Melanoma Differentiation-associated Gene 7/IL-24 Exerts Cytotoxic Effects by Altering the Alternative Splicing of Bcl-x Pre-mRNA via the SRC/PKCδ Signaling Axis*

Melanoma differentiation-associated gene 7 (MDA-7/IL-24) exhibits cytotoxic effects on tumor cells while sparing untransformed cells, and Bcl-x(L) is reported to efficiently block the induction of cell death by MDA-7/IL-24. The expression of Bcl-x(L) is regulated at the level of RNA splicing via alternative 5′ splice site selection within exon 2 to produce either the pro-apoptotic Bcl-x(s) or the anti-apoptotic Bcl-x(L). Our laboratory previously reported that Bcl-x RNA splicing is dysregulated in a large percentage of human non-small cell lung cancer (NSCLC) tumors. Therefore, we investigated whether the alternative RNA splicing of Bcl-x pre-mRNA was modulated by MDA-7/IL-24, which would suggest that specific NSCLC tumors are valid targets for this cytokine therapy. Adenovirus-delivered MDA-7/IL-24 (Ad.mda-7) reduced the viability of NSCLC cells of varying oncogenotypes, which was preceded by a decrease in the ratio of Bcl-x(L)/Bcl-x(s) mRNA and Bcl-x(L) protein expression. Importantly, both the expression of Bcl-x(L) and the loss of cell viability were “rescued” in Ad.mda-7-treated cells incubated with Bcl-x(s) siRNA. In addition, NSCLC cells ectopically expressing Bcl-x(s) exhibited significantly reduced Bcl-x(L) expression, which was again restored by Bcl-x(s) siRNA, suggesting the existence of a novel mechanism by which Bcl-x(s) mRNA restrains the expression of Bcl-x(L). In additional mechanistic studies, inhibition of SRC and PKCδ completely ablated the ability of MDA-7/IL-24 to reduce the Bcl-x(L)/(s) mRNA ratio and cell viability. These findings show that Bcl-x(s) expression is an important mediator of MDA-7/IL-24-induced cytotoxicity requiring the SRC/PKCδ signaling axis in NSCLC cells.

A number of studies have demonstrated that the overexpression of Bcl-x(L) in cells confers apoptosis resistance as well as cooperates with oncogenic factors (e.g. c-Myc) in tumorigenesis (1)(2)(3)(4)(5)(6)(7)(8)(9)(10). The regulation of Bcl-x(L) expression can be complex at times, consisting of both transcriptional and post-transcriptional processes. In regard to post-transcriptional processing/ alternative splicing, the BCL-x gene, via alternative 5Ј splice site (5ЈSS) 5 selection within exon 2, produces either the Bcl-x(s) isoform through activation of an upstream/proximal 5ЈSS or the Bcl-x(L) isoform through activation of a downstream/distal 5ЈSS. A number of studies have demonstrated that Bcl-x(s), in contrast to Bcl-x(L), promotes apoptosis (9,(11)(12)(13)(14). Hence, the alternative 5ЈSS selection of Bcl-x pre-mRNA emerged as a potential target for anti-cancer therapeutics. For example, Taylor et al. (15) demonstrated that Bcl-x 5ЈSS selection can be specifically modulated using antisense oligonucleotides specific against the Bcl-x(L) 5Ј splice site. Treatment of cells with these oligonucleotides induced an increase in the expression of Bcl-x(s) and a decrease in the expression of Bcl-x(L), resulting in sensitization of NSCLC cells to chemotherapeutic agents (15). These findings were also demonstrated by Kole and co-* This work was supported by research grants from the Veteran's Administration (VA Merit Review, I BX001792 (to C. E. C.) and a Research Career Scientist Award, 13F-RCS-002 (to C. E. C.)); from the Vietnam Education Foundation (fellowship to N. T. V.); from the National Institutes of Health via Grants HL125353 (to C. E. C.), HD087198 (to C. E. C.), CA117950 (to C. E. C.), CA154314 (to C. E. C.), RR031535 (to C. E. C.), CA192613 (to P. D.), CA097318 (to P. B. F.), CA127641 (to P. B. F.), P01 CA104177 (to P. B. F.), and a T32 Post-Doctoral Fellowship (Postdoctoral Training Program in Cancer Biology, CA085159) (to B. A. S.); from the United States-Israel Binational Science Foundation via Grant BSF#2011360 (to C. E. C.); and from the National Foundation for Cancer Research (to P. B. F.). The authors declare that they have no conflicts of interest with the contents of this article. The contents of this manuscript do not represent the views of the Department of Veterans Affairs or the National Institutes of Health. 1 Both authors contributed equally to this work. 2  workers (16) in additional cancer types as well as in vivo models. Thus, regulation of the 5ЈSS selection within the Bcl-x exon 2 is a critical factor in determining whether a cancer cell is susceptible or resistant to apoptosis in response to chemotherapy (15)(16)(17)(18)(19).
