Vegfr3 protein, human

EXAMPLE 5

Summary

OPCML, a GPI anchored tumor suppressor gene is inactivated by somatic methylation in multiple cancers. We previously identified this gene by LOH mapping and demonstrated that it was inactivated by somatic methylation in 80% of ovarian cancers. Restoring OPCML expression by stable transfection suppressed in-vitro growth and in-vivo tumorigenicity. We investigated the role of OPCML in growth signaling pathways. In SKOV-3 and PEO1, ovarian cancer cell lines with no expression of OPCML, we demonstrated that OPCML negatively regulates a specific repertoire of receptor tyrosine kinases (RTKs) EPHA2, FGFR1, FGFR3, HER2 and HER4, and reciprocally, OPCML siRNA knockdown in normal ovarian surface epithelial cells up-regulates these same RTKs. OPCML has no effect on the RTKs EPHA10, FGFR2, FGFR4, EGFR, HER3, VEGFR1 and VEGFR3. Example immunoprecipitation experiments revealed that OPCML binds to EphA2, FGFR1 and HER2 extracellular domains with no such interaction to EGFR, thus OPCML binds directly to RTKS that it negatively regulates. We demonstrate that OPCML is located exclusively in the raft membrane fraction and sequesters RTKs that it binds to the raft fraction, leading to polyubiquitination and proteosomal degradation via a cav-1 endosomal mechanism resulting in systems depletion of this specific RTK repertoire, that does not occur with RTKs that OPCML does not bind. We demonstrate that OPCML abrogates EGF mediated phosphorylation of FGFR1, HER2 and EGFR and the downstream phosphosignaling of pErk and pAKT.

A recombinant modified OPCML-like protein without a GPI anchor, signal peptide or glycosylation was constructed and expressed in E. coli. This rOPCML tumor suppressor protein therapeutic caused growth inhibition by apoptosis in 6/7 ovarian cancer cell lines tested, with no effect on OPCML expressing normal ovarian surface epithelium, by an identical mechanism to the transfected normal protein. rOPCML was then injected intraperitoneally twice weekly in two murine intraperitoneal models of ovarian cancer (nude mouse A2780 and SKOV3) and demonstrated profound inhibition of tumour weight, ascites volume and peritoneal dissemination compared with BSA control.

Mechanism of OPCML TSG Function

OPCML is a non-transmembrane, external lipid leaflet GPI-anchored protein, and is frequently lost from cells by somatic inactivation of the gene. We hypothesised that it may mediate its tumour suppressor properties via interactions with transmembrane signalling proteins, and so we analysed the effect of receptor tyrosine kinase (RTK) growth factor stimulation on OPCML gene expression. Treatment of 4/4 ovarian cancer cell lines with EGF or FGF 1/2 resulted in rapid OPCML RNA and concomitant protein expression (data not shown) suggesting that OPCML may be a putative suppressor-type immediate-early negative feedback regulator.

Stable transfection of OPCML in the basal unstimulated or ligand-stimulated SKOV-3 ovarian cancer cells, resulted in the profound protein down-regulation of a specific repertoire of RTKs: EPHA2; FGFR1; FGFR3; HER2 and HER4 (FIG. 16A) and this RTK down-regulation spectrum is reproducible by transient transfection of a different ovarian cancer cell line, PEO1 (FIG. 16B). These same RTKs were also reciprocally up-regulated when physiological OPCML was knocked down by siRNA in OSE-C2, a normal ovarian surface epithelial cell line (Davies et al, (2003) Experimental Cell Research 288: 390-402) (FIG. 16C). This specific inactivation by OPCML was not seen for other RTKs we have investigated so far including: EPHA10; FGFR2; FGFR4; EGFR; HER3; VEGFR1 and VEGFR3 (FIG. 16). The phenotypic consequences of these signalling effects were confirmed in growth assays in ligand-supplemented media where OPCML-transfectants were significantly growth-inhibited compared with vector control (data not shown).

Negative Regulation of Specific RTKs by OPCML is Related to Direct Protein Interaction

We further explored as examples EPHA2, FGFR1 and HER2, RTKs that are strongly inactivated at the protein level upon OPCML expression. We also analysed EGFR as an example of a protein that is unaffected by OPCML. Immunoprecipitation (IP) experiments demonstrated protein/protein interactions with EPHA2, FGFR1 and HER2, but no such binding to EGFR (FIG. 17A). These findings were further confirmed using a recombinant OPCML (GST-OPCML D1-3) pull-down assay (FIG. 17B) which was then used to determine that the extracellular domains (ECDs) of the RTKs FGFR1 and HER2 (as examples) were capable of interacting specifically with OPCML (FIGS. 17C&D), showing that the site of interaction lay within the ECD of the RTKs and domain 1-3 of OPCML, defining the site of OPCML action as extracellular.

