Malignant mesothelioma (MM) is a highly aggressive tumor that arises from the surface serosal cells of the pleura and, less frequently, the peritoneum. The incidence of the disease in the majority of Western nations is expected to steadily rise and peak over the next 2 decades.1 There is no known curative modality for MM and long-term survival is rare even with aggressive multimodal therapy.
A strong link has been established between exposure to asbestos and increased risk for MM.2 Subsequent laboratory research has confirmed the carcinogenicity of asbestos in mesothelial cell lines and in animal models. Despite recent restrictions on the use of asbestos, the incidence of MM is continuing to rise due to the long latency period for the development of MM after asbestos exposure. Simian virus 40 (SV40) has also been implicated in the pathogenesis of MM. SV40 may have synergistic oncogenic effects when combined with asbestos, although the exact role and importance of this virus in the development of MM is controversial.3 One recent study demonstrated that significantly lower amounts of asbestos were sufficient to cause MM in animals infected with SV40 compared with uninfected animals, supporting the possible role of SV40 as a co-carcinogen in this disease.4 Using a novel SV40 large T antigen (Tag) transgenic model of mesothelioma, Robinson et al.5 also demonstrated that asbestos-induced malignant transformation is SV40 dose-dependent, with higher copy numbers of the transgene increasing the rate of tumor formation.
Treatment of MM with surgery, chemotherapy, or radiation therapy is rarely curative. Clinical trials of single modality treatment with extrapleural pneumonectomy or pleurectomy, chemotherapy, or radiation therapy have not shown significant improvement in survival compared with supportive treatment. Median survival has ranged from 10–17 months. A multimodal approach, including surgery, chemotherapy, and radiation therapy, is thought to improve survival in selected patient, although this remains controversial. Sugarbaker et al.6 reported a 15% 5-year survival with multimodal therapy and a 25% 5-year survival in patients who underwent complete surgical resection. Recent trials of new-generation platinum- and pemetrexed-based regimens have reported encouraging results. In particular, a Phase III trial of pemetrexed plus cisplatin for MM reported a median survival of 12 months compared with 9 months after treatment with cisplatin alone.6
Despite these promising results, long-term survival with currently available treatments is rare. Novel therapies for MM are needed. Improvement in our understanding of the molecular biology of MM has already identified promising new therapies and pathways that are candidates for targeted therapies. In this review, we focus on the key molecular signaling pathways, including vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), Wnt, and cell cycle control genes p53, pRb, and blc2, which appear to play an important role in the pathogenesis of MM.
Vascular Endothelial Growth Factor
VEGF is a potent inducer of angiogenesis and its critical role in tumor progression is well established.7 VEGF-targeted therapy is a promising approach aimed at inhibiting tumor growth, invasion, and metastasis. VEGF up-regulation appears to play an important role in mesothelial cell transformation. High levels of VEGF have been observed in the serum of mesothelioma patients compared with the serum of normal subjects.8, 9 Also, elevated pleural effusion VEGF levels are associated with diminished survival in patients with MM.10 Similarly, Demirag et al.11 reported that VEGF overexpression on immunohistochemistry (IHC) independently predicts short survival in patients with MM (P = .0002).
Recent reports have suggested that SV40 may play a role in inducing VEGF expression in MM cells. One study demonstrated that the SV40 large tumor antigen (Tag) is involved in VEGF promoter activation and that it potently increases VEGF levels in several MM cell lines.12 Moreover, transfection of normal human mesothelial cells with full-length SV40 DNA resulted in a significant increase of VEGF levels, suggesting that SV40 may contribute to increased VEGF production by mesothelial cells.13 Evidence also suggests that interleukin-6 (IL-6), a cytokine that is frequently overexpressed in MM, might play a role in the induction of VEGF production. A recent study by Adachi et al.14 demonstrated that IL-6/IL-6R stimulation increased expression of VEGF in 80% of MM cell lines and that an anti-IL-6R antibody inhibited this expression.
In addition to stimulating the neovascularization essential for tumor growth and metastasis, recent evidence has demonstrated that VEGF may also act in a functional autocrine loop capable of directly stimulating the growth of malignant mesothelioma cells. Strizzi et al.11 reported that MM cell lines express elevated levels of both VEGF and the VEGF receptors, VEGFR-1 (Flt-1) and VEGFR-2 (KDR), compared with normal mesothelial cells. In that study they found that recombinant VEGF phosphorylated both Flt-1 and KDR and increased proliferation of all MM cell lines examined. In contrast, neutralizing antibodies against VEGF, Flt-1, or KDR significantly reduced MM cell proliferation. Similar autocrine signaling in MM cells was demonstrated with VEGF-C and its cognate receptor VEGFR-3.15 Simultaneous inhibition of VEGF and VEGF-C displayed a synergistic effect in reducing the viability of mesothelioma cells in vitro.