In cells, Bcl-x 5ЈSS selection is regulated by the generation of de novo ceramide in response to apoptotic stimuli such as the chemotherapeutic agent, gemcitabine (20,21). More recent studies by Zhou and co-workers (22) and Chang et al. (23) verified these early findings and extended the list of chemotherapeutic agents to emetine, a potent protein synthesis inhibitor, and amiloride, a potassium-conserving diuretic. Later studies from our laboratory identified the RNA splicing factor, SAP155, as a regulator of the 5ЈSS selection of Bcl-x pre-mRNA (24,25), and this RNA trans-factor was required for the effect of ceramide on the alternative 5ЈSS selection of Bcl-x pre-mRNA in NSCLC cells (24,25).
In the present study, the role of melanoma differentiationassociated gene-7/interleukin-24 (MDA-7/IL-24) was examined in the context of Bcl-x 5ЈSS selection. MDA-7/IL-24 is a cytokine classified as a member of the IL-10 gene family that was initially identified through a subtraction hybridization approach using a differentiation therapy model of human melanoma (26). MDA-7/IL-24 potently inhibits cell growth and induces apoptosis in various epithelial cancers both in vitro and in vivo, including lung cancers (27). In contrast, MDA-7/IL-24 has shown no lethality toward normal cells (28).
The ability of MDA-7/IL-24 to inhibit cell growth of tumor cells and to induce apoptosis in tumor cells has been attributed, in part, to modulation of the expression of Bcl-x(L) (27,29,30). Specifically, a potential functional role for changes in Bcl-x(L) expression in adenovirus-delivered MDA-7/IL-24 (Ad.mda-7)induced apoptosis was suggested by the finding that forced overexpression of Bcl-x(L) diminished the apoptotic effect of Ad.mda-7 in lung carcinoma cells (27,29). The possible link to Bcl-x 5ЈSS selection was suggested in this mechanism as the induction of ceramide production plays a decisive role in MDA-7/IL-24-mediated apoptosis (31,32).
In this study, we explored the hypothesis that MDA-7/IL-24 reduces the levels of Bcl-x(L) by modulating the 5ЈSS selection of Bcl-x pre-mRNA in a de novo ceramide-dependent manner. Indeed, we demonstrate that MDA-7/IL-24 induces the activation of the Bcl-x(s) 5Ј splice site, thereby lowering the Bcl-x(L)/ (s) ratio in NSCLC cells, and thus, instigating the down-regulation of Bcl-x(L). Surprisingly, this mechanism was ceramideindependent, but the loss of SAP155 expression was still observed. Furthermore, the expression of Bcl-x(s) mRNA was shown to be a major component in the ability of MDA-7/IL-24 to induce the loss of cell viability as well as induce the loss of Bcl-x(L) expression. Exploration of the signal transduction pathway mediating this distal mechanism in response to MDA-7/IL-24 identified the SRC/PKC␦ signaling axis as critical. These findings, therefore, suggest that induction of Bcl-x(s) mRNA may prove an effective therapeutic avenue to enhance the cancer-specific killing of MDA-7/IL-24 treatment, which may be an effective treatment for NSCLC lung tumors presenting with a low Bcl-x(L)/(s) ratio.

Results
Ad.mda-7 Induces a Loss of Cell Viability in NSCLC Cells-Previously, MDA-7/IL-24 was reported to induce cytotoxic effects on NSCLC cell lines without affecting non-transformed counterparts (27,28). Our initial studies confirmed this cytotoxic effect in regard to adenovirus-delivered MDA-7/IL-24 (Ad.mda-7) for several established NSCLC cell lines with differing oncogenotypes (A549, H838, and H1299) after 48 and 72 h (Table 1 and Fig. 1, A-C). Of note, no loss of viability was observed in these cell lines within Յ24 h of Ad.mda-7 treatment (data not shown). Importantly, Ad.mda-7 treatment had no significant effect on the viability of non-transformed, immortalized lung epithelial cells (HBEC-3KT cells; Fig. 1D). Hence, Ad.mda-7 elicits cytotoxicity in tumorigenic lung cells regardless of oncogenotype, while sparing non-cancerous lung cells as reported previously (27,28).