Downstream Signalling

Upon acute ligand stimulation, OPCML expression led to profound abrogation of phospho-FGFR1-Y766, phospho-HER2-Y1248 and, also, phospho-EGFR-Y1173. Whilst EGFR total protein down-regulation is NOT observed, presumably due to the absence of an RTK ECD physical interaction with OPCML, the consequence of OPCML mediated loss of the activating dimerisation partners of EGFR, (HER2 and HER4), coupled with the continuing availability of the HER3 family member (that results in an inhibitory dimerisation with EGFR), explain the down-regulation of EGFR signalling even though total EGFR levels are unaffected (FIG. 18A). Analysis of FGFR1 signalling showed a similar pattern of phospho inhibition relating to protein down-regulation (FIG. 18B).

Analysis of downstream signalling demonstrated abrogation of phospho-ERK 1 & 2 (T202 & T204) and phospho-AKT-S473 (FIG. 18C), suggesting that both pro-growth and pro-survival pathways are inhibited by OPCML re-expression, via a systems level abrogation of this specific RTK spectrum.

OPCML-Mediated RTK Degradation Mechanism

Using HER2 as a paradigm molecule of OPCML-RTK regulation, we found that the available HER2 in OPCML expressing cells was sequestered in the detergent resistant membrane (DRM) fraction. In the OPCML non-expressing line, HER2 was found equally distributed between the DRM and the detergent soluble (non-raft) fractions. The total level of EGFR was not affected by the expression of OPCML and its distribution showed a much less pronounced but discernible shift to the DRM fraction (FIG. 19A). These data indicate that OPCML expression leads to redistribution of HER2 to the DRM fraction in the plasma membrane (that broadly correlates with membrane “rafts”). IFM was employed to examine the trafficking of OPCML in cells; EEA-1 (a marker of the early endosome) and Caveolin-1 (a marker of the raft-caveolar pathway) were used to investigate this apparent redistribution. A decrease in HER2 co-localisation with EEA1 shows that the sequestration of HER2 to the DRM fraction decreases its endocytosis via clathrin-mediated pathways. While an increase in co-localisation with caveolin-1 was observed, the immunofluorescence pattern suggests this is a function of the redistribution of HER2 into the DRM fraction (housing lipid-raft domains) where caveolin is also localised, as HER2 did not appear to be exclusively localised to caveolae in the presence of OPCML expression. Furthermore, in the presence of OPCML the staining was organised into specific sub-cellular particles, suggestive of distinct vesicular compartments (FIGS. 19B&C).

This analysis demonstrated that OPCML expression was associated with increased ubiquitination of HER2 (that binds OPCML), which was strongly increased upon EGF stimulation (FIGS. 19D&E). Exposure to MG-132, a potent inhibitor of the proteasomal 26S proteinase, attenuated HER2 degradation with no such effect on EGFR expression (that does not bind OPCML). In contrast, chloroquine, a weak base that alkalinises the lysosome, showed no inhibition of HER2 degradation (FIGS. 19H&I). This suggested that the proteasomal pathway was preferentially utilised for OPCML-mediated HER2 degradation. Furthermore, disruption of cholesterol (a component of DRM fraction/lipid rafts) using methyl-β-cyclodextrin (Mβ-CD) also inhibited the degradation of HER2 and increased HER2 phosphorylation (FIG. 19J) suggesting that cholesterol-rich lipid-raft structures are important for OPCML-specific internalisation and degradation of HER2.

These findings suggest that OPCML-mediated negative regulation of this specific repertoire of RTKs is the result of direct binding of OPCML to the ECD of that RTK. These multiple but specific binding events result in ‘lipid-raft’ sequestration, enhanced ubiquitination, and a switch away from clathrin-mediated endocytosis to proteasomal degradation of those specific RTKs negatively regulating their signaling through reducing their protein level. Our data, in the context of very recent publications (Howes et al (2010) J. Cell Biol. 190(4): 675-91; Howes et al (2010) Curr. Opin. Cell Bio. 22(4): 519-527)), would suggest that CLIC/GEEC bulk internalization route is a strong candidate pathway for OPCML-mediated degradation of HER2 and that this is linked to RTK inactivation and the observable strong tumour suppressor phenotype of OPCML.

Recombinant OPCML (r-OPCML) inhibits tumour growth in vitro and in vivo Purified recombinant human OPCML domain 1-3 protein (r-OPCML) (FIG. 8) was produced from the bacterial expression vector (pHis-Trx) subcloned with domains 1-3 of OPCML, excluding the signal peptide and GPI anchor sequences (FIG. 8A). Addition of r-OPCML protein to growth media demonstrated a specific, dose-dependent inhibition of cell growth in OPCML non-expressing SKOV-3 ovarian cancer cells, without affecting normal ovarian surface epithelial cells, OSE-C2 (FIG. 20A). We have confirmed that r-OPCML profoundly inhibited cell growth in 6 of 7 additional OPCML non-expressing epithelial ovarian cancer cell lines; 2 of 2 breast HER2-positive and negative cells; and 5 of 5 lung cancer cell lines (FIG. 20B). To determine the mechanism of this pharmacological growth inhibition we performed Annexin V FACS apoptosis assay in SKOV-3 and A2780 demonstrating evidence of early apoptosis induced by r-OPCML at 2-6 hours post exposure depending on cell line (FIG. 20C). We then performed caspase-glo apoptosis assays across a concentration range in SKOV-3 and A2780 ovarian cancer cells and demonstrated that r-OPCML induces apoptosis in both these cell lines in a dose dependent fashion, demonstrating the underlying mechanism of the observed growth inhibition (FIGS. 20D&E). Immunoblotting confirmed that addition of r-OPCML protein to media potently downregulated the same spectrum of RTKs as seen by transfecting OPCML into cancer cells, as well as abrogating pERK and pAKT in both SKOV-3 and A2780 (FIG. 21). This suggests that pharmacological use of extracellular unanchored r-OPCML utilises the same mechanism of action as transfection induced intracellular re-expression of the normal GPI-anchored, glycosylated, OPCML protein. These data were confirmed by IFM for HER2 in SKOV-3, closely mirroring stable transfection of the normal protein in the same cell line (see, FIG. 9).

