Skip to main content

Identification, lead generation, structural biology and validation of targets for cancer therapy. an integrated methodological approach

Deliverables

These vectors, which contain donor and recipient markers for Flourescent Resonance and have been combined to the tow interacting proteins p53 and mdm2 will be useful in the design of a screen for small molecule or peptide based agents that interfere with this protein-protein interaction in cellular models. mdm2 binding to p53 is the pre-cursor signalling p53 for degradation via ubuiquitination. p53 is the most commonly de-regulated tumour suppressor in cancer. Interfering with this degradation signalling so often relied on by cancer cells has been an area of commercial interest for some time and several companies have been evaluating agents for their potential use clinically. Since FRET itself as a platform as yet is not suitable for high through put screening, these materials nonetheless would be useful in employing secondary cellular screens to help pharmaceutical and biotechnology companies optimise their leads. Companies already in this field include, Roche (who designed the published Nutlin compounds), Cyclacel, AstraZeneca, De Novo, Ascenta Therapeutics, Novelix Pharmaceuticals and the USA institutions, MD Anderson
We have shown that Ral is involved in the asymmetric localisation of neosynthesized proteins in polarised epithelial cells. This function is mediated by a direct link of Ral to Sec5, a protein that participates to the dynamic formation of the exocyst - or more accurately of variable exocyst complexes. Secondly, we have generated genetic data that place Ral in a circuit downstream of the Rap GTPase - a scheme that is conflicting with observations in ex vivo grown immortalised cell lines. A protein, sec5 that is a part of the exocyst was identified. This partnership was validated in vitro and in vivo (by co-immunoprecipitation): sec5 is a bona fide and most probably direct partner of activated Ral. What is the possible function of this interaction? We have explored two aspects of exocyst function. We have shown that Ral via interaction with the exocyst triggers secretion in chromaffin cells. In polarised cells (differentiated MDCK cells) where the exocyst is nearby tight junctions, we are finding that the interaction between Ral and the exocyst involves Ral in the establishment and/or maintenance of the apical /baso-lateral polarity barrier.
Assay is based on the immobilised fusion protein GST-Ras G12V and the Alexa dye-labelled Ras binding domain of Raf kinase. Besides the detection of altered interaction by compounds, this assay also proved to be suitable for the potential identification of inducers of Ras GTPase activity. The system is stable for 24 hours at 30°C, can be scaled down to volumes of 1 to 5 µl quantities and in collaboration with EVOTEC OAI AG a 60,000 compound library has been screened. The collaboration with EVOTEC OAI AG was successfully continued and using the assay described above more than 70.000 compounds were screened for their ability to either inhibit the Ras/Raf interaction or to induce GTPase activity of G12VRas. This work resulted in a patent (see also below), which claims to be capable of detecting reversible protein-protein interactions between GTPase negative forms of GTP-binding proteins and their effector's proteins. In addition the system is capable of detecting chemical compounds that inhibit the above interactions in the living cell. In addition, the rescue of blocked GTPase activity can be detected. The system is potentially applicable to any pair of high-affinity or medium-affinity binding partner proteins.
