The human embryonic lethal abnormal vision (ELAV)-like protein HuR is a messenger RNA (mRNA)-binding protein that controls the stability of certain transcripts, including cyclooxygenase2 (COX-2).
The human embryonic lethal abnormal vision (ELAV)-like protein HuR is a messenger RNA (mRNA)-binding protein that controls the stability of certain transcripts, including cyclooxygenase2 (COX-2).
To investigate a possible contribution of dysregulation of mRNA stability to the progression of cancer and to COX-2 over expression in mesothelioma, the authors studied expression of COX-2 and HuR in 5 mesothelioma cell lines (MSTO, NCI, Ist-Mes1, Ist-Mes2, and MPP89) and in a group of 29 human mesothelioma specimens that were characterized previously for COX-2 expression.
All 5 cell lines expressed HuR, whereas COX-2 was not detectable in MSTO or NCI cells. Treatment with cytokines induced a shift in systolic HuR protein levels in MPP89 and Ist-Mes2 cells that was accompanied by an increase in the expression of COX-2 mRNA and protein. In Ist-Mes1 cells, cytokine stimulation did not cause the passage of HuR from nucleus to cytoplasm, and the synthesis of COX-2 did not increase. In tumor tissues, immunohistochemistry revealed a positive, statistically significant correlation between high COX-2 expression and cytoplasmic localization of HuR (P = .016). Moreover, on univariate analysis, overall survival was found to be influenced strongly by cytoplasmic HuR localization (P = .004).
The current results suggested that HuR plays a role in tumor progression in mesothelioma and that COX-2 may be a target of its activity in neoplastic cells. Together, these observations indicate that strategies aiming toward the modulation of HuR may have a potential clinical benefit in mesothelioma. Cancer 2008. © 2008 American Cancer Society.
There is considerable evidence that arachidonic acid metabolites, particularly prostaglandins, participate both in normal growth responses and in aberrant growth, including carcinogenesis.1 Cyclooxygenase (COX) plays a role in the conversion of arachidonic acidto several metabolites with the isoform COX-2, leading to preferential formation of prostaglandin E2 (PGE-2). COX-2 has been implicated in tumor proliferation, increased angiogenesis, tumor invasiveness, and decreased cell-mediated immunity.2–6 Physiologically, COX-2 expression is regulated tightly and is highly inducible by inflammatory stimuli.7 In many malignant tumors, an increased expression of COX-2 has been described and is related to poor prognosis. Over expression of the inducible isoform COX-2 occurs in malignant mesothelioma (MM), and a connection between COX-2 expression and human MM has been established with the demonstration that nonsteroidal anti-inflammatory drugs specific for COX-2 have antiproliferative effects.8 Several studies examined COX-2 levels in resected MM specimens and correlated levels of COX-2 with clinicopathologic prognostic factors and determined that high COX-2 expression was correlated with poor survival.9, 10
The development of MM is associated in most patients with a history of asbestos exposure.11 Research has demonstrated that asbestos exposure generates reactive oxygen species and activates macrophages and other cell types to produce these compounds as well as cytokines and growth factors.12 Furthermore, the deposition of insoluble amphibole fibers results in a chronic inflammatory state in exposed individuals.13 Even if it has been proposed that alteration of the cell cycle-regulatory proteins could play a role in the progression of this deadly disease,14, 15 the exact mechanism by which asbestos induces MM still has not been characterized completely.16 The existence of inflammation has been associated with up-regulation of the inducible COX-217 and is associated with an increased risk of cancer.18 Molecular events leading to over expression of the COX-2 protein in MM have not been characterized definitively.