In view of these in-vitro findings, we proceeded to determine whether r-OPCML protein had potential and relevance as an in-vivo tumour suppressor therapy. Mice with either SKOV-3 or A2780 cancer cells injected intraperitoneally (IP), after tumour establishment, received twice-weekly IP injections of either 1 ml (10 μM) bovine serum albumin (BSA) or 1 ml (10 μM) r-OPCML. The experiment was terminated after 3 weeks due to obvious extensive IP tumour growth and deteriorating condition of BSA-treated control animals whereas r-OPCML treated mice remained well (FIG. 22A). r-OPCML significantly and profoundly suppressed both IP tumour growth and ascites formation in-vivo in both IP models (FIG. 22B-D), and in A2780 tumour bearing mice, profoundly inhibited the number of IP peritoneal deposits compared with BSA control (FIG. 22E). Western blotting of SKOV3 IP tumour recovered from BSA treated and r-OPCML treated animals clearly demonstrated the same spectrum of RTKs inhibited as predicted from the in-vitro analysis (FIG. 22F).

EXAMPLE 1

Summary

Epithelial ovarian cancer (EOC) is the leading cause of death from gynecologic malignancy. Its molecular basis is poorly understood but involves dysfunction of p53 (Hall et al (2004) “Critical evaluation of p53 as a prognostic marker in ovarian cancer”. Expert Reviews in Molecular Medicine 6: 1-20), BRCA1 and −2 (Radice (2002) “Mutations of BRCA genes in hereditary breast and ovarian cancer” J Exp Clin Cancer Res. 21(3 Suppl): 9-12), PI3K (Meng et al (2002) “Role of PI3K and AKT specific isoforms in ovarian cancer cell migration, invasion and proliferation through the p70S6K1 pathway” Cellular Signaling 18(12): 2262-2271), and growth factor and angiogenic signaling pathways (Maihle et al (2002) “EGF/ErbB receptor family in ovarian cancer” Cancer Treat Res. 107: 247-58; Le Page et al (2006) “Gene expression profiling of primary cultures of ovarian epithelial cells identifies novel molecular classifiers of ovarian cancer” British Journal of Cancer 94: 436-445; Birrer et al (2007) “Whole genome oligonucleotide-based array comparative genomic hybridization analysis identified Fibroblast Growth Factor 1 as a prognostic marker for advanced-stage serous ovarian adenocarcinomas” Journal of Clinical Oncology 25(16): 2281-2287; Trinh et al (2009) “The VEGF pathway and the AKT/mTOR/p70S6K1 signaling pathway in human epithelial ovarian cancer” British Journal of Cancer 100: 971-978; and Lafky et al (2008) “Clinical implications of the ErbB/epidermal growth factor (EGF) receptor family and its ligands in ovarian cancer” Biochim Biophys Acta. 1785(2): 232-65).

We previously identified opioid binding protein cell adhesion molecule (OPCML) as epigenetically inactivated in 83% of ovarian cancers and demonstrated that it was a functional tumor suppressor in vitro and in vivo (Sellar et al (2003) “OPCML at 11q25 is epigenetically inactivated and has tumor-suppressor function in epithelial ovarian cancer” Nat. Genet. 34(3): 337-43). Here, we show that OPCML interacts with and downregulates HER2 and FGFR1 proteins, leading to inhibition of those signaling pathways, with consequent inhibition of in-vitro growth in SK-OV-3 ovarian cancer cells. siRNA knockdown of physiologically expressed OPCML in OSE-C2 normal ovarian surface epithelial cells strongly upregulated HER2 and FGFR1. OPCML sensitized HER2 positive ovarian cancer cells to lapatinib and trastuzumab in vitro and was a good prognostic indicator in patients with HER2 positive ovarian cancer. The finding that OPCML actively mediates negative regulation of multiple RTK pathways opens novel research avenues in normal cell and cancer biology.

Experimental Procedures

Antibodies

The polyclonal goat and monoclonal mouse anti-OPCML antibodies were purchased from R&D. Anti-HER2 antibodies were purchased from Calbiochem (anti-ErbB2 (Ab-4) and (3B5) mouse MAbs). Anti-EGFR antibody was from R&D Systems. Anti-EGFR goat pAb-cat no AF-231. Phospho-specific EGFR and HER2 antibodies were purchased from AbCam. Anti-HA antibody was from Santa Cruz Biotechnology (Santa Cruz CA) HRP-conjugated secondary antibodies were from Dako. Alexa-Fluor 488 goat anti-rabbit IgG, Alexa-Fluor 555 goat anti-mouse were from Molecular Probes (Eugene, OR).