A Cdk knockout strain and a strain carrying the Cdk4R24C mutation found in human familial melanoma was generated as pre-existing know-how and used in the study. A mutant Cdk4 (R24C) will be analysed for susceptibility of tumour formation. These animals will be treated with different carcinogens and will be crossed with mice carrying mutations in the Ras or p53 pathways. Spontaneous or induced tumours will be analysed at the molecular level and the role of Cdk4 activity in tumour formation will be clarified using specific Cdk4 mutants and inhibitors. No spontaneous tumour formation was detected in mice deficient in Cdk4 (Cdk4 neo/neo). In order to determine tumour susceptibility in Cdk4 R24C mice, we aged a group of 180 knockin mice carrying the Cdk4 R24C allele either in homozygosity or heterozygosity. Mice were sacrificed at any sign of disease and tissue samples were recovered for histology and nucleic acid/protein analysis. Mice homozygous for the R24C mutation (Cdk4R24C/R24C) are born at the expected mendilan ratio, are fertile and develop normally. However a proportion of mutant mice die before the first 8 months of life and 80% of the mice are dead by month 15. Necropsy analysis shows a significant number and a wide spectrum of tumours in these animals, including malignancies of endocrine origin (55% incidency); epithelial non-endocrine (24% incidency); mesenchymatous (67%); and hematopoietic malignancies (3% incidency). Detailed pathology was performed on histological sections. Hemangiosarcomas develop in 56% of the animals and are the major cause of death in Cdk4R24C/R24C mice. They are frequently observed at multiple sites (spleen, liver, subcutaneous) or accompanied with metastasis (54% of the cases). This phenotype agrees with a significant number of reports showing Cdk4 amplification or over-expression in human sarcomas. Leydig cell tumors, on the other hand, are the most frequent type of malignancy in the Cdk4R24C/R24C males (62% incidency). Tumours of the pancreatic endocrine cells are also frequent (31%) and correspond mainly to b cell tumours, as detected by immunostaining with insulin-, PP-, or somatostatin antibodies. Tumours of the pituitary were observed in 22% of the Cdk4R24C/R24C mice. These tumours have also been observed in mice heterozygous for the retinoblastoma knockout allele, and in mice deficient in p18INK4c or p27Kip1. Interestingly, pituitary tumours in these models develop in the pars intermedia, whereas in the Cdk4 R24C mice are mainly of adenohypophysis (pars distalis) origin (80% of pituitary tumours). Pituitary tumours were also immunostained with antibodies against the pituitary hormones (ACTH, Prolactine, LH, FSH and GH). Most of pituitary tumours in Cdk4 R24C mice correspond to adenomas with a variety of patterns of positivity for hormones, although some carcinomas usually positive for prolactine were also found (about 15% of pituitary tumours). Cdk4 R24C mice develop a wide spectrum of other tumours affecting a variety of tissues such as liver, thyroid, lung, liver, gut, salivary gland, Harderian gland, kidney, and hematopoietic malignancies. A group of 50 mice were sacrificed at 14-16 months without any external sign of disease. Detailed necropsy and pathologic analysis of these animals indicated the existence of several tumours with a distribution similar to that described above. Sarcomas and pituitary tumours are significantly reduced in this group in agreement with the fact that these malignancies are the major cause of death in Cdk4 R24C mice. Since the Cdk4 R24C mutation occurs in heterozygosity in human melanoma, we also aged 25 Cdk4R24C/+ mice. These animals also succumbed to the same type of malignancies with a similar distribution to that described above, although with a slightly increased latency.
In their active, GTP bound forms, Rac and Cdc42 can interact with a fusion protein in which residues 75-18 of PAK are inserted in between the donor GFP (EBFP) and acceptor GFP (EGFP) and thereby change the FRET between the GFP fluorophores. The work on this system using purified proteins, which was described in the first year report, was written up and published. A key aim of this work is to be able to use the FRET system within living cells. We are attempting to demonstrate its utility in both intact bacterial and in mammalian cells. The former has the advantage of being easier to manipulate and allowing very high expression levels. It was felt that this might aid development of the appropriate technology for measurement of FRET and to understand the limiting sensitivity. Although, a bacterial system could be used as a drug screen, because the cell permeability differs from that of mammalian cells a mammalian cell system would be more acceptable. Accordingly, we have begun the cloning process to set up an intracellular bacterial assay to monitor the interaction between EGFP-PAK-EBFP and Rac via FRET. In order to allow independent