Regulation of gene expression by post-transcriptional modification of messenger RNA (mRNA) stability and translation is an important mechanism used in the adaptation of cells during inflammatory responses as well as in the control of cellular growth.19 AU-rich elements (AREs) in the 3′-untranslated region (3′UTR) of some mRNAs are involved in regulating the degradation of mRNAs and provide an effective way to control protein expression by regulating mRNA half-life as well as translation. An ARE has been identified within the 3′UTR of COX-2 mRNA that confers post-transcriptional regulation of expression by controlling both mRNA decay and mRNA protein translation.20, 21 A group of proteins capable of binding to AREs to control mRNA stability has been identified. Among these proteins, the human family of human embryonic lethal abnormal vision (ELAV)-like proteins consists of 4 members (Hel-N1/HuB, HuC, HuD, and HuR) that are involved in post-transcriptional regulation of mRNA turnover and mRNA stability.22 These human proteins are highly conserved and show homology to the Drosophila nuclear protein ELAV. Three of these proteins (Hel-N1/HuB, HuC, and HuD) are expressed mainly in neuronal tissue. In contrast, the fourth protein, HuR, is expressed in many cell types and stabilizes cellular mRNAs that contain AREs in their 3′UTR. Several studies have demonstrated that HuR binds to the COX-2 adenine- and uridine-rich element and stabilizes COX-2 mRNA, leading to increased expression of COX-2 protein.23–26 In cancer cells, it was demonstrated previously that altered post-transcriptional regulation of COX-2 is mediated by increased cytoplasmic mRNA binding of the mRNA stability factor HuR.23 Under normal conditions, HuR is localized primarily in the nucleus. However, in response to cellular signaling, HuR is translocated rapidly to the cytoplasm, where it influences mRNA stabilization and translational efficiency.27 It is noteworthy that the over expression of HuR in different cancer types, and particularly in its subcellular localization, recently was correlated with tumor aggressiveness.28, 29 On the basis of this finding, we sought to determine whether, in MM cells, changes in cytoplasmic trafficking of this factor could account for the observed COX-2 mRNA stabilization and protein expression promoted by cytokine mix. Drawing from this background, we decided to investigate COX-2 and HuR expression in several MM cell lines and in a group of well characterized human MM specimens.
The human MM cell lines MSTO-211H and NCI-H2452 were obtained from the American Type Culture Collection (ATCC) (Rockville, Md). Cells were cultured as monolayers in flasks using ATCC complete growth medium in a humidified atmosphere containing 5% carbon dioxide at 37°C. The human MM cell lines Ist-Mes1, Ist-Mes2, and MPP89 were obtained from Genova Institute Culture Collection. Ist-Mes1 and Ist-Mes2 cells were cultured in Dulbecco Modified Eagle Medium with piruvate supplemented with 10% fetal bovine serum (FBS), glutamine (2 mM), 1% nonessential amino acids, and antibiotics (0.02 IU/mL-1 penicillin and 0.02 mg.mL-1 streptomycin); and the established MM cell line, MPP89, was maintained in Hamm F10 medium with 15% FBS, and supplemented with glutamine (2 mM) and antibiotics (0.02 IU/mL-1 penicillin and 0.02 mg/mL-1 streptomycin) in a humidified atmosphere that contained 5% carbon dioxide at 37°C.
Subconfluent MM cells were stimulated with cytomix (14 ng/mL tumor necrosis factor alpha [Alexis Biochemical, Lausen, Switzerland], 1 ng/μL interleukin-1 beta [Alexis Biochemical], 27 ng/mL human interferon gamma [PBL Biochemical Laboratories, Piscataway, NJ], and 160 ng lipopolysaccharide [Alexis Biochemical]) for 3 hours. Then, supernatant fluids were harvested and frozen at −80° for PGE-2 detection, and the cells were removed by trypsin and divided into 2 identical sides for protein and mRNA extraction. Controls were treated in the same manner without cytomix. Experiments in each MM cell line were repeated 3 times.
PGE-2 levels were detected in medium from cell culture by using the Correlate-EIA high-sensitivity prostaglandin E2 enzyme immunoassay kit from Assay Designs (Ann Arbor, Mich) according to manufacturer's instructions. Each determination was made in triplicate.