Cell Culture

The SK-OV-3 derived OPCML expressing lines (SKOBS-3.5, BKS2.1 and empty vector SKOBS-V1.2) were described previously (Sellar et al, 2003). Stimulation time courses were undertaken with 50 ng/ml human recombinant epidermal growth factor (hrEGF-Promega) following serum-starvation overnight.

Plasmid Constructs

The OPCML cDNA expression plasmids in pcDNA3.1zeo previously described (Sellar et al, 2003) were used for transient transfections. The cDNAs encoding all three Ig domains and domains 1 and 2 were generated by PCR and introduced into the bacterial GST-fusion expression vector pGEX-6P-1 (GE-Healthcare) and sequenced to confirm their fidelity. Vector pIRES-AcGFP1 (Clontech) was employed in transient transfections of OPCML complete cDNA. The HA-tagged Ubiquitin pRK5-HA-Ubiquitin-WT was obtained from Dr. Luke Gaughan, Newcastle University, and the EGFR and HER2 cDNA in pcDNA-3.1zeo was provided by Prof. Bill Gullick, University of Kent. FGFR1 cDNA clones was provided by Prof. Graeme Guy, FGFR1 extracellular domain clones provided by Prof. Kyung Hyun Kim.

Expression of Recombinant OPCML and FGFR Ectodomain

Recombinant proteins were produced in the BL21 bacterial cell line (Promega) as described.

Solubilisation and Refolding of Inclusion Bodies

Inclusion bodies were solubilised in denaturation buffer (8 M Urea, 20 mM Tris-HCl, pH 8.0, 150 mM NaCl and 10 mM DTT) to a final concentration of 5 mg/ml. The suspension was centrifuged and filtered through 0.45 μm membrane filter. Refolding of proteins was undertaken by extensive dialysis against cold PBS in 10 kDa MWCO dialysis tubing. The suspension was then centrifuged and filtered to remove insoluble protein precipitates and soluble aggregates. Protein concentrations were monitored throughout the experiment with protein assay reagent (Bio-Rad Laboratories, California) using bovine serum albumen as a standard

RNA Extraction and cDNA Synthesis

Total RNA was extracted from cell pellets using TriReagent® (Sigma-Aldrich, Dorset, UK) following their protocol. Synthesis of cDNA was from 1 μg of RNA template with OligodT₁₅ primers (Promega, UK), by Moloney-Murine Leukaemia Virus Reverse Transcriptase (MMLV-RT) (Promega, UK) and cDNA was stored at −20° C.

qRT-PCR

Primers were designed using PerlPrimer v.1.14 open source software. Custom oligonucleotide synthesis was carried out by Invitrogen, UK. Quantitative reverse-transcription PCR (qRT-PCR) analysis of gene expression was carried out on an Applied Biosystems 7900HT thermal cycler using SYBR green I technology. Premixed qPCR reagent, Platinum® Quantitative PCR SuperMix-UDG with ROX (Invitrogen, UK), was used for amplification. The expression of specific genes was normalized to the expression of the endogenous control gene HPRT1.

Co-Immunoprecipitation and Pull-Down Assays

Cell layers were washed in PBS and incubated for 30 minutes in lysis buffer (1% TritonX-100, 10 mM Tris pH8.0, 150 mM NaCl, 2.5 mM MgCl₂, 5 mM EGTA, 1 mM Na₃VO₄, 50 mM NaF and protein inhibitor cocktail (Roche). Cell Lysates were then cleared by centrifugation at 13,000 rpm for 20 minutes at 4° C. and aliquots containing equal amounts of protein were incubated with the appropriate antibody before addition of secondary antibody conjugated to sepharose resin. Beads were then washed 3× with lysis buffer and eluted by heating for 5 minutes in 50 μl of SDS sample buffer.

Pull-down assays were performed using recombinant GST-OPCML fusion proteins bound to magnetic glutathione beads (Promega). Cell lysates prepared as for immunoprecipitation, proteins produced using TNT in vitro Rabbit reticulocyte lysate expression system (Promega) or expressed in bacteria were used analysed for interactions.

Immunofluorescent Microscopy

Cells grown on glass slides were fixed in 4% paraformaldehyde for 10 minutes at room temperature. Cells were then permeabilized for 20 minutes with PBS containing 0.2% Saponin prior to blocking in PBS containing 10% goat serum, 2% albumen 2% fetal calf serum for 1 h. Slides were incubated with appropriate combinations of mAb OPCML, mAb HER2 and pAb EGFR primary antibodies for 1 h at room temperature, followed by incubation for 1 h with animal anti-mouse Alexa-555 (OPCML), animal anti-rabbit Alexa 488 (HER2) before mounting and imaging on a Zeiss LSM 510 confocal microscope.

siRNA Knockdown

Endogenous OPCML was knocked down in OSE-C2 cells by transient transfection of a specific pool of 3 siRNAs (Stealth knockdown-Invitrogen) using lipofectamine RNAiMAX reagent.