Ad.mda-7 Induces Alterations in the 5Ј Splice Site Selection of Bcl-x Pre-mRNA-The loss of Bcl-x(L) expression is a required mechanism for MDA-7/IL-24-induced loss of cell viability in several cancer cell types (mesothelioma I-45xL, GBM glioblastoma, and prostate carcinoma cells) (29, 30 -32). The alternative splicing of Bcl-x pre-mRNA is one method of regulating Bcl-x(L) expression. Indeed, alterations in Bcl-x splicing are sensitive to ceramide production, and MDA-7/IL-24 is reported to increase ceramide synthesis (20,21,24,25,33). Furthermore, the ceramide-sensitive RNA trans-factor, SAP155, promotes the formation of Bcl-x(L) mRNA, and siRNA targeting SAP155 results in decreases in the Bcl-x(L)/ Bcl-x(s) mRNA ratio in NSCLC cells (25). Thus, we hypothesized that MDA-7/IL-24 treatment induces the down-regulation of SAP155 and the subsequent lowering of the Bcl-x(L)/ Bcl-x(s) mRNA ratio prior to the observed loss of viability (less than 48 h). Consistent with this hypothesis, A549 cells treated with Ad.mda-7 for 24 h exhibited a reduction in SAP155 protein levels ( Fig. 2A). Short-term treatment of NSCLC cells with Ad.mda-7 also induced a significant decrease in Bcl-x (L)/(s) mRNA ratios when compared with control Ad.CMV adenovirus ( Fig. 2A). This effect of Ad.mda-7 was both concentrationdependent and stable for Ͼ36 h (Fig. 2, B and C). Ad.mda-7 also altered Bcl-x alternative splicing in H838 cells (Fig. 2D), demonstrating translatability to other NSCLC cell lines of differing oncogenotypes. To determine whether the effect was specific to NSCLC cells, the ovarian cancer cell lines (SKOV and DOV) were treated with either Ad.CMV or Ad.mda-7 (Fig. 2D). These cell lines also demonstrated significant changes in Bcl-x splicing. Importantly, Bcl-x alternative splicing was not affected by Ad.mda-7 in non-transformed HBEC-3KT cells, correlating with a lack of cytotoxicity induced by MDA-7/IL-24 (Fig. 2E). Lastly, we examined whether the effect of MDA-7/IL-24 on RNA splicing was specific for Bcl-x pre-mRNA. In this regard, we tested the capacity of MDA-7/IL-24 to affect the splicing ratios of Mcl-1(L)/(s) and CD44, but no effect on the alternative splicing of these pre-mRNAs was observed (Fig. 2F). Hence, MDA-7/IL-24 induces a reduction in the Bcl-x(L)/Bcl-x(s) mRNA ratio, which was not due to a generalized effect on constitutive RNA splicing.
Ablation of Bcl-x(s) "Rescues" Ad.mda-7-elicited Cytotoxicity-As shown in this study, Ad.mda-7 reduces the cell viability on NSCLC cells and decreases the ratio of Bcl-x(L)/(s) mRNA. Although these data intriguingly correlate, they do not determine whether or not the Bcl-x splicing is a required mechanism for Ad.mda-7-induced killing or simply secondary to its pro- apoptotic effect. Furthermore, we posited the question: is it the loss of Bcl-x(L) or the production of Bcl-x(s) that plays a role in MDA-7/IL-24-induced loss of cell viability? To answer this question, we again treated NSCLC cells with Ad.mda-7, and first examined Bcl-x(L) expression. Indeed, it is currently known that MDA-7/IL-24 treatment reduces Bcl-x(L) expression, and that cells ectopically expressing Bcl-x(L) are resistant to MDA-7/IL-24-induced cytotoxicity (29). In line with these published observations, we observed that MDA-7/IL-24 reduced Bcl-x(L) protein levels in A549 cells after a 48-h incubation (Fig. 3A). The Ad.mda-7-elicited reduction in Bcl-x(L) protein was also evident in H838 cells, but not HBEC-3KT cells (Fig. 3, B and C). Surprisingly, Ad.mda-7 was not able to stim-ulate the detectable expression of Bcl-x(s) protein (Fig. 3, A-C). Other laboratories have also reported difficulty detecting endogenous Bcl-x(s), and typically, either 10-fold more protein extract or high ectopic expression of Bcl-x(s) is required to detect the protein (9,15,21,34). Thus, it was unclear whether Bcl-x(s) mRNA or protein was sufficient to mediate the cytotoxic effects of MDA-7/IL-24.
To address whether Ad.mda-7 exerts its cytotoxic effect via the production of Bcl-x(s), A549 cells were treated with the Bcl-x(s)-targeted siRNA followed by treatment with either Ad.mda-7 or Ad.CMV. As anticipated, MDA-7/IL-24-stimulated Bcl-x(s) mRNA levels were reduced by Bcl-x(s) siRNA (Fig. 4A). To examine whether the expression of Bcl-x(s) mRNA was a mediator of MDA-7-induced cell death, NSCLC cells were again treated with Bcl-x(s) siRNA, followed by either Ad.CMV or Ad.mda-7 for 48 and 72 h. Central to this study, NSCLC cells treated with Bcl-x(s) siRNA exhibited a significant increase in viability following Ad.mda-7 treatment when compared with NSCLC cells treated with control siRNA after 48 and 72 h (Fig. 4B). These data demonstrate that Ad.mda-7 induces cell death, at least in part, by the expression of Bcl-x(s) mRNA.
