Regulation of receptor tyrosine kinase signaling by protein tyrosine phosphatase-1B.

Receptor tyrosine kinases (RTKs) are key regulators of cellular homeostasis. Based on in vitro and ex vivo studies, protein tyrosine phosphatase-1B (PTP1B) was implicated in the regulation of several RTKs, yet mice lacking PTP1B show defects mainly in insulin and leptin receptor signaling. To address this apparent paradox, we studied RTK signaling in primary and immortalized fibroblasts from PTP1B(-/-) mice. After growth factor treatment, cells lacking PTP1B exhibit increased and sustained phosphorylation of the epidermal growth factor receptor (EGFR) and the platelet-derived growth factor receptor (PDGFR). However, Erk activation is enhanced only slightly, and there is no increase in Akt activation in PTP1B-deficient cells. Our results show that PTP1B does play a role in regulating EGFR and PDGFR phosphorylation but that other signaling mechanisms can largely compensate for PTP1B deficiency. In-gel phosphatase experiments suggest that other PTPs may help to regulate the EGFR and PDGFR in PTP1B(-/-) fibroblasts. This and other compensatory mechanisms prevent widespread, uncontrolled activation of RTKs in the absence of PTP1B and probably explain the relatively mild effects of PTP1B deletion in mice.

relay molecules containing src homology-2 and/or phosphotyrosine binding domains, most of which also are RTK substrates. The resultant complexes lead to activation of downstream signaling, including the Ras-Erk and PI-3K/Akt pathways. Ultimately, downstream signaling pathways stimulate new transcription and changes in cellular behavior and/or state (reviewed in Ref. 1).
Because of their pleiotropic actions, RTKs must be regulated carefully. Abnormally increased RTK activity can have dire consequences, including developmental abnormalities (reviewed in Ref. 2), cancer (reviewed in Ref. 3), or fibrosis (reviewed in Ref. 4). Classic PTPs comprise a large family of receptor-like and non-receptor enzymes that share a highly conserved catalytic (PTP) domain that is absolutely specific for phosphotyrosine hydrolysis (reviewed in Ref. 5). Because tyrosyl phosphorylation is reversible, PTPs probably play important roles in the regulation of RTKs and/or their substrates (reviewed in Refs. 5,6).
Based largely on experiments in which wild-type PTPs or their catalytically impaired (dominant-negative) mutants were overexpressed, multiple PTPs, including LAR (7), DEP-1 (8), TC-PTP (9, 10), Shp-1 (11,12), Shp-2 (13), and PTP1B (14 -17) have been implicated in the dephosphorylation of various RTKs. Which, if any, of these PTPs regulate RTK signaling under physiologically relevant conditions of expression has remained largely unclear. Also unknown is the extent to which RTK dephosphorylation, as opposed to other down-regulatory mechanisms such as RTK degradation and/or inhibitory seryl phosphorylation, plays the (a) key role in receptor inactivation.
PTP1B is a widely expressed non-receptor PTP that is localized on intracellular membranes via a hydrophobic C-terminal targeting sequence (18,19). A role for PTP1B in the regulation of many cellular functions has been suggested, including integrin (20 -23), cadherin (24,25), and cytokine receptor signaling (26 -28), cell cycle regulation (29 -31), and the response to cellular stress (32). Multiple studies indicated that PTP1B dephosphorylates the EGFR (16,17) and the insulin receptor (IR) (14,(33)(34)(35). Analysis of PTP1B Ϫ/Ϫ mice confirmed that the IR is a key physiological target of PTP1B (36,37). However, these mice lack any obvious signs of increased activity of the EGFR or PDGFR, such as increased tumor incidence or fibrosis.