MM cells were lysed in ice-cold lysis buffer (20 mM Tris [pH 8.0], 1% NP40, 10% glycerol, 137 mM NaCl, 10 mM ethylenediamine tetracetic acid [EDTA], and inhibitor of protease) for 20 minutes followed by centrifugation at 4°C for 15 minutes. Cytoplasmic lysates were obtained by resuspending cell pellets in ice-cold hypotonic lysis buffer (10 mM N-[2-hydroxethyl] piperazine-NA-2-ethanesulfonic acid [HEPES] [pH 7.9], 1.5 mM MgCl2, 0.5 mM dithiothreitol, 50 mM NaF, and 10 mM KCl) containing protease inhibitors and incubating on ice for 10 minutes. Cells were lysed by 3 passages through a 27-gauge needle, and cytoplasmic extracts were separated from nuclei by centrifugation at 14,000 revolutions per minute (rpm) for 10 seconds. Nuclear lysates were obtained by resuspending pellets in ice-cold lysis buffer (10 mM HEPES [pH 7.9], 1.5 mM MgCl2, 0.2 mM EDTA, 25% glycerol, and 420 mM NaCl) containing protease inhibitors and incubating on ice for 20 minutes. After centrifugation at 14,000 rpm for 5 minutes, the nuclear extracts were collected. Protein concentration was measured using Bradford Protein Assay (Bio-Rad Laboratories; Hercules, Calif).
Lysates were denatured and separated on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels. After electrophoresis, proteins were transferred to polyvinylidene difluoride membranes and probed with antibodies against COX-2 (COX-2 monoclonal antibody: Cayman Chemical, Ann Arbor, Mich), HuR (3A2; Santa Cruz Biotechnology, Santa Cruz, Calif), monoclonal anti-γ–tubulin and actin (Sigma Chemical Company, St. Louis, Mo). and anti-α–tubulin mouse antibody (Calbiochem, Darmstadt, Germany). Goat antimouse immunoglobulin G horseradish peroxidase-conjugated secondary antibodies (Bio-Rad Laboratories) were used. Then, the blots were reacted with ECL Western blotting detection reagents (Amersham Pharmacia, Uppsala, Sweden) according to the manufacturer's instructions. Ovine COX-2 standard (Cayman Chemical) was used as a positive control. The expression of COX-2 was assayed in 80 μg of total lysates, and HuR was detected in 10 μg of total, cytoplasmic, and nuclear lysates. The appropriate separation of cytoplasmic and nuclear fraction was demonstrated by labeling for α-tubulin.
Total RNA was prepared from cultured MSTO-211H, NCI-H2452, Ist-Mes1, Ist-Mes2, and MPP89 cells using TRIzol Reagent (Invitrogen Life Technologies, Paisley, UK) according to the manufacturer's protocols. Reverse transcription of RNA for first-strand combinational DNA (cDNA) synthesis was performed using 4 μg total RNA, 0.5 μg oligo(dT)12-18 primer (Invitrogen Life Technologies), and 10 mM dinucleotide triphosphate mix in a final volume of 12 μL. The reaction was incubated at 70°C for 10 minutes and immediately chilled on ice. Primer extension was then performed 10 minutes at temperature room and at 42°C for 2 minutes after addition of First-Strand Buffer, 10 mM dithiothreitol, and 40 U RNase OUT Recombinant Ribonuclease Inhibitor (Invitrogen Life Technologies) in a final volume of 19 μL. Next, 1 μL (200 U) SuperScript II Reverse Transcriptase (Invitrogen Life Technologies) was added, and the mixture was incubated at 42°C for 50 minutes. Finally, the reaction was inactivated by heating at 70°C for 10minutes, and cDNA was stored at −20°C.
Real-time polymerase chain reaction (PCR) analysis was conducted in a volume of 25 μL containing 40 ng cDNA (1:100 dilution of reverse transcriptase mixture), 1.25 μL COX-2 primer, and 12.5 μL TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, Calif) in the following sequence: 2 minutes at 50°C, denaturation for 10 minutes at 95°C followed by 40 cycles of the amplification step at 95° for 15 seconds (denaturation), and then at 60°C for 60 seconds (annealing/extension) in 96-well plates with the ABI PRISM 7000 sequence Detection System (Applied Biosystems). Real-time PCR for the endogenous control glyceraldehyde-3-phosphate dehydrogenase (GADPH) was carried out under the same conditions using a GADPH Assay on Demand (Applied Biosystems). A standard curve for the COX-2 gene was constructed using serial dilutions (200 ng, 40 ng, 8 ng, and 1.6 ng) of a pool of cDNA from the cell lines MSTO, NCI, Ist-Mes1, Ist-Mes2, and MPP89. Results were analyzed using the Applied Biosystems analysis software, and expression levels were calculated from a linear regression analysis of the standard curve. Results are given as COX-2 expression versus GADPH expression (COX-2 relative expression) to correct for differences in the quantity of cDNA used in the PCR reaction. All real-time PCR reactions for each sample were performed in triplicate.
