Interfering with RAS–effector protein interactions prevent RAS-dependent tumour initiation and causes stop–start control of cancer growth

Oncogene (2010) 29, 6064–6070 & 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 www.nature.com/onc SHORT COMMUNICATION Interfering with RAS–effector protein interactions prevent RAS-dependent tumour initiation and causes stop–start control of cancer growth T Tanaka and TH Rabbitts Leeds Institute of Molecular Medicine, Section of Experimental Therapeutics, St James0 s University Hospital, University of Leeds, Leeds, UK RAS mutations are the most common gain-of-function change in human cancer and promise to be a critical therapy target. As a new approach, we have used a surrogate to drug the ‘undruggable’ (that is, RAS-effector protein–protein interactions inside cancer cells) in preclinical mouse models of RAS-dependent cancers. Using this novel reagent, we have specifically targeted RAS signalling in a transgenic mouse model of lung cancer by directly blockading RAS-effector interactions with an antibody fragment that binds to activated RAS, and show that the interaction of RAS and effectors, such as phosphoinositide 3-kinase and RAF, is necessary for tumour initiation. Further, interference with oncogenic RAS–effector interactions result in control of tumour growth in human cancer cells but, crucially, does not necessarily cause tumour regression. These findings support the concept that ablating RAS-dependent signalling in cancer will have chemo-preventive effects that confer a chronic state in cancer and suggest that mutant RAS-targeted therapies may require conjoint targeting of other molecules and/or current cancer therapeutic strategies (for example, radiotherapy and chemotherapy) to be curative. In this context, our findings suggest that the oncogene addiction model is not universally correct in its central thesis that cancer cell death is inevitable after loss of oncogenic protein function. Oncogene (2010) 29, 6064–6070; doi:10.1038/onc.2010.346; published online 6 September 2010 Keywords: antibody fragment; cancer; lung tumourigenesis; RAS signalling; therapy Introduction A future strategy for cancer treatment is to target specific mutant proteins with drugs (genotype-specific therapy) that interfere with protein interactions by mutant proteins inside cells. A prerequisite of molecular target drug development is confirmation that a critical Correspondence: Professor TH Rabbitts, Section of Experimental Therapeutics, Leeds Institute of Molecular Medicine, St James’s University Hospital, Wellcome Trust Brenner Building, Beckett Street, Leeds LS9 7TF, UK. E-mail: Received 23 April 2010; revised and accepted 5 July 2010; published online 6 September 2010 mutant molecule is a drug target by demonstrating susceptibility in preclinical mouse models. Further, the oncogene addiction model (Weinstein, 2002; Weinstein and Joe, 2008) posits that tumours rely on the continued expression of mutant oncogenes for maintenance of neoplasia such that tumour regression might be expected to result from ablating mutant protein function. However, few studies have been able to assess specific mutations in the context of the complex genetics and cell biology of tumour cells because there are few effective reagents. Further, there have been limited therapeutic assessments of specific functions of mutant oncogenic proteins for the same reason, other than limited examples (for example, targeting for HER-2, EGFR and BCR–ABL). Gene targeting (Shirasawa et al., 1993) or siRNA-mediated knock-down (Brummelkamp et al., 2002) have been used in some cases to assess efficacy of, for instance, the RAS target, but these cannot address specific functionality, such as protein–protein interactions. RAS proteins are involved in cellular signal transduction by a diverse set of protein–protein interactions, including effector molecules, such as RAF, phosphoinositide 3-kinase (PI3K) and Ral guanosine diphosphate dissociation stimulator (Downward, 2003; Herrmann, 2003). RAS gene mutations (KRAS, NRAS and HRAS) are common in a wide variety of human cancers (Forbes et al., 2006) resulting in constitutive activation of RAS and RAS–effector interactions and consequent aberrant signalling to the nucleus (Karnoub and Weinberg, 2008). In addition, highly frequent mutations in epidermal growth factor receptor (EGFR) tyrosine kinase, PI3K p110a catalytic subunit (PI3KCA) and BRAF are found in various human cancers (Weir et al., 2004). In mouse models, numerous studies have demonstrated that Ras mutants and Ras effector mutants contribute to tumour formation and maintenance (Johnson et al., 2001; Dankort et al., 2007) and inducible transgenic models show that tumours in lung and melanoma harbouring mutant RAS can regress when mutant RAS expression is suppressed (Chin et al., 1999; Fisher et al., 2001). Moreover, disruption of RAS– PI3K interaction by mutation on RAS binding domain of PI3K p110a can inhibit mutant Ras-mediated lung tumourigenesisis in mice (Gupta et al., 2007). These data are encouraging the belief that inhibitors of RAS signalling pathways will be valuable in cancer therapy. As a unique approach to evaluate that mutant RAS– effector interaction on RAS signalling would be an Blockading RAS signalling controls cancer growth T Tanaka and TH Rabbitts 6065 effective target for cancer therapy, we have recently isolated a single immunoglobulin heavy chain variable domain fragment that specifically binds to the activated form of RAS (GTP-bound RAS) (Tanaka et al., 2007) and interferes with oncogenic RAS function inside the cell by competitively preventing RAS-effector interactions. This novel reagent, therefore, acts as drug surrogate, targeting the activated mutant RAS-specific protein interactions. The single domain (designated intracellular Domain antibody iDab#6-memb) binds with high affinity and specificity to the RAS switch I region, which is where the RAS effectors bind (Vetter and Wittinghofer, 2001). In the current study, we pEF-iDab#6-memb - pcDNA3-HRASV12 employed this specific iDab inhibitor of RAS-effector interactions to examine the potential role of mutant RAS-dependent signalling pathways in the initiation events of cancer and in the sustained growth of tumours that have mutations of RAS together with other oncogenes. We show that inhibiting mutant RASeffector interactions is sufficient to prevent tumour initiation and control cancer cell growth, but it is not remedial in the models used. The anti-RAS iDab#6-memb binds specifically to the switch I region of the GTP-bound RAS (Tanaka et al., 2007) at locations where RAS effectors, such as PI3K and RAF, bind (Nassar et al., 1995; Pacold + + - + + FLAG panRAS ERK p-ERK AKT p-AKT HCT116 500 400 300 200 100 0 1 2 4 5 Days 6 300 200 100 1 2 3 4 5 Days 300 200 100 1 2 3 400 HT1080 400 0 400 0 7 Cell number (x 104) Cell number (x 104) 500 3 SW480 500 Cell number (x 104) Cell number (x 104) 600 6 7 4 5 Days 6 7 6 7 DLD1 300 200 100 0 1 2 3 4 5 Days Figure 1 Effects of blocking RAS–effector (...truncated)