We used fibroblasts derived from wild-type (WT) and PTP1B Ϫ/Ϫ mice to address the role of PTP1B in the regulation of RTK signaling. We find that PTP1B Ϫ/Ϫ cells exhibit enhanced and sustained tyrosyl phosphorylation of both RTKs. Despite increased RTK phosphorylation, EGF-and PDGF-evoked Akt activation is not enhanced, whereas Erk activation is only minimally increased. Our results show that PTP1B is a bona fide regulator of EGF and PDGF signaling in vivo, but other cellular regulatory mechanisms, including other PTPs, can largely compensate for loss of PTP1B regulation of these RTKs. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cell Stimulation, Immunoprecipitation, and Immunoblotting-MEFs were seeded at 2.5 ϫ 10 5 cells per 60-mm tissue culture plate for 24 h, starved in serum-free Dulbecco's modified Eagle's medium plus antibiotics for 48 h, and stimulated with EGF (50 ng/ml) or PDGF-BB (50 ng/ml). At the indicated times after stimulation, cells were lysed in Nonidet P-40 buffer (1% Nonidet P-40, 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM sodium orthovanadate, and protease inhibitor mixture (final concentrations, 20 g/ml phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 1 g/ml aprotinin, 1 g/ml of pepstatin, and 1 g/ml of antipain)). Lysates were clarified by centrifugation at 13,000 rpm for 10 min and protein concentrations were determined using a bicinchoninic acid protein assay kit (Pierce Chemical). For immunoprecipitations, lysates were incubated with the appropriate antibodies for 3 h overnight at 4°C. Immune complexes were collected onto protein A-Sepha-rose beads, washed extensively, resolved by SDS-PAGE and transferred onto Immobilon-P membranes (Millipore, Bedford, MA). Immunoblots were blocked with 5% bovine serum albumin or Carnation non-fat dry milk in 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 0.05% Tween 20. After incubation with appropriate primary and secondary antibodies, blots were visualized using enhanced chemiluminescence reagents (ECL; Amersham Biosciences). Quantification was carried out using NIH Image Pro Software Version 1.62; data are expressed as relative units of phosphorylated protein per total protein for each band. Anti-PTP1B and anti-PDGFR␤ antisera were used at 1/1000 for immunoblotting. Anti-human PTP1B monoclonal antibodies (FG6) were used at 1 g/ml for immunoprecipitations and immunoblotting. All other antibodies were used at concentrations as recommended by the supplier.
In-gel Phosphatase Assays-Cells were starved overnight in 0.5% fetal calf serum, treated with 2 mM H 2 O 2 at 37°C for 30 min, and stimulated with 50 ng/ml PDGF at room temperature for 10 min or left untreated. Anti-PDGFR immunoprecipitates were prepared from lysates of one 75-cm 2 tissue culture flask. Control immunoprecipitations were carried out using normal rabbit IgG (Santa Cruz). In-gel phosphatase assays were performed as described by Burridge and Nelson (41) with minor modifications. For substrate, poly(Glu 4 Tyr 1 ) n was radiolabeled with [␥ 32 P]ATP (PerkinElmer Life Sciences) using recombinant human p60c-Src, expressed as a glutathione S-transferase-fusion protein in Escherichia coli. The substrate was cast in a 10% SDS-polyacrylamide gel (acrylamide/bisacrylamide, 30:0.8). Instead of Tween 40, Tween 20 was used throughout, and the final renaturation step was carried out overnight in a buffer containing 50 mM Tris-HCl, pH 7.4, 0.3% 2-mercaptoethanol, 1 mM EDTA, and 0.04% Tween 20.

