Colorectal cancer and RASSF family—A special emphasis on RASSF1A

Authors

  • Maria Sofia Fernandes,

    1. Institute of Molecular Pathology and Immunology of the University of Porto (IPATIMUP), Porto, Portugal
    Search for more papers by this author
  • Fátima Carneiro,

    1. Institute of Molecular Pathology and Immunology of the University of Porto (IPATIMUP), Porto, Portugal
    2. Faculty of Medicine, University of Porto, Porto, Portugal
    Search for more papers by this author
  • Carla Oliveira,

    1. Institute of Molecular Pathology and Immunology of the University of Porto (IPATIMUP), Porto, Portugal
    2. Faculty of Medicine, University of Porto, Porto, Portugal
    Search for more papers by this author
  • Raquel Seruca

    Corresponding author
    1. Institute of Molecular Pathology and Immunology of the University of Porto (IPATIMUP), Porto, Portugal
    2. Faculty of Medicine, University of Porto, Porto, Portugal
    • IPATIMUP—Cancer Genetics, Rua Dr. Roberto Frias, s/n, 4200-465 Porto, Portugal
    Search for more papers by this author
    • Tel.: +351-22-5570700, Fax: +351-22-5570799


Abstract

The RAS-association domain family, commonly referred to as RASSF, is a family of 10 members (RASSF1-10) implicated in a variety of key biological processes, including cell cycle regulation, apoptosis and microtubule stability. Furthermore, RASSFs have been implicated in tumorigenesis and several family members are now thought to be tumor suppressors. As opposed to the KRAS oncogene, for which mutational activation is frequent in colorectal cancer (CRC), RASSFs are found to be silenced mainly by aberrant promoter methylation. In particular, RASSF1A, RASSF2 and RASSF5 methylation has been associated with CRC development, though the mechanisms of action remain poorly understood. This review focus on the current knowledge of RASSF inactivation in CRC, particularly RASSF1A, and on the implications RASSFs may have as potential biomarkers and for the development of new targeted therapies for CRC.

Colorectal cancer (CRC) is one of the leading causes of cancer mortality worldwide.1 Developing as a multistep process, CRC involves both genetic and epigenetic alterations in a number of oncogenes and tumor suppressor genes.2 Of particular relevance is the high frequency of activating mutations in KRAS and in two of the best characterized KRAS effectors, BRAF and phosphatidylinositol 3-kinase (PI3K). Detected in ∼30–40% and 15% of CRC, respectively, mutations in KRAS and BRAF promote aberrant cell proliferation through the mitogen-activated protein kinase (MAPK) cascade.3–7 However, other KRAS downstream effectors are involved, some of which promoting growth inhibitory effects, including differentiation, apoptosis, cell cycle arrest and senescence but the mechanisms underlying such growth inhibitory actions remain poorly understood.8–10 RAS-association domain family (RASSF) members belong to a recently identified family of putative tumor suppressor RAS effectors for which epigenetic silencing by promoter methylation has been shown to occur throughout the progression of cancer including CRC.11, 12 In this article, we summarize the role of RASSFs in CRC, its mechanisms of inactivation, associated cellular effects and its implications in terms of prognosis and therapeutic stratification of colorectal patients.

Colorectal Cancer and RAS Signaling

Colorectal cancer can be subdivided in different molecular subsets. Indeed, CRC can either hold microsatellite instability (MSI), characterized by a defective mismatch repair system representing 15% of CRC, or as it happens for the majority of the cases (85%) present microsatellite stability (MSS) which is characterized by having chromosomal instability (CIN).13 Although in most cases (about 70%) CRC occurs sporadically (MSI and MSS) a minority of cases develop in an hereditary context, being the most common form the hereditary nonpolyposis CRC (HNPCC).14

Activating mutations in genes of the MAPK and PI3K pathways, including KRAS, BRAF and PIK3CA, are frequent events in CRC.3, 15 Interestingly, KRAS and BRAF mutation frequencies and patterns differ among the different subsets of CRC, namely in MSS, MSI, sporadic and hereditary forms. Whereas KRAS mutations are found in both MSS and MSI sporadic and hereditary tumors, BRAF mutations show a more restricted pattern being mostly found in MSI sporadic CRC.4, 7, 15–17 A considerable proportion of patients with CRC (∼30–40%) exhibit KRAS mutations mostly affecting Codons 12 and 13 of Exon 2 and these have been shown to occur early in carcinogenesis.3–5, 7, 15 Mutations in BRAF occur in ∼15% (5–45%) of CRC, with the mostly reported mutation a valine-to-glutamic acid amino-acid (V600E) substitution on Exon 15.3–5, 15

