Overexpression of the c-myc Oncogene Inhibits Nonsense-mediated RNA Decay in B Lymphocytes*

Background: The Myc transcription factor plays an important role in physiology and cancer. Results: High expression of Myc inhibits nonsense-mediated RNA decay (NMD), leading to the stabilization and up-regulation of mRNAs. Conclusion: The repression of NMD by Myc contributes to the Myc ability to regulate genes. Significance: This novel function of Myc may play an important role in Myc-dependent tumorigenesis. The Myc transcription factor plays a vital role in both normal cellular physiology and in many human cancers. We have recently demonstrated that nonsense-mediated RNA decay (NMD), a mechanism that rapidly degrades select mRNAs, is inhibited by the stress-induced phosphorylation of translation initiation factor eIF2α, and this inhibition stabilizes many transcripts necessary for tumorigenesis. Here, we demonstrate that NMD is inhibited by high Myc expression. We show that the phosphorylation of eIF2α, likely due to the ability of Myc to generate reactive oxygen species and augment endoplasmic reticulum stress, is necessary for the inhibition of NMD by Myc. The inhibition of NMD both stabilizes and up-regulates multiple Myc targets, suggesting that the inhibition of NMD may play an important role in the dynamic regulation of genes by Myc.

The c-myc gene plays an important role in normal cellular growth, proliferation, and differentiation, and myc dysregulation plays a causal role in many human malignancies. The Myc transcription factor performs these cellular functions by regulating gene expression through a variety of mechanism. For example, Myc transcriptionally up-regulates genes involved in cancer, including those involved in cell cycle activation, protein synthesis, and amino acid and carbohydrate metabolism (for review, see Refs. 1,2). Myc has also been demonstrated to repress a select group of transcripts (3). However, in contrast to the well described role of Myc in gene transcription, the role of Myc in mRNA stability has not been extensively explored.
Nonsense-mediated RNA decay (NMD) 2 is a well established mechanism to rapidly degrade mutated mRNAs responsible for many human genetic diseases (for review, see Ref. 4). NMD has also been implicated in tumorigenesis (for review, see Ref. 5). A systematic analysis of mutations in human genes has revealed that although most mutation in oncogenes are missense mutations, tumor suppressor genes exhibit a disproportionate number of NMD-provoking mutations (6), and although NMD has traditionally been thought of as a mechanism to protect an organism from deleterious dominant negative or gain-of-function effects of truncated proteins that arise from mutated transcripts, NMD has recently been found to regulate up to 10% of nonmutated transcripts, many of which play roles in tumorigenesis (7)(8)(9).
It has also been determined recently that NMD is a regulated process that can dynamically alter gene expression. NMD is inhibited by a variety of cellular stresses that commonly occur in the tumor microenvironment (7)(8)(9). This inhibition of NMD promotes tumorigenesis, likely in part by stabilizing several transcripts important for the cellular response to stress, including the transcription factor ATF-4 (7)(8)(9). Myc also promotes the metabolic adaptation to hypoxia and other stresses, in part by generating reactive oxygen species (ROS) which stabilizes the hypoxia-inducible transcription factor-1␣ (10 -12) (for review, see Ref. 13). Because ROS generation leads to the phosphorylation of translation factor eIF2␣ (14) and we have recently demonstrated that the phosphorylation of eIF2␣ inhibits NMD (8 -9), we hypothesized that oncogenic stress resulting from Myc dysregulation inhibits NMD. We therefore investigated whether Myc inhibits NMD, whether this occurs via phosphorylation of eIF2␣, and the significance of NMD inhibition on the stabilization and up-regulation of Myc targets. lines (CB30, CB33, JW20, DH20) were supplied by D. Araten (New York University). All other cell lines and growth conditions have been described previously (8,15). To suppress Myc expression P493-6 cells were treated with 1 g/ml tetracycline for 3 days; for moderate Myc expression 1 M estradiol was added for 8 h, or cells were washed three times and resuspended in tetracycline-deficient medium. ROS were diminished by adding 5 mM N-acetylcysteine daily 2 days prior to assessment. To assess RNA stability the RNA polymerase II inhibitor 5,6dichlorobenzimidazole-1-␤-D-ribofuranoside (DRB) was added at 100 g/ml to suppress transcription. RNA was then serially collected at the indicated times, and gene expression was determined on cDNA by quantitative PCR as detailed below. RNA expression in the absence of synthesis reflects RNA stability. All time points are referenced to a normalized time 0.
