Expression of phosphorylated raf kinase inhibitor protein (pRKIP) is a predictor of lung cancer survival

Raf-1 kinase inhibitor protein (RKIP) is a member of a conserved group of proteins called phosphatidylethanolamine-binding proteins (PEBP). RKIP was first identified by Yeung, et al., [1] and was reported to function by inhibiting the Raf-1/MEK/ERK and NF-κB proliferative and survival signaling pathways [1‐3]. Based on modulation of these and other pathways, RKIP is thought to function in a number of physiological and pathological processes [4]. For example, the importance of RKIP in metastasis was demonstrated by the finding that the restoration of RKIP expression inhibits prostate cancer metastasis in a murine model [5, 6] and, hence, RKIP was identified as a metastasis suppressor gene. In addition, over-expression of RKIP reverses tumor cell resistance to apoptosis by both chemotherapeutic drugs [7] and by TRAIL [8]. RKIP has also been implicated as an immune surveillance cancer gene in these studies [8]. The expression level of RKIP is down-regulated in a number of human cancers including highly metastatic prostate carcinoma [6], breast carcinoma [9], colon cancer [10] and hepatocellular carcinoma [11]. RKIP was also shown to be a prognostic marker in the pathogenesis of human prostate cancer [5], and in other cancers [10, 12, 13]. The mechanism of RKIP dysregulation in such malignancies is not clear. Recent findings demonstrated that Snail, a transcription factor overexpressed in many cancers and a metastasis inducer gene product (reviewed in [14, 15], negatively regulates RKIP transcription and expression [16].

The inhibitory activity of RKIP on the Raf-1/MEK/ERK pathway is, at least in part, regulated by PKC-induced phosphorylation of RKIP at serine 153 [17]. The PKC family of serine/threonine kinases is a key mediator of several physiological processes including growth, differentiation, and transformation (e.g., see review [17]). Mutant RKIP that has serine 153 substituted with valine failed to associate with Raf-1 and was not phosphorylated following PKC stimulation. It has also been reported that pRKIP binds to GRK-2 and, thus, inhibits GRK-2-mediated phosphorylation of G-protein coupled receptors (GPCRs) resulting in the inhibition of receptor internalization and cell signaling integrity [18].

In the present study, we have examined the expression levels of total RKIP and pRKIP in human non-small cell lung cancers (NSCLC) on a population basis using a high-density lung tissue microarrays (TMA). Surprisingly, we found that the expression of total RKIP was similar in non-malignant bronchial epithelium, primary tumors and metastatic lesions. Moreover, RKIP neither predicted metastatic potential nor disease-specific death. In contrast, pRKIP expression was a strong predictor of outcome with relatively higher levels of pRKIP predicting a better survival compared to relatively lower expression.

Reductions in RKIP expression has been shown to be an indicator of metastatic spread in numerous cancers [5, 6, 9‐13, 22‐26] as well as a predictor of poor outcome in colon, gastric, and prostate cancer [5, 10, 12]. Recent evidence has also shown RKIP functionality can be modulated through phosphorylation as well [27‐29]. Here, we examined the expression of RKIP and phospho-RKIP (pRKIP) in lung cancer to assess the predictive and/or prognostic power of these proteins. We first evaluated 3 lung cancer cell lines by Western blot analysis to ascertain the relative levels of RKIP and pRKIP (Figure 1). Both RKIP and pRKIP were expressed in all three cell lines, but interestingly, the level of expression was different among the cell types.

To further evaluate the in situ expression of total RKIP and pRKIP in normal and diseased lung on a population basis, we turned to a high-density lung tissue microarray (TMA). The properties of the lung TMA have been described previously [19] and are summarized in Table 1. Of the 696 surgical specimens obtained from 671 patients, 372 primary cases of NSCLC were marker informative and linked with outcome information (disease-specific survival). The TMA consisted of a total of 5,109 spots of benign and malignant histopathologies, of which 3,881 were informative for pRKIP and RKIP. The expression of RKIP and pRKIP was localized primarily in the cytoplasm with some nuclear staining. Figure 2 show representative images of pRKIP and RKIP staining, respectively. Antibody specificity was confirmed based on negative staining with an appropriate non-immune antibody, a concentration-dependent titration of staining intensity, a lack of extracellular staining, and specific competition of antibody binding by a pRKIP peptide (see Figure S1 in Additional file 1). For this study, we focused on the cytoplasmic expression of RKIP and pRKIP. Expression was quantified using integrated measure of frequency and intensity of relevant cells of in a given spot as described in Materials and Methods.

We compared the levels of RKIP and pRKIP expression across histopathologies and found no statistically significant group difference among each category of NSCLC (data not shown). We next examined RKIP and pRKIP expression in normal, primary and metastatic lesions. RKIP expression remained constant for normal versus invasive cancer and metastatic lesions (Figure 3A). However, as shown in Figure 3B, when we examined pRKIP expression, we did find a slight, albeit highly statistically significant group difference (P < 0.0001) with pRKIP expression, in general, decreasing in expression from non-malignant (mean integrated intensity, 0.961) to invasive cancer (mean integrated intensity, 0.954) to lymph node (mean integrated intensity, 0.932) and distant metastases (mean integrated intensity, 0.864).

