Leptomycin B, an Inhibitor of the Nuclear Export Receptor CRM1, Inhibits COX-2 Expression*

Cyclooxygenase (COX)-2, the inducible prostaglandin synthase, is overexpressed in cancer and chronic inflammatory diseases. Post-transcriptional regulation ofCOX-2 mRNA is important in controlling the expression of the COX-2 gene. Here, we report that leptomycin B (LMB), a specific inhibitor of the nuclear export factor CRM1 potently inhibits the stabilization of COX-2 mRNA in MDA-MB-231 human mammary cancer cells. However, COX-2 promoter-driven reporter gene expression is not inhibited by LMB, suggesting that LMB acts at the post-transcriptional level. Subcellular fractionation experiments indicate that LMB inhibited the time-dependent export of COX-2 mRNA into the membrane-bound polysomal compartment at the endoplasmic reticulum. LMB suppressed COX-2 expression by interleukin-1β in HT-29 human colon cancer cells and in human umbilical vein endothelial cells but had no effect on COX-2 expression induced by Escherichia coli lipopolysaccharide in monocytic THP-1 cells. These data suggest that the nuclear export of COX-2 mRNA may be rate-liming in a cell-specific manner. LMB may be useful to control COX-2 expression in various human diseases in which COX-2 plays a pathogenetic role.

Two isoforms of COX have been identified and cloned in eukaryotic cells. COX-1, the constitutively expressed isoform, is involved in the maintenance of physiological functions. In contrast, COX-2 is inducible by extracellular stimuli such as tumor promoters, pro-inflammatory cytokines, mitogens, oncogenes, and growth factors in a variety of cells (1). Overexpression of COX-2 is observed in human cancers such as colon and breast (4). Gene deletion of COX-2 suppressed the developmental of intestinal polyps in APC gene deleted mice, suggesting that COX-2 expression is critical for tumorigenesis (5). In addition, overexpression of COX-2 in the mammary glands of transgenic mice results in enhanced tumorigenesis in multiparous mice (6). These data strongly suggest that unregulated expression of COX-2 is an important event in tumorgenesis. The control of COX-2 gene expression is regulated at the level of transcriptional and post-transcriptional mechanisms (1,3,7). However, the mechanisms by which COX-2 expression is controlled are not fully understood.
Leptomycin B (LMB) was originally discovered as a potent anti-fungal antibiotic from Streptomyces sp. (8). However, recent data showed that LMB causes G 1 cell cycle arrest in mammalian cells and is a potent anti-tumor agent against murine experimental tumors (9,10). Studies in Scizosaccharomyces pombe identified the cellular target of LMB as the CRM1 (chromosomal region maintenance)/exportin 1 protein (11), which is critical for the export of RNA and proteins containing a nuclear export sequence (NES). In the case of RNA export, CRM1 binds to ribonuclear proteins containing the NES motif (12)(13)(14)(15). For example, the export of Rev-response element-containing human immunodeficiency virus (HIV) RNA is inhibited by LMB (13). Studies from Steitz and co-workers (16) indicated that LMB disrupted the interaction of the RNA binding protein HuR with its protein partners, resulting in blockage of nuclear export of c-Fos mRNA and perhaps other immediate-early gene mRNAs that bind to HuR (16,17). We and others have recently showed that the COX-2 mRNA binds specifically to HuR at the 3Ј-untranslated region (18). 2 Based on these observations, we speculated that LMB might regulate COX-2 gene expression by controlling nuclear export of its mRNA.
In the present study, we investigated whether LMB can regulate the expression of COX-2 by modulating the nuclear export of COX-2 mRNA.
Preparation of Whole Cell Lysates-Cells were washed twice with ice-cold phosphate-buffered saline (PBS) and lysed in modified RIPA lysis buffer (50 mM Tris-Cl (pH 7.4), 150 mM NaCl, 1% Triton X-100, 1% Nonidet P-40, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixtures (1ϫ, Roche Molecular Biochemicals)). After 10-min incubation on ice, cells were collected by cell scraper and cen-* This work was supported by National Institutes of Health Grants HL49094 and CA95181 and an Established Investigator Grant from the American Heart Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Northern Blot Analysis-Total RNA was isolated using the RNA-STAT60 reagent (Tel-test) according to the instructions provided by the manufacturer. Northern blot analysis was done as described previously (20).
