Malignant Pleural Mesothelioma–Targeted CREBBP/EP300 Inhibitory Protein 1 Promoter System for Gene Therapy and Virotherapy

Gene therapy and virotherapy are one of the approaches used to treat malignant pleural mesothelioma. To improve the efficiency of targeting malignant mesothelioma cells, we designed a novel system using the promoter of the CREBBP/EP300 inhibitory protein 1 (CRI1), a gene specifically expressed in malignant pleural mesothelioma. Four tandem repeats of the CRI1 promoter (CRI1^{−138 4x}) caused significantly high promoter activity in malignant pleural mesothelioma cells but little promoter activity in normal mesothelial cells and normal fibroblasts. The recombinant adenoviral vector expressing proapoptotic BH3-interacting death agonist or early region 1A driven by the CRI1^{−138 4x} promoter induced cell death in malignant mesothelioma cells but not in normal cells. Moreover, these viruses showed antitumor effects in a mesothelioma xenograft mouse model. Here, we describe a novel strategy to target malignant mesothelioma using the CRI1^{−138 4x} promoter system. [Cancer Res 2008;68(17):7120–9]

Malignant pleural mesothelioma is an aggressive tumor of mesenchymal origin and is increasing worldwide as a result of widespread exposure to asbestos that was widely used in industrialized countries until approximately 1970. There is substantial interest in this disease because millions of people have been exposed to asbestos fibers, and there are more than 3,000 cases of mesothelioma seen annually in the United States (1). The median survival of patients with mesothelioma from time of diagnosis ranges between 1 and 2 years (2, 3). The mortality is expected to increase, at least until 2020, which is mainly due to the long latency (30–50 years) of the disease (4).

Despite considerable advances in the understanding of its pathogenesis and etiology, malignant mesothelioma remains largely unresponsive to standard modalities of cancer therapy (5). In some cases, extrapleural pneumonectomy can prolong the median survival time of more than 2 years; however, this approach is suitable for only a few patients. Most surgical intervention is often impossible because of intrapleural spread. Although chemotherapy can ameliorate the symptoms of the disease, including pain and breathlessness with pleural effusion, no regimen for mesothelioma has proven curative (6, 7). Moreover, the diffuse nature of pleural mesothelioma, which often covers most of the lung and the interlobular fissures, is the principal limitation to radiotherapy (8). Long-term survival (>5 years) with any treatment modality is exceedingly rare in malignant pleural mesothelioma. Thus, there is an urgent need for new therapeutic options for mesothelioma.

The first human gene therapy trial approved in the United States as a primary cancer treatment was aimed at mesothelioma. At least four gene therapy trials have been carried out in mesothelioma patients using different vector systems (adenovirus and vaccinia virus) and transgenes [herpes simplex virus thymidine kinase (HSV-tk) combined with ganciclovir, interleukin-2, and IFN-β; ref. 9]. However, no significant clinical responses were observed, indicating that improvements are critically needed in gene therapy approaches for the treatment of mesothelioma (10).

Another attractive method to treat malignant pleural mesothelioma is to induce apoptosis by the introduction of proapoptotic genes (11, 12) or cell lysis by the replicative oncolytic adenovirus (13). To use these approaches, the development of a specific promoter system targeting mesothelioma but not normal cells is essential considering the side effects of proapoptotic genes and the replicating adenovirus. To design a system for specific gene therapy and virotherapy to treat malignant mesothelioma, we evaluated the specificity of the promoters of four different mesothelioma-specific genes: calretinin (14), Wilms' tumor suppressor gene (WT1; ref. 15), mesothelin (16), and CREBBP/EP300 inhibitory protein 1 (CRI1; ref. 17). Transient transfection assay showed that the CRI1 promoter is highly active in malignant pleural mesothelioma cells (H2452, MSTO-211H, H2052, and H28) but much less active in normal mesothelial cells and pleural cells. However, the other three promoters of mesothelioma-specific genes (calretinin, WT1, and mesothelin) showed high promoter activity not only in the mesothelioma cells but also in normal cells.

In the present study, we have assessed the capability of adenovirus-mediated transgene expression induced by the CRI1 promoter specifically in mesothelioma cells in vitro and the feasibility of targeting malignant mesothelioma in both in vitro cell culture and in vivo in a mesothelioma xenograft mouse model.