Plasmids and Virus Production and Infection-Lentiviruses were constructed expressing ␤-globin sequences (8) into a pCCL retrovirus backbone in a 3Ј to 5Ј orientation, so that introns were not removed during viral processing, and genomic globin sequences were integrated into the host genome (data not shown). Retroviruses and lentiviruses were generated in 293T cells, and target cells were infected and selected as described previously (8).
Immunoblotting and RNA Analysis-RNA isolation, cDNA generation, real time (quantitative) PCR, PCR (for XBP splicing), protein isolation, and immunoblots were performed using techniques, antibodies, and primers described previously (8,15) except primary antibody detection was assessed with fluorescent antibodies and an Odyssey infrared imager (LiCor). For mRNA quantitation, PCR primers were designed from adjoining exons. For pre-mRNA quantitation, primers were designed to include an intronic sequence. All mRNA expression was normalized to 18 S RNA. Primer sequences not reported (8) are available upon request.

RESULTS
Myc Inhibits NMD-To investigate the effect of Myc expression on NMD activity, we first utilized the P493 cell line which contains an inducible Myc expression system. In these cells tetracycline treatment tightly represses ectopic Myc protein expression and results in low endogenous Myc expression ( Fig.  1A) (16). This cell line also expresses an Epstein-Barr virus EBNA2 gene as an estrogen-activated construct (ER-EBNA2), and in the presence of estradiol-forced cell proliferation leads to intermediate expression of Myc. As a first step to assess NMD activity, we engineered these cells to express either a wild-type ␤-globin gene or a mutated ␤-globin gene that carries a single nucleotide substitution that leads to a premature termination codon (␤-globin PTC 39) and thus encodes a transcript that is degraded by NMD. NMD activity was then determined by inhibiting transcription with the RNA II polymerase inhibitor DRB and serially assessing globin transcript expression under conditions resulting in low Myc (tetracycline), high Myc (no tetracycline), and intermediate Myc levels (cells either grown without tetracycline for 8 h or grown in the presence of tetracycline and estradiol for 8 h) (Fig. 1A). Myc expression in P493 cells in the absence of tetracycline was equivalent to that seen in Burkitt lymphoma cell lines (data not shown).
The stability of wild-type ␤-globin mRNA was comparable under all three Myc expression levels (Fig. 1B). In contrast, the stability of the PTC 39 ␤-globin mRNA was dramatically increased in cells with high Myc expression and resembled that of the wild-type ␤-globin mRNA stability. Under the two conditions that led to moderate levels of Myc, the stability of the PTC 39 ␤-globin mRNA was slightly increased compared with cells with low Myc expression. When expressed as a ratio of stability in high Myc-expressing cells to stability in low Mycexpressing cells, the ␤-globin PTC 39 mRNA was ϳ4-fold stabilized (Fig. 1B, right). These data suggest that high Myc expression, as seen in many malignancies, can inhibit NMD.