Interestingly, significant arterial vascularization is rarely exhibited in MM. Whereas at least 2 studies16, 17 reported a significant correlation between VEGF expression by IHC and intratumoral microvascular and microlymphatic density, other studies failed to find a correlation.10, 11 Strizzi et al.11 proposed that competitive binding of VEGF in favor of rapidly proliferating MM tumor cells might make VEGF less available to stimulate the endothelial cells necessary to promote angiogenesis.
These reports suggest that VEGF may play a key role in MM tumor progression by regulating both angiogenesis and tumor cell proliferation, thus raising interest in VEGF as a therapeutic target in MM. Bevacizumab (Avastin, Genentech, South San Francisco, Calif), a monoclonal antibody that targets VEGF, in combination with cisplatin produces an inhibitory effect in an in vivo model of human MM. A Phase II clinical trial of bevacizumab combined with gemcitabine-cisplatin chemotherapy in patients with malignant mesothelioma is under way, with encouraging preliminary results.18 A Phase II clinical trial at the University of Chicago of SU-5416, a selective inhibitor of VEGF-R2, in patients with MM is also under way.19
Epidermal Growth Factor Receptor
The role of epidermal growth factor receptor (EGFR) signaling in MM has become an active area of research in large part due to the success of EGFR-targeted therapies in lung cancer. EGFR is a member of the erbB family of tyrosine kinase receptors, which also includes erbB-2, -3, and -4. Multiple ligands bind to and activate EGFR, including amphiregulin, betacellulin, EGF, and transforming growth factor (TGF)-α. The latter 2 are believed to be the most important ligands for EGFR. Ligand binding with EGFR leads to receptor dimerization and internalization followed by autophosphorylation of the receptor tyrosine kinase domains. This subsequently results in the activation of signaling pathways that are involved in cell proliferation, differentiation, and survival.20 Overexpression of EGFR has been recognized to play an important role in the pathogenesis and progression of a variety of malignancies, including colon cancer, nonsmall-cell lung cancer (NSCLC), and breast cancer.
The involvement of EGFR in malignant pleural mesothelioma was first reported by Dazzi et al.,21 who found expression of the receptor in 68% of fixed paraffin-embedded MM tissue specimens. A similar recent study found EGFR immunoreactivity in 55.7% of MM samples and no immunoreactivity in normal pleura.22 Both studies found that EFGR expression is not an independent prognostic factor for patient survival. One recent IHC study of 168 tumor sections identified EGFR expression in 44% of cases, with EGFR positivity associated with improved survival (P = .01).23 The authors suggest that the favorable prognosis associated with EGFR positivity may be partly explained by the greater expression of EGFR in the epithelioid cell type, which is known to have a better prognosis than the sarcomatoid cell type. Thus, the correlation between EGFR expression with improved survival does not negate EGFR as a potential therapeutic target in MM.