Tissues and cell lines. The human malignant mesothelioma cells H2452, MSTO-211H, H2052, and H28, the human lung adenocarcinoma cells H322 and A549, and the human breast cancer cells MCF7 were obtained from the American Type Culture Collection (ATCC) and grown in Ham's F12 (A549 cells), RPMI 1640 (H2452, MSTO-211H, H2052, H28, and H322 cells), or DMEM high glucose (MCF7) supplemented with 10% heat-inactivated fetal bovine serum (FBS). Normal pleural rat cells 4/4 R.M.-4 obtained from ATCC were grown in Ham's F12 K supplemented with 15% heat-inactivated FBS. The human hepatoblastoma cells Hep3B obtained from ATCC were grown in Eagle' MEM supplemented with 1 mmol/L sodium pyruvate, 0.1% nonessential amino acids, and 10% heat-inactivated FBS. The normal human lung fibroblasts (NHLF) obtained from Clonetics and the normal human mesothelial cells obtained from Dominion Pharmakine were grown in culture medium supplied by the manufacturer. All cell lines were cultured in 10% CO₂ at 37°C. Additionally, normal human pleura specimens were obtained from a consenting patient undergoing treatment for diseases other than mesothelioma (84-y-old male). The lysates of normal human lung protein were obtained from Chemicon International.

Plasmids. The human CRI1, calretinin, WT1, and mesothelin promoters were obtained from purified human genomic DNA (Clontech) by PCR. The position of the transcription initiation site (+1) was determined by the Ensembl Human Genome browser.⁴

The luciferase reporter construct pGL.Calretinin⁻²¹⁷⁹ was generated by subcloning the promoter region of calretinin −2179/+70 from the genomic DNA using PCR primers (5′-tttggtacc^KpnIaatacttttacacaaagcgtggcctg and 5′-aaa^HindIIIaagcttgagctggtgcgaccaccacacccgggagccggcgg). The PCR-generated fragment was digested with KpnI and HindIII and subcloned directly into the pGL3-Basic luciferase reporter construct (Promega). pGL.WT1⁻¹⁸⁸⁷, pGL.Mesothelin⁻²³¹⁰, and pGL.CRI1⁻²⁵⁸⁶ were constructed in the same manner using the following primers: mesothelin, 5′-tttctcgagggggtcaggcttgtgctcccgggagtcctg and 5′-aaaaagcttggtctgtgtctgtggagggcagtgactcag; WT1, 5′-tttggtacctgtctcgagagtcctttctccactcaaaaa and 5′-aaaacgcgtttgctgcggctcagacccggacgccccgc; and CRI1, 5′-tttacgcgtaacatgtagcaagtggctatctaaca and 5′-tataagcttgagggctgctgggccccc. The CRI1 promoter deletion constructs were generated by subcloning multiple parts of the CRI1 promoter region −1849/+84, −1674/+84, −1587/+84, −1083/+84, −766/+84, −567/+84, −366/+84, −296/+84, −138/+84, −74/+84, and +1/+84, amplified from pGL.CRI1⁻²⁵⁸⁶, using the following oligonucleotides, digested with NheI and XhoI and subcloned into pGL3-Basic: 5′-ttt^NheIgctagc⁻¹⁸⁴⁹tggaatcatgcgggtgacaaataacactgg, 5′-ttt^NheIgctagc⁻¹⁶⁷⁴gcacagaactaatggaatatatacatatata, 5′-ttt^NheIgctagc⁻¹⁵⁸⁷tgtatatatatgtaaaggggagttcaataag, 5′-ttt^NheIgctagc⁻¹⁰⁸³gatattgcaaatgacggtggccttaaaagta, 5′-ttt^NheIgctagc⁻⁷⁵²taataataatagaaataaagtgcaga, 5′-ttt^NheIgctagc⁻⁵⁶⁷atacatgtaaatgcaaataaaaagaattcaa, 5′-ttt^NheIgctagc⁻³⁶⁶atatgctgatctgtggttatggatgt, 5′-ttt^NheIgctagc⁻²⁹⁶taaattgcccaggggaagaggaa, 5′-ttt^NheIgctagc⁻¹³⁸cgaatcgatggaaacgtagctcaaaaggcga, 5′-ttt^NheIgctagc⁻⁷³aaccacagtggcgcgccaagtaggag, and 5′-ttt^NheIgctagc⁺¹gcctttgcgcacgcgcacgaacgcac. The human CRI1 region −138/−1 was generated from pGL.CRI1⁻¹⁰⁸³ and subcloned into pGL.CRI1⁻¹³⁸(−138 1x), termed pGL.CRI1−138 2x. The plasmids pGL.CRI−138 3x and pGL.CRI−138 4x were constructed in the same manner. All recombinant plasmids were sequenced to ensure the identity and proper orientation of the insert.

Transient transfection reporter assays. All transfections were carried out in six-well plates. Cells were seeded 24 h before transfection at the following densities: 5 × 10⁵ per well for NHLF and normal human mesothelial cells and 3 × 10⁵ per well for all other cells. Transfections were carried out with Lipofectin (Invitrogen Life Technologies) in accordance with the manufacturer's protocol. Transfected cells were harvested 24 h after lipofection. The results of one representative experiment are presented as fold induction of relative light units normalized to β-galactosidase activity relative to that observed for the control vectors. Each experiment was repeated at least thrice. Error bars indicate the SD from the average of the triplicate samples in one experiment.