The Expression of Bcl-x(s) mRNA Restrains the Expression of Bcl-x(L)-Our above data suggest that the expression of Bclx(s) is an important mechanism for MDA-7/IL-24-induced cell death. As mentioned previously, the Bcl-x(s) protein was not detected after Ad.mda-7 treatment. However, we did find that the levels of Bcl-x(L) protein were dramatically preserved by the down-regulation of Bcl-x(s) mRNA (Fig. 5A). Specifically, A549 cells were again treated with Ad.mda-7 in the presence or absence of Bcl-x(s) siRNA. Remarkably, NSCLC cells treated with Bcl-x(s) siRNA exhibited significantly enhanced levels of Bcl-x(L) protein when compared with control siRNA. These data demonstrate that Bcl-x(s) siRNA (i.e. specific removal of Bcl-x(s) mRNA) rescued the loss of Bcl-x(L) protein expression induced by Ad.mda-7 (Fig. 5A). This effect was not mediated at the mRNA level for Bcl-x(L), as the expression of Bcl-x(L) mRNA did not mimic the observed effects on the protein (data not shown). Thus, Bcl-x(L) protein expression is negatively regulated by Bcl-x(s) expression.
To determine whether the expression of the coding sequence of Bcl-x(s) mRNA was mediating this effect on Bcl-x(L) expression, Ad.Bcl-x(s), which contains only the coding mRNA, was used to express this cDNA in A549 cells. The expression of Bcl-x(s) coding mRNA induced a significant (58%) reduction in the levels of Bcl-x(L) protein (Fig. 5B). More importantly, cotreatment with Bcl-x(s) siRNA completely blocked this effect, with no effect observed for Bcl-x(s) siRNA or coding cDNA on the levels of Bcl-x(L) mRNA (data not shown). These data demonstrate, for the first time, a novel mechanisms by which MDA-7/IL-24 induces cytotoxicity in tumor cells, at least in part, by generating Bcl-x(s) mRNA, which reduces Bcl-x(L) protein levels, and thereby, limits the cytoprotective effects of Bcl-x(L).

MDA-7/IL-24-induced Alterations in
Bcl-x Pre-mRNA Are Independent of Ceramide-generating Pathways-As mentioned previously, MDA-7/IL-24 is known to induce ceramide synthesis via the de novo pathway (24,25,33), and Bcl-x 5ЈSS selection has been previously reported by our laboratory to be responsive to de novo ceramide production elicited by gemcitabine (20,21,24,25). Therefore, we hypothesized that MDA-7/IL-24-induced reductions in the Bcl-x(L)/Bcl-x(s) splicing ratio required de novo ceramide production. Surprisingly, incubation of MDA-7/IL-24-treated A549 cells with fumonisin B 1 or myriocin, two inhibitors of de novo ceramide synthesis, were unable to inhibit the effect of MDA-7/IL-24 on the Bcl-x(L)/ Bcl-x(s) splicing ratio (Fig. 6, A and B). Additional inhibitors of sphingolipid synthesis such as acid and neutral sphingomyelinase did not affect the ability of MDA-7/IL-24 to reduce the Bcl-x(L)/Bcl-x(s) splicing ratio (data not shown). These data demonstrate that MDA-7/IL-24 affects the alternative splicing of Bcl-x pre-mRNA via a ceramide-independent pathway.

MDA-7/IL-24-induced Alterations in
Bcl-x pre-mRNA Require the SRC/PKC␦ Signaling Axis-Because the mechanism by which MDA-7/IL-24 induced the activation of the Bclx(s) 5Ј splice site was independent of ceramide signaling, we undertook a broad-based approach to identify the signaling pathway required for MDA-7/IL-24 to affect Bcl-x RNA splicing. To identify a target pathway, we subjected A549 cells to an array of small molecule inhibitors ( Table 2) targeting critical factors in signaling pathways related to MDA-7/IL-24-induced cell death (e.g. ER stress, SRC kinase, and protein kinase C (PKC) signaling pathways). Of these, the SRC inhibitor, Src-1, and the broad spectrum PKC inhibitor, Gö 6983, completely inhibited the ability of MDA-7/IL-24 to reduce the Bcl-x(L)/(s) mRNA ratio (Fig. 7, A and B). Interestingly, Inoue et al. (35) demonstrated that SRC signaling pathways were involved in MDA-7/IL-24 signaling, and other laboratory groups have demonstrated that the novel PKC isoform, PKC␦, is a downstream mediator of SRC signaling (70). For this reason, we treated NSCLC cells with rottlerin, a reported inhibitor of PKC␦ (36), and specific PKC␦ and SRC siRNA. Both the small molecule inhibitor of PKC␦ and down-regulation of the enzyme by siRNA blocked the ability of MDA-7/IL-24 to reduce the Bcl-x(L)/(s) mRNA ratio and to rescue MDA-7-mediated cytotoxicity (Fig. 7, B and C). Taken together, these data indicate that MDA-7/IL-24-induced alterations in Bcl-x pre-mRNA splicing are dependent on the SRC/PKC-␦ signaling axis (Fig.  7D), but independent of de novo ceramide generation.