MTT Proliferation Assay

Cell proliferation assays were carried out in quadruplicate using the thiazolyl blue tetrazolium bromide (MTT) assay. Cells were plated out in 96-well plates at a density of 2,000 cells/well and cultured in low serum medium (0.25% FCS) or low serum medium supplemented with 50 ng/ml EGF. At appropriate time points, the medium was removed from cells and replaced with 100 μl PBS and 11 μl of 5 mg/ml MTT (w/v). Cells were incubated in this solution for 2 hours at 37° C. and the purple fomazan product was solubilised in 100 μl DMSO, resuspended and read on plate reader at 540 nm.

Statistical Analyses

Data are expressed as mean±SEM. Differences were analysed by Fishers exact or Student's t test. P<% 0.05 was considered significant. Progression-free survival curves were estimated using the Kaplan-Meier method and analysed by the log-rank test. Correlation between the mRNA expression indices of genes was analysed using Pearson's correlation analysis.

Statistical Analysis and Mining of Tothill Data

Gene expression data on the 251 epithelial ovarian cancers within 285 ovarian tumors (published by Tothill et al (2008) Clinical Cancer Research 14: 5198) were obtained from the Gene Expression Omnibus (GEO). OPCML, EGFR and ERBB2 gene expression Pearson correlation coefficients were computed for all probe-sets. For survival analyses included all patients followed up to 5-years, and excluded patients with borderline/low malignant potential histology in view of their distinct natural history compared to invasive tumors. The effect of gene expression (probe: OPCML 206215_at, ERBB2 210930_s_at) on survival was assessed as a continuous variable using Cox-regression, and after transformation to categorical variables by median dichotomization or quartiles using Kaplan-Meier curves and the log-rank test.

Results

OPCML is Rapidly Induced by EGF and FGF 1/2

Serum starved SK-OV-3 cells (low OPCML expression) {Sellar, 2003 #2} were stimulated with 50 ng/ml EGF or 10 ng/ml FGF. EGF induced OPCML rapidly, achieving maximal mRNA expression at 30 min, with return to basal levels of expression by 60 min (FIG. 1A(i)), with maximal OPCML protein at 60 min (FIG. 1A(ii)). Similarly, FGF1/2 also induced OPCML mRNA, by 15 minutes (FIG. 1B(i)) with protein peaking at 90 minutes (FIG. 1B(ii)). These data were replicated for several other cell lines (data not shown). Specifically, induction of OPCML expression in a panel of ovarian cancer cell lines upon EGF stimulation (50 ng/ml) demonstrated consistent induction of OPCML mRNA by 5 to 10-fold, with varied timescale of peak induction.

OPCML Interacts with HER2 and FGFR1 Via Different Binding Sites

To determine if OPCML interacted with RTKs, co-immunoprecipitation (co-IP) using an OPCML polyclonal antibody was performed in a SK-OV-3 cell lines stably transfected with OPCML (BKS2.1) and vector-only controls (SKOBS-V1.2). Other OPCML stable transfected clones have been reported previously and behave identically as BKS2.1 (Sellar et al, 2003). Immunoblotting with anti-HER2 and anti-EGFR demonstrated that both interacted with OPCML, however reciprocal Co-IP using anti-HER2 and anti-EGFR antibodies confirmed the Co-IP only for OPCML with HER2 and not with EGFR (FIG. 2A(i & ii)). We further used GST/OPCML domain fusion proteins in pull-down experiments with either SK-OV-3 cell lysates (expressing HER2 and EGFR) or with purified TnT HER2 ECD fragments (structures shown in FIG. 2B). HER2 interacted with a full length OPCML extracellular domain (ECD) fused to GST (GST-OPCML D1+2+3) but not the truncated OPCML ECD lacking Ig domain 3 (GST-OPCML D1+2) from SK-OV-3 lysates (FIG. 2C(i)), in addition to in vitro translated HER2 ECD (FIG. 2C(ii)), suggesting that the third (juxtamembrane) Ig domain (Ig-Ill) of OPCML is crucial for interaction with HER2. We then explored whether OPCML interacted with the fibroblast growth factor receptors 1 and 2 (FGFR1 & 2). Co-IP of SKOBS-V1.2 and BKS2.1 with OPCML antibody revealed that OPCML bound to FGFR1, confirmed by reciprocal co-IP (FIG. 2A(ii)). We used the GST/OPCML fusion proteins in pull-downs using cell lysates transiently transfected with full length FGFR1 (FIG. 2C(i)) and separately in in vitro studies with His-tagged FGFR1 (FIG. 2C(ii)). These experiments showed that both GST-OPCML D1+2+3 and GST-OPCML D1+2 interacted with FGFR1, therefore domain 3 was not essential for FGFR1 binding, implying that FGFR1 and HER2 bound to different sites on OPCML. Further experiments showed that GST-OPCML D2+3 interacted with FGFR1 but not GST-OPCML D3, showing that domain 2 is essential for FGFR1 binding (data not shown).