RESULTS AND DISCUSSION
EGFR and PDGFR Are Hyperphosphorylated in PTP1B Ϫ/Ϫ MEFs-Previous reports indicated that PTP1B dephosphorylates the EGFR, at least when one or both of these proteins are overexpressed. To investigate whether the EGFR is a PTP1B target under physiological conditions, primary MEFs were generated from PTP1B (exon 1) Ϫ/Ϫ mice and WT control mice. Starved MEFs were stimulated with EGF (50 ng/ml) for various times or left unstimulated. Total cell lysates or EGFR immunoprecipitates were resolved by SDS-PAGE and subjected to anti-phosphotyrosine immunoblotting (Fig. 1A). As expected, in starved cells, there was no detectable tyrosyl phosphorylation of the EGFR in total cell lysates or EGFR immunoprecipitates from either WT or PTP1B Ϫ/Ϫ MEFs. However, after stimulation, EGFR tyrosyl phosphorylation was enhanced and sustained in PTP1B Ϫ/Ϫ MEFs compared with control Total cell lysates (TCL) and immunoprecipitates (IP) of either EGFR or PDGFR␤ were subjected to SDS-PAGE in 8% gels and immunoblotted for either anti-phosphotyrosine (pTyr) or receptor levels, as indicated. C, immunoblots of PTP1B, confirming its absence in the lysates of MEFs derived from PTP1B Ϫ/Ϫ mice. 0.13, and 1.55 Ϯ 0.21 at 1, 5, 15, and 30 min, respectively, as quantified by scanning densitometry). Reprobing these blots with anti-EGFR antibodies revealed that the difference in EGFR tyrosyl phosphorylation was not caused by increased EGFR expression in PTP1B Ϫ/Ϫ MEFs, but instead reflected increased specific tyrosyl phosphorylation. Together with the earlier PTP1B overexpression/dominant-negative mutant studies, these data show that PTP1B plays a role in dephosphorylating the EGFR, at least in MEFs. Although to our knowledge PTP1B has not been shown to regulate the PDGFR, PDGFstimulated PDGFR tyrosyl phosphorylation also was increased in the absence of PTP1B (Fig. 1B), (KO/WT ratio Ϯ S.E., 1.53 Ϯ 0.22, 1.72 Ϯ 0.12, 1.94 Ϯ 0.35, and 2.1 Ϯ 0.44 for 1, 5, 15, and 30 min, respectively), suggesting that PTP1B also regulates the PDGFR. Anti-PTP1B immunoblots confirmed the absence of PTP1B protein in the PTP1B Ϫ/Ϫ MEFs (Fig. 1C).
Downstream Signaling Pathways Are Affected Minimally in PTP1B Ϫ/Ϫ MEFs-Although EGFR and PDGFR tyrosyl phosphorylation are increased in the absence of PTP1B, mice lacking PTP1B exhibit no obvious hypermorphism of EGFR or PDGFR pathways. To begin to understand this apparent paradox, we examined signaling pathways downstream of these RTKs in WT and PTP1B Ϫ/Ϫ MEFs. Activation of the Erk pathway, as assessed by immunoblotting with phospho-specific anti-Erk antibodies, was enhanced slightly in response to EGF stimulation of PTP1B Ϫ/Ϫ , compared with WT MEFs (KO/WT ratio Ϯ S.E., 1.02 Ϯ 0.08, 1.31 Ϯ 0.1, 1.36 Ϯ 0.25, and 1.39 Ϯ 0.04 at 1, 5, 15, and 30 min, respectively). Erk activation was enhanced to an even greater extent in PDGF-stimulated PTP1B Ϫ/Ϫ MEFs compared with WT ( Fig. 2A) (KO/WT ratio Ϯ S.E., 1.13 Ϯ 0.15, 1.2 Ϯ 0.07, 1.58 Ϯ 0.14, and 1.76 Ϯ 0.36 at 1, 5, 15, and 30 min, respectively). In contrast, growth factorstimulated Akt activation (as assayed by pAkt immunoblotting) was not increased in PTP1B Ϫ/Ϫ MEFs (compared with WT MEFs) in response to either growth factor, in fact there was a tendency for decreased activation (Fig. 2B)  To begin to understand the molecular basis for these effects, we focused on more immediate events after PDGFR activation. PDGF-stimulated tyrosyl phosphorylation of Shp2 (Fig. 3A) and Shc (Fig. 3B), two signal relay molecules implicated in Erk activation (reviewed in Refs. 42,43), was enhanced significantly in PTP1B Ϫ/Ϫ MEFs compared with WT. Association of Shp2 with the PDGFR is probably increased as well (Fig. 4). Activation of the Akt pathway is mediated, at least in part, by recruitment of phosphatidyl inositol-3 kinase (PI-3K) to the PDGFR (reviewed in Ref. 44). Consistent with the lack of enhanced Akt activation in PDGF-stimulated PTP1B Ϫ/Ϫ MEFs, association of the p85 (targeting) subunit of PI-3K with the PDGFR was not increased in the absence of PTP1B (Fig.  3C). In addition, PDGFR-associated PI-3K activity was not altered significantly in PTP1B Ϫ/Ϫ , compared with WT cells (data not shown).