Construction of the recombinant adenoviral and lentiviral vectors. The plasmids pCalretinin⁻²¹⁷⁹/GFP³, pCRI1⁻²⁵⁸⁶/GFP, and pCRI1⁻²⁵⁸⁶/HA-BID were constructed by ligating green fluorescent protein (GFP) or HA-BID into pGL.Calretinin⁻²¹⁷⁹, pGL.CRI1⁻²⁵⁸⁶, and pGL.CRI1^{−138 4x} after excising the luciferase gene. The cDNA of adenovirus type 5 early region 1A (E1A) was synthesized by reverse transcription-PCR from total cellular RNA of human embryonic kidney 293 cells using specific primers 5′-ttt^HindIIIaagcttctgaaaatgagacatattatctgccacggaggtgt and 5′-aaa^EcoR5gatatcttatggcctggggcgttta. The recombinant adenovirus vectors Ad-Calretinin⁻²¹⁷⁹/GFP, Ad-CRI1⁻²⁵⁸⁶/GFP, Ad-CRI1^{−138 4x}/GFP, Ad-CRI1^{−138 4x}/HA-BID, and Ad-CRI1^{−138 4x}/E1A were generated by homologous recombination (18, 19). The viral titer for each vector was determined by plaque assay and the optimal multiplicity of infection (MOI) was determined by infecting each cell line with Ad-CMV/GFP and assessing the expression of GFP by flow cytometric analysis. H2452 cells were infected with the recombinant adenoviral vectors at a MOI of 50 plaque-forming units (pfu)/cell, and all other human cells were infected at a MOI of 20 pfu/cell. For infection of the conditionally replicating adenovirus (CRAd), H2452 cells were infected with Ad-CRI1^{−138 4x}/E1A at a MOI of 10 pfu/cell, and all other human cells were infected at a MOI of 4 pfu/cell. Ad-CRI1^{−138 4x}/GFP was used as control. The human CRI1 short hairpin RNA (shRNA) lentiviral transfer vector for human CRI1 gene was obtained from Sigma (MISSION shRNA Bacterial Glycerol Stock). The transformed human embryonic kidney 293T cells (1 × 10⁶) were plated in a 10-cm dish and cotransfected the following day with 26 μL of the Lentivirus Packaging Mix (Sigma) and the shRNA transfer vector (2.6 μg) by lipofection. Twenty-four and 48 h later, viral supernatants were collected in serum-free medium and filtered through 0.45-μm pore size filters. The nontarget shRNA lentiviral (Sigma) vector was used as a control. The viral titer was measured by HIV p24 Antigen ELISA kit (ZeptMetrix; refs. 20–22). Cells were infected with the recombinant shRNA lentiviral vector at a concentration of MOI of 2.5 transducing units/cell.

Immunoblot analysis. Tissues and cells were washed once with ice-cold PBS containing 5 mmol/L EDTA and 1 mmol/L sodium orthovanadate, homogenized, and lysed in ice-cold lysis buffer [1% Triton X-100, 20 mmol/L Tris-HCl (pH 8.0), 137 mmol/L NaCl, 10% (v/v) glycerol, 2 mmol/L EDTA, 1 mmol/L (v/v) sodium orthovanadate, 1 mmol/L phenylmethylsulfonyl fluoride, 10 μg/mL aprotinin, 10 μg/mL leupeptin]. Cell lysates were clarified by centrifugation (10 min at 15,000 × g at 4°C) and protein concentration was determined using the detergent-compatible protein assay (Bio-Rad). Equal amounts of protein were separated on a SDS-PAGE gel. The gel was electrophoretically transferred to a Hybond polyvinylidene difluoride transfer membrane (Amersham). The membrane was incubated with primary and secondary antibodies according to the SuperSignal West Pico chemiluminescence protocol (Pierce) to detect secondary antibody binding. Antibody specific for calretinin was purchased from BD Transduction Laboratories. Antibodies against CRI1 and GFP were obtained from Abcam. Antibody specific for WT1 was obtained from Invitrogen Life Technologies. Antibody specific for adenovirus type 5 E1A was obtained from PharMingen. Anti-mesothelin was kindly provided by Dr. Ira Pastan (Center for Cancer Research, National Cancer Institute, NIH, Bethesda, MD; ref. 23). Anti-actin was obtained from Sigma. Secondary horseradish peroxidase–conjugated goat anti-rabbit antibody and anti-mouse antibody were obtained from Jackson ImmunoResearch Laboratories.

