of Shp2 [84]. The ITIM-containing receptor PECAM-1 also recruits Shp1 and impacts neutrophil migration. In response to IL-8, PECAM-1 and Shp1 coassociate at the leading edge of migrating neutrophils. Shp1 localization in PECAM-1-deficient neutrophils is diffuse; they fail to show increased Shp1 phosphatase activity in response to IL-8 and migrate poorly [85].

In contrast to the previous inhibitory receptors, loss of CEACAM-1, which is expressed highly on neutrophils, has no effect on migration, demonstrating that ITIM-mediated signaling through inhibitory receptors that recruit Shp1 is specific for different activating signals. However, CEACAM-1-deficient neutrophils have increased phospho-Syk levels and produce elevated levels of superoxide and IL-1β in response to TLR2 and -4 ligands, whereas production of TNF-α and IL-6 is not affected. CEACAM-1 recruits Shp1 to limit inflammasome activation caused by an LPS-stimulated coassociation of TLR4 and Syk [86].

It has been proposed that the YXXL motifs in the cytoplasmic region of death domain-containing receptors, such as Fas, TNF, and TRAIL, are phosphorylated and may act as atypical ITIMs that can recruit Shp1 in neutrophils [87]. Treatment of neutrophils with FasL leads to cleavage of Caspases 3 and 7, which is lost in FasL-treated neutrophils from spin mice [67]. Signaling through these receptors can also inhibit survival signals initiated by G-CSF, GM-CSF, or IFN-γ, implying that neutrophils lacking Shp1 would survive longer.

Stadtmann et al. [77] propose a mechanism for Shp1 control of integrin inside-out signaling. E-Selectin engagement of neutrophils or stimulation with the chemokine CXCL1 was found to increase Shp1 phosphorylation on S591 but not Y536, which results in decreased phosphatase activity. This leads to increased phosphorylation of phosphatidylinositol-4-phosphate 5-kinase type 1 gamma (PIPKIγ), with which Shp1 coassociates, and accumulation of PI(4,5)P₂, which in turn, increases integrin adhesiveness, potentially through activation of small GTPases, such as Rap1b. Shp1-deficient neutrophils have higher PI(4,5)P₂ levels, resulting in increased integrin affinity and avidity, leading to the hyperadhesive phenotype seen in these cells. Rap1b has been implicated through studies of neutrophils deficient in this enzyme. Kumar et al. [33] found that neutrophils lacking Rap1b show enhanced migration and chemotaxis and reduced integrin-dependent superoxide production in vitro and enhanced recruitment to the lung in response to LPS challenge in vivo. Rap1b is thought to activate Shp1, although the mechanism is not known, suggesting a potential feedback loop to regulate integrin adhesion.

Limited studies of TLR signaling in me and mev neutrophils, as well as contradictory results between TLR-stimulated neutrophils from spin mutant mice vs. neutrophil-specific Shp1-deficient animals, suggest that TLR signaling in neutrophils is complex. Extrapolation from studies in other cell types (see DC and macrophage sections) may implicate phosphatase-dependent and -independent mechanisms for how Shp1 impacts TLR signaling. Reduced inflammation is observed in spin mice that express a mutant form of MyD88 and neutrophils from these mice show reduced responses to TLR stimulation [10]. However, neutrophils from lineage-specific Shp1-deficient mice do not show enhanced TLR-stimulated MAPK signaling, and these mice are not rescued by neutrophil-specific deletion of MyD88. The fact that loss of Shp1 in neutrophils leads to inflammatory disease and that spin mice are protected from inflammatory disease when the neutrophil number is reduced clearly points to a key role for neutrophils in inflammation [64, 67]. It is interesting that global loss of IL-1R or IL-1α protects spin mice from inflammation, as does loss of RIP1 kinase in hematopoietic cells [10, 78]. Lukens et al. [78] suggest that Shp1 downmodulates a signaling response involving RIP1 kinase-mediated NF-κB and MAPK activation, leading to IL-1α production in response to mild skin injury or other inflammatory insults. However, it is not clear which cells are producing or responding to IL-1α in this model, and these experiments need to be performed in the context of neutrophil-specific deletion of Shp1. Hence, the role of Shp1 regulation of TLR (and IL-1R) signaling in neutrophils remains unclear.