Of significance, KRAS and BRAF mutations are preferentially found as alternative molecular alterations and thus mutations in these genes are not frequently found in the same tumor. This suggests that KRAS and BRAF oncogenes may play different roles in the development and progression of CRC.15, 18 In contrast, activating mutations involving hotspots on Exons 9 and 20 of PIK3CA are found in about 15% of sporadic and hereditary CRC, and occur concomitantly with KRAS or BRAF mutations.3, 6, 19

Deregulation of RAS signaling is critical as members of the RAS superfamily of GTPases play a key role in controlling a complex network of signaling pathways in response to extracellular stimuli. As transducers of signals from cell surface receptors to intracellular specific effectors, RAS proteins regulate a wide range of biological functions such as cell proliferation, differentiation, survival and death.9 A particular feature of RAS GTPases is their ability to switch between the inactive GDP-bound and the active GTP-bound conformational states with concomitant change in the affinity for downstream effectors. The transition between the on and off states is regulated by other proteins. Upon stimulation, guanine nucleotide exchange factors (GEFs) promote the activation of RAS by stimulating GDP for GTP exchange. Conversely, GTPase-activating proteins (GAPs) accelerate RAS-mediated GTP hydrolysis.20 In their active state, RAS proteins interact and activate their effectors and stimulate downstream signaling pathways. The classical RAS signal transduction pathway, critical in controlling cellular proliferation, differentiation and survival, is the RAS/RAF/MAPK kinase (MEK)/extracellular signal-regulated kinase (ERK). The serine/threonine kinase RAF (in humans A-RAF, B-RAF and C-RAF) is one of the best characterized RAS effectors. RAF phosphorylates MEK1/2, which in turn phosphorylates and activates ERK1/2.21 Another well-known cascade of RAS activated signaling is the PI3K pathway. The activation of the lipid RAS effector PI3K activates the serine/threonine protein kinase AKT and other proteins which then regulate cell survival and death.22

Abbreviations

APC: anaphase-promoting complex; ATM: ataxia telagectasia mutated; C1: protein kinase C conserved region; CIN: chromosomal instability; CRC: colorectal cancer; DAG: diacylglycerol/phorbol ester binding domain; ERK: extracellular signal-regulated kinase; GAPs: GTPase-activating proteins; GEFs: guanine nucleotide exchange factors; HNPCC: hereditary nonpolyposis colorectal cancer; LOH: loss of heterozygosity; MAPK: mitogen-activated protein kinase; MEK: mitogen-activated protein kinase kinase; MOAP1: modulator of apoptosis 1; MSI: microsatellite instability; MSS: microsatellite stability; NORE1: novel RAS effector 1; PAR-4: prostate apoptosis response protein 4; PI3K; phosphatidylinositol 3-kinase; RA: RAS-association domain; RASSF: RAS-association domain family; SARAH: SAV/RASSF/Hpo domain; SCRC: sporadic CRC; TNM: tumor-node-metastasis.

The RAS-Association Domain Family (RASSF) of Proteins

The RASSF family of proteins comprises 10 members encoded by different genes, RASSF1 to RASSF10, plus additional isoforms as a result of alternative promoter usage or splice site variations.11 The first family member, NORE1 (novel RAS effector 1, also termed RASSF5), was identified in mouse as a RAS-GTP-binding protein.23 Shortly thereafter, the human RASSF1 gene was cloned and mapped to 3p21.3, a locus that frequently suffers loss of heterozygosity (LOH) in lung tumors, and reported to be a potential tumor suppressor gene highly inactivated by methylation.24, 25 The other homologous were subsequently identified.

The predominant feature of this family is a RAS-association domain (RA) that potentially associates with the RAS family of GTPases (Fig. 1). However, evidence of a true functional RAS binding domain is controversial and the ability of most RASSF family members to associate with RAS has yet to be clearly established. According to the localization of the RA domain, C- or N-terminal, RASSF family members are classified into classical (RASSF1-6) and N-terminal (RASSF7-10), respectively.11 A SARAH (SAV/RASSF/Hpo) domain, involved in protein-protein interactions and in activation of the Hippo pathway, is also present at the C-terminus of RASSF1-6.11, 26 RASSF1 and RASSF5 contain an additional conserved DAG (diacylglycerol/phorbol ester) binding domain near its N-terminal, known as C1 (protein kinase C conserved region) and some isoforms of RASSF1 possess an ATM (ataxia telagectasia mutant) consensus sequence.11, 27, 28

Figure 1.