Luciferase Assay-MDA-MB-231 cells grown in 24-well plates were transfected for 4 h with luciferase DNA construct containing a COX-2 promoter (21,22). After 48-h post-transfection, cells were grown for 6 h in the presence or absence of serum along with or without LMB. Cell lysates were assayed with the luciferase assay system (Promega) using a TD 20/20 luminometer (Turner Designs Instruments, Sunnyvale, CA).
RNase Protection Assay-The RNase protection assay was performed according to the instructions provided by the manufacturer (Ambion) with some modifications. Briefly, RNA (10 g) was hybridized with 32 P-labeled antisense riboprobes of human COX-2 (500 bp) and human GAPDH (100 bp). The hybridization reaction was then treated with RNases A and T to remove unhybridized probes. The protected and hybridized mRNA was resolved in 6% polyacrylamide gel containing 8 M urea. The gel was dried and exposed on a Kodak film at Ϫ80°C. The bands were visualized by autoradiography or PhosphoImager (Amersham Biosciences).

Leptomycin B Potently Inhibits the Expression of COX-2 mRNA and Protein Induced by Serum Withdrawal in MDA-
MB-231 Cells-Previously we showed that COX-2 mRNA and protein were induced by serum withdrawal in MDA-MB-231 human breast carcinoma cells (20). We recently found that the interaction of COX-2 mRNA with the RNA-binding protein HuR is critical for stabilization of COX-2 mRNA. 2 Steitz and co-workers (16) have shown that one of the functions of HuR is to facilitate the nuclear export of immediate-early mRNAs into the cytoplasm. Furthermore, they showed that HuR ligands bind to CRM1 and that HuR-binding mRNA export is inhibited by LMB, a specific inhibitor of CRM1-dependent nuclear export (16). We determined the effect of LMB to see if COX-2 mRNA export is critical in the regulation of expression. Northern and Western analysis showed that LMB potently and effectively suppressed the expression of COX-2 mRNA and protein in a concentration-dependent manner (Fig. 1, A and B). Extremely low concentrations of LMB, 10 ng/ml (18 nM), were sufficient to completely inhibit the serum withdrawal-induced COX-2 mRNA (Fig. 1A).
Leptomycin B Does Not Inhibit the Transcription of the COX-2 Gene-Serum withdrawal-induced expression of COX-2 gene in MDA-MB-231 cells were previously characterized to require the stabilization of the COX-2 mRNA via signaling of the p38 stress-activated kinase pathway (20). Nuclear run-on analysis did not indicate changes in transcription rate of the COX-2 gene (20). Thus, the effect of LMB most likely is at the post-transcriptional level. To confirm this point, we measured the expression of COX-2 promoter-driven luciferase expression in transiently transfected cells. As shown in Fig. 2, LMB had no effect on COX-2 transcription. Neither serum addition nor withdrawal modulated COX-2 transcription, suggesting that LMB inhibits COX-2 expression by an effect at the post-transcriptional level.
Subcellular Localization of COX-2 mRNA-To better define the subcellular localization of the COX-2 mRNA, we conducted biochemical fractionation and in situ hybridization experiments. Cells were extracted with Nonidet P-40 buffer to isolate the cytosolic fraction. The remaining nuclear/RER pellet was extracted with EDTA to remove the membrane-bound polysomal RNA. The EDTA-resistant fraction was extracted with high salt to isolate nuclear RNA. As shown in Fig. 3A, U6 RNA was found predominantly in the nuclear fraction and GAPDH RNA was found in the cytosolic and RER compartments. This validates the biochemical extraction scheme that was employed. As expected, COX-2 mRNA was induced only when cells were serum-deprived. In contrast to the fractionation of U6 and GAPDH mRNAs, COX-2 mRNA showed a predominantly RER/ polysomal distribution, consistent with the fact that COX-2 is a protein located in the lumen of the ER. A significant proportion of the COX-2 mRNA was also found in the nuclear fraction. Since transcription rates are similar in 10 and 0% serum treated cells, these data suggest that both nuclear and ERassociated COX-2 mRNA species accumulate under serumdeprivation induced condition.