OPCML Downregulates HER2 and FGFR1, and Abrogates Phosphorylation of HER and EGFR, Together with Downstream Signaling of the MEK-ERK Cascade

We then explored the functional consequences of these OPCML-RTK interactions. OPCML expressing BKS2.1 demonstrated strong downregulation of HER2 but not EGFR protein as compared with SKOBS-V1.2 (FIG. 3A(i)), implying that OPCML specifically regulates HER2 protein. We extended our investigations to the FGF receptor family and demonstrated downregulation of FGFR1 but not FGFR2 in BKS2.1 (FIG. 3A(ii)). Immunofluorescence microscopy (IFM) confirmed that OPCML expression in BKS2.1 dramatically reduced the levels of HER2 and FGFR1 but not EGFR or FGFR2 (FIGS. 3B(i & ii) and 3E).

We explored the impact of OPCML on cellular RTK phospho-activation and signaling in ovarian cancer cells. Phosphorylation of 2 analogous autophosphorylation sites, HER2-Y1248 and EGFR-Y1173 was abrogated in BKS2.1 (FIG. 3C(i)) (an independent OPCML stable transfectant, previously described (Sellar et al (2003)) (data not shown), and OSE-C2 expressing physiological levels of OPCML (data not shown). Similarly, FGF mediated phosphorylation of FGFR1-Y766 (known to transactivate phospho lipase Cγ) was abolished in BKS2.1 cell lines (FIG. 3C(i)). In both EGFR and FGFR signaling systems, we noted inhibition of phospho-PLC. and phospho-ERK 1& 2 (T202 & Y204—FIG. 3C(ii)) but not phospho-Akt S473 or T308 (FIG. 3C(iii)) suggesting that OPCML principally affected the MEK-ERK cascade. These signaling findings were phenotypically confirmed in growth assays; BKS2.1 and SKOBS-3.5 lines and OSE-C2 were profoundly growth-inhibited compared with vector control SKOBS-V1.2 (p<0.0001, student's t-test) (FIG. 3C(iv)).

To explore the physiological role of OPCML, normal epithelial cell line OSE-C2 (OPCML expressing) was transfected with OPCML siRNA, which abolished OPCML protein. This resulted in a strong induction of HER2 and FGFR1 (but not EGFR or FGFR2) and phospho activation of HER2-Y1248 and EGFR-Y1173 levels (FIG. 3D(ii)). Since OPCML expression and physiological function seems to be regulated by growth factor signaling and it is downregulating at least two members of two different families of RTKs, we decided to extend our analysis to other RTKs. SiRNA against OPCML was used to verify whether RTKs appearing downregulated in the OPCML-expressing lines would show reciprocal upregulation if OPCML is knocked down in OSE-C2 cells. From this analysis, in addition to HER2, HER4 also appears downregulated in both SKOBS-3.5 and BKS2.1 cells, whereas FGFR1 and FGFR3 appear downregulated in predominantly the BKS2.1 line expressing five times more OPCML than SKOBS-3.5 (FIG. 3D(i)). The reciprocal analysis of looking at RTK expression after OPCML knockdown revealed HER2, HER4, FGFR1, all showing substantive upregulation in siRNA lane. FGFR3 exhibits a slight increase in expression level with knockdown (FIG. 3D(ii)). In contrast OPCML does not affect EGFR, HER3, FGFR2, FGFR4, EPHA50, VEGFR1 and VEGFR3.

OPCML Prevents HER2/EGFR Hetero Dimer Formation and Reduces EGF Receptor Availability.

SKOBS-V1.2 and BKS2.1 cell extracts were subjected to Co-IP and immunoblotted with antibodies as shown in FIG. 4A, demonstrating loss of hetero-dimerisation in the presence of OPCML. Further, OPCML reduced EGF receptor availability (FIG. 4B).

OPCML is Localized in the Detergent—Resistant (Raft) Membrane Fraction and Co-Localizes with EGFR and HER2 in Ovarian Cancer Cells.

To define the mechanism of OPCML-based RTK degradation, we used HER2 as a paradigm for further study. Initially, we investigated the influence of OPCML expression upon the mode of HER2 degradation linked to immunofluorescent confocal microscopy (IFM) analysis to examine the trafficking of OPCML and HER2 in cells. It has been previously reported that GPI-anchored proteins are sequestered in the detergent insoluble ‘lipid-raft’ membrane microdomain of cells (Sangiorgio et al (2004) Ital J Biochem 53(2): 98-111). To examine the localisation of OPCML (a GPI anchored protein) within lipid rafts, purified membrane of OPCML negative (SKOBS-V1.2) and positive (BKS-2.1) were subjected to solubilisation in 1% Triton X100 (for detailed method see Materials and Methods) and samples subjected to ultracentrifugation to separate detergent solubilised and insoluble proteins (FIG. 5A(i)). This experiment revealed that the majority of OPCML was localized within the detergent insoluble fraction, along with Caveolin-1 (a marker of caveolae—a distinct form of lipid raft domain). Interestingly, HER2, in the OPCML-expressing line, was reduced as previously shown in FIGS. 3A and 3B but also sequestered in the detergent insoluble fraction when compared to the OPCML negative line, where HER2 was equally distributed. The distribution of EGFR was only marginally effected by the expression of OPCML. IFM was employed to examine the trafficking of OPCML in cells; EEA-1 (a marker of the early endosome) and caveolin-1 (a marker of the raft-caveolar pathway) were used to distinguish between Clathrin-coated pit and caveolar endocytic vesicles. These studies revealed that the majority of the internalized protein co-localized with Caveolin-1 compared to EEA-1 (OPCML+cell line: cav-1 co-localisation=23%, EEA-1=7.5% of total HER2; OPCML-cell line: cav-1 co-localisation=4.5%, EEA-1=32% of total HER2). Furthermore, vesicular staining was seen to be markedly different within the cell, consistent with these representing distinct compartments (FIG. 5A(ii)). IFM also confirmed that OPCML co-localized with EGFR and HER2 in ovarian cancer cells (FIG. 5A(iii)).