Asbestos fibers are involved in the activation of EGFR in mesothelioma. Pache et al.24 demonstrated that asbestos fibers cause aggregation and increased immunoreactivity of the EGFR protein in human mesothelial cells in vivo and that asbestos-induced EGFR autophosphorylation may lead to the induction of the AP-1 family members, c-fos and c-jun.25, 26 Crocidolite and erionite asbestos fibers were found to induce EGFR expression in rat pleural mesothelial cells.27 Intense patterns of EGFR protein expression were linked to mesothelial cells phagocytosing longer asbestos fibers. Previous studies have shown a relation between asbestos fiber length and carcinogenicity.28
It is uncertain whether there is a role for EGFR inhibitors in the treatment of MM. A recent in vitro study by Cole et al.29 demonstrated that inhibition of EGFR signaling with the small molecule tyrosine kinase inhibitor (TKI) PD153035 suppresses MM cell motility and invasion. Similarly, Janne et al.30 found that gefitinib (Iressa, AstraZeneca, Wilmington, Del), another EGFR-TKI, significantly inhibits EGFR-dependent cell signaling including AKT phosphorylation and extracellular signal-regulated kinases 1 and 2 in all MM cell lines examined. This resulted in marked antiproliferative effects of these MM cells in vitro. In another study, gefitinib increased tumor responsiveness to radiation therapy in an animal model of MM.31 Despite the promising in vitro activity of gefitinib in MM cell lines, the recently published Cancer and Leukemia Group B (CALGB) Phase II clinical trial of gefitinib treatment in 43 previously untreated mesothelioma patients found that gefitinib was not effective in this disease.32 This was in spite of the fact that 97% of patients displayed EGFR overexpression. Interestingly, a recent study found that common EGFR tyrosine kinase domain mutations conferring sensitivity to gefitinib in lung cancer are not prevalent in MM.33
p53 and pRB Pathways
Functional inactivation of the p53 and retinoblastoma protein (pRB) pathways appears to be a critical requirement in the development of several human cancers, including lung cancer. These pathways play key roles in apoptosis and cell cycle regulation.34 The tumor suppressor protein pRB limits cell proliferation by inhibiting entry into the S-phase of the cell cycle. Another key tumor suppressor protein, p53, accumulates and is activated in response to cellular stresses such as hypoxia, DNA damage, and oncogene activation. Activated p53 initiates transcription of genes that can lead to cell cycle arrest and apoptosis. Mutations and deletions the p53 and pRb tumor suppressor genes occur frequently in many human cancers. Such mutations, however, are extremely rare in MM, and several studies have demonstrated that pRb35 and p5336 remain genetically intact in all MM specimens examined. In fact, 1 recent IHC study37 found overexpression p53 in 81% of patient mesothelioma samples.
In contrast, 1 of the predominant genetic abnormalities in human MM is homozygous deletion of the 9p21 region, occurring at a frequency of greater than 70% in MM cell lines.38 The INK4a/ARF locus within the 9p21 chromosome band plays a critical role in the regulation of both the pRb and p53 tumor suppressor pathways (Fig. 1). This locus encodes 2 distinct proteins, p16INK4a and p14ARF, translated from alternatively spliced mRNA. p16INK4a exerts its tumor suppressive effect by inhibiting the cyclin D-dependent kinases (CDKs), thus preventing CDK-mediated hyperphosphorylation and inactivation of pRB and leading to G1-phase cell cycle arrest. p14ARF acts by directly binding to and promoting the degradation of Mdm2, thus leading to the stabilization of p53.39 Therefore, a single mutational event at the INK4a/ARF locus has the potential to disrupt both of these 2 key growth control pathways.
Xio et al.40 found deletion of this locus, by fluorescence in situ hybridization, in greater than 70% of MM. Homozygous deletion of the p16INK4a gene has been reported in 85% of MM cell lines and in 22% of primary tumor specimens examined.41, 42 It has been suggested that the greater frequency of p16 deletions in MM cell lines than in tumor samples may be either because 1) this alteration occurs in the culturing process and is an in vitro phenomenon or 2) the tumor specimens may be contaminated with a small population of nonmalignant cells making it difficult to detect genetic abnormalities.43 At the protein level it has been demonstrated that p16INK4A is abnormally expressed in all cell lines and all primary MM specimens examined.35 These findings also suggest that epigenetic mechanisms might play a role in silencing gene expression in cells with an absence of mutations in the INK4a/ARF locus. Indeed, Wong et al.44 detected methylation of p16INK4a in 10% of cell lines and 27% of primary tumors. Treatment of these methylated MM cell lines with an inhibitor of DNA methylation (5′-aza-2′deoxycytidine) resulted in re-expression of the p16INK4a protein. Wong et al.44 hypothesize that methylation of p16INK4a may be a possible therapeutic target in MM and suggest that responses to treatment of mesothelioma with the demethylating agent DHAC in clinical trials may in part be due to re-expression of p16INK4a. However, due to the nonspecific actions of demethylating agents it is difficult to link any specific gene to the clinical response.
Recent studies have shown loss of p16 to be associated with poor survival in MM patients. A microarray analysis of 99 MM specimens found homozygous deletion of p16 to be a significant independent adverse prognostic factor in pleural mesotheliomas, with a median survival of 10 months for p16-deleted cases vs 34 months for nondeleted cases (P = .001).45 This is consistent with a prior study by Borczuk et al.46 that identified loss of p16 immunoreactivity to be an independent predictor of poor survival in peritoneal mesotheliomas.