PTP1B Regulation of PDGFR Tyrosyl Phosphorylation Is Cell-autonomous and Direct-Conceivably, increased RTK activation in PTP1B
Ϫ/Ϫ MEFs could reflect an indirect effect of PTP1B on fibroblast differentiation, rather than an effect of PTP1B on RTK signaling per se. To exclude this possibility, we restored PTP1B expression to PTP1B Ϫ/Ϫ cells (Fig. 4A). A large T Ag-immortalized PTP1B Ϫ/Ϫ fibroblast cell line was generated and reconstituted with human PTP1B (hPTP1B-WT), the PTP1B substrate-trapping mutant D181A (hPTP1B-D/A), or the parental pWZL expression vector (see "Experimental Procedures"). Immunostaining revealed that Ͼ90% of hPTP1B-WT-and hPTP1B-D/A-infected, hygromycin-resistant cells expressed hPTP1B protein (data not shown); thus, these pools can be studied directly, avoiding potential clone-to-clone variation. Notably, PDGF-stimulated (but not basal) PDGFR tyrosyl phosphorylation was greater in the vector control-infected PTP1B Ϫ/Ϫ cell pool, compared with the cells reconstituted with hPTP1B-WT (Fig. 4B). Restoration of hPTP1B-WT expression in immortalized PTP1B Ϫ/Ϫ fibroblasts diminished PDGFevoked Erk activation but left PDGF-evoked Akt activation unaffected (Fig. 4B). The effects of restoring hPTP1B expression on PDGFR tyrosyl phosphorylation were observed only at FIG. 2. Minimal alteration of downstream signaling after growth factor stimulation of PTP1B ؊/؊ MEFs. MEFs were stimulated with either EGF or PDGF␤␤ for the indicated times. Lysates were subjected to SDS-PAGE and immunoblotted with either phospho-specific Erk (pErk), followed by anti-Erk-2 (Erk-2) antibodies (A) or phospho-specific AKT (pAkt) followed by anti-total Akt (Akt) antibodies (B).
submaximal doses of PDGF. The reason for this is not clear, but may be caused, at least in part, by the increased level of PDGFR expression that accompanied immortalization (data not shown). The resultant increase in the concentration of activated PDGFR may allow access to a secondary PDGFR phosphatase with a higher K m value (see below).
These data indicate that PTP1B regulates PDGFR tyrosyl phosphorylation in a cell-autonomous manner. To seek further evidence that PTP1B directly regulates PDGFR tyrosyl phosphorylation, we asked whether the PDGFR could form a stable complex with PTP1B-D/A, which is impaired catalytically, but retains substrate-binding ability (16,43,45). PTP1B immunoprecipitates prepared from starved or PDGFR-stimulated, immortalized PTP1B Ϫ/Ϫ cells reconstituted with hPTP1B-WT, hPTP1B-D/A, or the vector controls were immunoblotted with anti-phosphotyrosine antibodies. An ϳ180-kDa band corresponding in size to the PDGFR coimmunoprecipitated with hPTP1B-D/A but not hPTP1B-WT (Fig. 4C); this interaction was detected only upon PDGF stimulation. Unfortunately, available anti-PDGFR antibodies were insufficiently sensitive to unambiguously identify this band as the PDGFR. However, this is highly likely; using a highly sensitive fluorescence resonance energy transfer approach, we recently demonstrated complex formation between the PDGFR and PTP1B in hPTP1B-D/A-expressing cells (46). Previous studies reported associations between the C215S and/or D181A substrate trapping mutants of PTP1B and the PDGFR (17) as well as the EGFR (16) and IR (15), respectively. Unlike the inducible interaction we observed, in these earlier studies, constitutive complex formation was observed. This most probably reflects the high level of overexpression of PTP1B trapping mutant and/or the RTK used in those studies.