We next transfected both OPCML-expressing and non-expressing cell lines with a HA-tagged ubiquitin construct to analyze the levels of receptor ubiquitination +/−OPCML. Twenty four hours post transfection, cells were serum starved and subjected to acute stimulation with EGF (50 ng/ml) for 60 minutes. Consistent with the significant reduction in receptor levels, OPCML expression was associated with enhanced ubiquitination of HER2, which was strongly increased upon EGF stimulation (FIG. 4B(i&ii)). IFM and quantification of co-localisation demonstrated that OPCML expression induced a shift in proportion of the HER2 into caveolin-1 positive vesicles compared to a predominant co-localisation with EEA-1 in the OPCML negative cell line (+OPCML: HER2/CAV-1, 22.863%±1.859; HER2/EEA-1 8.767±1.852.−OPCML: HER2/CAV-1, 4.767%±1.559; HER2/EEA-1 30.667±3.756) (FIG. 5C(i&ii)). Transmembrane proteins in the EEA-1 compartment can enter either the late-endosome-lysosome for degradation, or the Rab11-positive recycling endosome. Whilst Caveolin-1 positive vesicles have been reported to be non-recycling and result in proteasomal degradation of their cargo (Di Guglielmo et al (2003) Nat Cell Biol 5: 410-421). Consistent with degradation by the proteasome, chloroquine (CQ), a weak base that alkalinises the lysosome, was ineffective, but MG-132, a potent antagonist of the proteasomal 26S proteinase, inhibited HER2 degradation in the OPCML expressing cell line with no effect on EGFR expression found (FIG. 5D(i&ii)). Furthermore, disruption of cholesterol using methyl-R-cyclodextrin (MP-CD) also inhibited the degradation of HER2 and increased the phosphorylation at Y1248 (FIG. 5D(iii)) suggesting an important role for the lipid-raft in the OPCML-specific regulation and degradation of HER2. In conclusion, OPCML binds specifically to HER2, sequesters the receptor in lipid-rafts, enhancing caveolar-based endocytosis, ubiquitination and subsequent proteasomal degradation of the oncogenic receptor.

OPCML Regulates/Predicts Response to Lapatinib in Ovarian and Breast Cancer.

The finding that OPCML could regulate activity of HER2 and EGFR led us to explore whether OPCML might influence the efficacy of anti-EGFR/HER2 therapeutics. OPCML transfected and control cells were pre-incubated with lapatinib, trastuzumab, cituximab, erlotinib and gefitinib. We then used EGF induced phospho-ERK activation as an assay to define the effectiveness of therapeutic inhibition. The dual inhibitor of EGFR and HER2 tyrosine kinases, lapatinib, exhibited strong OPCML mediated sensitization, reducing the effective concentration of lapatinib required to abolish the phospho-ERK signal by 10-fold for BKS2.1 compared with SKOBS-V1.2 (FIGS. 6A(i) and 6C). We noted enhanced down-regulation of phospho-AKT in OPCML expressing cells. Lesser sensitization than for lapatinib was observed with trastuzumab (FIG. 6A(ii)). Notably, cetuximab, erlotinib and gefitinib, inhibitors of EGFR showed no sensitization in OPCML transfected cells (data not shown) consistent with the hypothesis that OPCML interacts with HER2 and not EGFR.

We then investigated whether siRNA knockdown of physiological OPCML expression in normal OSE-C2 cells could affect sensitivity to lapatinib. We observed that the lapatinib-mediated reduction in phospho-ERK signal strength was significantly reversed by OPCML siRNA knockdown in these normal ovarian surface epithelial cells (FIG. 6B). These data demonstrate that OPCML modulates sensitivity to lapatinib through regulation of the level of HER2, however the mechanism of this finding remains to be clarified.

We next tested whether OPCML could be used to predict response to lapatinib in ovarian and breast cancer. Histology was obtained by new biopsy of recurrent disease and TTP (time in months to progression from start of therapy until progression) assessed. Docetaxel and anthracyclines were administered for a maximum of 6 cycles and capecitabine was administered until disease progression or unacceptable toxicity. HER2 immunohistochemistry (IHC) was performed using the Dako Herceptest kit: 3+ in all cases. The results are shown in Table 1 below.