The high frequency of abnormalities in the INK4A/ARF locus in MM presents an attractive diagnostic marker or target for therapeutic intervention. It has been proposed that detection of homozygous deletion of this locus by FISH might be an efficient approach for improving the cytologic diagnosis of MM from body cavity effusions.47 Another study48 reported that mice treated with p16INK4a-based adenoviral vector gene therapy demonstrated prolonged survival and even a potential cure. In MM cell lines and mouse xenografts, re-expression of p16INK4a led to cell cycle arrest, inhibition of cell growth, and apoptosis.49
Similarly, we recently evaluated the potential therapeutic efficacy of p14ARF expressed in an adenoviral vector (Adp14ARF) in human mesothelioma cell lines that lack p14ARF expression.50 Overexpression of p14ARF resulted in increased levels of p53 and the p21WAF protein as well as dephosphorylation of the retinoblastoma protein. Infection with the Adp14ARF also led to cell cycle arrest, cell growth inhibition, and apoptotic cell death. p21WAF, a p53 transcriptional target, is the universal cyclin-dependent kinase inhibitor. p21WAF inhibits cdk-mediated phosphorylation of pRB, thus leading to cell cycle arrest at the G1-phase checkpoint. In this way, p14ARF can affect both the p53 and pRb pathways. p14ARF gene therapy, by restoring and increasing p53 activity, might also sensitize mesothelioma cells to ionized irradiation and chemotherapeutic agents.
In addition to abnormalities in the INK4a/ARF locus, SV40 is also a key candidate for the disruption of the p53 and pRb tumor suppressor pathways. SV40 is a powerful carcinogen that has been shown to transform human mesothelial cells in vitro and to induce mesothelioma development in animal models.3 The SV40 large T antigen (Tag) has been shown to bind and inactivate both p53 and pRB in human MM cells.51, 52
The Wnt signal transduction pathway plays an important role in the pathogenesis of MM. The Wnt pathway plays a critical role in cell fate determination, proliferation, and patterning during embryogenesis. Although this signaling is essential for normal developmental processes, aberrant activation of the Wnt pathway has been closely associated with tumorigenesis. Activation of the canonical Wnt pathway, via binding of the Wnt ligands to the frizzled transmembrane receptors, leads to the stabilization and accumulation of β-catenin in the cytoplasm. β-Catenin subsequently becomes translocated into the cell nucleus, where it interacts with TCF/LEF transcription factors to promote the expression of Wnt-responsive genes including the oncogenes c-myc and cyclin D (Fig. 2).
In a recent IHC study using serous effusions and pleural biopsies, Dai et al.53 found staining of β-catenin confined to the cell membrane in all normal and reactive nonneoplastic mesothelium. In contrast, they found reduced membranous staining and markedly increased nuclear and cytoplasmic staining of β-catenin in 26 of 33 advanced stage and 7 of 9 early stage MM. These findings are consistent with a similar study54 that found all MM tumors examined to be positive for cytoplasmic β-catenin and 19% to also have nuclear β-catenin, whereas no reactive mesothelial hyperplasia samples stained with either cytoplasmic or nuclear β-catenin. These findings strongly suggest that the Wnt-β-catenin pathway is abnormally activated in MM.
Dysregulation of β-catenin signaling by activating mutations in exon-3 of the β-catenin gene is an important event in the genesis of several human cancers but this has not been found in MM.54, 55 Instead, it appears that genes more upstream in the Wnt pathway, including extracellular signaling components, may play a more important role. Our group previously demonstrated that the Wnt signaling pathway is activated in MM through the overexpression of disheveled proteins. Additionally, we have recently found that endogenous extracellular Wnt antagonists, such as the secreted frizzled related proteins (sFRPs) and Wnt inhibitory factor-1 (Wif-1), are frequently down-regulated in MM cell lines and primary tumors.56, 57 A monoclonal antibody or siRNA directed against Wnt-1 and Wnt-2 induced apoptosis in cancer cells overexpressing Wnt. Furthermore, MM cells treated with these antibodies displayed down-regulation of key downstream Wnt signaling effectors such as disheveled and β-catenin.58, 59 These findings suggest that Wnt-targeted therapies may be useful in the treatment of MM.