DCs

DCs are key APCs that direct T cell responses. They express many cell surface receptors that deliver activating and inhibitory signals to tailor cytokine production, ensuring T cells respond appropriately, promoting immune responses to foreign antigens and tolerance to self-antigens. Early characterization of mev mice reported that numbers of DCs in thymus, skin, spleen, and peripheral lymph nodes are reduced, but increased numbers of monocyte-derived DCs are present in the lungs and liver. Similar to neutrophils from these mice, it is unclear whether this is the result of an intrinsic defect or an abnormal inflammatory environment [13]. BMDCs and splenic CD11c⁺ DCs from mev mice show enhanced TNF-α production but decreased IFN-β production upon stimulation with the TLR3 and -4 ligands, poly (I:C), and LPS, respectively [61]. However, splenic CD11c⁺ DCs from mice with the milder spin mutation do not show any differences in LPS-stimulated MAPK or NF-κB activation [67]. Ramachandran et al. [88] inhibited Shp1 activity in wild-type BMDCs using RNA interference, treatment with the Shp1 inhibitor SSG, or overexpression of a dominant-negative mutant of Shp1 (C453S, which renders Shp1 phosphatase dead) to study the role of Shp1 in DC function. In the absence of any developmental effect, Shp1 plays a negative role in IFNGR, IL-10R, CCR7, and TLR4-mediated NF-κB activation. This results in the altered production of cytokines in response to LPS stimulation; Shp1 inhibition enhances MAPK activation and increases production of IL-12, IL-1β, and IL-6, but TNF-α levels do not change. We [64] and others [89] generated DC-specific, Shp1-deficient mice, which develop profound autoimmunity but no skin and little lung inflammation. Kaneko et al. [89] found that splenic CD11c⁺ DCs from these mice up-regulate CCR7 and the activation marker CD86 and show increased migration to draining lymph nodes in vivo. LPS stimulation of these cells leads to increased IL-6, IL-1β, and IL-10 production, but contrary to earlier studies, levels of TNF-α are also increased, whereas IL-12 levels are reduced [64, 89]. LPS-induced survival of BMDCs is negatively regulated by Shp1, perhaps through its inhibition of Akt activation [88]. Inhibition of Shp1 in DCs enhances T cell stimulation, which is likely responsible for the autoimmune phenotype of DC-specific, Shp1-deficient mice [88, 89].

Shp1 modulation of signaling downstream of TLR receptors in DCs is at least partly responsible for the autoimmune phenotype seen in DC-specific, Shp1-deficient mice, as the phenotype is partially rescued in DCs that also lack MyD88 [64]. The mechanism of this modulation, however, is unclear. An et al. [61] found that Shp1 can bind to and inhibit the activity of IRAK1 in a phosphatase-independent manner, perhaps mediated by an ITIM-like domain found within the kinase domain of IRAK1 (known as a kinase tyrosine-based inhibitory motif, or KTIM, domain) [90]. Shp1 can also associate with IRAK4 in DCs [88]. As with neutrophils, phosphatase-dependent and -independent roles of Shp1 may account for the contradictory studies observed between cells expressing different forms of Shp1.

Like neutrophils, DCs express many ITIM-containing inhibitory receptors on their cell surface that recruit Shp1 to modulate signaling, including Pirb, Sirpα, PEACAM-1, FcγRIIb, and the C-type lectin receptors, Clec4a2 (DCIR) and MICL (Clec12a). These receptors are often paired with activating receptors, but again, how these inhibitory receptors each recruit Shp1, yet show specificity for inhibiting different activating signals, is unclear. Pirb recruits Shp1 in DCs stimulated by the CCR7 ligands, CCL19 and CCL21, to down-regulate chemokine signals that lead to MAPK activation, actin polymerization, and migration [32]. Pirb-deficient mice show defects in DC maturation, which skews T cell responses [81]. Sirpα may be key for DC survival; mice deficient in Sirpα have reduced numbers of cDCs [91]. Loss of Sirpα in DCs leads to similar responses to those observed following Shp1 inhibition, namely, increased IL-12 production, up-regulation of DC activation markers, increased survival, and enhanced T cell priming [92]. However, mice expressing a mutant form of Sirpα that lacks the cytoplasmic tail, which cannot recruit Shp1, are resistant to experimental autoimmune encephalomyelitis and fail to develop collagen-induced arthritis [93, 94]. Similar to Shp1-deficient DCs, PEACAM-1-deficient DCs show increased migration from tissues to the draining lymph nodes. Engagement of PEACAM-1 on wild-type DCs promotes a tolerogenic response, blocking LPS-induced NF-κB activation and IL-6 production, but PEACAM-1 engagement of Shp1-deficient DCs has no effect [95]. Many C-type lectin receptors are expressed on DCs, and of these, Clec4a2 (DCIR) and MICL (Clec12a) contain ITIMs that can associate with Shp1, (reviewed in Kanazawa [96]), but the role of Shp1 in their function is unclear.

Subcellular localization is key to controlling Shp1 function in DCs. Siglec G, an ITIM-containing receptor present on phagosomal membranes, can negatively regulate cross-presentation of CD8⁺ DCs by recruiting Shp1, which dephosphorylates p47^phox. This inhibits production of reactive oxygen species and prevents the alkalinization of phagosomes that is required for antigen degradation and subsequent presentation [97]. The intracellular metabolic enzyme IDO is expressed in DCs and contains an ITIM that can engage Shp1 and Shp2. TGF-β treatment of pDCs leads to phosphorylation of the IDO ITIM and formation of a complex between IDO and Shp1 that activates Shp1 phosphatase activity. This is thought to inhibit IRAK1 and activate the noncannonical NF-κB pathway, up-regulating type 1 IFN production. The immunomodulatory effect of IDO is independent of its catalytic activity. Lack of Shp1 in pDCs ultimately impairs the ability of TGF-β-stimulated pDCs to promote the generation of regulatory T cells [98].