Simplified modular structure of RASSF1A. The functional domains of RASSF1A are indicated. C1/DAG, protein kinase C conserved region/diacylglycerol/phorbol ester binding domain; ATM, ataxia telangiectasia mutated phosphorylation site consensus sequence; RA, RAS-association domain; SARAH, Sav/RASSF/Hpo interaction domain. Conserved protein domains and amino acid positions were predicted using Prosite (http://prosite.expasy.org/). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

To date, the best characterized RASSF family member is RASSF1. RASSF1 gene codes for eight exons and generates seven tissue-specific transcripts (RASSF1A-G) by differential promoter usage and alternative splicing. The two major forms, RASSF1A and RASSF1C, are transcribed from two distinct CpG island promoters and were shown to be ubiquitously expressed in normal tissues.29, 30 Interestingly, RASSF1A loss of expression was shown to occur in a number of human cancers and such loss was correlated with methylation of the CpG-island promoter sequence of RASSF1A.12, 28 In addition to RASSF1A, other RASSF family members are frequently inactivated in human cancers such as RASSF5, RASSF2, RASSF4 and RASSF6.11 However, despite the increasing number of reports on this novel family of proteins their biological effects remain unclear.

RASSF1A Silencing and Colorectal Cancer

Since its identification in 2000, the tumor suppressor RASSF1A has been extensively investigated. Transcriptional silencing of RASSF1A by inappropriate promoter methylation has been frequently found in a number of human cancers including lung, breast, colorectal, gastric, cervical and many others.28 Therefore, in addition to the well-known mutations in KRAS, BRAF and PIK3CA that constitutively activate the MAPK and PI3K signaling pathways in colorectal carcinogenesis, RASSF1A methylation represents an alternative mechanism of aberrant RAS signaling.

Though somatic mutations of RASSF1A have been identified, they are uncommon.26, 30 Only one frame-shift mutation at codon 277 and one missence mutation at codon 201 (within the RA domain) have been detected in a vast array of carcinomas and cell lines. A number of polymorphisms have also been identified, many of which locating in the functional domains of RASSF1A.24, 31–34 The functional significance of these alterations remain to be established, but many of them were proven to impair RASSF1A role.26

Methylation-associated inactivation of RASSF1A has been described in a vast number of colorectal samples. Cancerous and precancerous lesions and synchronous normal colorectal mucosa have been evaluated for RASSF1A promoter methylation status. Since RASSF1A has been linked to RAS signaling pathway, association studies have been performed to address the relationship between RASSF1A methylation and KRAS mutation status as well as the relation with clinico-pathological parameters.

As described to date, RASSF1A methylation was reported to range from 1235 to 81%36 in CRC. In colon primary tumors, Yoon et al. detected methylation at the CpG island of the RASSF1A gene in 3 out of 26 (12%) tumor tissues while none of the available normal tissues were methylated.35 A number of other studies found a higher percentage of RASSF1A methylation in CRC samples. Wagner et al. observed RASSF1A promoter methylation in 45% (13/29) of the primary colorectal cancers as well as in 80% (4/5) of CRC cell lines analyzed.37 In addition, RASSF1A methylation was also detected in 36% (4/11) matching normal colorectal mucosa samples available for tumors with RASSF1A methylation.37 In a study with 222 sporadic CRC samples, van Engeland et al. found RASSF1A methylation in 20% (45/222) of the samples and a mutually exclusive relationship with the presence of KRAS mutations was suggested.38 Indeed, only 5% (11/222) of the samples presented concomitant KRAS mutation and RASSF1A methylation and inactivation of RASSF1A occurred predominantly in CRC without KRAS mutation (15%, 34/222), pointing to an alternative pathway for affecting RAS signaling.38 Similarly, in a distinct set of CRC samples for which RASSF1A methylation was observed in 20% (40/202) of the samples, only 5% (10/202) showed both KRAS mutation and RASSF1A methylation.39 A number of other studies reported RASSF1A methylation in 17% (8/47),40 36% (26/73)41 and 47% (17/36)42 of CRC samples. In a series of flat CRC samples, RASSF1A methylation was detected in a remarkable percentage of the cases (81%, 39/48); RASSF1A methylation was also detected in 49% (19/39) of matching normal colonic mucosal tissues.36 Again, a mutually exclusive relationship between KRAS mutations and RASSF1A methylation was found, with the concordant occurrence of KRAS mutation and RASSF1A methylation detected only in 6% (3/48) of the cases.36 An additional study compared the methylation status of RASSF1A in polypoid-type and flat-type early stage CRC samples.43 Interestingly, and in contrast to the high percentage of RASSF1A methylation found in flat CRC by Sakamoto et al., Noda et al. found RASSF1A methylation in only 20% (6/30) of the flat-type samples.43 Furthermore, RASSF1A methylation was similar (14%, 6/43) in polypoid-type samples, with an overall methylation of 16% (12/73) indicating that RASSF1A methylation is common to both subsets of CRC.43