Here we present evidence that the expression levels of total RKIP was not significantly different between normal epithelium, primary non-small cell lung cancer, or metastatic lesions. Moreover, total level of RKIP had no predictive value for lung cancer patient outcome. In contrast, analyses of the expression levels of pRKIP revealed, for the first time, prognostic significance in patients with NSCLC. Specifically, individuals whose tumors had relatively higher pRKIP expression survive longer as compared to patients with relatively lower levels. It is also interesting to note that the prognostic power of pRKIP is somewhat stronger in early stage cancer (stage 1) and in patients 65 years of age or over. Although we do not yet know the mechanistic basis for this subgroup bias, we do acknowledge that such stratification might be relevant for future targeted therapy and/or early predictions of lung cancer survival. Notably, this is the first time that pRKIP has been examined in any cancer and the first indication of clinical correlate. The present findings in NSCLC are distinguished from previous findings in other cancers whereby low levels of RKIP expression suggested a poor outcome or a greater likelihood of metastasis. These findings emphasize the importance of examining the expression of both the non-phosphorylated active RKIP and phosphorylated inactive RKIP in different cancers.

The physiological significance of RKIP and pRKIP is suggested by the findings that RKIP is involved in the inhibition of the Raf-1/MEK/ERK and NF-κB cell survival signaling pathways. The interplay between these pathways and RKIP expression levels has been implicated at many steps in tumor formation and/or progression. For example, several lines of evidence demonstrate that overexpression of RKIP results in inhibition of the constitutive activation of the Raf-1/MEK/ERK and NF-κB pathways [30, 31]. Further, overexpression of RKIP results in the inhibition of metastasis and invasiveness in various tumor models [6, 16, 32]. RKIP has also been shown to regulate drug resistance in certain systems. One clear indication of this was shown by the reversal of resistance to several chemotherapeutic drugs following overexpression of RKIP [8, 31]. Consistent with this, drug-sensitive tumors were rendered resistant by knockdown of RKIP [7]. RKIP was also shown to regulate immune resistance. Overexpression of RKIP sensitized TRAIL-resistant tumor cells to apoptosis by TRAIL [8].

The underlying mechanism of the differential expression of pRKIP and RKIP is not known. Accordingly, we had expected that relatively lower levels of pRKIP might correlate with a residual higher RKIP levels that would inhibit the Raf-1 and NF-kB pathways and presumably result in a better prognosis. However, the opposite was found. Likewise, we expected relatively higher levels of pRKIP to correlate with the residual lower level of active RKIP resulting in minimal inhibition of survival pathways thus resulting in poorer prognosis. However, the opposite was found. Thus, interestingly, our present data are not concordant with these expectations.

Our findings are in apparent contradiction with other proposed functional activities of RKIP. For example, Corbit et al., have reported that RKIP phosphorylation at serine 153 dissociates RKIP from Raf-1, reversing its inhibitory function [17]. Moreover, Lorenz et al. have reported that pRKIP binds to GRK-2 thus inhibiting GRK-2-mediated phosphorylation of G-protein coupled receptors (GPCRs) [18]. The resultant inhibition of receptor internalization has been predicted to promote cell growth and survival by maintaining appropriate extracellular signaling stimulation [18]. That relatively lower levels of pRKIP apparently correlates with a more aggressive tumor outcome in our patient population raises the intriguing question of whether a relaxation of receptor internalization - predicted with lower pRKIP - might in fact promote cell growth and survival. This interesting possibility is currently being examined. Finally, RKIP has been reported by Eves et al. with regard to its role in the regulation of spindle checkpoints [33]. In this study, pRKIP was found associated with the mitotic centrosomes and kinetochomes in a variety of cell types, thus implicating RKIP/pRKIP in the process of mitosis [33]. RKIP modulates the mitotic spindle assembly checkpoint by controlling Aurora B kinase activity through the Raf/MEK/ERK signaling pathway. Depletion of RKIP results in the inhibition of Aurora B kinase due to the elevated MAP kinase activity. Interestingly, several studies have shown that the excessive activation of Raf-1 MAP kinase inhibits the cell cycle leading to upregulation of cyclin-dependent kinase inhibitors and resulting in cell cycle arrest or senescence [34, 35]. Nevertheless, our results and those of others [36] strongly suggest that, at least in the case of human NSCLC and melanoma, the functional interplay and balance of RKIP and pRKIP and various signaling pathways, may be more complex than seen in in vitro systems or other malignancies. Such tissue-specific mechanisms remain to be elucidated, and certainly should be further clarified prior to considering RKIP as a therapeutic target.

In summary, this study examines for the first time, the expression profile of RKIP and phospho-RKIP in lung cancer. Significantly, we found that phospho-RKIP was a predictive indicator of disease-specific death.