More recently, activating receptors have been shown to inhibit signals using ITAMs to recruit Shp1, and this specificity may be achieved by the nature of ligand engagement. Mincle (Clec4a) is expressed in DCs and signals via the ITAM domain of the FcRγ chain. However, in the absence of Mincle, treatment of DCs with Leishmania leads to reduced tyrosine phosphorylation of Shp1 and increased DC activation, implicating an inhibitory role for Mincle. Shp1 associates with the ITAM domain of Mincle, allowing Leishmania to suppress DC function [99].

The maintenance of the balance of activating and inhibitory signals in DCs is key to controlling the response of the adaptive immune system. Further understanding of the complex role of Shp1 in this process (Fig. 4B) will give valuable insights into the potential use of Shp1 activity modulators to control immune responses to pathogens and/or tumors.

MACROPHAGES/MONOCYTES

Initial characterization of me and mev mice revealed that macrophages accumulate in the lungs, which combined with the increase in myelopoiesis seen in these mice, suggests that macrophages and monocytes contribute to the me phenotype, in particular, the inflammatory lung disease that results in the death of these mice [7, 66]. Eosinophilic crystals containing Ym1 protein, known to be produced by alternatively activated macrophages, accumulate in the me lungs [100]. It is unclear whether hyperactive alveolar macrophages are the primary drivers of the me lung inflammation or whether the crystal formation is a macrophage response to lung inflammation caused by other cell types. The development of macrophages in mev mice is clearly disrupted. Increased numbers of macrophage progenitors and monocytes are found in the BM and spleen. Several tissues, including the lung, contain increased numbers of macrophages, but the development of certain types of macrophages, such as those found in lymphoid tissues, is impaired [13]. As noted earlier for neutrophils and DCs, it is hard to distinguish between cell-intrinsic macrophage phenotypes vs. those that arise as a result of an abnormal inflammatory environment in Shp1-deficient mice. Furthermore, the monocyte–macrophage lineage is heterogeneous, which makes investigating the role of these cells in the me mouse difficult. Studies that are typically carried out using myelomonocytic cell lines, such as RAW264.7, HL60, THP1, and U937, vs. BMMs or primary tissue macrophages derived from Shp1 mutant mice often produce conflicting results. Moreover, given the heterogeneity in the monocyte–macrophage lineage, there is no perfect Cre-expressing mouse strain available that can specifically address the function of gene deletion in these cells.

It is clear that many signaling pathways are activated in Shp1-deficient macrophages (Fig. 4C). BMMs from me mice show increased proliferative responses to GM-CSF but not to G-CSF or CSF1, although macrophages lacking Shp1 show increased survival upon CSF1 withdrawal. Surprisingly, GM-CSF-induced tyrosine phosphorylation of the GM-CSFR or the downstream signaling molecules Jak2, Stat5, or MAPK is not increased [101]. Likewise, BMMs from wild-type and me mice stimulated with CSF1 show similar patterns of tyrosine phosphorylation, except for hyperphosphorylation of p62Dok [40], implying that as with neutrophils and DCs, disruption of Shp1 function in macrophages leads to the phosphorylation of specific substrates. IL-3-dependent, proliferative responses in U937 cells are increased when a dominant-negative mutant of Shp1 (C453S) is expressed [102], and IL-4 treatment of BMMs from mev mice results in increased Stat6 activation [103]. IFN-γ treatment of an immortalized macrophage cell line generated from me mice increases iNOS expression and NO production [104]. However, in these cells and in BMMs from spin mutant mice, phosphorylation of JAK2 and Stat1 is not affected by IFN-γ treatment [10]. Shp1 constitutively associates with the IFNAR; BMMs from me mice treated with IFN-α show increased tyrosine phosphorylation of Jak1 and Stat1 but not Tyk2 or Stat2 [59]. The mechanism of Shp1 control of cytokine signaling is not clear.

As in neutrophils, integrin signaling is clearly activated in macrophages with reduced Shp1 function, given their hyperadhesive phenotype [76, 102]. However, no difference is observed in chemotaxis to fMLF, C5a, or opsonized zymosan in U937 cells overexpressing a C453S dominant-negative Shp1 [102]. One potential mechanism that Shp1 may use to regulate integrin signaling is through limiting activation of membrane-bound PI3K, thus limiting phosphatidylinositol (3,4)-bisphosphate and phosphatidylinositol (3,4,5)-trisphosphate generation [76]. CD11b-mediated recruitment of Shp1 has also been shown to inhibit IL-4-stimulated Stat6 activation in adipose tissue macrophages [105].

