Background The epidermal growth factor (EGF) stimulates rapid tyrosine phosphorylation of EGF-receptor (EGFR). This event precedes signalling from both the plasma membrane and from endosomes, and it is essential for recruitment of an ubiquitin ligase, CBL, that sorts activated receptors to endosomes and degradation. Because hyper-phosphorylation of EGFR is involved in oncogenic pathways, we performed an unbiased screen of siRNA oilgonucleotides targeting all human tyrosine phosphatases. Results We report the identification of PTPRK and PTPRJ (DEP-1) as EGFR-targeting phosphatases. DEP-1 is a tumour suppressor that dephosphorylates, thereby stabilizes EGFR by hampering its ability to associate with the CBL-GRB2 ubiquitin ligase complex. DEP-1 silencing enhanced tyrosine phosphorylation of endosomal EGFRs and, accordingly, increased cell proliferation. In line with functional interactions, EGFR and DEP-1 form physical associations, and EGFR phosphorylates a substrtae trapping mutant of DEP-1. Interestingly, the interactions of DEP-1 and EGFR are followed by physical segregation: whereas EGFR undergoes endocytosis, DEP-1 remains confined to the cell surface. Conclusions EGFR and DEP-1 physically interact at the cell surface and maitain bidirectional enzyme-substrate interactions, which are relevant to their respective oncogenic and tumor suppressive functions. These observations highlight the emerging roles of vesicular trafficking in malignant processes.

The tyrosine phosphatase PTPROt is a suggested tumor suppressor (TS) in B-cell chronic lymphocytic leukemia (CLL), and its expression is reduced in this disease. In order to examine how reduced PTPROt expression affects CLL in vivo we induced CLL in PTPROt-targeted mice. Unexpectedly, loss of both Ptprot alleles delayed disease detection and progression and lengthened survival relative to mice carrying two intact alleles, indicating that PTPROt fulfills a novel tumor-promoting role in CLL. Tumor cells from mice lacking PTPROt exhibited reduced B-cell receptor (BCR)-induced signaling, as well as increased apoptosis and autophagy. Inhibition of BCR/Src signaling in CLL cells induced their apoptosis, indicating that these findings are linked causally. These results suggest a cell-autonomous mechanism for the weakened CLL phenotype of PTPROt-deficient mice and uncover non-redundant roles for PTPROt in support of BCR signaling and survival of CLL cells. In contrast, loss of only one Ptprot allele induced earlier detection and progression of CLL and reduced survival, consistent with a tumor-suppressing role for PTPROt. Tumor cells from mice lacking one or both Ptprot allele exhibited increased interleukin-10 (IL-10) expression and signaling, factors known to support CLL; cells lacking one Ptprot alleles exhibited normal BCR signaling and cell death rates. We conclude that loss of one Ptprot allele promotes CLL, most likely by activating IL-10 signaling. Loss of both Ptprot alleles also reduces BCR signaling and increases cell death rates, offsetting the IL-10 effects and reducing the severity of the disease. PTPROt thus functions as an obligate haploinsufficient TS in CLL, where its expression levels determine its role as a promoter or inhibitor of the tumorigenic process in mice. Partial loss of PTPROt generates the strongest disease phenotype, suggesting that its intermediate expression levels in CLL are selected for.

Bone resorption by osteoclasts is essential for bone homeostasis. The kinase Src promotes osteoclast activity and is activated in osteoclasts by the receptor-type tyrosine phosphatase PTPROt. In other contexts, however, PTPROt can inhibit Src activity. Through in vivo and in vitro experiments, we show that PTPROt is bifunctional and can dephosphorylate Src both at its inhibitory residue Tyr527and its activating residue Tyr416. Whereas wild-type and PTPROt knockout mice exhibited similar bone masses, mice in which a putative C-terminal phosphorylation site, Tyr399, in endogenous PTPROt was replaced with phenylalanine had increased bone mass and reduced osteoclast activity. Osteoclasts from the knock-in mice also showed reduced Src activity. Experiments in cultured cells and in osteoclasts derived from both mouse strains demonstrated that the absence of phosphorylation at Tyr399caused PTPROt to dephosphorylate Src at the activating site pTyr416. In contrast, phosphorylation of PTPROt at Tyr399enabled PTPROt to recruit Src through Grb2 and to dephosphorylate Src at the inhibitory site Tyr527, thus stimulating Src activity. We conclude that reversible phosphorylation of PTPROt at Tyr399is a molecular switch that selects between its opposing activities toward Src and maintains a coherent signaling output, and that blocking this phosphorylation event can induce physiological effects in vivo. Because most receptor-type tyrosine phosphatases contain potential phosphorylation sites at their C termini, we propose that preventing phosphorylation at these sites or its consequences may offer an alternative to inhibiting their catalytic activity to achieve therapeutic benefit.

The erbb-2 gene receptor is often over-expressed in human cancer and its overexpression is accompanied by worse prognosis. Targeting erbb-2 gene with antibodies is an effective approach to curtail the progression of erbb-2 gene-expressing cancer types. Two monoclonal antibodies, L-26 and N-12, previously generated in our laboratory, have shown effective tumor inhibition in mice, especially when used in combination. Here, we describe novel peptide mimics of erbb-2 gene protein epitopes, also called mimotopes, that were selected from a constraint random 12-mer peptide phage library, specific for the antibodies L-26 and N-12. Initial sequencing analyses revealed little sequence conservation among the peptide mimotopes, and no sequence homology with the erbb-2 gene protein. However, computational analyses of the two groups of peptides, specific for L-26 and N-12, suggested different epitopes on the erbb-2 gene extracellular domain. In vitro assays showed that the phage displayed peptide mimotopes were specific to their respective antibodies. Selected cyclic peptide mimotopes, but not their corresponding linear equivalents, were able to inhibit binding of the antibodies L-26 and N-12 to the surface of erbb-2 gene-expressing cancer cells in a concentration-dependent manner. In line with this observation, phage-displayed cyclic peptides successfully competed in vitro with recombinant erbb-2 gene protein for binding to their respective antibodies L-26 or N-12. Consistent with the antibody inhibition experiments, we detected specific anti-erbb-2 gene antibodies following vaccination with KLH-coupled cyclic peptides but not with multiple antigenic linear peptides. Potentially, the selected peptides could serve as a starting point for the development of a vaccine against erbb-2 gene over-expressing cancer.