Although coimmunoprecipitation of the PDGFR and WT-PTP1B was not observed by immunoblotting, PDGFR immunoprecipitates contained an ϳ50-kDa PTP activity detectable by a highly sensitive in-gel phosphatase assay (Fig. 4D). This activity is almost certainly PTP1B, because it is absent in PDGFR immunoprecipitates prepared from PTP1B Ϫ/Ϫ cells. Interestingly, PTP1B/PDGFR association was detected only when cells were pretreated with hydrogen peroxide (data not shown). Much recent work suggests that specific PTPs, including PTP1B, undergo physiological oxidation (and thus transient inactivation) by peroxide (reviewed in Ref. 47). Conceivably, PDGFR/PTP1B association requires a conformational change in PTP1B that is induced upon oxidation. Alternatively, transient PDGFR/PTP1B interactions may be "trapped" by the peroxide-induced generation of one or more intermolecular disulfide bonds. A peroxide-dependent association between the EGFR and PTP1B also was detected using similar assays (data not shown).
Although our data show clearly that PTP1B plays a role in regulating EGFR and PGDFR phosphorylation, we have been unable to detect any effect of PTP1B deficiency on the biological activities of these growth factors, at least in MEFs. Consistent with the normal growth factor-induced AKT stimulation in PTP1B Ϫ/Ϫ MEFs, we observed no increase in PDGF-induced ruffling in PTP1B Ϫ/Ϫ MEFs compared with WT, whereas EGF did not evoke significant ruffling in either WT or PTP1B Ϫ/Ϫ MEFs. Likewise, there was no significant difference in the ability of PTP1B Ϫ/Ϫ and WT MEFs to fill an artificial "wound" created by scraping off a strip of cells from a MEF monolayer (data not shown). We were unable to study direct effects of PTP1B deficiency on EGF-or PDGF-stimulated mitogenesis, because neither of these growth factors acts as a sole mitogen in MEFs.
There are several possible reasons why lack of PTP1B (although it has some biochemical effects on EGFR and PDGFR signaling) fails to reach the "threshold" for important biological effects. In immortalized fibroblasts, PTP1B is required for adhesion-evoked Erk activation (Ref. 26). 2 In our experiments, "growth factor"-evoked Erk activation was carried out on at-2 F. G. Haj and B. G. Neel, unpublished observations.   FIG. 3. Phosphorylation of PDGFR targets in wild type and PTP1B ؊/؊ MEFs. Lysates of MEFs, stimulated with PDFGF␤␤ for the indicated times, were immunoprecipitated with anti-Shp2 (A), anti-Shc (B), or anti-PDGFR␤ (C) antibodies, and immune complexes were resolved by SDS-PAGE. Shp2 and Shc immunoprecipitates were immunoblotted with anti-phosphotyrosine (pTyr), -Shp2, and -Shc antibodies, as indicated. PDGFR immunoprecipitates were immunoblotted for PDGFR and for the p85 subunit of PI-3K. tached cells; thus, the requirement of PTP1B for adhesion signaling may limit the increase in Erk activation that results in