Alveolar macrophages from mev mice show enhanced phagocytosis of dead cells [106], but phagocytosis of opsonized zymosan was not increased in U937 cells overexpressing the C453S dominant-negative Shp1 [102]. FcR-mediated phagocytosis of IgG-opsonized RBCs is negatively regulated through Sirpα-mediated recruitment of Shp1 [107]. Similar to DCs, Shp1 is recruited to phagosomes during phagocytosis and is required for phagosome acidification [108]. This indicates that Shp1 also functions downstream of specific phagocytic receptors in macrophages.

Most studies agree that Shp1 clearly downmodulates signaling from TLRs in macrophages, reducing the production of proinflammatory cytokines; however, there are some differences, as outlined below, that suggest different pathways may be more or less sensitive to the level of Shp1 activity. Overexpression of wild-type Shp1 in RAW264.7 cells reduces TNF-α and iNOS production in response to LPS stimulation [109]. Conversely, LPS stimulation of BMMs from me mice produce increased levels of TNF-α, but there is no change in IL-1β and reduced IL-6 production [110]. LPS stimulation of both splenic and alveolar macrophages from me mice also leads to increased TNF-α production [111, 112]. Knockdown of Shp1 in wild-type splenic and PT-elicited macrophages increases LPS-stimulated TNF-α production; re-expression of Shp1 in me splenic macrophages reduces LPS-stimulated TNF-α to wild-type levels [61, 111]. Microglia from me mice also release increased TNF-α, NO, and IL-1β in response to LPS [113]. The inflammatory macrophages infiltrating the CNS of me mice following Theiler’s murine encephalomyelitis virus infection show a more inflammatory profile of gene expression than from wild-type macrophages [114]. Despite these data, there are contradictory results from BMMs generated from other Shp1 mutant mice. BMMs from mev mice stimulated with the TLR ligands LPS and Pam3CSK4 show defects in IL-12 production but produce wild-type levels of TNF-α and IL-6 [63]. Croker et al. [10, 67] report that PT-elicited macrophages and BMMs from spin mutant mice show wild-type levels of LPS-induced TNF-α production. A comprehensive study comparing different types of macrophages expressing different Shp1 mutants is required to determine whether these differences are a result of experimental procedures, macrophage heterogeneity, and/or phosphatase-dependent or -independent control of different signaling pathways.

As noted earlier, how a tyrosine phosphatase is able to down-regulate TLR signaling in macrophages is unclear. Shp1 can be recruited and activated by ITIM-containing receptors, and downstream readouts of MAPK and NF-κB activation are increased in Shp1-deficient cells, but exactly what the substrates of Shp1 are is not known. Shp1 is known to interact with IRAK1 and can reduce its kinase activity in a phosphatase-independent manner [61]. Shp1 has also been shown to associate with IRAK4 in DCs [88] and TRAF6 in osteoclasts [115]. TRAF6 is a key signaling molecule downstream of MyD88-transduced TLR signals, and the interaction of Shp1 with TRAF6 is another possible mechanism allowing Shp1 to dampen TLR signals, but this has not been examined in macrophages.

If mev or spin mutant macrophages seem relatively unaffected by inflammatory stimuli, perhaps macrophage signaling pathways are less sensitive to Shp1-mediated regulation. Indeed, mice lacking Shp1 in various tissue macrophage types (afforded by crossing Shp1-floxed mice to LysM-cre or CX3cr1-cre strains that mediate deletion in some populations of monocytes and tissue macrophages [116]) fail to develop obvious inflammation [unpublished results]. Likewise, Shp1-floxed CD11c-cre mice that lack Shp1 in alveolar macrophages, in addition to DCs, do not succumb to the lethal pneumonitis seen in me or mev mice [64, 89]. Hence, there may be no physiologic consequence in vivo resulting from loss of Shp1 in macrophages.

Like neutrophils and DCs, macrophages express many different inhibitory receptors on their cell surface that can recruit Shp1 and down-regulate signaling pathways. Sirpα and Pirb are the key ITIM-containing receptors in macrophages that recruit Shp1 [117, 118]. These proteins bind preferentially to substrate trapping mutants of Shp1 and are hyperphosphorylated in macrophages from mev mice [20]. It is interesting to note that these proteins are hypophosphorylated in me macrophages, indicating that Shp1 may protect some tyrosine phosphorylated sites. Sirpα plays a key role in the “don’t eat me” signal by recruiting Shp1 in response to engagement with its ligand CD47, which is expressed on RBCs. This transmits an inhibitory signal down-regulating the FcR- and complement-mediated phagocytic response of macrophages [107, 119]. In the lung, surfactant proteins SpA and SpD interact with Sirpα and recruit Shp1 to down-regulate phagocytosis in alveolar macrophages [106]. Shp1 can also mediate activating signals triggered by Sirpα; ligation of Sirpα by antibodies or soluble CD47 triggers production of NO, a process that is dependent on JAK2 and Stat1, and requires the recruitment of Shp1 [120]. Sirpα is also implicated in mediating Shp1 inhibition of TLR pathways. The engagement of Sirpα with a mAb reduces MyD88-dependent TNF-α production in response to LPS in THP1 and U937 cells [121]. In RAW264.7 cells, a Sirpα mutant lacking the cytoplasmic domain, which can no longer bind Shp1, fails to increase TNF-α production in response to LPS [62]. By contrast, there is no difference in cytokine production from Pirb-deficient macrophages stimulated with TLR ligands LPS or Pam3CSK4 [122]. Pirb is constitutively phosphorylated in primary macrophages and associates with Lyn [21], indicating that Pirb may be more important in controlling signals in the basal state. Like mev macrophages, Pirb-deficient macrophages are also hyperadhesive, indicating that recruitment of Shp1 by Pirb negatively regulates macrophage integrin signaling [76, 80]. Pirb signaling through Shp1 may also regulate macrophage polarization, as Pirb-deficient macrophages tend to polarize toward an M1-like phenotype [123].

