Negative Signaling by Inhibitory Receptors

NK cell tolerance toward normal cells is derived from "negative signaling" that originates through cell surface inhibitory receptors, which detect MHC-I molecules on the surface of normal cells. The MHC-I-binding inhibitory receptors expressed on NK cells include killer cell Ig-like receptors (KIR; human), Ly49 (mouse), NKG2A/CD94 (human and mouse), and ILT2/LIR1 (CD85j; human). The engagement of NK cell inhibitory receptors with MHC-I molecules on normal target cells causes them to coaggregate with activating receptors that are simultaneously interacting with ligands at the target cell interface. The magnitude of inhibition is proportional to the degree of MHC-I engagement. If sufficiently engaged, the inhibitory receptors efficiently and dominantly block downstream signals that are initiated by the activating receptors. Accumulated evidence indicates that this inhibition is primarily mediated through recruitment of two PTPs, named SH2 domain-containing protein tyrosine phosphatase-1 (SHP-1) and SHP-2, to the cytoplasmic domains of the inhibitory receptors. Numerous early activation signaling events are abolished upon inhibitory receptor engagement with MHC-I, most notably intracellular calcium mobilization [28,82].

As mentioned above, the physical contact interface between an NK cell and a normal MHC-I-expressing cell that is resistant to attack has been termed the inhibitory NKIS (iNKIS). The accumulation of receptors and signaling molecules at the iNKIS is referred to as the supramolecular inhibition cluster (SMIC). Target cell contact at the iNKIS still initiates transient LFA-1-mediated adhesion and the rapid polarization of talin toward the target cell [167]. The accumulation of talin and Lck quickly dissipates from the iNKIS, however, and adhesion is disrupted within minutes, in sharp contrast with the stable adhesion and accumulation of these molecules at the cNKIS [25, 167]. Furthermore, raft polarization, actin polymerization, and MTOC reorientation are lacking in the iNKIS, whereas these events can last more than 15 min in the cNKIS [54,104,142,167, 169]. PKC-6, PLCy, Itk, ZAP-70, SLP-76, and BLNK do not accumulate at the iNKIS [169]. In fact, the only signaling molecules known to be recruited to the c-SMIC within 1 min of target cell conjugation are Lck and SHP-1 [167]. A major distinguishing feature of the iNKIS is early SHP-1 accumulation at the c-SMIC, which is also the site at which inhibitory KIR or Ly49 interacts with MHC-I [47,53,167]. One report noted that CD45, CD43, and ezrin are lacking at the iNKIS, whereas all three are evenly distributed within the cNKIS [112]. Further analysis in that report revealed that the narrow gap between human NK cell and target cell surfaces within the iNKIS (about 15 nm) corresponds to the height of interacting KIR/HLA-C extracellular domains, whereas CD45 and CD43 are taller, which may explain their physical exclusion from this narrow interface at the c-SMIC [112].

NK cell inhibitory receptors function through immunoreceptor tyrosine-based inhibitory motifs (ITIMs) in their cytoplasmic domains. ITIMs are (I/V)xYxx(L/V) sequences that, when phosphorylated on NK cell inhibitory receptors, become specific binding sites for SHP-1 and SHP-2 [21,24,28,101, 102,113, 188]. SHP-1 and SHP-2 contain tandem SH2 domains that interact with the tyrosine-phosphorylated ITIMs. Inhibitory Ly49, KIR, ILT2/LIR1, and NKG2A have been shown to recruit SHP-1 and/or SHP-2 to varying degrees via phosphorylated ITIMs [48,81,94,113]. KIR enrichment at the iNKIS is delayed on truncation of the cytoplasmic domain or in the presence of high doses of an inhibitor of actin polymerization, indicating roles for the cyto-plasmic domain and actin in the efficiency of KIR clustering toward a resistant target cell [149]. KIR with mutant ITIM tyrosines can still accumulate at the iNKIS but are unable to inhibit lipid raft polarization [54], and dominantnegative SHP-1 can block KIR-mediated inhibition of raft polarization [104].

The following discussion will focus primarily on studies with KIR, which are the best-characterized MHC-I-binding inhibitory receptors. SHP-1 bind ing to the cytoplasmic domain of KIR occurs only when both ITIMs become tyrosine-phosphorylated. SHP-2, however, can bind when only one ITIM is phosphorylated [21, 58,188,189]. Mutational analysis has shown that SHP-2 binds with higher avidity to the phosphorylated amino-terminal ITIM on KIR, and SHP-2 binding avidity correlates well with inhibitory capacity of various mutant receptors [21,188,189]. The PTP recruitment patterns from these studies form the basis of the model shown in Fig. 4. Surprisingly, mutation of both KIR ITIM tyrosines to phenylalanine has been shown to result in a receptor that weakly associates with SHP-2 and is still slightly inhibitory [188, 189]. Furthermore, an unphosphorylated peptide encompassing the amino-terminal ITIM readily binds SHP-2, suggesting that SHP-2 may con-stitutively interact with KIR in the unphosphorylated, resting state [189]. One KIR, named KIR2DL5, uniquely possesses an altered carboxy-terminal ITIM (TxYxxL), which results in a receptor that inhibits NK cell cytotoxicity in a SHP-2-dependent manner [190]. In addition to direct recruitment of SHP-1 and SHP-2 to the iNKIS by inhibitory receptors, binding of these PTPs to phosphorylated ITIMs releases them from a self-inhibiting conformation in the cy-tosol where the amino-terminal SH2 domain blocks the catalytic domain [70, 161]. This autoinhibition has been shown to be abrogated in vitro by binding of the SH2 domains to tyrosine-phosphorylated KIR ITIM peptides [24,28,131].