To further validate the biochemical fractionation data, we conducted an in situ hybridization experiment with COX-2 and GAPDH probes to detect the respective transcripts. As shown in Fig. 3B, GAPDH signals were diffusely localized throughout the cytoplasm of MDA-MB-231 cells regardless of treatment of the cells with 10% or 0% serum. However, COX-2 mRNA showed a dramatic change when cells were treated with 0% serum for 24 h. A strong signal around the nuclear membrane and ER was observed only under 0% serum treatment, whereas background signals were seen in cells cultured with 10% serum. These data confirm the biochemical fractionation experiments that COX-2 mRNA accumulates in membrane-bound polysomal (ER-associated) and nuclear compartments. Predominant accumulation of the COX-2 mRNA at the nuclear membrane (both EDTA-extractable and high salt-extractable) suggest that nuclear transport of COX-2 mRNA may be rate-limiting.
Effect of LMB Treatment on Subcellular Localization of COX-2 mRNA-To determine whether the LMB effect on the suppression of the COX-2 mRNA occurs in the nuclear compartment or ER compartment or both, MDA-MB-231 cells were treated with LMB, and at various times thereafter COX-2 and GAPDH mRNAs were analyzed by biochemical fractionation followed by RNase protection experiments. As shown in Fig.  4A, 24-h serum starvation induced significant levels of COX-2 mRNA in both ER-associated and nuclear compartments (lanes  3 and 4). Continued treatment with 0% FBS further induced COX-2 mRNA, and at 24 h thereafter, maximal levels of COX-2 mRNA accumulated at the membrane-bound ER fraction, whereas a smaller fraction was found in the nucleus. Treatment with LMB from 0 -24 h resulted in pronounced suppres- sion of COX-2 mRNA (but not the GAPDH mRNA), especially at the 24-h time point (lanes 15 and 16). Interestingly, both membrane-bound ER and nuclear COX-2 mRNA species were suppressed.
To further confirm the time-dependent effect of LMB, cells were treated with 0% FBS at time 0 h and subsequently treated or not with LMB from 8 -24 h (Fig. 4B). As it can be seen, LMB treatment from 8 -24 h suppressed both nuclear and RERlocalized COX-2 transcript levels. However, membrane-bound ER-localized COX-2 transcripts were profoundly suppressed by LMB. Although nuclear COX-2 mRNA was suppressed by LMB, it was not suppressed to the same degree as the cytosolic counterpart. These data suggest that COX-2 mRNAs accumulate at the nuclear pore complex in MDA-MB-231 cells and that inhibition of the CRM1 pathway with LMB results in degradation of COX-2 mRNA in both the intranuclear and extranuclear (membrane-bound ER) compartments of the nuclear pore complex. Indeed, time course analysis of COX-2 mRNA in membrane-bound ER and nuclear compartments after LMB treatment indicates that COX-2 mRNAs in both compartments decay with similar rates (data not shown). Thus, LMB appears to decouple the COX-2 mRNA from the nuclear export pathway and target it for degradation in both cytosolic and membranebound ER compartments.
Cell-type and Stimuli-specific Inhibition of COX-2 Expression by LMB-We also tested whether LMB can inhibit the expression of COX-2 by known COX-2 inducers such as IL-1␤ or LPS in other cell types including HT-29, a human colon carcinoma cell line, HUVEC, or THP1, a human monocytic cell line. The data showed that the expression of COX-2 induced by IL-1␤ in HT-29 cells is inhibited by LMB in a concentrationdependent manner (Fig. 5A). The COX-2 expression induced by IL-1 in HUVEC was also partially inhibited by LMB at 10 ng/ml concentration (Fig. 5B). However, LPS-stimulated COX-2 expression in THP1 cells was not inhibited by LMB at even 20 ng/ml (Fig. 5C). COX-2 regulation in various cell systems is regulated at both transcriptional as well as post-transcriptional levels (1,3,7). Since LMB is a specific regulator of CRM1-dependent nuclear export, it is likely that it would only block expression of genes that are controlled by CRM1-dependent export of mRNA. These data suggest that LMB inhibits COX-2 expression in a cell-type and agonist-specific manner.
In summary the data presented here demonstrate for the first time that LMB potently inhibits COX-2 expression induced by serum withdrawal in a human breast cancer cell line MDA-MB-231. Furthermore, our data suggest a possible mechanism by which LMB inhibits COX-2 expression by showing that LMB blocks nuclear export of COX-2 mRNA, which may be an important control point of COX-2 gene expression. Finally we provide some evidence that LMB may be a new and potent inhibitor of COX-2 expression with a cell-and/or agonist-type specificity.