Shp1-mediated inhibition of signaling downstream of ITAMs or ITAMi signaling is also observed in macrophages. The FcαR signals through the ITAM-containing FcRγ subunit; engagement of the FcαR can inhibit IgG-mediated phagocytosis in human monocytes [124]. Whereas a multivalent ligand capable of cross-linking the receptor promotes an activating signal, a monovalent ligand leads to ITAM-dependent and ERK-dependent Shp1 recruitment. The human FcγRIIa receptor is another example of an ITAM-containing receptor that recruits Shp1 to down-regulate responses. However, clustering of FcγRIIa in a complex with Syk, p85 PI3K, and p62Dok is required to increase phosphatase activity [125]. A weak association between Shp1 and a peptide corresponding to one of the ITAMs found in the receptor is observed, but the clustering may lead to corecruitment of the ITIM-containing FcγRIIb, which could mediate Shp1 recruitment [51]. However, Ben Mkaddem et al. [53] found that engagement of FcγRIIa could inhibit MCP-1-induced chemotaxis and LPS-stimulated proinflammatory cytokine production in cells that expressed no ITIM-containing FcRs. They show that Shp1 interactions with IRAK1 and Vav are important to down-regulate these macrophage responses. Clearly, the recruitment of Shp1 by ITIM- and ITAM-containing receptors is complex.

Osteoclasts originate from a monocyte/macrophage lineage and depend on RANKL and M-CSF for their differentiation. Disruption of these signaling pathways leads to defects in bone homeostasis. Shp1 is implicated in signaling downstream of these ligands. The bone-resorbing ability of osteoclasts from mev mice is increased, which results in lower bone density [126, 127]. Several mechanisms have been suggested to explain how Shp1 can dampen osteoclast activity. Shp1 can interact directly with the CSF1R to reduce phosphorylation of downstream signaling molecules [56]. RANKL signals through its receptor via TRAF6 to activate NF-κB. Shp1 can interact with TRAF6 in a phosphatase-independent manner; overexpression of C453S Shp1, which lacks phosphatase activity, leads to increased RANKL-stimulated phosphorylation of p85 PI3K and Akt and increased NF-κB activation [115]. Kim et al. [128] observed that Lyn is also a negative regulator of osteoclast function. Stimulation with RANKL leads to Lyn, Shp1, and Gab2 association with RANK, causing dephosphorylation of Gab2 and reduced NF-κB activation. Lyn-deficient cells have reduced phosphorylation of Shp1 and enhanced NF-κB activation. Finally, osteoclasts deficient in the ITIM-containing receptors Sirpα, Pirb, and PEACAM-1 show up-regulation of bone resorption in response to RANKL and M-CSF [24, 129, 130]. All of these receptors could recruit Shp1 to mediate inhibitory signals downstream of RANKL and CSF1 stimulation.

MAST CELLS AND BASOPHILS

Mast cells play a key role in the me phenotype, particularly in the lungs. This was demonstrated by crossing me mice with 2 different mouse strains that lack mast cells; Kit^W/Wv mice have a point mutation in the c-Kit receptor but have additional phenotypes [131, 132] and Kit^W-Sh mice [133], where the phenotype is more restricted to mast cell deficiency. Both me mice crossed to Kit^W/Wv mice and mev mice crossed to Kit^W-Sh mice have significantly reduced lung inflammation and hence, prolonged survival. Further analysis of the lungs in mev mice reveals a Th2-like inflammatory response, including eosinophilia, increased Th2 cytokines IL-4, IL-5, and IL-13 but not IFN-γ, and increased Stat6 phosphorylation [134]. This Th2 skewing was also seen in the upper airway and nasal mucosa [135]. mev mice that also lack Stat6 or IL-4 have significantly reduced lung inflammation. It is possible that Th2 skewing results from the presence of Shp1-deficient mast cells that over produce IL-5, causing eosinophil activation. Mast cells are also required for the enhanced LPS-induced airway inflammation and allergen-induced anaphylaxis seen in mev mice [136]. However, loss of Shp1, specifically from mast cells using Shp1-floxed mice crossed to a strain expressing a mast cell-specific Cre (Mcpt5-cre), do not exhibit the lung inflammation seen in me mice [137], suggesting that Shp1-deficient mast cells alone are not sufficient to cause the lethal pneumonitis, and interactions among several different cell types lacking Shp1 are required.