Flow cytometric analysis for apoptosis. Cells were plated in 24-well plates at a density of 2 × 10⁵ per well 1 d before the recombinant adenoviral vector infection. After 72 h, cells were harvested and washed once with PBS. Cells were resuspended in PBS containing 0.2% Triton X-100 and 1 mg/mL RNase for 5 min at room temperature and then stained with propidium iodide at 50 μg/mL to determine subdiploid DNA content using a FACScan. Doublets, cell debris, and fixation artifacts were gated out, and sub-G₀-G₁ DNA content was determined using CellQuest version 3.3 software.

Animal experiments. The experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of Okayama University Graduate School of Medicine and Dentistry. Human mesothelioma xenografts were established in 6-wk-old female BALB/c nude mice (Charles River Laboratories, Inc.) by s.c. inoculation of 2.5 × 10⁶ MSTO-211H cells into the dorsal flank. The mice were randomly assigned into four groups (n = 8 per group). Each group of the mice was injected with 200 μL solution containing PBS, Ad-CRI1^{−138 4x}/GFP, Ad-CRI1^{−138 4x}/HA-BID, or Ad-CRI1^{−138 4x}/E1A by 27-gauge needle for the first 3 d. Animals were then observed closely and survival studies were performed. Tumors were measured two to three times a week, and tumor volume was calculated as a × b² × 0.5, where a and b were large and small diameters, respectively.

Histochemical study. Tissue or tumor sectioning and staining were performed in the Histology Laboratory in the Department of Gastroenterological Surgery in Okayama University Graduate School of Medicine and Dentistry. For immunohistochemical analysis of the E1A protein, tumors were fixed in 20% formalin, embedded in paraffin, and then cut into 4-μm sections. To retrieve antigens, the sections were baked, deparaffinized, and heated in citrate buffer [10 mmol/L citric acid (pH 6.0)] in a steamer. After endogenous peroxidase was inactivated with 1.5% H₂O₂/methanol for 10 min, the sections were incubated with rabbit anti-E1A polyclonal antibody (1:50 dilution; PharMingen) or rabbit anti-GFP polyclonal antibody (1:500 dilution; Abcam) for 1 h and then biotinylated goat anti-rabbit IgG antibody (DAKO) for 30 min. The specific binding was visualized with an avidin-biotin-peroxidase reagent and its substrate diaminobenzidine tetrachloride (DAKO) and subsequent counterstaining with Mayer's hematoxylin.

Statistical analysis. All of the in vitro experiments in Figs. 1 to 5 were performed at least thrice. The results of one representative experiment are presented. For quantitative results, error bars indicate the SD from the average of at least triplicate samples in one experiment. For the in vivo experiments in Fig. 6, statistical differences were determined using unpaired two-tailed Student's t tests or ANOVA.

Analysis of CRI1, calretinin, WT1, and mesothelin protein expression in normal cells and in tumor cells. To investigate the expression of the mesothelioma markers (calretinin, WT1, and mesothelin) and CRI1, immunoblot analysis was performed using nine kinds of thoracic neoplasm, normal cells, and tissues, including malignant pleural mesothelioma cells. As shown in Fig. 1A, CRI1 was detected in all types of malignant mesothelioma cells (H2452, MSTO-211H, H2052, and H28) and A549 pulmonary adenocarcinoma cells, whereas no expression was seen in H322 pulmonary adenocarcinoma cells, normal mesothelial cell, normal pleura, or normal lung tissue extracts. Calretinin expression was detected in H2452, MSTO-211H, and H2052 malignant mesothelioma cells, A549 pulmonary adenocarcinoma cells, and normal mesothelial cells. WT1 expression was strongly detected in all types of malignant mesothelioma cells (H2452, H2052, MSTO-211H, and H28) and normal mesothelial cells. Mesothelin expression was seen in H2052 malignant mesothelioma cells, A549 pulmonary adenocarcinoma cells, normal mesothelial cells, normal pleura, and normal lung extracts. Significantly, the expression of calretinin, WT1, and mesothelin was detectable in normal human mesothelial cells and extracts of normal human pleura and lung, whereas the expression of CRI1 was not seen in these normal human cells and tissues. These results suggest that CRI1 is the most specific malignant mesothelioma marker that distinguishes malignant pleural mesothelioma cells from normal mesothelial cells.

