Although it is well documented that n-3 PUFAs may have a range of health benefits including improved cognitive function  and suppression of carcinogenesis , the mechanisms by which these effects are mediated are still far from clear. However, PUFAs are known to affect gene transcription by a number of routes that remain imperfectly understood . In order to obtain new insight into potential novel signaling pathways by which PUFAs may act, we investigated changes in gene expression in MetaCore following exposure to EPA using microarray analysis. GeneGo MetaCore analysis of the data was performed to identify pathways modulated by EPA treatment to try and clarify mechanisms by which a high fish diet may exert benefit. The highest scoring network identified by GeneGo MetaCore was for the Ephrin receptor. This network is particularly associated with cell morphogenesis including factors associated with cytoskeletal structure, formation of cell contacts and adhesion to the extracellular matrix; all factors important in both cancer prevention and cognitive function. Three key genes were identified by microarray, EphB4, EphA1 and Vav2, on the associated pathway as having significantly increased gene expression as a result of EPA treatment. These three genes were chosen for further gene expression analysis by real-time RT-PCR and protein analysis by western blotting in order to substantiate the array findings. The Eph receptor family is one of the largest groups of receptor protein kinases . The Ephs (receptors) and their ligands (ephrins) can be divided into two subclasses, A and B, on the basis of sequence homology, structure and binding affinity . Altered expression of Eph receptors and their ephrin ligands has been reported in a large variety of human cancers including epithelial cancers from the colon and ovary [29–33].
Altered expression of Eph receptors and their ephrin ligands has been reported in a large variety of human cancers including epithelial cancers from the colon and ovary [29–33]. This was therefore an area that we were interested to look at to determine whether treatment with EPA could result in perturbation of these pathways and to elucidate whether the mechanism by which high fish consumption is associated with a reduced incidence of colorectal cancer could be via Eph receptor pathways. The apparent increase in EphA1 expression seen in the arrays experiment as a result of EPA treatment was not replicated in the second experiment and so no further analysis of this gene and protein were undertaken. However, EphB4 gene expression was significantly higher in EPA treated cells than in untreated cells by 8 hr after media change but EphB4 protein expression was significantly down-regulated at this time point with respect to both untreated and AA treated HT29 cells. This apparently contradictory result could be explained by the observation that at earlier time point's treatment with EPA led to a down-regulation in EphB4 expression as compared to untreated cells. Allowing time for this to be translated into protein production the down-regulation in EphB4 expression by EPA treatment at 8 hr is expected. However, the AA data do not support such an explanation as this is even more down-regulated initially but there is no difference in protein expression. These data suggest that EphB4 protein is not well correlated with the mRNA levels for this gene and we postulate that there must be a considerable degree of post-translational control.
In the intestinal epithelium EphB receptors are Wnt signaling target genes that control cell compartmentalization along the crypt axis . EphB4 expression in normal healthy colon is low and outlines the membranes of intestinal 'precursors' at the base of the crypt. It has also been found to be consistently over-expressed in tumor cells of early adenomas when compared to normal tissue [30, 35], with expression shown to be high (50-100% positive cells) in dysplastic aberrant crypt foci and small adenomas. This high level of expression is shown to be lost during colorectal cancer progression at the adenoma-carcinoma transition and was absent in advanced colorectal tumors . This down-regulation during cancer progression was also observed for EphB2. This would suggest a role for EphBs as a tumor suppressor. The causal role of EphB silencing in colorectal cancer progression is supported by the observation that a reduction in EphB activity exacerbates colorectal tumorigenesis in ApcMin/+ mice . It has also been reported that EphB2 has a role in the maintenance of normal tissue architecture in the prostate and mutational inactivation is present in a significant fraction of prostate tumors, suggesting a role for EphB2 in the progression and metastasis of prostate cancer . Thus it is feasible that EPA might suppress carcinogenesis by preventing the down-regulation of EphB activity, although this may well not be directly as a consequence of modified gene expression. The different expression levels of the EphBs in tumor progression and our observed inconsistency between gene and protein expression, highlight the complexity of the pathways involved. Initially increased expression and activation of the pathways appear to benefit tumor growth but ultimately this increased expression imposes restrictions on subsequent tumor progression . Interestingly, it has recently been reported that tumor cells expressing EphB receptors were restricted to large homogeneous clusters by the ligands ephrin-B1 and ephrin-B2 . This response was strongest for ephrin-B1 +EphB2 and ephrin-B1 + EphB3 with the response from ephrin-B1 and EphB4 only producing minor modifications in cell distribution and compartmentalization in the DLD1 colorectal adenocarcinoma and Co115 colon carcinoma cells lines used in that study. This cell compartmentalization subsequently led to a suppression of tumor progression beyond the earliest stage. HT29 cells have very low levels of EphB3 and EphB2 , whether the increase in EphB4 protein expression after exposure to EPA was sufficient to induce the compartmentalization response would require further work However this could suggest a mechanism by which the increase in EphB4 protein expression after exposure to EPA in HT29 cells might be beneficial in terms of halting cancer progression from adenoma through to invasive carcinoma.
The routes by which PUFAs might mediate EphB4 expression remain a matter of conjecture. Transcriptional regulation analysis in MetaCore gave no clear evidence of any one particular transcription factor being involved in the control of gene expression by the tested PUFAs in this study (data not included), but it is recognized that EphB4 expression is regulated by one of the HOX family of proteins (HOXA9) and in turn these have been shown to be involved in signaling by other lipophilic molecules, namely steroids and retinoic acid .
Vav2 protein expression increased over the 24 hr post media change for all treatments while gene expression initially dropped relative to the initial time point. This suggests that Vav2 expression is sensitive to the stress associated with media change. Protein expression was also higher at 8 hr in EPA treated cells but again these changes were not reflected at the gene transcript level. Vav2 is the second member of the Vav oncogene family and unlike Vav1, which is restricted to hematopoietic cells, Vav2 has been found to be expressed ubiquitously . It is a Rho family guanine nucleotide exchange factor and Vav2 has been shown to have a role in growth factor signaling to the cytoskeleton  and is required for integrin-dependant activation of Rac during cell spreading . It has recently come to light that in breast cancer tissue the levels of Vav2 expression as compared to normal or hyperplasic epithelium are down-regulated . Vav2 acts downstream of the Eph A and B receptors in the receptor pathway and its perturbation in breast cancer progression is very relevant to the work we undertook in HT29 cells. Vav2 is important in mediating cytotoxic lymphocyte activity  against target cells including cancer cells, but studies in epithelial cells using siRNA knock-down of Vav2 suggest an important role in stimulating cell migration . It may therefore be that one of the mechanisms by which a diet high in n-3 PUFAs may inhibit the development of colorectal cancer is by preventing the down-regulation of Vav2 levels and thereby limiting cancer progression.
In general we saw little difference in response to EPA and AA over the first 6 h after changing the media, although the response to AA was if anything stronger than to EPA for both EphB4 and Vav2. These changes closely follow the changes in redox status mentioned previously and so may be a response to a transitory increase in oxidative stress. However, this association is by no means conclusive and the observed effects could equally well be associated with similar affinities for some nuclear receptors, for example RXR . Alternatively the early transitory responses may reflect the disappearance of EPA from the medium. Peak intracellular concentrations of EPA are found at 6 h , and thus it may be that we see an initial short-term exposure effect, followed by a more subtle response to the smaller but more sustained differences in PUFA content found after 8-14 hr. At these later times, responses to EPA and AA were dissimilar, such that the expression of both the EphB4 and Vav2 genes was generally higher in those cells exposed to EPA n-3 PUFA than in those exposed to AA n-6 PUFA.