Discussion
A number of studies have been published demonstrating that alternative splicing is often dysregulated in cancer (24,25,(37)(38)(39)(40)(41)(42)(43)(44)(45). For example, modifications in the alternative splicing of apoptotic factors such as Bcl-x, caspase 9, and Mcl-1 have been linked to the resistance to chemotherapy of neoplastic cells (20,21,24,25,37,38,44,45). In this study and for the first time, we show that MDA-7/IL-24 treatment of NSCLC cells alters the 5ЈSS selection of Bcl-x pre-mRNA, thereby increasing the pro-apoptotic Bcl-x(s) mRNA relative to anti-apoptotic Bcl-x(L) in A549 cells. The effect of MDA-7/IL-24 treatment on alternative splicing was specific to Bcl-x pre-mRNA, as Ad.mda-7 treatment had no effect on the ratio of Mcl-1(L)/ Mcl-1(s) as well as any of the numerous CD44 splice variants. The effect of MDA-7/IL-24 on activating the Bcl-x(s) 5ЈSS also appeared to be independent of the oncogenotype (Table 1), as Ad.mda-7 also decreased the Bcl-x(L)/(s) ratio in H838 cells. The effect of MDA-7/IL-24 on Bcl-x 5ЈSS selection was also evident in other epithelial tumor cells such as the SKOV3 ovarian cancer cell line (p53 deletion, normal Ras), although to a lesser degree. Hence, the specific effect of MDA-7/IL-24 on the activation of the Bcl-x(s) 5ЈSS translates to multiple oncogenotypes as well as various solid tumor types, suggesting a broadbased mechanism that can be exploited for cancer therapeutics.
In regard to the therapeutic efficacy of MDA-7/IL-24, our findings in this study led to the hypothesis that the alternative 5ЈSS selection of Bcl-x pre-mRNA was an important mecha-nism for MDA-7/IL-24 to simply lower the expression of Bclx(L) and enhance the cytotoxic effects of this cytokine on cancer cells. In contrast to this hypothesis, we observed that siRNAs specifically targeting Bcl-x(s) significantly blocked the MDA-7/IL-24-induced cytotoxicity to NSCLC cells. These data initially suggested that the expression of Bcl-x(s), and not just the down-regulation of Bcl-x(L), induced by activation of the proximal 5ЈSS in exon 2, was important in the cytotoxic effects elicited by MDA-7/IL-24. What was most surprising in regard to the above findings is the lack of observed expression for Bclx(s) protein, although an increase in the mRNA is observed. Hence, the expression of Bcl-x(s) protein is unlikely to be mediating the cytotoxic effects of MDA-7/IL-24, and this study begins to clarify a decade-old question regarding the function of Bcl-x(s). In particular, previous studies suggested that Bcl-x(s) functions by dimerizing with Bcl-x(L) to sequester Bcl-x(L) from Bax association (34). Although the studies examining this interaction in vitro are compelling, albeit on occasion in oppo- sition of earlier studies on Bcl-x(s) function (9,34), the Bcl-x(s) protein itself is rarely observed in most cell types unless ectopically expressed (9, 15, 21, 34). In addition, 10-fold more protein extract has been required to detect Bcl-x(s) at the protein level in previously reported studies (15,21). The findings from our laboratory over the years may even suggest that the Bcl-x(s) protein may have been misidentified as a separate Bcl-x splice variant (Fig. 5B) (see for example Refs. 24,25). Regardless, the current study indicates that the expression of the Bcl-x(s) mRNA, not the protein, regulates the biological effects of MDA-7/IL-24-induced loss of cell viability by reducing the protein levels of Bcl-x(L). For example, NSCLC cells treated with Bcl-x(s) siRNA significantly rescued the reduction in Bcl-x(L) expression induced by Ad.mda-7 as well as treatment with Ad.Bcl-x(s). These data suggest that Bcl-x(s) mRNA functions to inhibit the expression of Bcl-x(L) rather than the Bcl-x(s) protein, directly antagonizing the function of the Bcl-x(L) protein.