The human CRI1 promoter generates high promoter activity in pleural malignant mesothelioma cells. To analyze the promoter activity of the mesothelioma markers and CRI1, 5′ flanking regions of these marker genes were cloned into pGL3-Basic luciferase reporter constructs (Fig. 1B). The ability of these constructs to promote the expression of the luciferase gene was measured in four malignant mesothelioma cell lines and two pulmonary adenocarcinoma cells in transient transfection reporter assay (Fig. 1C). The pGL.CRI1⁻²⁵⁸⁶/Luc generated significantly high transcriptional activity in all kinds of malignant mesothelioma cells (H2452, MSTO-211H, H2052, and H28; 42.01- to 86.77-fold) and A549 pulmonary adenocarcinoma cells (22.79-fold), with low activity in H322 pulmonary adenocarcinoma cells (9.22-fold). The pGL.Calretinin⁻²¹⁷⁹/Luc generated significantly high promoter activity in H2452, MSTO-211H, and H2052 malignant mesothelioma cells and A549 pulmonary adenocarcinoma cells (15.90- to 45.38-fold). The pGL.WT1⁻¹⁸⁷⁷/Luc construct generated significant transcriptional activity in all types of malignant mesothelioma cells (H2452, MSTO-211H, H2052, and H28; 25.76- to 51.57-fold) with less activity in H322 and A549 pulmonary adenocarcinoma cells (4.57- to 4.87-fold). The pGL.Mesothelin⁻²³⁵⁵/Luc construct showed significantly high promoter activity in H2052 malignant pleural mesothelioma cells and A549 pulmonary adenocarcinoma cells (22.25- to 47.24-fold) but little activity in other kinds of cells (H2452, MSTO-211H, H28, and H322; 8.29- to 12.74-fold). These results suggest that the 5′ flanking region of the CRI gene generates higher promoter activity than that of the other three mesothelioma marker genes (calretinin, WT1, and mesothelin) in all types of malignant pleural mesothelioma cells.

The human CRI1 promoter generates little promoter activity in normal mesothelial cells and pleural cells. To determine the tumor specificity of the transcriptional activity of the 5′ flanking region of the mesothelioma marker genes, we also performed a transient transfection reporter assay in normal human mesothelial cells, normal rat pleural 4/4 R.M.-4 cells, and NHLF cells (Fig. 1D). In these normal cells, the promoter activity of pGL.CRI1⁻²⁵⁸⁶/Luc was less than (8.29- to 9.02-fold) that of the reporter constructs pGL.Calretinin⁻²²⁵⁰/Luc, pGL.WT1⁻²²⁸⁸/Luc, and pGL.Mesothelin⁻²³⁵⁵/Luc (10.26- to 27.19-fold). These data suggest that the 5′ flanking region of the CRI1 promoter generated the least promoter activity in normal mesothelial cells and pleural cells among the four different mesothelioma marker genes.

CRI1 is involved in cell viability of MSTO-211H mesothelioma cells. To determine the function of CRI1 in mesothelioma cells, cell viability was observed in the presence/absence of CRI1. CRI1 expression was suppressed by lentiviral vector expressing shRNA for human CRI1 (LV/shCRI1) in MSTO-211H mesothelioma. As shown in Supplementary Fig. S1A, immunoblot analysis showed that endogenous CRI1 expression level was suppressed after 4 days of LV/shCRI1 infection compared with lentiviral vector expressing nontarget shRNA (LV/shNon-target). The 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt (MTS) assay (24) showed that MSTO-211H cells infected with LV/shCRI1 lentivirus had decreased cell viability compared with control LV/shNon-target lentivirus (Supplementary Fig. S1B). These data suggest that CRI1 is essential for MSTO-211H mesothelioma cell viability.

Identification of the region in the CRI1 promoter that confers mesothelioma-specific transactivation. To identify the region of the CRI1 promoter that confers mesothelioma-specific transactivation, we made deletion constructs of the CRI1 promoter (Fig. 2A) and compared their promoter activities. These constructs were used in transient transfection reporter assays in MSTO-211H and H2452 human mesothelioma cells, A549 and H322 human pulmonary adenocarcinoma cells, normal human mesothelial cells, and NHLF (Fig. 2B). The ideal construct would have minimal activity in nonmesothelioma cells but high activity in MSTO-211H and H2452 human mesothelioma cells. In all types of cells, the proximal 1,849-bp fragment of the CRI1 promoter exhibited higher activity than the 2,586-bp fragment. The proximal 1,587-bp fragment showed lower transcriptional activity than the 1,674- and 1,083-bp fragment. The proximal 766-bp fragment of the CRI1 promoter exhibited the highest activity (100.29-fold) in MSTO-211H mesothelioma cells. Promoter deletions of −567, −366, −296, −138, and −73 showed stepwise decreases in transcriptional activity in all types of cells indicated. The +1-bp fragment (pGL.CRI1 +1) conferred no activity above control levels in all types of cells. The −138 construct showed the most promise as a mesothelioma-specific promoter as it had high transcriptional activity in mesothelioma cells with little activity in the other types of cells, including normal mesothelial cells.