All patients were treated at the Second University of Naples between 1980 and 1996. Clinical data were obtained by retrospective chart review. Survival was determined from the date of initial surgery. Indeed, surgery/biopsy was the first step in diagnosis for all patients. Consequently, surgery/biopsy and diagnosis overlapped. Follow-up was available for all patients. Two patients who died of causes other than mesothelioma during follow-up were excluded from the study. All patients underwent at least cytoreductive surgery, and 13 patients also received radiotherapy or chemotherapy. A clear history of asbestos exposure was reported only for 3 patients. This does not mean that all other patients were not exposed to asbestos. In fact, all patients were living in Campania, an Italian region where, unfortunately, asbestos contamination is very high. Tissues from 29 MM specimens (16 epithelioid, 6 sarcomatoid, and 7 mixed mesotheliomas) that were obtained from open biopsies or pleurectomies were collected and fixed in 10% formalin before they were embedded in paraffin.
The formalin-fixed, paraffin-embedded samples were sectioned at 5 μm and stained with hematoxylin and eosin. The histologic diagnosis was reexamined by a pathologist (A.B.) according to the 1981 World Health Organization criteria. In addition, the most representative blocks were selected to be cut into new 5-μm-thick sections for immunohistochemical studies.
All 29 tumors were assessed by immunohistochemistry for the presence of HuR. Sections from each specimen were cut at 5 μm, mounted on glass, and dried overnight at 37°C. All sections were then deparaffinized in xylene, rehydrated through a graded alcohol series, and washed in phosphate-buffered saline, which was used for all subsequent washes and for antibody dilution. Endogenous peroxidase activity was blocked by 5% hydrogen peroxide. The mouse monoclonal antibody for HuR (3A2; Santa Cruz Biotechnology Inc.) was applied at 4°C for 12 hours at 1:1000 dilution after antigen retrieval in citrate buffer in a pressure cooker for 5 minutes. The optimal working dilution was defined on the basis of titration experiments. Then, the sections were immunostained with the streptavidin-biotin system (Dako, Carpenteria, Calif) using diaminobenzidine as the final chromogen and hematoxylin as the nuclear counterstain. Negative controls for each tissue section were prepared by leaving out the primary antibody. A suitable positive control was run with each set of slides. All samples were processed under the same conditions. The expression of HuR was evaluated as absent, nuclear, or nuclear and cytoplasmic. The experimental conditions for the immunohistochemical analysis of COX-2, have been described elsewhere.10
The Fisher exact test was used to assess the relation between ordinal data (correlation matrix between immunostaining parameters). A univariate overall survival analysis was conducted for each prognostic variable according to the Kaplan-Meier method.30 The terminal event was death attributable to cancer. The statistical significance of the differences in survival distribution among the prognostic groups was evaluated by the log-rank test.31 P values <.05 were regarded as statistically significant in 2-tailed tests. SPSS software (version 10.00; SPSS Inc., Chicago, Ill) was used for statistical analysis.
Figure 1 shows the levels of COX-2 and HuR expression in 5 mesothelioma cell lines by Western blot analysis. HuR was expressed in all cell lines, whereas COX-2 protein was detectable only in Ist-Mes1, Ist-Mes2, and MPP89 cells. To analyze the effects of cytokine stimulation, all 5 cell lines were incubated for 3 hours with cytomix and the effect of stimulation was evaluated by measuring the increased COX-2 mRNA in the treated cells compared with the untreated cells. Whereas, in MSTO, NCI, and Ist-Mes1 cells, no differences in COX-2 mRNA were observed; in MPP89 and Ist-Mes2 cells, we observed increases of 5-fold (P = .0003) and 6-fold (P = .0001), respectively, compared with COX-2 mRNA controls (Fig. 2A).