In view of the strong tumor suppressor role of OPCML and these findings, we explored whether its expression was related to ovarian cancer prognosis. We used a recently published expression microarray dataset of 251 ovarian cancers (Tothill et al, 2008) with full clinical annotation and follow-up of patients for progression free survival (PFS). The relationship between OPCML mRNA expression and PFS was examined for all 251 ovarian cancer patients with epithelial ovarian cancers in the dataset. Overall high OPCML expression demonstrated a significant association with better survival, as shown by the Kaplan-Meier curve in FIG. 7A(i) (Log-rank p=0.061 Breslow test p=0.034), although the difference was of modest magnitude. However, because the findings described herein suggested that OPCML's tumor suppressor role was related to repression of HER2 and FGFR1 function we specifically analysed the patient cohort according to their HER2 RNA expression and explored the impact of OPCML expression on this group of patients. We found that OPCML mRNA level was strongly prognostic only for patients expressing top quartile HER2 mRNA levels; patients with above median expression of OPCML had 27 months median PFS (FIG. 7A(ii)) and Table 1 & 2) compared with top quartile HER2 and below median OPCML expression with 13 month median PFS (Log-Rank p=0.004). In contrast, bottom quartile HER2 expressing patients showed similar survival regardless of OPCML expression. This validated our hypothesis that OPCML impacts on an intact HER2 pathway. No significant association was observed for EGFR or FGFR1 and OPCML with PFS in this dataset (data not shown).

A possible explanation for this clinical data is that strong OPCML expression (in the context of strong HER2 expression) regulates HER2 protein level/activity and abrogates HER2 pro-oncogenic signaling with consequent better patient prognosis, whereas tumors with weak OPCML expression and strong HER2 expression have unrestrained HER2 pro-oncogenic signaling and consequently poor prognosis.

EXAMPLE 2

To complement the findings described in Example 1, we expressed and purified recombinant human OPCML and assessed its affect on in vitro tyrosine kinase signaling and cell growth. The results are in agreement with those in Example 1.

FIG. 8 shows purification of recombinant human OPCML expressed in E. coli. Expressed OPCML was present in inclusion bodies and was successfully refolded by dialysis into PBS. Recombinant OPCML is a 272 amino acid polypeptide whereas physiologically synthesized OPCML is a 345 amino acid polypeptide. Post-translationally modified OPCML (including N-linked glycosylation) is 55 kDa, whereas without glycosylation, the signal peptide region and the GPI anchor region, OPCML is 31 kDa. The protein and polynucleotide sequences of recombinant human OPCML correspond to SEQ ID Nos: 5 and 6 respectively.

FIG. 9 shows cellular uptake of recombinant OPCML Cells which took up exogenous OPCML demonstrate downregulation of HER2, as confirmed by IFM.

FIG. 10 shows that administration of exogenous OPCML inhibits receptor tyrosine kinase signaling in vitro. Specifically, administration of exogenous OPCML downregulates HER2 and ERK protein levels.

FIGS. 11A and 11B show that administration of exogenous OPCML inhibits SKOV-3 cell growth in vitro, as assessed by a MTT cell growth assay.

FIG. 11C shows that OPCML inhibits growth of a range of ovarian and breast cancer cell lines, while sparing normal ovarian surface epithelial cells. Interestingly, growth of HER2+ and HER2− breast cancer cell line was profoundly inhibited, suggesting that the mechanism is not solely restricted to HER2+ cells but may be mediated through FGFR pathways or other EGFR components that interact with HER2 and are phospho-inactivated as part of OPCML therapy. Only one cell line showed resistance to OPCML (PEA2).

EXAMPLE 3

Having established that OPCML is a prognostic factor in strongly HER2 expressing ovarian cancer (see Example 1), we assessed whether its expression was related to the prognosis of other cancers. This was done by a Kaplan Meier analysis of overall survival according to OPCML dichotomized survival.

FIG. 12A (first graph—A1) shows that OPCML expression is a good prognostic factor in EGFR/HER2 positive node negative breast cancers in patients receiving no adjuvant systemic therapy. OPCML expressing tumors show better survival, with a 10 year relapse free survival of 72% vs 52%. (‘cum survival’=cumulative survival). Analysis was based on a published dataset of 149 node negative breast cancer patients (Wang et al 2005, Lancet 365(9460): 671). The second two graphs (A2 and A3) show that OPCML expression is a particularly good prognostic factor in ER-breast cancer.

FIG. 12B shows that OPCML is a prognostic factor in lung cancer. OPCML expressing tumors show better overall survival (OS). Analysis was based on a published dataset of 115 lung cancer patients (Takeuchi et al 2006, J Clin Oncol 24(11): 1679-1688).

FIGS. 12C and 12D show that OPCML is a prognostic factor in brain high grade gliomas. OPCML expressing tumors show better overall survival (OS). Analysis was based on a published dataset of high grade glioma patients (Phillips et al 2006 Cancer Cell 9(3): 157-73).