loss of negative regulation of RTK signaling by PTP1B. However, in PTP1B Ϫ/Ϫ MEFs, integrin-evoked Erk activation is normal, 2 so additional compensatory mechanisms must limit the increase in RTK-evoked Erk activation in primary cells. Other PTPs may be able to substitute for PTP1B, thereby limiting the effect of PTP1B deficiency on EGFR and PDGFR signaling. Consistent with this notion, using the in-gel phosphatase assay, we detected multiple PTPs, including PTP1B, in PDGFR immunoprecipitates (Fig. 4D). One of these activities, corresponding to an ϳ70-kDa species, was increased in PDGFR immunoprecipitates from PDGF-stimulated PTP1B Ϫ/Ϫ compared with WT cells. We have identified, by means of immunodepletion, this increased PTP activity as Shp2 (data not shown). Although some evidence suggests that Shp2 has a positive (i.e. signal-enhancing) function in PDGFR signaling (48 -51), other work indicates a negative regulatory role (52,53). Alternatively, the increase in Shp2 tyrosine phosphorylation might help explain the increased Erk activation in PTP1B Ϫ/Ϫ cells (Fig. 2). Activities of ϳ110 kDa, which represents PTP-PEST, and possibly 48 kDa, which represents TC-PTP, 3 associate with the PDGFR under our conditions. Because receptor-like PTPs tend to be detected poorly by in-gel assays (41), this approach may underestimate the number of PTPs that play a role in PDGFR tyrosyl dephosphorylation. Taken together, however, our results support the idea that one reason for the mild effect of PTP1B deficiency on PDGFR (and, probably, EGFR) signaling is that multiple PTPs can compensate for the lack of PTP1B. These other PTPs are likely to be less efficient (at dephosphorylating the PDGFR and EGFR) than PTP1B; thus, PDGFR (and EGFR) tyrosyl phosphorylation are somewhat increased in PTP1B Ϫ/Ϫ MEFs.
Although dephosphorylation has a role in EGFR and PDGFR inactivation, additional mechanisms may be of equal or greater importance. Chief among these is receptor degradation via lysosomal targeting, promoted by the Cbl family of ubiquitin ligases (reviewed in Ref. 54). Classic studies show that the half-lives of the EGFR and PDGFR decrease dramatically after ligand stimulation. Although recent work indicates that dephosphorylation precedes degradation under normal conditions (46,55), it is not clear whether the former is obligate for the latter. If not, (or if alternative PTPs can substitute for this function of PTP1B), the EGFR and/or PDGFR may be targeted for degradation even if the receptor remains phosphorylated for an abnormally long time after growth factor stimulation in PTP1B Ϫ/Ϫ cells. Comparing EGFR and PDGFR turnover rates in WT and PTP1B Ϫ/Ϫ MEFs should help address these issues. We also cannot exclude the possibility that other RTK-inactivating mechanisms, such as inhibitory seryl phosphorylation events (56), help compensate for the absence of PTP1B.