Figure 2B shows that the increased expression of COX-2 protein in MPP89 and Ist-Mes2 cells after stimulation with cytokines was accompanied by a clear translocation of HuR protein from the nucleus to the cytoplasm that was not associated with increased expression of HuR protein in the total lysates. A different situation was observed in Ist-Mes1 cells, which, after stimulation with cytokines, did not display an increase in COX-2, and this corresponded with the failure of HuR to translocate from the nucleus to the cytoplasm. In MSTO and NCI cells, neither HuR translocation nor increased COX-2 expression was noted after stimulation with cytomix compared with controls.
Finally, we evaluated COX-2 activity by determining PGE-2 expression in the same cells. Figure 2C shows that the MPP89 and Ist-Mes2 cells incubated with cytomix, respectively, produced 2-fold (P = .03) and 1.5-fold (P = .01) more PGE-2 than controls. In stimulated Ist-Mes1, MSTO, and NCI cells, we did not observed any increase in COX-2 activity compared with controls. These results indicate that, in cells that are sensitive to cytokine stimulation, the expression of COX-2 increases only if there is translocation of HuR from the nucleus to the cytoplasm and that this contributes to COX-2 mRNA stabilization and protein synthesis.
Table 1 presents characteristics of the patients who were enrolled in this study and summarizes the results from immunohistochemical analysis of the 29 mesothelioma specimens. Histologically, tumor specimens included 16 epithelioid, 6 sarcomatoid, and 7 mixed mesotheliomas. HuR staining was either negative, or nuclear, or nuclear and cytoplasmic. Figure 3 presents some samples of typical immunohistochemical staining for HuR. COX-2 immunostaining always was cytoplasmic, as previously reported elsewhere.10
|Patient No.||Histology||COX-2 Score*||HuR Score†||Survival, mo|
By using a rank correlation matrix, a positive, statistically significant correlation was recorded between high COX-2 expression levels and cytoplasmic localization of HuR (P = .016). Conversely, no correlation was recorded between immunohistochemical parameters and mesothelioma histology type. These results are summarized in Table 2.
On univariate analysis, overall survival appeared to be influenced by HuR expression. The median survival of patients who had absent or nuclear HuR expression was longer than the median survival of patients who had cytoplasmic HuR expression (P = .004). Indeed, when we statistically analyzed the immunohistochemical specimens by determining the percentage of positive cells, we did not produce any significant results, most likely because of the small number of patients enrolled. The correlation between high COX-2 expression and poor survival has been described elsewhere.10 It is interesting to note that the histologic tumor type did not influence overall survival in our patient population. These data are summarized in Table 3. Figure 4 depicts Kaplan-Meier survival plots for all patients and demonstrates a statistically significant association between cytoplasmic expression of HuR and poor outcome (P = .004). It is important to note that chemotherapy and radiotherapy did not appear to have any impact on overall survival in univariate analysis, as depicted in Table 4.
|Variable||Median Survival (95% CI), Weeks||P|
|Variable||Median Survival (95% CI), Weeks|
MM is one of the most lethal human tumors, and its incidence is expected to increase in Western countries within the next 20 years.32 The prognosis for patients with MM is poor, and they have a median survival of few months whether they are treated or untreated.33, 34 Recent randomized studies on the treatment of mesothelioma with combined chemotherapy demonstrated a survival benefit when a combination of cisplatin and antifolate drugs was used35, 36 and when surgery was supplemented by postoperative radiotherapy in patients who underwent incomplete resection.37 Unfortunately, none of those forms of treatment had a significant impact on the progression and outcome of mesothelioma, and new therapeutic approaches must be investigated for a more successful treatment of this disease.