To further elucidate the potential role of RASSF1A methylation in MSI CRC, our group has studied a total of 51 MSI CRC (31 sporadic and 20 HNPCC). In addition to RASSF1A methylation, samples were analyzed for KRAS and BRAF mutations.12 Overall, MSI CRC both sporadic and hereditary, showed RASSF1A promoter methylation in 43% (22/51) of the samples.12 Interestingly, no significant differences in RASSF1A methylation were observed between sporadic MSI CRC (52%, 16/31) and HNPCC carcinomas (30%, 6/20) (Fig. 2a). However, and despite the similar frequencies of RASSF1A methylation, we demonstrated that MSI sporadic CRC accumulated significantly more epigenetic/genetic alterations in RASSF1A and KRAS/BRAF than MSI HNPCC. We demonstrated that sporadic CRC with RASSF1A methylation accumulated more KRAS/BRAF mutations than MSI HNPCC carcinomas.12 Indeed, in MSI sporadic CRC, 36% (11/31) of the samples analyzed had RASSF1A methylation in accumulation with mutations in KRAS (3%, 1/31), BRAF (26%, 8/31) or both KRAS and BRAF (6%, 2/31) (Fig. 2b). In HNPCC, the majority of samples showed a single genetic or epigenetic alteration with accumulation of RASSF1A methylation and mutations in KRAS observed only in 10% (2/20) (Fig. 2b), which could be due, at least in part, to the lack of BRAF mutations.12 Overall, our study shows that a considerable fraction of MSI sporadic CRCs harbors concomitant RASSF1A methylation and BRAF and/or KRAS mutations, suggesting a synergistic effect between the silencing of the tumor suppressor gene RASSF1A and activation of the oncogenes KRAS or BRAF12 (Fig. 2b).

Figure 2.

Frequency of RASSF1A promoter methylation (a) and its association with KRAS and BRAF mutations (b) in MSI sporadic CRC (SCRC) and HNPCC (Mut, mutation). Adapted from Ref. 12.

Another study found RASSF1A methylation in 67% (2/3) of aberrant crypt foci with KRAS mutations and only 20% (4/20) without KRAS mutations, although the sample size was limited.44 In a set of primary CRC and synchronous liver metastasis, RASSF1A methylation was observed in 35% and 85% of the samples, respectively.45 Though RASSF1A methylation frequencies were unrelated to KRAS/BRAF mutations, RASSF1A content methylation was higher in liver metastasis than primary CRC only in KRAS wild type suggesting a specific role of this gene in metastatic site.45

The stratification of CRC patients according to the MSI status was also performed by Ahlquist et al.46 In a series of 65 sporadic MSI and MSS CRC samples, RASSF1A methylation was found in 31% (18/59) of the tumors. However, and in contrast to our results,12 RASSF1A methylation was found in only 24% (7/29) of the MSI subset and the occurrence of RASSF1A methylation in the presence of KRAS and/or BRAF mutations did not show any trend.46

More recently, Derks et al. also investigated the relationship between chromosomal instability, microsatellite instability and promoter methylation of CRC-associated genes including RASSF1A. RASSF1A methylation was observed in 27% (19/71) of CRC samples and although not statistically significant, RASSF1A methylation occurred more frequently in MSI CRCs (46%) than in MSI negative (26%).47 These results corroborate our data for which RASSF1A methylation was found in 52% of sporadic MSI CRC.12 Interestingly, promoter methylation of RASSF1A was significantly (positively) related to chromosomal gain at 8q23-qter.47