Analysis of the effect of a tandem human CRI1 promoter. To enhance the level of transcriptional activity sufficient for the use of gene therapy, we tested the effect of additional tandem copies of the CRI1 −138/−1 promoter region to the construct pGL3.CRI1−138/+84 in a transient transfection reporter assay (Fig. 3A). In MSTO-211H cells, transcriptional activity increased significantly (169.7-fold) with the addition of one extra CRI1 cassette −138/−1 (pGL.CRI1−138 2x). Moreover, with double to triple tandem copies (pGL.CRI1−138 3x to 4x), the activity increased up to 216.1- to 251.1-fold. In H2452 mesothelioma cells, with single to triple tandem copies (pGL.CRI1−138 2x to 4x), the activity increased up to 99.6- to 181.3-fold. However, in H322 and A549 pulmonary adenocarcinoma cells, normal human mesothelial cells, and NHLF, these tandem copies showed no enhanced transcriptional activity (1.52- to 8.24-fold; Fig. 3B). These results show that four tandem copies of the CRI1 cassette −138/−1 (pGL.CRI1−138 4x) generate the highest mesothelioma-specific transactivation.

Ad-CRI1^{−138 4x}/GFP expresses GFP in MSTO-211H malignant mesothelioma cells but not in normal human mesothelial cells or NHLFs. To evaluate gene expression mediated by the CRI1−138 4x promoter in adenovirus-mediated delivery system, we examined GFP expression induced by a recombinant adenoviral vector that expressed GFP driven by the CRI1^{−138 4x} promoter system (Ad-CRI1^{−138 4x}/GFP; Fig. 3C) in MSTO-211H mesothelioma cells, normal human mesothelial cell, and NHLF. Recombinant adenoviral vectors that expressed GFP driven by promoters of cytomegalovirus (CMV), calretinin⁻²¹⁷⁹, and CRI1⁻²⁵⁸⁶ were used as controls (termed Ad-CMV/GFP, Ad-Calretinin⁻²¹⁷⁹/GFP, and Ad-CRI1⁻²⁵⁸⁶/GFP; Fig. 3C). As shown in Fig. 3D, Ad-CMV/GFP strongly induced GFP in all three kinds of cells. Ad-Calretinin⁻²¹⁷⁹/GFP induced GFP not only in MSTO-211H mesothelioma cells but also in normal human mesothelial cells. Ad-CRI1⁻²⁵⁸⁶/GFP weakly induced GFP in MSTO-211H cells and little expression was seen in normal human mesothelial cells and NHLF. On the other hand, Ad-CRI1^{−138 4x}/GFP did not show any significant fluorescent population in normal human mesothelial cells and NHLF but strongly induced GFP in MSTO-211H mesothelioma cells. These data suggest that the CRI1^{−138 4x} promoter is able to induce target gene expression in mesothelioma cells and does not target normal cells, including normal mesothelial cells, in the adenovirus-mediated delivery system.

Ad-CRI1^{−138 4x}/HA-BID induced BH3-interacting death agonist expression and cell death/apoptosis in the malignant pleural mesothelioma. To induce cell death in the mesothelioma by the CRI1^{−138 4x} promoter system, the proapoptotic BH3-interacting death agonist (BID) gene was inserted into the CRI1^−138 4x system in place of GFP (Ad-CRI1^{−138 4x}/HA-BID; Fig. 4A; refs. 19, 25, 26). As shown in Fig. 4B, Ad-CRI1^{−138 4x}/HA-BID strongly induced exogenous BID in H2452 and MSTO-211H mesothelioma cells. On the other hand, there was little expression in normal human mesothelial cells, NHLF, nonmesothelioma Hep3B hepatocellular carcinoma cells, and MCF7 breast cancer cells. The cell death/apoptosis caused by the Ad-CRI1^{−138 4x}/HA-BID construct was examined by sub-G₀-G₁ DNA content using propidium iodide staining and flow cytometry 72 h after Ad-CRI1^{−138 4x}/HA-BID infection. As shown in Fig. 4C, infection with the Ad-CRI1^{−138 4x}/GFP control caused little DNA fragmentation (range, 0.50–3.79%) in all types of cells as indicated (top). On the other hand, Ad-CRI1^{−138 4x}/HA-BID infection caused a marked increase in sub-G₁ DNA content only in H2452 (25.91%) and MSTO-211H mesothelioma cells (45.77%) but not in the other types of cancer cells or normal cells (1.24–7.62%; bottom). These results are consistent with the BID expression in Fig. 4B. These data suggest that Ad-CRI1^{−138 4x}/HA-BID induces exogenous BID expression and cell death/apoptosis only in mesothelioma cells.