Several groups have investigated signaling in BMMCs that lack Shp1. These cells depend on IL-3 for their generation and survival. Zhang et al. [133] report reduced proliferation of mast cells generated from mev mice, but they are more resistant to apoptosis when IL-3 is removed. This is accompanied by more sustained and increased ERK phosphorylation [138]. Mast cells express FcεRI, engagement of which initiates downstream signals by the FcRγ ITAM-mediated recruitment of Syk kinase, leading to degranulation and cytokine release. Slp76 and LAT are key adaptor proteins involved in FcεRI signaling; their tyrosine phosphorylation is increased in BMMCs from me mice [139]. This results in increased MAPK activation and higher TNF-α and IL-6 production. BMMCs from mev mice also show slight increases in degranulation in response to FcεRI signaling [133]. Challenging mev BMMCs with oxidant stress by treating with H₂O₂ or LPS leads to increased production of IL-4 and IL-13 compared with wild-type mast cells [133]. Shp1-deficient BMMCs are also resistant to antigen-induced cell death as a result of prolonged MAPK signaling and expression of the anti-apoptotic protein Bcl-xL [140]. Treatment of wild-type BMMCs with IL-4 leads to decreased expression of FcεRI, which is dependent on Stat6 phosphorylation. Unlike BMMs, loss of Shp1 in BMMCs does not affect Stat6 phosphorylation, but it prevents the down-regulation of FcεRI in response to IL-4 [141].

Shp1 can associate with the IL-3R and FcεRI [58, 142]. Kimura et al. [142] show that Shp1 is constitutively associated with FcεRI and is phosphorylated in response to receptor engagement; however, the interaction is not mediated by Shp1 SH2 domains. Mast cells express several different ITIM-containing receptors, including Pirb, gp49B1, FcγRIIB, and MAFA (Fig. 4D), and it is likely that these are involved in recruiting Shp1 to down-regulate signaling (reviewed in Kawakami and Galli [143]).

It has been suggested that Shp1 plays a positive role in basophil development. Zhou et al. [136] show an increase in basophil differentiation from mev BM cells. However, no obvious inflammatory phenotype was observed in mice when Shp1 deficiency was restricted to basophils [unpublished results].

MYELOID LINEAGE-SPECIFIC DELETION OF Shp1

The Shp1-floxed mouse has proven to be an indispensible tool for dissecting the role of different myeloid cells in the complex phenotype of the me mouse (summarized in Table 3). However, this approach has only been as good as currently available Cre-expressing strains allows. Specific deletion of Shp1 in neutrophils is achieved using the MRP8-cre strain and therefore, clearly demonstrates a necessary and sufficient role for this cell type in inflammation [64]. Likewise, we [64] and others [89] achieved specific deletion of Shp1 in DCs, mediated by the CD11c-cre strain, to show a role in autoimmune disease. Available Cre strains do not achieve specific deletion of Shp1 in macrophages or monocytes. The LysM-cre strain is often used to mediate deletion in myeloid cells, including macrophages and neutrophils. Although we detected good deletion of Shp1 in tissue macrophages using this strain, activation of the LysM promoter occurred later in neutrophil development and did not lead to Shp1 deletion in mature neutrophils, which explains why we [unpublished results] did not see any inflammatory disease in this model. Given that mast cells are required for inflammatory lung disease in the me mouse [131–133], it is somewhat surprising that specific deletion of Shp1 in mast cells produces no phenotype. Clearly, the me phenotype arises from crosstalk among different subsets of Shp1-deficient myeloid cell populations.

HOW PATHOGENS HIJACK Shp1 IN MYELOID CELLS TO EVADE THE IMMUNE RESPONSE

Pathogens have developed several different ways to hijack host myeloid cells to evade the immune response that would cause their destruction. Often, these mechanisms give us further insight into the function of cellular proteins. Leishmania inhibits several macrophage functions, in particular, NO production, to survive and propagate. It has been known for some time that Shp1 is activated in Leishmania-infected macrophages, and resistance to infection is enhanced in Shp1-deficient mice [144, 145]. The Shp1 inhibitor SSG is used to treat leishmaniasis in humans [146]. Leishmania elongation factor-1 alpha has been identified as a Shp1-binding and -activating protein, causing a reduction in JAK2/Stat1 signaling and MAPK activity, leading to reduced macrophage NO production [147]. The Leishmania virulence factor GP63 is a metalloprotease that cleaves Shp1, leading to activation of its phosphatase activity and reduced JAK2 phosphorylation [148]. However, mev mice infected with Leishmania in their rump instead of the footpad, i.e., an area less affected by spontaneous inflammation, and infected PT-elicited macrophages do not show a difference in survival of the parasite [149]. Hence, further investigation into Shp1-regulated macrophage defense against Leishmania is warranted.