The poor response of MM to conventional therapeutic agents is partly due to this tumor's resistance to apoptosis. The bcl-2 family of genes is known to play a major role in the intrinsic apoptotic pathway. Its members include both proapoptotic proteins such as Bax, Bak, and Bad and antiapoptotic proteins such as Bcl-2, Bcl-xL, and Mcl-1. The proapoptotic Bcl-2 family members are thought to induce apoptosis by promoting permeability of the mitochondrial membranes. This increased permeability results in the release of caspase activators such as cytochrome c and subsequently leads to the activation of downstream caspases that cause cellular demolition and apoptotic morphology.60 In contrast, antiapoptotic Bcl-2 and Bcl-xL are thought to block this mitochondrial permeabilization. Bcl-2, in particular, has been demonstrated to protect neoplastic cells from chemotherapy and radiation-induced apoptosis.61 Additionally, a low bax expression has been linked to higher resistance to chemotherapy and a poor prognosis in breast carcinoma.62 It has been hypothesized that by altering the ratio of expression of various proapoptotic and antiapoptotic bcl-2 family genes, apoptosis can either be induced or inhibited.63
Expression of bcl-2 is rare in MM, unlike in tumors such as breast and endometrial adenocarcinomas, which frequently display up-regulation of bcl-2.64–67 In contrast, elevated levels of Bcl-xL mRNA and protein have been detected in all mesothelioma cell lines and tumor samples examined.66, 68 Interestingly, Narasimhan et al.65 also found expression of proapoptotic Bax in all 14 of 14 mesothelioma cell lines, including those that were highly resistant to proapoptotic stimuli. This suggests that overexpression of Bcl-xL may be necessary to counteract the proapoptotic effect of Bax. Because uniform Bcl-xL overexpression in malignant mesothelioma could account for resistance to apoptosis even in the presence of wildtype p53, Bcl-xL may be a possible oncogenic candidate in this tumor type.
Pharmacologic inhibition of Bcl-xL expression by exposure to a histone deacetylase inhibitor, sodium butyrate (NaB), has been shown to lead to apoptotic cell death in malignant mesothelioma.69 Antisense oligonucleotides targeting the bcl-xL gene product have been shown to similarly facilitate apoptosis in mesothelioma.68, 70 Additionally, a recombinant adenoviral vector that can transduce human cells with the proapoptotic bax gene has been shown to effectively induce apoptosis in mesothelioma.71 Inhibition of bcl-xL displayed a synergistic effect when combined with the adenoviral proapoptotic gene therapy vectors AdBak and AdBax.63
Of interest, 1 recent study found that long-term recurrent exposure to asbestos affects human immune cells, leading to Bcl-2 enhancement in CD4+ lymphocytes in vitro.72 The same study found Bcl-2 expression in CD4+ peripheral blood lymphocytes to be significantly increased in MM patients compared with healthy volunteers (P = .0153) and asbestosis patients with no signs of malignancy (P = .129). The clinical significance of this increased Bcl-2 expression remains to be addressed.
Further investigation on the role of the bcl-2 family members in the pathogenesis of MM, especially in relation to improved therapeutic strategies, is needed.
Over the past decade there have been significant advances in our understanding of the molecular pathways involved in MM carcinogenesis. It is likely that the implementation of high-throughput proteomic and genomic technologies such as microarray analysis will further accelerate research in this field. Despite this continuing progress in our understanding of the biology of MM, progress in treatment of this fatal tumor has been slow and challenging, with most treatments having no survival advantage over palliative care.
Perhaps the greatest and most crucial challenge now is to translate our growing knowledge of MM pathogenesis into more effective ways to treat this devastating disease. MM is a complex tumor that results from accumulation over many years of multiple genetic alterations, many of which remain to be uncovered. Thus, it is overly simplistic to expect that any single “magic bullet” may reverse the malignant phenotype. Rather, elucidating the several key pathways involved will hopefully result in a rational treatment in which several molecularly targeted agents will be combined with effective chemotherapeutic regimes. Early diagnosis in high-risk populations might also be a desirable goal as treatment effectiveness has been shown to be much greater the earlier the clinical stage of the tumor at diagnosis. Timely integration of laboratory-based studies into standard clinical practice will require continued cooperation between basic scientists and clinical investigators working toward improved survival and quality of life in patients with this aggressive and fatal disease.
Supported by a grant from Supported by the Larry Hall Memorial Trust and National Institutes of Health Grant (RO1 CA 093708-01A3) (both to D.M.J.). A.Y.L. is supported by the UCSF Medical Student Research Committee.