regulation of the co-expression of Rac and EGFP-PAK-EBFP suitable compatible vectors must be selected. Two such vectors are the pPROTet and pPROLar vectors supplied by Clontech that are inducible using tetracycline and IPTG/arabinose, respectively. However, in our hands these low copy number vectors are difficult to handle and, despite numerous attempts, we have had problems cloning Rac and PAK into them and then obtaining significant expression. In contrast, using standard expression vectors, e.g. pGEX and pET, the expression of both proteins is very high. Hence, we have decided to rethink the strategy and to select a further two compatible vectors, probably based on pGEX and pBAD expression systems, which have a good track record for expression. In parallel, we have been setting up the same FRET system in mammalian cells. EGFP-PAK-EBFP was cloned into the mammalian expression vector, pcDNA-4 (Invitrogen) to generate a His-tagged fusion. HEK293T cells were transiently transfected with this vector. 24-86h later cells were harvested and the lysate analysed by Western Blot analysis using an antibody against the His-Tag and by recording fluorescence emission spectra. The blots showed a single band of the expected mobility for the construct and no indication of any degradation. This is of significance as there was concern that the relatively unstructured linker region containing PAK might be particularly susceptible to proteolysis. The transfected cells were significantly fluorescent as compared to mock-transfected cells. A key issue with the transiently transfected cells is the variable level of expression from one cell to another. A lot of time has been spent in optimising the transfection procedure, but it is felt that this is intrinsically too variable to allow effective cotransfection with Rac. Hence we have initiated work on expression using the BacMam system, based on viral transduction. Work on a more artificial system based on the Q61L constitutively activated Rac was completed. Furthermore we have decided to use a construct in which the CAAX motif has been deleted; this will both serve to prevent translocation to the plasma membrane and to prevent prenylation and hence interaction with RhoGDI. This would likely prevent interaction with EGFP-PAK-EBFP. The Rac has been cloned into pcDNA-4 and expression trials will begin shortly. The same construct will also be put into the BacMam system for co-expression with EGFP-PAK-EBFP. The final experiments involved the titration of Q61LRac.GTP into each of the four proteins and measuring the two "FRET" ratios. The data obtained with cPAK were very similar to that we obtained previously (Graham et. al., Anal. Biochem. (2001) 296, 208-217 giving the anticipated Kd. The R(509/444) decreased from 3.9 to a minimum value of 2.3 at a concentration of Rac of about 4mM (EC50 1mM), whereas thrombin cleavage caused a larger decrease to a ratio of 1.0. Thus, Rac binding does not result in a complete separation of the GFP fluorophores. Addition of up to 20mM Rac.GTP did not cause any effect on the fluorescence properties of either EGFP-PAK-Rac-TCS-EBFP or EGFP-PAK-EBFP-Rac. If the interaction between Rac and PAK on separate molecules (i.e. intermolecular) was the only protein: protein interaction present, simulation to a Competition model showed that 10mM Rac should be sufficient to block this homodimerization.
CD43 or leukosialin is a trans-membrane sialoglycoprotein, whose extra-cellular domain participates in cell adhesiveness and the cytoplasmic tail regulates a variety of intracellular signal transduction pathways involved in cell proliferation. CD43 is abundantly expressed on the surface of hematopoietic cells, but CD43 expression is also frequently found in the tumour cells of nonhematopoietic origin. In the early stages of some tumours, the accumulation of tumour suppressor protein p53 has been described. Partner 15 has now shown that the expression of CD43 causes the induction of functionally active p53 protein. Moreover, they found that the activation of p53 by CD43 is mediated by tumour suppressor protein ARF.
The p53 protein induced by CD43 is transcriptionally active stimulating transcription from p53-responsive promoters. CD43 does not induce p53 in cells lacking ARF, but ARF expression in these cells restores the ability of CD43 to induce p53. CD43 is also able to induce endogeonous ARF in p53 negative cells. Thus, the p53 response could be activated due to the possible mitogenic stimuli caused by CD43 expression. CD43 over-expression in human cancer cells hindered the detection of the Fas death receptor on the cell surface and thereby helped to evade Fas-mediated apoptosis. When both p53 and ARF proteins were present, CD43 over-expression activated p53 and suppressed colony formation due to induction of apoptosis. These observations suggest potential contribution of CD43 to tumour development by promoting cell growth and evading apoptosis.