Several pathogens take advantage of ITIM-containing receptors to hijack Shp1 activity. Opa proteins of Neisseria bacteria bind to CEACAM-1 in neutrophils, recruiting Shp1 to suppress the host inflammatory response [73]. One might predict that mice with a neutrophil-specific deletion of Shp1 would be protected against infection from this pathogen. Sialic acids on group B Streptococcus and Klebsiella pneumoniae bind to hSiglec 9 (or mSiglec E) on neutrophils, resulting in reduced oxidative burst, reduced neutrophil extracellular trap formation, and increased bacterial survival [150, 151]. In macrophages, these bacterial interactions with Siglec 9 result in increased production of IL-10 [152]. A high molecular weight hyaluronan, expressed by group A Streptococcus, also binds Siglec 9 [153]. Staphylococcus aureus binds to Pirb on macrophages, which suppresses inflammasome activation and results in reduced IL-6 and IL-1β production [154]. The enteropathogenic Escherichia coli protein Tir is inserted into the host cell plasma membrane, such that an ITIM found in its C-terminus can interact with Shp1. This leads to increased binding of Shp1 to TRAF6, reduced activation of NF-κB and MAPKs, and reduced production of TNF-α and IL-6 from infected macrophages [155].

There are also examples of pathogens taking advantage of signaling pathways mediated by ITAMi recruitment of Shp1. Leishmania leads to dampened DC activation, as detailed earlier, which Iborra et al. [99] have recently shown to be the result of a ligand produced by Leishmania that binds to the receptor Mincle and sends an inhibitory signal through the FcRγ ITAM to activate Shp1. Likewise, E. coli binds to FcγRIII, which signals through FcRγ ITAM to recruit Shp1, resulting in reduced PI3K activity to reduce phagocytosis of the bacteria through the scavenger receptor, MARCO [156].

Pathogens express several virulence factors that can interfere with host cell responses. Anaplasma phagocytophilum infects and survives in neutrophils, escaping killing. The AnkA protein virulence factor is phosphorylated by SFKs in a Glu-Pro-Ile-Tyr-Ala (EPIYA) motif early during infection, which recruits and activates Shp1 to block neutrophil responses [157]. Mycobacterium tuberculosis infects monocytes and has a virulence factor Lipoarabinomannan; a cell wall glycolipid that promotes tyrosine phosphorylation and activation of Shp1, leading to reduced LPS-stimulated TNF-α and IL-12 production; and increased bacterial survival [158]. The Bordatella pertussis virulence factor CyaA binds integrin αMβ2 on neutrophils and macrophages to increase cAMP signaling and activate PKA, which leads to the activation of Shp1 by an unknown mechanism, causing decreased NO production and increased survival of the pathogen [159]. These examples highlight the importance of controlling Shp1 activation to modulate innate immune responses.

Shp1 AND MYELOID CELLS IN HUMAN DISEASE

Cancer

The large number of protein tyrosine kinases that are found activated in human tumors indicates that cells must carefully balance their levels of protein tyrosine phosphorylation to maintain homeostasis. Consequently, there are also several examples of Shp1 dysregulation in human tumors (reviewed in Watson et al. [2], Wu et al. [160], and Demosthenous et al. [161]). Several myeloid tumors modify Shp1 signaling to “release a brake” and promote their growth. AML uses the ITIM-containing receptor leukocyte-associated Ig-like receptor 1 (LAIR1) to maintain cell survival by recruiting Shp1. However, Shp1 acts in a phosphatase-independent manner as an adaptor protein to bind calcium/calmodulin-dependent protein kinase type 1 (CAMK1) and activate the CREB pathway in this situation [162]. Activation of JAK/Stat signaling often occurs in myelodysplastic syndromes and myeloid malignancies [163]. Shp1 is an important regulator of this signaling pathway, and there are several examples where Shp1 regulation of this pathway is reduced. There are examples of PTPN6 promoter methylation, thus reducing Shp1 expression to drive cell growth in a wide variety of B cell lymphomas and leukemias [164]. Aberrant splicing of PTPN6 is seen in AML patients, which may lead to reduced protein expression [165]. AML patients often have mutations in FLT3, and cell lines transformed with FLT3 mutants show decreased Shp1 mRNA expression [166, 167]. Loss of Shp1 was found to correlate with a worse prognosis for myelodysplastic syndrome progression into acute leukemia [168]. Some patients with severe congenital neutropenia are predisposed to AML and have mutations in G-CSFR that prevent Shp1 from down-regulating Stat activation in response to G-CSF [71]. Other myeloid malignancies driven by activated tyrosine kinase signaling also demonstrate mechanisms for reducing the ability of Shp1 to affect these pathways. Mast cell leukemic cell lines, driven by gain-of-function mutations in the Kit receptor, down-regulate Shp1 through ubiquitination [169]. Given that Shp1 can dephosphorylate Bcr-Abl, decreased Shp1 is observed in advanced-stage chronic myeloid leukemia patients, mediated by a post-transcriptional modification [170, 171]. Most recently, Demosthenous et al. [161] reported the presence of missense mutations in PTPN6 in 5% of diffuse large B cell lymphomas. These mutations resulted in loss of phosphatase function, leading to JAK3-mediated deregulation of STAT3 signaling to promote tumor cell proliferation. Several nonhematopoietic tumor cell lines show increased Shp1 protein levels, but whether this is driving transformation by an unknown mechanism or is simply a cellular response to dampen signaling through tyrosine kinases is unclear.