In this scenario, the RNA-binding protein HuR may represent a suitable candidate as an alternative therapeutic strategy for this deadly neoplasm. Indeed, mutated HuR has never been identified in cancer, but it influences the expression of target mRNAs that codify for genes fundamental in the acquisition of a tumor phenotype, such as the genes involved in cell cycle and apoptosis regulation, in angiogenesis and migration, and in evasion of immune recognition.38 HuR is present predominantly in the nucleus of unstimulated cells, but it can be exported to the cytoplasm in response to a variety of agents, notably, proliferative and stressful signals.27 Several stress-activated signaling pathways have been reported that modulate the cytoplasmic abundance of HuR and, thus, its RNA-binding function (for a review of the literature, see Lopez de Silanes et al38).
It is noteworthy that HuR has been identified as an RNA-binding protein that recognizes COX-2 AREs and exerts effects on COX-2 expression by controlling rapid mRNA decay and translational efficiency.23–26 Because recent results indicate that COX-2 molecules may be involved in the progression of mesothelioma and may be potential targets for therapy,39–42 we decided to investigate HuR and its regulation of COX-2 in mesothelioma cells. To the best of our knowledge, this is the first study investigating HuR in mesothelioma. We observed that HuR was expressed in all the 5 MM cell lines tested, whereas COX-2 was detectable only in 3 of 5 cell lines. Cytokine treatment of these cells mobilized HuR from the nucleus to the cytoplasm and produced consequent COX-2 mRNA stabilization and augmented protein expression. Our findings strongly argue for a role of HuR in the stress-mediated stabilization of COX-2 mRNA. It also is important to note that the greater mRNA COX-2 levels in MPP89 cells (compared with protein expression in the absence of stimulation with cytokines) suggest that COX2 mRNA in this cell line may be regulated after transcription. After cytokine stimulation followed by HuR translocation from the nucleus to the cytoplasm, the increased synthesis of mRNA COX-2 leads to an increase in metabolically active protein, as demonstrated by the levels of PGE-2. Remarkably, the Ist-Mes1 cell line has very stable mRNA, it is not subject to stimulation with cytokine and protein expression of COX2, and its product PGE-2 remain unchanged. Although the regulation of mRNA stability appears to be the most important regulatory step for COX-2 expression, several studies have reported that other mechanisms, such as transcriptional control or hypermethylation,20 also are involved in the regulation of COX-2 expression. We believe that the finding that, in some MM cell lines (MSTO, NCI and Ist-Mes1), there is a deteriorating increase in COX-2 expression after cytokine stimulation may be explained best by these additional regulatory mechanisms for COX-2 expression. Nevertheless, there are several other potential mechanisms for cytokines to affect COX-2 expression (such as activation of nuclear factor-κB); therefore, the finding of a correlation between HuR localization and COX-2 expression across cell lines will need further functional studies to support a mechanistic correlation.
To investigate whether these in vitro observations may have some clinical relevance, we decided to analyze HuR expression in a group of samples from patients with MM that already had been characterized for COX-2 expression.10 We observed that HuR was expressed in the great majority of MM samples as a nuclear protein; moreover, in a group of samples, we also detected cytoplasmic HuR expression. It is noteworthy that the cytoplasmic expression of HuR was correlated significantly with high COX-2 expression and with poor survival. These observations strongly suggest that HuR plays a role in tumor progression in MM and that COX-2 may be a target of its activity in the neoplastic cells. It must be noted, however, that the data described here are retrospective and essentially observational in nature; therefore, they cannot explain the functional mechanisms by which COX-2 actually promotes tumor growth. Therefore, additional studies performed with larger, independent groups of patients are needed to decipher the interactions between HuR, and COX-2 expression in MM.
In conclusion, to our knowledge, this is the first study of the relation between COX-2, HuR, and survival in patients with MM. The data obtained strongly suggest that cytoplasmic expression of HuR protein may be part of a regulatory pathway that controls the mRNA stability of several biologically important tumor targets and that, among these targets, COX-2 also could be considered. Therefore, we propose that HuR may be a promising predictive marker and a possible target for molecular tumor therapy for patients with MM. Considering the finding that at least some COX-2 inhibitors have significant side effects related to cardiovascular toxicity,42, 43 targeting HuR may be an important alternative therapeutic strategy for patients with MM who over express COX-2. Further studies urgently are required both at the molecular level and at the clinical level to confirm these observations and eventually to propose HuR as a concrete target for MM therapy.