Myc Inhibits NMD via Generation of ROS and Phosphorylation of eIF2␣-Because we recently demonstrated that NMD is inhibited by phosphorylation of eIF2␣, which occurs as a result of a variety of cellular stresses (9), we investigated the possibility that eIF2␣ phosphorylation is enhanced by high expression of Myc. eIF2␣ phosphorylation is mediated by a variety of kinases, including the endoplasmic reticulum (ER)-residing kinase PERK which is activated when unfolded proteins accumulate in the ER, a condition stimulated by the glycosylation inhibitor tunicamycin. Although there was only a minimal increase in base-line eIF2␣ phosphorylation in high Myc-expressing P493 cells, we noted a marked robust phosphorylation of eIF2␣ with tunicamycin treatment in those cells with elevated Myc in contrast to cells with low Myc (Fig. 2A). eIF2␣ phosphorylation promotes the translation of the stress response transcription factor ATF-4, and we observed a rapid induction of ATF-4 and its target CHOP when cells with elevated Myc were treated with tunicamycin compared with cells with low Myc expression treated with tunicamycin (Fig. 2B, top). Processing of the XBP-1 transcription factor by the ER-residing IRE1 endonuclease serves as a surrogate marker of ER stress, and splicing of XBP-1 also occurred dramatically in response to tunicamycin in cells expressing Myc highly, and much slower in cells with low Myc expression (Fig. 2B, bottom). Together, these data suggest that high Myc expression sensitizes cells to enhanced ER stress and phosphorylation of eIF2␣.
The generation of ROS leads to PERK-mediated phosphorylation of eIF2␣ (14). We assessed whether the ability of Myc to generate ROS (10 -11) plays an important role in its inhibition of NMD. As expected, cells expressing high levels of Myc exhibited a significant increase of ROS compared with cells express-ing lower levels of Myc, and this effect decreased when cells were cultured with the ROS scavenger N-acetylcysteine (Fig.  2C). Similarly, the concentration of intracellular H 2 O 2 was increased with elevated Myc expression, as assessed by oxidation of 10-acetyl-3,7-dihydroxyphenoxazine (Amplex Red) (Fig. 2D). N-Acetylcysteine did not change the stability of the wild-type ␤-globin mRNA, but the PTC 39 ␤-globin mRNA was no longer stabilized in high Myc cells in the presence of N-acetylcysteine (Fig. 2E).
To determine whether eIF2␣ phosphorylation is necessary for Myc inhibition of NMD (as it is for the inhibition of NMD in hypoxic or tunicamycin-treated cells) (8,9), we expressed Myc or a control retrovirus in eIF2␣ wild-type MEFs and in MEFs in which eIF2␣ cannot be phosphorylated due to the presence of mutated eIF2␣ alleles (eIF2␣ S51A MEFs) (Fig. 2F). These cells also expressed either wild-type or the PTC 39 ␤-globin construct. In wild-type eIF2␣ MEFs, Myc overexpression selectively stabilized the PTC 39 ␤-globin transcript ϳ2-fold compared with the control infected eIF2␣ wild-type cells (Fig. 2G). In contrast, Myc overexpression did not significantly alter the stabilization of the ␤-globin PTC 39 transcript in the eIF2␣ S51A MEFs, indicating that eIF2␣ phosphorylation is required for the inhibition of NMD by Myc. When Myc was overexpressed in wild-type PERK MEFs, we again noted stabilization of the ␤-globin PTC 39 transcript (Fig. 2H). Such stabilization was not observed in PERK Ϫ/Ϫ MEFs. Together, these data indicate that high Myc expression can promote eIF2␣ phosphorylation and inhibit NMD through a ROS-and PERK-dependent mechanism.
Myc Stabilizes Endogenous NMD Transcripts-Although NMD was originally thought to be responsible for degrading only mutated transcripts, a number of studies have identified nonmutated transcripts also degraded by NMD (7)(8)(9). Because the NMD-targeted PTC39 ␤-globin transcript is stabilized by high Myc expression, we next investigated whether endogenous NMD-targeted transcripts, including transcripts vital for the cellular response to stress, are stabilized by Myc. We found that the transcripts for ATF-4 and several other NMD targets were significantly stabilized in high Myc-expressing P493 cells compared with low Myc-expressing cells (Fig. 3A), thus confirming that NMD is inhibited by Myc.