The mechanism by which Bcl-x(s) coding mRNA elicits this effect remains unclear, although the data suggest that the effect occurs at the level of Bcl-x(L) protein synthesis or turnover/ stability. One can surmise that the removal of the portion of exon 2 sequence encoding for the Bcl-x(L) mRNA induces the formation of a new RNA cis-element or hairpin structure that competes for the association of an RNA trans-factor or RNAbinding protein important in the synthesis of the Bcl-x(L) protein. Indeed, cytosolic polyadenylation binding proteins such as the CPEB family members play roles in regulating cytoplasmic polyadenylation, and thus, regulating the translation of proteins in response to cellular stress. These proteins bind specific RNA cis-elements, and Bcl-x(s) may simply act as a scavenger for an activating CPEB2, such as CPEB2B, which has roles in driving anoikis resistance and metastasis in triple negative breast cancer (43). Although this is a plausible mechanism, the effect of Bcl-x(s) mRNA on Bcl-x(L) protein expression may also occur at the post-translational level as there is a dramatic and rapid loss of Bcl-x(L) protein observed in response to Ad.mda-7. Indeed, Bcl-x(s) mRNA may also bind/sequester factors that stabilize the Bcl-x(L) protein, leading to the degradation of the protein. In support of this possibility, Fisher and co-workers (29) have shown that MDA-7/IL-24 can induce the loss of Bcl-x(L) at the post-translational level. Lastly, another possibility, albeit remote, exists in that Bcl-x(s) mRNA acts as a

List of Inhibitors
Signaling pathways and factors examined for involvement in the alternative splicing of Bcl-x pre-mRNA modulated by MDA-7/IL-24 as analyzed by small molecule inhibitors are shown. The table depicts the inhibitors and their respective concentrations utilized in this study. Signaling pathways were examined for effects on the ratio of Bcl-x(L)/(s) mRNA utilizing small-molecule inhibitors at doses well characterized in the scientific literature and previously utilized in studies on A549 cells (43)(44)(45)(46)(47). mTOR, mammalian target of rapamycin. N/C, no change; iPLA2, intracellular phospholipase A2.  ribozyme for removal of the Bcl-x(L) protein. Although unlikely and not previously recorded for mammalian cells, several types of RNA mechanisms first thought to be limited to lower organisms have now been reported in mammalian cells such RNA trans-splicing and cytosolic RNA splicing (e.g. XBP-1 pre-RNA associated with the unfolded protein response) (44,45). Overall, the ability of Bcl-x(s) mRNA to regulate the expression of Bcl-x(L) is novel and warrants future studies to identify this intriguing new mechanism.

Method of action Concentration Outcome
In this study, we also delved into the signaling mechanism regulating the 5ЈSS selection of Bcl-x pre-mRNA in response to MDA-7/IL-24. Specifically, we show that SAP155, an RNA trans-factor reported by us as a major regulator of the 5ЈSS selection of Bcl-x pre-mRNA, was down-regulated by this cytokine (24,25). Chalfant and co-workers (24,25) and Fisher and co-workers (33) identified ceramide generation and subsequent activation of ceramide-induced protein phosphatases as a major mechanism by which certain chemotherapies reduced the survival of NSCLC cells. Furthermore, MDA-7/IL-24 induced cell death via a ceramide-dependent pathway in some cancer cell types (32,33). Based on these studies, we examined the role ceramide in MDA-7/IL-24-induced reductions in Bclx(L)/Bcl-x(s) pre-mRNA ratios, but surprisingly, the ceramide synthase inhibitor, fumonisin B 1 , was unable to block MDA-7/ IL-24-elicited alterations in Bcl-x alternative splicing. Furthermore, myriocin, which inhibits ceramide synthesis at the level of sphingolipid biosynthesis, had no effect on MDA-7/IL-24induced changes in Bcl-x(L)/Bcl-x(s) mRNA ratios. Additional inhibitors of ceramide generation also had no effect. Therefore, MDA-7-induced changes in Bcl-X splicing must occur in a ceramide-independent manner.
The ceramide-independent nature of the signaling mechanism regulating Bcl-x RNA splicing in response to Ad.mda-7 led to a more broad-based approach and identified the SRC/ PKC␦ signaling pathway responsible for the modulation in Bcl-x RNA splicing in response to MDA-7/IL-24 (35). This was a novel finding as this is in contrast to previous signaling pathways reported to regulate Bcl-x RNA splicing in cancer such as the PI3K/PKC pathway or PKC␣ signaling in non-cancerous cells (see for example Refs. 24,25,71,72). One may surmise that MDA-7/IL-24 logically acts to drive terminal differentiation pathways, explaining the SRC activation paradigm as well as PKC␦, which are implicated in a number of differentiation processes. Indeed, our laboratories have previously shown that anoikis-resistant cancer cells induced SRC activation in response to MDA-7/IL-24, which correlated with significantly higher cell death (46). Thus, although SRC signaling is usually linked to pro-survival/oncogenic pathways, MDA-7/ IL-24 may reinstate the normal cell cycle checkpoints, allowing for additional oncogenic stress (1SRC activation) to induce cellular senescence and subsequent loss of viability for the cancer cell. Validation of this hypothesis may aid in explaining why SRC inhibitors did not show strong clinical effectiveness for breast and lung cancer as a single agent (47,48). Overall, this study allows the postulation that multiple signaling pathways exist to modulate specific RNA trans-factors to repress or activate the Bcl-x(s) 5ЈSS, suggesting that this splicing event is a key distal point for numerous cell signaling pathways, and thus, may have broad roles in a number of biological functions as well as be exploited to enhance the killing of NSCLC cells.