Ad-CRI1^{−138 4x}/E1A induced viral proliferation and cell death only in mesothelioma cells. The CRAd that spreads and replicates only in cancer cells is suitable for diffused or metastasized pleural mesothelioma (13). To test the applicability of CRAd in combination with the mesothelioma-specific CRI1^{−138 4x} promoter system, we also made Ad-CRI1^{−138 4x}/E1A that expresses cytotoxic E1A (adenoviral replication-programming protein) driven by the CRI1^{−138 4x} promoter system (Fig. 4A). As shown in Fig. 5A, Ad-CRI1^{−138 4x}/E1A induced cell death/lysis in H2452 and MSTO-211H mesothelioma cells but not in normal human mesothelial cells or NHLF 4 days after infection. The control vector (Ad-CRI1^{−138 4x}/GFP) did not induce any cell death in these cells. A cell viability assay revealed that Ad-CRI1^{−138 4x}/E1A drastically induced cell death/lysis after 3 to 4 days of infection in H2452 and MSTO-211H cells but not in normal human mesothelial cells or NHLF (Fig. 5B). Immunoblot analysis showed that E1A expression was drastically increased in H2452 and MSTO-211H mesothelioma cells by Ad-CRI1^{−138 4x}/E1A after 96 h of infection. However, E1A expression was minimal in NHLF and normal human mesothelial cells. No significant increase of GFP expression was seen in MSTO-211H mesothelioma cells between 24 and 96 h after Ad-CRI1^{−138 4x}/GFP infection (Fig. 5C). These results suggest that Ad-CRI1^{−138 4x}/E1A (a CRAd) replicates and induces cell death/lysis in mesothelioma cells.

Ad-CRI1^{−138 4x}/E1A or HA-BID suppressed mesothelioma xenograft tumors and extended the survival of mesothelioma xenograft mice. To analyze the therapeutic efficacy of Ad-CRI1^{−138 4x}/E1A or HA-BID against human tumor cells in vivo, we established mesothelioma xenograft tumors derived from MSTO-211H cells in nude mice. After the tumors had reached a diameter of about 0.5 cm, 5 × 10⁷ pfu of Ad-CRI1^{−138 4x}/E1A, Ad-CRI1^{−138 4x}/HA-BID, Ad-CRI1^{−138 4x}/GFP, or PBS were administered intratumorally for 3 days. Tumor growth was significantly suppressed by Ad-CRI1^{−138 4x}/E1A or Ad-CRI1^{−138 4x}/HA-BID injection compared with Ad-CRI1^{−138 4x}/GFP or PBS (P < 0.001; Fig. 6A). The mouse group treated with PBS or Ad-CRI1^{−138 4x}/GFP died within 75 days after tumor inoculations (Fig. 6B), whereas seven of eight mice from the mouse group treated with Ad-CRI1^{−138 4x}/HA-BID and all mice from the mouse group treated with Ad-CRI1^{−138 4x}/E1A were still alive. Adenoviral E1A expression was increased and spread in Ad-CRI1^{−138 4x}/E1A–administered tumors, whereas GFP expression was localized (no spread) in Ad-CRI1^{−138 4x}/GFP–administered tumors at 14 days after treatment (Fig. 6C). These results indicate that Ad-CRI1^{−138 4x}/E1A (a CRAd) replicates and induces cell lysis in a mesothelioma-specific manner in vivo.

Several protein markers for diagnosing mesothelioma have recently become available (27). One of the calcium-binding proteins, calretinin is the most sensitive mesothelioma-positive marker and is frequently expressed in all the logic types of mesothelioma (27). However, Lugli and colleagues (28) reported that calretinin is also expressed in normal tissues, including Leydig cells of the testis, neurons of the brain, theca-lutein and theca interna cells of the ovary, and normal mesothelium. The other mesothelioma marker mesothelin is a 40-kDa cell surface glycoprotein that is highly expressed in epithelioid mesothelioma. Mesothelin is useful to diagnose epithelioid mesothelioma and sarcomatoid mesothelioma (16, 23). Hassan and colleagues reported that serum mesothelin levels are elevated in patients with mesothelioma compared with normal healthy volunteers. Serum mesothelin is decreased after surgical dissection of mesothelioma (29). However, mesothelin expression was also detectable in normal mesothelium (30, 31). Another mesothelioma marker, WT1 is a zinc finger DNA-binding protein and acts as a transcription activator or repressor depending on the cellular or chromosomal context. WT1 is useful to diagnose epithelioid mesothelioma (27). However, the expression of WT1 is also seen in normal tissues, including normal mesothelium (32, 33). Taken together, these mesothelioma markers are pathologically important to distinguish mesothelioma from the other kinds of tumors (e.g., pulmonary adenocarcinoma; ref. 27), but they are also expressed in normal pleura and mesothelium. Gordon and colleagues (17) profiled the gene expression pattern of malignant pleural mesothelioma, normal lung, and pleural tissues using cDNA microarrays. They reported that the CRI1 gene was one of the malignant pleural mesothelioma-specific markers. In our study, CRI1 protein was observed specifically in mesothelioma cells but not in normal mesothelium cells (Fig. 1A), which agrees with the microarray data (17).