Of 2100 transcripts found to be up-regulated by Myc in human B (P493) cells (17) and 694 transcripts stabilized by the pharmacological, molecular, and environmental inhibition of NMD in human osteosarcoma (U2OS) cells (9), 63 genes were common to both sets. Of note, this number of NMD targets also up-regulated by Myc is likely a low estimate because the data bases compared were derived from different cell lines and Myc targets are cell context-dependent (18). We arbitrarily chose eight of these transcripts and confirmed that all eight were stabilized with the addition of the translational inhibitor emetine, a well established method to inhibit NMD pharmacologically ( Fig. 3B and data not shown). Six of these transcripts also displayed increased stabilities in cells with either high or intermediate Myc expression compared with low Myc expression (Fig.  3B), indicating that elevated Myc expression can also suppress the NMD-mediated degradation of endogenous Myc targets.
The regulation of steady-state gene expression is a complex function of both synthesis and degradation, and we next explored whether the inhibition of NMD by Myc can contribute to the up-regulation of Myc targets. We assessed pre-mRNA (mRNAs containing introns) expression as an indicator of newly synthesized mRNA, and processed mRNA (mRNAs without introns) expression as an indicator of both transcription and post-transcriptional processes. When we examined four Myc targets that are not targeted by NMD, three genes displayed elevation of both pre-mRNAs and mRNAs to similar extents in response to high Myc expression, and none of the genes at the pre-mRNA or mRNA level was significantly affected by emetine (Fig. 3C). One Myc target, GPR54, was only minimally elevated at the pre-mRNA level despite a marked induction at the mRNA level. This transcript has not been identified as a NMD target and indeed was not induced by emetine, suggesting that another mechanism may play a role in its posttranscriptional up-regulation. In contrast, seven transcripts that have been reported to be up-regulated by both Myc and NMD demonstrated significant increases in steady-state mRNA expression with either high Myc expression or emetine treatment. As expected, emetine treatment alone did not increase transcription (i.e. increase the pre-mRNAs) of any of these genes. Although two of these transcripts (HSPA14, SNHG8) showed significant induction of pre-mRNA with high Myc expression, the rest showed minimal increases of pre-mRNA and/or mRNA induction out of proportion to pre-mRNA induction. Together, these data suggest that the inhibition of NMD by Myc plays a contributing role to the upregulation of a set of Myc targets.
eIF2␣ Is Phosphorylated and NMD Targets Stabilized in Cell Lines Derived from Burkitt Lymphoma-We next investigated whether NMD is also inhibited in cell lines derived from Burkitt lymphoma, a tumor marked by high Myc expression, compared with a variety of other B cell lymphoma cell lines with lower Myc expression (Fig. 4A and data not shown). Consistent with our model that the inhibition of NMD by Myc requires eIF2␣ phosphorylation, we observed a high correlation of eIF2␣ phosphorylation with Myc expression in these cell lines (Fig. 4A).
We then assessed the stability of several NMD targets in these cell lines. ATF-4 mRNA was significantly (p Ͻ 0.05) stabilized Ͼ2-fold in all Burkitt lymphoma cell lines compared with its stability in the low-Myc non-Burkitt cell lines (Fig. 4B). Similarly, the stabilities of five NMD-degraded mRNAs that were stabilized by Myc in P493 cells (Fig. 3B) were also significantly stabilized in the Burkitt lymphoma cell lines (Fig. 4C). In addition, published expression analyses of genes overexpressed in Burkitt lymphomas compared with reactive lymph nodes (19) showed a significant (p Ͻ 0.002) overlap with mRNAs we have reported previously to be targeted by NMD (Fig. 4D). Specifically, there was an almost 2-fold increase in NMD targets up-regulated in these cell lines compared with what would be expected randomly. Together, these data suggest that the inhibition of NMD noted with exogenous Myc overexpression may also be observed in human tumors with high Myc expression.

DISCUSSION
Myc is best known for affecting gene expression by directly transactivating and repressing genes. However, many genes upregulated by Myc are not direct targets, i.e. have Myc bound at their promoters (18). In addition, a recent study noted that many of the transcripts up-regulated by Myc are not induced at the transcriptional level, suggesting that post-transcriptional processes may play an important role in their up-regulation (20). Consistent with these findings, Myc has recently been appreciated to regulate microRNAs that can then affect mRNA stability (21). Our data indicate that an additional mechanism by which Myc regulates gene expression is by inhibiting NMD and stabilizing a select set of transcripts. This not only represents an important new aspect of Myc biology, but also increases the significance of NMD regulation, particularly in tumorigenesis.