Why PTP1B deficiency has a larger effect on Erk than on Akt activation also remains to be determined. PTP1B might prefer certain phosphorylation sites on the PDGFR and EGFR; thus, 3 B. Markova and F. D. Bohmer, unpublished observations. FIG. 4. Reconstitution of PTP1B in immortalized murine PTP1B ؊/؊ fibroblasts. Immortalized PTP1B Ϫ/Ϫ cells were generated and then reconstituted with hPTP1B-WT, hPTP1B-D/A, or the parental pWZL vector, as described under "Experimental Procedures." A, lysates from randomly growing cells were subjected to SDS-PAGE and then immunoblotted for hPTP1B and Shp2. B, lysates of PDGF-stimulated PTP1B Ϫ/Ϫ and hPTP1B-WT fibroblasts were subjected to SDS-PAGE and then immunoblotted with either anti-phosphotyrosine (Ptyr), anti-phospho-Erk (pMAPK), anti-total Erk-2 (ERK 2), anti-phospho-Akt (pAKT) and anti-total Akt (AKT) antibodies, respectively. C, lysates of unstimulated and PDGF-stimulated (15 min) PTP1B Ϫ/Ϫ , hPTP1B-WT-, and hPTP1B-D/A-reconstituted cells were immunoprecipitated with anti-hPTP1B antibodies and then subjected to anti-phosphotyrosine (Ptyr) immunoblotting. Right panel, multiple PTPs, including PTP1B, associate with the PDGFR in murine fibroblasts. Lysates of immortalized PTP1B Ϫ/Ϫ fibroblasts, reconstituted with hPTP1B or the parental vector, were subjected to immunoprecipitation with anti-PDGFR antibodies or normal rabbit IgG as control, as indicated. The immunoprecipitates were separated by SDS-PAGE in 10% gels containing 32 P-tyrosine-phosphorylated poly(Glu 4 Tyr 1 ), and PTPs were visualized as activity bands (white) after renaturation. Identification of PTPs is based on immunodepletion experiments and comparison of wild type and PTP1B Ϫ/Ϫ cells. Note that several fragments of PTP1B are detectable. A component of about 40 kDa is present in the PTP1B Ϫ/Ϫ cells and is therefore not a PTP1B fragment. Different parts of the same gel have been exposed for different lengths of time to optimize PTP detection. Lower panels show controls for PDGFR immunoprecipitation (PDGFR-IP) and PDGFR tyrosine phosphorylation (pTyr).
PTP1B deficiency could have greater consequences for downstream signaling pathways emanating from those sites. However, a previous study indicated that the isolated catalytic domain of PTP1B had no preference for individual sites on the PDGFR (49). Compensatory PTPs also could prefer certain PDGFR/EGFR sites. Negative regulatory mechanisms for the Akt pathway could be more robust than those for the Erk pathway, such that a mild increase in upstream RTK activation has a bigger effect on the latter than the former. A final, intriguing possibility is that Akt and Erk activation are directed by RTKs present within different cellular compartments. For example, Akt activation may involve activated RTKs at the plasma membrane, whereas full activation of the Erk pathway occurs inside the cell. This model is consistent with the observation that blocking endocytosis using a dominant negative mutant of dynamin impairs only some signals emanating from the EGFR, notably including Erk activation (57,58). Likewise, preventing internalization of NGF has no effect on TrkA-stimulated Akt activation, whereas Erk activation is impaired (59). Notably, we have shown recently that PDGFR and EGFR dephosphorylation by PTP1B requires RTK endocytosis and occurs on the surface of the endoplasmic reticulum (46). Conceivably, PTP1B deficiency preferentially affects the Erk pathway because PTP1B only dephosphorylates endocytosed RTKs. Most of the above models are not mutually exclusive, and a combination of explanations may account for our findings.
Our results have important implications for understanding the physiological function of PTP1B. PTP1B Ϫ/Ϫ mice are, in general, healthy; although they are hypersensitive to insulin (36,37) and leptin (27,28), they show no evidence of increased EGFR or PDGFR activity. Our results clearly show that although PTP1B plays a role in regulating both of these receptors, other regulatory mechanisms come into play when it is missing. These redundant signaling mechanisms may be designed to prevent the dire consequences of a generalized increase in RTK signaling that would occur were a single PTP-or a single RTK inactivation mechanism-to regulate the EGFR, PDGFR, and/or other RTKs. If this explanation can be generalized to other RTK/PTP interactions, it might explain why no classic PTPs have been found to be tumor suppressor genes. Why, then, is the IR different? That is, why is PTP1B a biologically limiting regulator of IR signaling? An important clue may be provided by the different fates of the respective ligand-stimulated RTKs. Whereas most activated EGFRs and PDGFRs are targeted for degradation (60), the vast majority of IRs are recycled (61). A recycling receptor may be especially dependent on dephosphorylation for its inactivation. Further work is needed to determine whether differential trafficking accounts for the differential effects of PTP1B deficiency on RTKs.