In regard to the RNA trans-factor modulated by the SRC/ PKC␦ pathway, this study does not conclusively demonstrate that SAP155 is the mediator of the activation of the Bcl-x(s) 5Ј splice site in response to MDA-7, but several key pieces of data implicate this RNA trans-factor in this paradigm. First, the levels of SAP155 are decreased in response to MDA-7 prior to an effect on Bcl-x RNA splicing, and our laboratory and others showed that this level of SAP155 reduction in cells would induce the activation of the Bcl-x 5Ј splice site (25,49). Second, MDA-7 did not significantly affect the levels of SAM68, hnRNP H, hnRNP K, and SRSF1, RNA trans-factors reported to affect Bcl-x RNA splicing in other cell systems (data not shown) (50,51). Thus, SAP155 is a strong candidate as a regulating RNA trans-factor in this signaling cascade. Whether PKC␦ directly affects the expression of SAP155 or another regulatory RNA trans-factor via phosphorylation or indirectly via additional downstream members of the signaling cascade is unknown. A study by White et al. (52) suggests that PKC␦ and the RNA trans-factor, hnRNP K, are linked in another type of apoptosis mechanism. Although hnRNP K levels were not affected by MDA-7, one can also hypothesize that PKC␦ may directly phosphorylate hnRNP K in response to MDA-7 to affect Bcl-x 5Ј splice site selection. It is also plausible that PKC␦ phosphorylates SAP155 as regulation of this RNA trans-factor by protein kinases and phosphatases has been reported (53)(54)(55)(56). Furthermore, PKC␦ has also been shown to translocate to the nucleus in response to apoptotic agonists (57), which localizes the enzyme in proximity to RNA splicing factors. Indeed, Cocco and co-workers (58) demonstrated that phosphoinositide-specific phospholipase C␤ 1 (PI-PLC␤1) will translocate to the nucleus with stimuli and is associated with RNA trans-factors (e.g. SRSF3). Thus, the enzyme that generates the activating second messenger lipid of PKC␦, diacylglycerol, is also localized to appropriate intracellular areas to mediate the direct regulation of RNA splicing. Future studies will focus on delineating the direct nature of this mechanism.
One open question from this study is whether MDA-7/IL-24 is driving cell death via receptor-mediated events using the IL-20R1/IL-20R2 and IL-22R1/IL-20R2 receptors or via intracellular pathways. For example, NSCLC cell lines have been reported to lack canonical cell surface IL-20/IL-22 receptor pairs for MDA-7/IL-24 such as the H1299 and A549 cell lines used in this study (59,60). Accordingly, the MDA-7/IL-24 paracrine loop is not active in these NSCLC cells, and alternative intracellular signaling pathways involving ER stress have been implicated for Ad.mda-7 effects in other cancer cell types (e.g. prostate cancer cell lines) lacking the surface receptors (46,59). Indeed, PKC␦ activation has been implicated in response to ER stress (57), suggesting that Ad.mda-7 in NSCLC cells is inducing the activation of the Bcl-x(s) 5Ј splice site via this pathway. On the other hand, our data demonstrating that inhibitors of ER stress pathways had no effect on Bcl-x RNA splicing ( Table  2) argue against these pathways mediating the effects of Ad.mda-7, and thus, suggest a receptor-mediated pathway. Indeed, there is a lack of knowledge as to whether the MDA-7/ IL-24 expressed intracellularly can act in an autocrine fashion OCTOBER 7, 2016 • VOLUME 291 • NUMBER 41 outside of a paracrine mechanism due to differences in affinity for IL-20R homo-dimers. Furthermore, the MDA-7/IL-24 receptors may be sequestered internally in NSCLC, much like the FAS receptor in prostate cancer cells as shown by Norris and co-workers (61). Thus, a comprehensive study is necessary to delineate the current conundrum of whether the MDA-7/ IL-24 receptors are sequestered intracellularly and active, and thus, can bind MDA-7/IL-24 expressed internally to mediate cell signaling events.

MDA-7/IL-24 Alters Bcl-x RNA Splicing
In conclusion, an important mechanism by which MDA-7/ IL-24 elicits lethality to NSCLC cells has been identified (i.e. via altering the alternative splicing of Bcl-x pre-mRNA to increase Bcl-x(s) mRNA at the expense of Bcl-x(L) mRNA). Our data further suggest that the expression of Bcl-x(s) mRNA reduces the Bcl-x(L) available to bind Bax, or stabilize mitochondrial membranes, enhancing the cytotoxic effects of MDA-7/IL-24 on NSCLC cells. As several clinical trials for MDA-7/IL-24 are ongoing for the treatment of various cancers, our findings suggest that the modulation of Bcl-x 5ЈSS selection may promote the therapeutic efficacy of this cytokine. Furthermore, assaying the Bcl-x(L)/(s) mRNA ratio in patient tumors undergoing MDA-7/IL-24 therapy may serve as a biomarker for determining the efficacy of this therapy during treatment.