Gene therapy approaches using mesothelioma-specific promoters combined with suicide genes have previously been tried. Inase and colleagues (34) constructed a system that expressed the HSV-tk gene driven by the calretinin promoter. However, in our study, Ad-Calretinin⁻²¹⁷⁹/GFP expressed GFP in normal human mesothelial cells. Thus, HSV-tk driven by the calretinin promoter might be expressed in normal mesothelial cells and induce side toxicity in normal cells as well as mesothelioma cells. The ideal promoter system would have minimal activity in nonmesothelioma cells but high activity in mesothelioma cells. In the present study, the CRI1^{−138 4x} promoter generated the most specific transcriptional activity in malignant pleural mesothelioma cells compared with the other three promoters of calretinin, mesothelin, and WT1 genes. Thus, the CRI1^{−138 4x} promoter should be useful for targeting mesothelioma cells. The molecular mechanism by which the 5′ flanking region of the CRI1 promoter −138/+84 confers mesothelioma-specific activity is not understood yet. Mesothelioma-specific transcription factors and/or cofactors might also be involved in the CRI1 transcriptional regulation.

Mesothelioma cells are more resistant to apoptosis than normal mesothelial cells, although most of them have the wild-type p53 tumor suppressor (35). To explain this paradox, Cao and colleagues (36) postulate that the antiapoptotic protein Bcl-xL might play a role in increased resistance to apoptosis by mesothelioma cells, as Bcl-xL seems to have increased expression in malignant mesothelioma cells. In this study, overexpression of the proapoptotic BID protein by Ad-CRI1^{−138 4x}/HA-BID induced apoptosis in mesothelioma cells, indicating that the overexpressed BID escaped the antiapoptotic effect by Bcl-xL in the mesothelioma cells. The use of BID as a suicide gene may thus be useful for localized mesothelioma gene therapy.

A problem for mesothelioma gene therapy is the inefficient vector and gene delivery systems. In advanced stages, pleural mesothelioma invades the chest wall or mediastinal tissues or structures (e.g., esophagus, trachea, great vessels, and lymph nodes) and could penetrate diaphragmatic muscle, peritoneum, retroperitoneal space, opposite pleura, or lymph nodes outside the chest (37). For the treatment of advanced-stage mesothelioma, the efficiency of nonreplicating adenovirus vector, including Ad-CRI1^{−138 4x}/HA-BID, by local injection might be limited. In recent cancer gene therapy studies, the CRAd has been considered as an approach to target metastasized cancer, including mesothelioma, because it replicates and chases the metastasized tumors (38–42). In our study, Ad-CRI1^{−138 4x}/E1A, a CRAd, showed mesothelioma-specific cell death/lysis in vitro and a strong antitumor effect in a mesothelioma xenograft tumor mouse model, suggesting that Ad-CRI1^{−138 4x}/E1A is useful for the treatment of advanced-stage mesothelioma.

Another problem for mesothelioma gene therapy with viral vectors, including CRAds, is a possibility that the therapeutic virus might be removed by host immunity, which might possibly limit CRAd replication. However, Robert and colleagues (43) have recently reported a new hexon-chimeric adenovirus vector that circumvents preexisting antivector immunity. In the use of the hexon-chimeric adenovirus vector, selected promoters, including the CRI1 promoter system, are required to limit the cell killing in mesothelioma but not surrounding normal cells. The combination of CRI1^{−138 4x}/E1A system with the hexon-chimeric adenovirus would be an attractive approach to treat immunocompetent mesothelioma.

To enhance the efficacy of CRAds, the combination of CRAds with volume reduction surgery and/or chemotherapy has been reported to be more effective than single treatment of surgery or chemotherapy (44, 45). To reduce the side effect to normal cells by chemotherapy or surgical stress, the use of Ad-CRI1^{−138 4x}/E1A that selectively targets mesothelioma is desirable for the combination therapy.

In this study, a mesothelioma-specific CRI gene promoter was cloned and characterized as the most mesothelioma-specific promoter compared with the promoters of mesothelioma markers, including calretinin, WT1, and mesothelin. The element of the promoter that confers mesothelioma specificity (CRI1 −138/−1) was identified by deletion mapping and the element was tandemarized (CRI1^{−138 4x}) to enhance the promoter activity to further target mesothelioma by gene therapy. The CRI1^{−138 4x} promoter-driven E1A or HA-BID in the adenovirus induced cell death only in human mesothelioma cell lines but not in normal mesothelial cells or NHLF. The efficacy of the mesothelioma cell death was further confirmed by an in vivo xenograft model. Thus, our adenovirus-mediated gene therapy system expressing HA-BID or E1A gene driven by the mesothelioma-specific CRI1 promoter (a CRAd) has great clinical potential for the treatment of mesothelioma.

No potential conflicts of interest were disclosed.

Grant support: The Ministry of Education, Science, and Culture, Japan.

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.

We thank T. Nakai and T. Yamanishi for technical advice, F. Inoue and M. Takahashi for kindly providing normal human pleura, and Dr. Ira Pastan for providing antibody against mesothelin.