Inflammatory disorders

Alterations in the PTPN6 gene have been reported in several cases of skin inflammatory disorders (neutrophilic dermatoses), including Sweet syndrome and Pyoderma gangrenosum [79]. These mutations appear to result in reduced Shp1 protein, and the phenotype is reminiscent of mice with Shp1 deficiency in neutrophils. Recent expression-profiling experiments from the meB2 Shp1 mutant strain highlight the involvement of inflammatory gene expression, in particular, IL-1β, in the inflammatory neutrophilic dermatoses in mice [172].

Autoimmune disease

Reduced Shp1 expression is observed in PBMCs and PBMC-derived macrophages from MS patients, resulting in increased tyrosine phosphorylation of Stat1 and -6, activation of NF-κB signaling, and increased TNF-α and IL-6 production [173]. This is consistent with the observation that Shp1-deficient mice, infected with a virus that induces demyelinating disease, show an increase in inflammatory macrophage infiltration in the CNS and more severe disease than wild-type mice [114]. Christophi et al. [174] suggest that Shp1 is the target for IFN-β-mediated control of disease in MS patients; IFN-β treatment leads to transcriptional activation of the Shp1 promoter and an increase in Shp1 protein.

Unanswered questions

The me mouse was described over 40 y ago, and it has remained an important model for studying the role of Shp1 in myeloid cell biology. Despite much progress, there are still many unanswered questions regarding the role of Shp1 in myeloid cells. It is clear that Shp1 can impact signaling in all myeloid cell types. Loss of Shp1 in specific myeloid cells is even sufficient to cause certain aspects of the me phenotype. However, there are aspects of the me phenotype that are not entirely explained; in particular, the lethal pneumonitis seen in me and mev mice is not observed in other models. It is likely that this phenotype arises because of interactions among several cell types, but this remains to be elucidated. The phosphatase vs. nonphosphatase functions of Shp1 are also not fully understood. This is a particularly important question in the context of designing Shp1 inhibitors to treat human disease, and it may explain contradictory results observed using cells obtained from the different Shp1 mutant mouse models. The effects of Shp1 deficiency in adult mice in the absence of developmental abnormalities is unknown and is another important consideration for the use of Shp1 inhibitors in humans. A clearer understanding of the key Shp1 substrates involved in different myeloid cells will enhance our understanding of the signaling pathways impacted by Shp1 and could be useful biomarkers for monitoring Shp1 activity in human disease. As a key regulator of innate immune cell function, it will be exciting to see if we can modulate Shp1 to treat human disease in the future.

AUTHORSHIP

C.A.L. and C.L.A. wrote the review.

Glossary

AML

acute myeloid leukemia

BM

bone marrow

BMDC

bone marrow-derived dendritic cell

BMM

bone marrow-derived macrophage

BMMC

bone marrow-derived mast cell

cDC

conventional dendritic cell

CEACAM-1

carcinoembryonic antigen-related cell adhesion molecule 1

DC

dendritic cell

DCIR

dendritic cell immunoreceptor

FasL

Fas ligand

Hox

homeobox

hSiglec

human Siglec

IFNAR

IFN-α/β receptor

IFNGR

IFN-γ receptor

IL-1R

IL-1 receptor

IRAK

IL-1 receptor-associated kinase

ITAMi

inhibitory ITAM

MAFA

mast cell function-associated antigen

me

motheaten

meB2

motheaten B2

mev

motheaten viable

MICL

myeloid C-type lectin-like receptor

MRP8

myeloid-related protein 8

MS

multiple sclerosis

mSiglec

mouse Siglec

Pam3CSK4

palmitoyl-3-cysteine-serine-lysine-4

pDC

plasmacytoid dendritic cell

PI(4,5)P₂

phosphatidylinositol 4,5-bisphosphate

Pirb

paired Ig-like receptor B

PKA/C/D

protein kinase A/C/D

PT

peritoneal

Ptpn6

protein tyrosine phosphatase, nonreceptor type 6

pTyr

phosphotyrosine

RAG-1

recombination-activating gene 1

RANKL

receptor activator for NF-κB ligand

RIP1

receptor-interacting serine/threonine-protein kinase 1

SFK

Src-family kinase

SH2

Src homology 2

Shp1

Src homology region 2 domain-containing phosphatase 1

Sirpα

signal regulatory protein α

spin

spontaneous inflammation

SSG

sodium stibogluconate

TRAF

TNFR-associated factor

DISCLOSURES

The authors declare no conflicts of interest.