Experimental Procedures
Materials DMEM, RPMI, FBS, and penicillin/streptomycin (100 units/ ml penicillin G sodium, and 100 g/ml streptomycin sulfate) were obtained from Invitrogen Life Technologies. NSCLC cell lines of various oncogenotypes (Table 1), A549, H838, and H1299 cells, were obtained from ATCC (Manassas, VA). HBEC-3KT cells were a gracious gift of Drs. John Minna and Jerry Shay (University of Texas-Southwestern, Dallas, TX) (62). The Bcl-x(s) adenovirus, which was propagated as reported previously (34), was a gracious gift of Dr. Gabriel Nunez (University of Michigan, Ann Arbor, MI).

Cell Culture
A549, H838, and H1299 cells were grown in 50% RPMI 1640 and 50% Dulbecco's modified Eagle's medium supplemented with L-glutamine, 10% (v/v) fetal bovine serum, 100 units/ml penicillin G sodium, and 100 g/ml streptomycin sulfate. All NSCLC cell lines were used within 6 months and verified by the company via characteristic morphology consistent with epithelial origin, a positive result for epithelial cell marker cytokeratin 18, and where applicable, by mutational analysis and genotyping. HBEC-3KT cells were used within 6 months of receipt from the Minna and Shay laboratory, and were validated as described previously by them (62). These cells were cultured with keratinocyte serum-free medium containing bovine pituitary extract and recombinant epidermal growth factor (Life Technologies). All cells were maintained at less than 80% confluency under standard incubator conditions (humidified atmosphere, 95% air, 5% CO 2 , 37°C).

Quantitative/Competitive RT-PCR Assays
Competitive qRT-PCR-Total RNA from cell lines was isolated using the RNeasy Mini Kit (Qiagen Inc., Valencia, CA) according to the manufacturer's protocol. Total RNA (1 g) was reverse-transcribed using Superscript III reverse transcriptase (SuperScript TM First-Strand Synthesis System for RT-PCR, Invitrogen TM ), and PCR was performed for Bcl-x splice variants as described previously (20,21,24,25). The final PCR products were resolved by 5% Tris borate-EDTA acrylamide gel electrophoresis, stained with SYBR Gold (Invitrogen TM ), and visualized using a Molecular Imager FX (Bio-Rad) with a 488-nm EX (530-nm BYPASS) laser. This assay is quantitative in regard to comparing the ratio of Bcl-x(L) to Bcl-x(s) mRNA between different samples.
Standard Real-time qRT-PCR-Total RNA and reverse transcription were undertaken as described above for quantitative/ competitive RT-PCR. For real-time quantitative PCR, the levels of Bcl-x(L) and Bcl-x(s) mRNA were individually determined using quantitative PCR in accord with the TaqMan technology (Applied Biosystems) specific to Bcl-x(L), Bcl-x(s), and 18S RNA as described previously by us (24,25). The cDNA was amplified using ABI 7900HT.

WST-1 Assay
Cells (1.0 ϫ 10 4 ) were plated into each well of a 96-well plate in 0.1 ml of medium. After 24 h at standard incubator conditions, the cells were transfected with siRNA and/or transduced with adenovirus as described previously (24,25,37,38). After the designated incubation times, WST-1 reagent (Roche Applied Science) was added to the cells (final 1:10 dilution), and the cells were incubated along with the reagent for 30 min. The plates were then read against a blank using a microplate (ELISA) reader at 420 -480 nm with the reference wavelength Ͼ600 nm.

Statistical Analysis
Statistical differences between treatment groups were determined by either analysis of variance or a Newman-Keuls multiple comparison test. p values less than or equal to 0.05 were considered significant.
Author Contributions-N. T. V. and B. A. S. conceived and designed experimental studies, developed novel methodologies, acquired data, analyzed and interpreted data, and performed the initial writing of the manuscript. M. D. S., J. A. M., J. C. S., and M. M. G. acquired and analyzed RNA-seq (RNA sequencing) data, performed biostatistical analyses, verified all presented data, and provided technical support. A. Y. provided technical support and generated data for the initial MDA-7/IL-24 treatments. P. D., P. B. F., M. A. P., and C. E. C. acquired data, provided funding support, provided technical support, and performed data analysis/interpretation as related to the use and cell signaling of MDA-7/IL-24. C. E. C., M. A. P., and P. B. F. provided overall study supervision, provided funding support, generated hypotheses, conceived and designed the study, and edited and reviewed the manuscript.