PhospholTIM: N+C

ITIM Domains



Inhibition: STRONG



weak weak

Fig. 4 A model of the patterns of recruitment of SHP-1 and SHP-2 to phosphorylated amino (N)- and carboxy (C)-terminal ITIMs of KIR and their inhibitory consequences. This model is based on biochemical and functional studies described in the text that examined mutant KIR in which the cytoplasmic ITIM tyrosines were selectively changed to phenylalanine, which cannot be phosphorylated


SHP-1 and SHP-2 share 60% sequence identity and very high homology in secondary and tertiary structures. Despite the high homology, they generally play very different roles in vivo, with SHP-1 acting as a negative regulator for many inhibitory receptors (including CD22, CD72, PIR-B, and CD5) [147, 192], whereas SHP-2 is primarily a positive regulator (PDGFR, EGFR, ICAM-1, PAR-2, andleptin receptor) [11,18,56,57,134,166,187,192]. Nonetheless, examples of positive effects on signaling by SHP-1 [117, 182] and negative influences mediated by SHP-2 [41, 86, 95,110,185] have also been reported. It is important to note that electrostatic charge differences surrounding the catalytic clefts influence substrate recognition by the two PTPs, with SHP-2 expected to prefer phosphotyrosines flanked by more acidic residues [184]. Accordingly, numerous examples of differential substrate specificities between SHP-1 and SHP-2 catalytic domains have been reported [116,125,157, 172,183]. Therefore, it is tempting to speculate that these two PTPs may play different functional roles at the iNKIS as a consequence of both recruitment differences dependent on phosphoITIM status of inhibitory receptors and their distinct substrate recognition capacities.

Recruitment and activation of SHP-1 and SHP-2 by phosphorylated inhibitory receptors is believed to prevent NK cell activation by dephospho-rylating numerous signaling intermediates at the iNKIS. When human NK cells engage with MHC-I-expressing target cells, tyrosine phosphorylation has been shown to be abrogated in a number of substrates, including Src family PTKs, PLCy, ZAP-70, Vav, SLP-76, LAT, Grb2, PI3K, and the ITAMS of Z [17,150,165]. It is unclear which of these are direct substrates of SHP-1 or SHP-2 or whether the reduced phosphorylation of some is a consequence of upstream dephosphorylation events. The inhibitory impact, however, appears to be at the level of and downstream from Syk family PTKs [17].

Two studies have provided evidence for direct SHP-1 substrates in human NK cells using substrate-trapping forms of SHP-1. First, Binstadt et al. specifically isolated tyrosine-phosphorylated SLP-76 from NK cell lysates in vitro [17]. Second, Stebbins et al. specifically isolated Vav-1 as a substrate on engaging a chimeric KIR/SHP-1 receptor during conjugation of an NK cell line with MHC-I-expressing target cells [150]. Although both of these studies offer important mechanistic insight, alternative evidence suggests that other relevant substrates exist. For instance, NK cells from SLP-76-deficient mice exhibit normal natural cytotoxicity and ADCC [133], suggesting that SLP-76 is not the key substrate that blocks activation. Futhermore, SHP-1 was shown previously to physically associate with Vav-1 via SH2/SH3 domain interactions [87], suggesting that this interaction might have enhanced capture in the substrate trapping experiments of Stebbins et al. Nonetheless, Vav activation is clearly an important early event in development of the cNKIS, and reversal of its activation by SHP-1-mediated dephosphorylation would indeed abrogate key activation events within the iNKIS.

SH2 domain-containing inositol 5'-phosphatase-1 (SHIP-1) is another negative effector enzyme that can be recruited to ITIMs on several inhibitory receptors, including the B cell receptor, FcyRIIb [67]. By cleaving the 5'-phosphate on PIP2 and PIP3 in the plasma membrane, SHIP-1 has the capacity to deplete the substrate for PLCy and eliminate PI3K-generated binding sites for certain proteins containing PH domains. Gupta et al. have provided convincing evidence that SHIP-1 is not involved in inhibitory KIR function [67]. Wang and colleagues, however, have shown that SHIP-1 can associate with certain inhibitory Ly49 receptors, notably Ly49A and Ly49C, which have broad capacity to bind most MHC-I ligands in mice [48,171]. Interestingly, a substantially greater number of the NK cells that develop in SHIP-1-deficient mice express Ly49A and Ly49C and survive longer, presumably because of enhanced Akt recruitment to elevated PIP3 in their plasma membranes [171]. SHIP-1 can also transiently localize within lipid rafts at the plasma membrane of NK cells during ADCC responses, apparently through associations with Z and an adaptor named Shc [61,62]. Overexpression of SHIP-1 was also shown to suppress ADCC responses, indicating that it can negatively impact upon CD16 function [61]. Therefore, a growing body of evidence suggests that SHIP-1 also plays important negative regulatory roles in NK cell functions.

Was this article helpful?

0 0

Post a comment