RASSF1A hypermethylation was also found in 30% (12/40) of serrated adenomas,48 in 6% (1/17) of adenomatous polyps and 26% (5/19) of ulcerative colitis.42 In a series of 120 colorectal adenomas, Harada et al. observed RASSF1 methylation in only 3% (3/120) of the cases and two out of three adenomas with RASSF1A methylation also showed methylation of RASSF249 (see Supporting Information Table 1 for more information). Notably, the different studies reporting RASSF1A methylation at different stages, as well as in corresponding normal mucosa adjacent to the tumor, point to a role of the tumor suppressor gene RASSF1A early in the development of colorectal carcinogenesis. It may mean that methylation possibly occurs in proximal sites of cancer cells due to field effect phenomena but the mechanism is not yet determined. As for CRC, RASSF1A hypermethylation has been found in normal tissue adjacent to the tumor in other cancer types, namely breast, suggesting that aberrant methylation may also play a role in the development of the tumor.50, 51

RASSF1A and Associated Clinico-Pathologic Features in Colorectal Cancer

To further characterize the subgroup of CRC patients with inactivation of RASSF1A, the relationship between RASSF1A methylation and clinico-pathological parameters was also explored in a number of studies. Wagner et al. investigated whether RASSF1A methylation correlated with tumor-node-metastasis (TNM) status, a widely accepted cancer staging system, but did not detect any significant associations.37 As observed for KRAS mutations, RASSF1A methylation was not associated with gender, Dukes stage and location of the tumor in van Engeland et al. study, except for the age at diagnosis that was slightly higher in RASSF1A methylated CRC.38 In our series of MSI sporadic CRC samples, no association was found between RASSF1A methylation and age, gender, tumor location, wall invasion, presence of lymph node metastases and staging of the tumor, except for poor degree of differentiation of CRC.12 In a recent study with 36 CRC samples, a significant correlation was observed between RASSF1A methylation (47%) and gender, with RASSF1A being more frequently methylated in females.42 Interestingly, in other studies RASSF1A methylation levels were significantly higher in the distal than the proximal CRCs46, 52 as well as in the normal-appearing mucosae52 (see Supporting Information Table 2 for more information).

Despite many efforts, as the above mentioned reports, additional studies are required for a better characterization of the subsets of patients with RASSF1A promoter methylation and subsequently a better understanding of the role of RASSF1A in CRC development. Figure 3 is a simplified scheme summarizing the available data regarding RASSF1A methylation in different subsets of CRC.

Figure 3.

Simplified overview of RASSF1A methylation pattern in the different subsets of CRC. For detailed information see text and Additional Supplementary Information.

Other RASSF Family Members and Colorectal Cancer

Similarly to RASSF1A, other RASSF family members have been shown to be inactivated by epigenetic methylation in a number of carcinomas, including colorectal. In particular, RASSF5 and RASSF2 inactivation were shown to be associated with CRC suggesting they may play a critical role in the progression of the tumors.

For instance, the two major transcripts of RASSF5 (NORE1A and NORE1B), derived from different promoter usage,53 were shown to be downregulated in 39 and 31% of primary CRC, respectively and regulation of NORE1A was shown to be tightly associated with promoter CpG sites methylation.54 Abnormal reduction of RASSF5 was more commonly observed in advanced stage and high grade tumors compared to early and low grade tumors.54 Hesson et al. found NORE1A promoter methylation in a number of cancer cell lines including a CRC cell line.55

RASSF2 is another RASSF family member suggested to be a putative tumor suppressor gene frequently inactivated in CRC by promoter methylation. In a series of primary colorectal tumors and adenomas, RASSF2 methylation was found in 42% (51/122) and 43% (21/49) of the cases, respectively. Interestingly, specimens with RASSF2 methylation showed KRAS/BRAF mutations significantly more frequently than those without RASSF2 methylation.56 In contrast, in a study of sporadic CRC, the methylation status of RASSF2A promoter was reported to be negatively associated with mutations in KRAS with 75% of colorectal tumors with RASSF2A methylation having no KRAS mutations.57 In this study, RASSF2A promoter region CpG island was found to be hypermethylated in 89% of colorectal tumor cell lines (8/9), 70% in primary colorectal tumors (21/30), and 88% in colon adenomas (7/8) supporting the idea of an early event, while DNA from matched normal mucosa was found to be unmethylated. The epigenetic and genetic changes occurred in a mutually exclusive manner providing evidence of alternative pathways affecting RAS signaling.57