It has long been appreciated that the expression levels between individual Myc targets vary. Although the reasons for this observation are complex and include the presence of cooperating transcription factors and epigenetic modifications, our data suggest that the inhibition of NMD can play an important role in augmenting the expression of Myc targets that are also targeted by NMD. An analysis of mRNAs stabilized by the inhibition of NMD (9) found a significant (p Ͻ 0.5 ϫ 10 Ϫ4 ) overlap with mRNAs bound by Myc (18), suggesting that even mRNAs transcriptionally activated by Myc may be stabilized by Myc inhibition of NMD for maximum induction (data not shown).
We found that although Myc stabilizes most of the NMD targets we examined, not all NMD targets are up-regulated in cells expressing high Myc. Indeed, many NMD targets (e.g. ATF-4) have not been defined as classic Myc targets, highlighting the complex regulation of Myc targets by several distinct and perhaps competing mechanisms. Consistent with this finding, we have found that many transcripts stabilized by NMD inhibition are not necessarily up-regulated with NMD inhibition (8,9). For example, the ATF-4 transcript is a bona fide NMD target, but even when it is stabilized by the hypoxic inhibition of NMD its expression is unchanged from its expression in normoxic cells. However, emphasizing the importance of NMD in regulating ATF-4, without the stabilization of the transcript by the inhibition of NMD, the ATF-4 transcript is downregulated in hypoxic cells (8).
The phosphorylation of eIF2␣ in response to cellular stress is a dynamic process and dependent on the coordinated actions of multiple eIF2␣ kinases and eIF2␣ phosphatases, some of which themselves activated by stress (for review, see Ref. 22). Although we did not note a dramatic base-line increase in basal, steady-state eIF2␣ phosphorylation with high expression of Myc in P493 cell, eIF2␣ phosphorylation correlated with Myc expression level in a variety of lymphoma cell lines. We also observed that high Myc expression renders cells more sensitive to ER stress and eIF2␣ phosphorylation in P493 cells. There are several potential mechanisms by which Myc could promote phosphorylation of eIF2␣ and thus inhibit NMD. First, Myc induces protein synthesis severalfold, which could lead to the accumulation of unfolded proteins in the ER, activation of PERK, and phosphorylation of eIF2␣ (23). Second, Myc promotes rapid proliferation which leads to replication stress and/or DNA damage that can promote eIF2␣ phosphorylation (for review, see Refs. 12,24). Because of the Myc ability to generate ROS we favor a third possibility, that direct generation of ROS by Myc could either promote unfolded proteins in the ER and/or activate ROS-responsive eIF2␣ kinases (14). Because PERK appears necessary for Myc-induced phosphorylation of eIF2␣, it is likely that unfolded ER proteins do accumulate with high Myc expression. ROS scavengers, such as N-acetylcysteine, could blunt several of these Myc phenotypes, and we are unable to distinguish between these possibilities at this time.
Although Myc is primarily known as a pro-proliferative oncogene, Myc also plays important roles in angiogenesis and tumor metabolism and enhances three-dimensional growth (for review, see Ref. 13). We have determined previously that the inhibition of NMD also stabilizes and up-regulates several stress response genes, improves cellular survival to ER stress, and augments three-dimensional tumor growth (8,9). The inhibition of NMD is therefore an additional mechanism by which Myc may promote tumorigenesis. NMD also degrades many mutated tumor suppressor transcripts, including p53 and BRCA1 (for review, see Ref. 5), and the stabilization of these transcripts could lead to truncated proteins, some of which exhibit dominant negative properties. Thus, the role Myc inhibition of NMD plays in tumorigenesis is potentially extensive and deserves further study.