In a study by Park et al., aberrant methylation of RASSF2A was shown to frequently occur in primary CRC (73%, 106/146) and a positive significant correlation was observed between RASSF2A methylation and KRAS/BRAF mutations in MSS (but not significant in MSI) and distal CRC.58 Furthermore, all colorectal adenomas (16/16) showed RASSF2A methylation, whereas most of the corresponding normal tissues were unmethylated,58 consistent with previous reports57 and suggesting that RASSF2A methylation is an early event in CRC. In a series of 120 colorectal adenomas, Harada et al. observed RASSF2 methylation in 25% (30/120) of the cases and adenomas with RASSF2 methylation often carried KRAS/BRAF mutations simultaneously (73%, 22/30) suggesting that these genetic and epigenetic alterations are mutually correlated and may work synergistically.49 RASSF2 methylation was also significantly more observed in large adenomas. In addition, 65 sporadic CRC samples were analyzed and concomitant RASSF2 methylation and KRAS/BRAF mutations was observed in 58% (11/19) of the lesions in the proximal colon, 5% (1/19) of those in the distal colon and 22% (6/27) of those in the rectum. In the proximal colon, the proportion of tumors with both RASSF2 methylation and KRAS/BRAF mutations was significantly higher in CRCs (58%) than in adenomas (27%).49

In addition to these studies, Nosho et al. also profiled early colorectal tumors for RASSF2 methylation and mutations in genes of the MAPK and PI3K signalling pathways.59 Methylation of RASSF2 was detected in 44% (136/307) of early colorectal tumors and a significant correlation was observed with tumors of larger size. Interestingly, RASSF2 methylation was correlated significantly with KRAS, BRAF or PIK3CA mutation and RASSF2 methylation with KRAS, BRAF or PIK3CA mutation was also correlated significantly with lymphatic invasion, venous invasion and lymph node mestastasis.59 Additionally, as shown by Hesson et al. and Park et al., RASSF2A methylation was found to be more frequently methylated in older patients.57, 58 In a recent study of 243 colorectal cancer tissues, RASSF2A methylation was detected in 58% of the samples and it was suggested that RASSF2A methylation might be a valuable biomarker in plasma for the early detection of colorectal cancer.60

Functional Role of RASSFs

The RAS-mediated signaling pathway is essential in controlling a variety of cellular functions. It is widely accepted that RAS has a central role in the regulation of cell growth through both RAF/MEK/ERK activation, promoting cell proliferation and by interacting with PI3K promoting antiapoptotic effects.9 However, in contrast to the RAF and PI3K oncoproteins, RASSFs are RAS effectors that function as tumor suppressors by modulating some of the growth inhibitory/proapoptotic responses mediated by RAS9, 10 (Fig. 4).

Figure 4.

Schematic simplified overview of the proposed mechanisms of RAS-mediated signaling pathways. Upon binding of growth factors to membrane receptors, as receptor tyrosine kinases, subsequent activation of RAS regulates several signaling cascades. The classical signaling pathways involve the MAPK- (RAS-RAF-MEK-ERK) and the PI3K-pathway (PI3K-Akt) implicated in cell growth, proliferation and survival. Additionally, RAS mediates a pro-apoptotic pathway involving RASSF family members. Of note, that not all downstream effects of RASSF family members are mediated through RAS/RASSF association. RASSF family members, activated by distinct mechanisms, regulate apoptosis (RASSF1A, RASSF5), cell cycle (RASSF1A, RASSF2) and proliferation (RASSF5). (*) denotes molecules frequently altered (genetically or epigenetically) in CRC. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

The implication of RASSF family members as RAS effectors and tumor suppressors has been described by many. Indeed, RASSF1, RASSF2, RASSF4 and RASSF5 were shown to bind RAS in a GTP-dependent manner.23, 61–64 Interestingly, RASSF1A and RASSF5 heterodimerization was shown to be required for interaction with RAS.65 In a number of studies, RASSF1A suppressed growth in vitro and in vivo supporting its role as a tumor suppressor gene.24, 32, 33, 66 More specifically, exogenous expression of RASSF1A in lung cells reduced colony formation, suppressed anchorage-independent growth and inhibited tumor formation in nude mice.24 Burbee et al. also demonstrated that exogenous expression of RASSF1A in a cell line lacking expression decreased in vitro colony formation and in vivo tumorigenicity.32 In two independent studies, mice were generated lacking RASSF1A and interestingly these mice showed an increased susceptibility to develop spontaneous tumors.67, 68

The role of RASSF1A in promoting apoptosis, cell cycle arrest and microtubule stabilization has been observed in a vast number of studies. RASSF1A has been shown to modulate different apoptotic pathways involving distinct set of pro-apoptotic proteins. For instance, RASSF1A, as well as RASSF5 and RASSF2, were shown to interact with the proapoptotic kinases of the Hippo pathway, MST1 and/or MST2, known to activate the SAPK-JNK signaling pathway.69–73 The proapoptotic scaffold protein CNK1 also forms a complex with the RASSF1A (RASSF1A-MST1 complex) increasing the ability to promote RASSF1A-induced cell death.74 In addition, RASSF1A can also activate bax, by binding to modulator of apoptosis 1 (MOAP1), a bax-binding protein.75–77 Thus, RASSF1A seems to have a pivotal role in recruiting a number of proteins to promote cell death.

RASSF1A is also involved in the regulation of the cell cycle. RASSF1A was shown to inhibit cyclin D1 accumulation and to induce cell cycle arrest at G1-S.78 The RASSF1A-induced arrest at G1 was latter associated with suppression of the JNK pathway79 and with the interaction with the transcription factor p120E4F that is then able to suppress cyclin A2.80 Other studies reported that RASSF1A is able to inhibit cell cycle progression at other stages as G2-M81 and prometaphase.82 The prometaphase arrest has been suggested to occur through RASSF1A interaction with Cdc20, an activator of the anaphase-promoting complex (APC), resulting in the inhibition of APC activity necessary for mitotic progression.82 Other studies, however, could not confirm such interaction.83

Other members of the RASSF family have been linked to growth inhibition and cell death. Overexpression of RASSF2 in human cancer cells was shown to promote apoptosis and cell cycle arrest.56, 64 Recently, RASSF2 was shown to form a complex with prostate apoptosis response protein 4 (PAR-4) tumor suppressor and this interaction was shown to be essential for the apoptotic effects of PAR-4.84 RASSF4 overexpression also induced RAS-dependent apoptosis and inhibited the survival of human tumor cell lines.61 In addition to the reported apoptotic effect mediated by the NORE-MST1 complex,70 RASSF5 was shown to mediate growth inhibition and apoptosis in a RAS dependent manner in 293-T cells85 and to suppress growth by modulation of p21CIP1 via p53.86 Moreover, association of NORE1A with cytoskeletal elements was found to be essential for NORE1A-induced growth suppression through the ERK signaling pathway.87 RASSF6 was shown to activate Bax, induce cytochrome c release and trigger both caspase-dependent and -independent pathways of apoptosis.88 Furthermore, overexpressed RASSF6 inhibited the survival of tumor cell lines and induced apoptosis.89 Anchorage-independent growth was also reduced in RASSF8 transfected cells.90

The ability of RASSF1A to promote cell cycle arrest is intimately related with its capacity to associate with microtubules.28 RASSF1A was shown to be a microtubule binding protein that interacts with and stabilizes microtubules.81, 91–94 Moreover, RASSF1A is known to modulate tubulin dynamics and therefore RASSF1A may control cell motility and invasion.28 Overexpression of RASSF1A was shown to inhibit cell migration and RASSF1A inhibition reduced cell–cell adhesion,95 demonstrating its importance in suppressing cellular effects associated to tumor progression.

Conclusions and Perspectives

The MAPK signaling pathway, through activated RAS, is commonly associated with proliferation and cell survival. However, it is now well established that RAS also induces growth inhibitory effects and apoptosis. In recent years, compelling evidence has emerged indicating that members of the RASSF family function as tumor suppressors and modulate, at least in part, the growth inhibitory responses mediated by RAS. In contrast to other RAS effectors, epigenetic control of RASSF family members, as RASSF1A, RASSF2 and RASSF5, occur through promoter methylation and inactivation of RASSFs, particularly RASSF1A has been recognized as a major contributor to the development of CRC having a role early in carcinogenesis. The available data suggests that determining the methylation status of RASSF1A may be useful as a molecular biomarker for CRC susceptibility. However, additional studies are required to further clarify and better stratify patients based on RASSF methylation in combination with KRAS/BRAF/PI3K mutations, MSI status and clinico-pathological features.

Ancillary