Improved detection methods for diagnosis of asymptomatic malignant mesothelioma (MM) are essential for an early and reliable detection and treatment of this type of neoplastic disease. Thus, focus has been on finding tumor markers in the blood that can be used for noninvasive detection of MM. Ninety-four asbestos-exposed subjects defined at high risk, 22 patients with MM, and 54 healthy subjects were recruited for evaluation of the clinical significance of 8-hydroxy-2′-deoxyguanosine (8OHdG) in WBCs and plasma concentrations of soluble mesothelin-related peptides (SMRPs), angiogenic factors [platelet-derived growth factor β, hepatocyte growth factor, basic fibroblast growth factor, and vascular endothelial growth factor β (VEGFβ)], and matrix proteases [matrix metalloproteinase (MMP) 2, MMP9, tissue inhibitor of metalloproteinase (TIMP) 1, and TIMP2] for potential early detection of MM. The area under receiver operating characteristic (ROC) curves indicate that 8OHdG levels can discriminate asbestos-exposed subjects from healthy controls but not from MM patients. Significant area under ROC curve values were found for SMRPs, discriminating asbestos-exposed subjects from MM patients but not from healthy controls. Except for platelet-derived growth factor β, the hepatocyte growth factor, basic fibroblast growth factor, and VEGFβ can significantly differentiate high-risk individuals from healthy control and cancer groups. No diagnostic value was observed for MMP2, MMP9, TIMP1, and TIMP2. In addition to the diagnostic performance defined by the ROC analysis, the sensitivity and specificity results of markers with clinical significance were calculated at defined cutoffs. The combination of 8OHdG, VEGFβ, and SMRPs best distinguished the individual groups, suggesting a potential indicator of early and advanced MM cancers. The combination of blood biomarkers and radiographic findings could be used to stratify the risk of mesothelioma in asbestos-exposed populations. (Cancer Epidemiol Biomarkers Prev 2008;17(1):163–70)

Malignant mesothelioma (MM) is an aggressive tumor of serosal cavities, which is resistant to conventional therapy, surgery, or radiation. MM remains a universally fatal disease of increasing incidence worldwide (1). Patient survival from presentation is <12 months (2).

Occupational hazard, particularly asbestos exposure, is the main factor involved in MM pathogenesis (3, 4). Asbestos-induced cell damage is mediated to some extent by iron-catalyzed formation of toxic oxygen radicals (5, 6), which induce DNA strand breaks (7) and oxidant-induced base modifications. 8-Hydroxy-2′-deoxyguanosine (8OHdG), a major product of such oxidative damage (8), causes G→T and A→C transversions (9, 10). These substitutions have been reported as the sites of spontaneous oncogene expression and may be largely responsible for the onset of carcinogenesis and cell proliferation, ultimately leading to cancer manifestation (8-11). Most of the necessary mutations occur early during cancer development, also resulting in processes such as chronic inflammation, together providing the environment to expand and select malignant clones.

Tumor growth and metastasis are angiogenesis-dependent events. Like other tumors, MMs induce the vascular stroma to grow (12). Current results indicate that the switch to the angiogenic phenotype depends on a net balance between positive and negative angiogenic factors released by the tumor cells (13). To date, many angiogenic factors, such as the hepatocyte growth factor (HGF), the fibroblast growth factor family members (FGF), the vascular endothelial growth factor (VEGF), or the platelet-derived growth factor (PDGF), have been identified and shown to be produced by a variety of different tumor cells, including those of MM (14-17). Functionally active growth factors induce mesothelioma cell migration and matrix metalloproteinase (MMP) production (14, 18). MMPs belong to the group of extracellular matrix degradation enzymes. The balance of secreted MMPs and their specific inhibitors (TIMPs) plays an important role in maintaining connective tissue homeostasis in normal tissue (19). Activated MMP2 and MMP9 are described to have an effect in MM carcinogenesis (20). Thus, determination of these mediators in blood plasma could be used for noninvasive early diagnosis of MM.

Recently, soluble mesothelin-related peptides (SMRPs) have been suggested as a promising biomarker for MM (21-23). Mesothelin is normally expressed at low levels in mesothelial cells. Its overexpression was observed in cancers including MM (21-24). SMRPs can be detected in serum and are highly increased in the blood of patients with MM (21-23) and ovarian tumors (23).

In the present study, the levels of 8OHdG in circulating WBCs, plasma concentrations of SMRPs, a panel of angiogenic factors [PDGFβ, HGF, basic FGF (bFGF), and VEGFβ], as well as MMPs (MMP2 and MMP9) and their inhibitors (TIMP1 and TIMP2) were evaluated in a cohort of asbestos-exposed, high-risk subjects, in patients with MM, and in healthy controls. The potential clinical relevance of these markers was evaluated for detection of the cancer risk and thus may help in the prevention and management of the disease.

Study Groups

Three groups of subjects were included in the study: 94 asbestos-exposed subjects (high risk), 54 aged-matched controls (no exposure to asbestos), and 22 patients with MM. All subjects filled in a questionnaire including their informed consent and provided a blood sample. The study was carried out according to the Helsinki Declaration and approved by the Ethical Committee of the Polytechnic University of Marche. The demographic characteristics are shown in Table 1. WBC 8OHdG as well as plasma angiogenic factors (VEGFβ, HGF, bFGF, and PDGFβ), MMPs (MMP2 and MMP9) and their inhibitors (TIMP1 and TIMP2), and SMRP levels were determined in the patients' samples.

Table 1.

Demographic characteristics of recruited subjects

CharacteristicsControl group (n = 54)Asbestos-exposed group (n = 94)MM group (n = 22)
Age (y) 63.0 ± 7.8 61.3 ± 7.4 68.7 ± 7.9 
Sex (M/F) 33/21 94/0 18/4 
Smoking history    
    Nonsmokers 41 (80%) 34 (36%) 6 (27%) 
    Ex-smokers 5 (9%) 49 (52%) 11 (50%) 
    Smokers 8 (15%) 11 (12%) 5 (23%) 
CharacteristicsControl group (n = 54)Asbestos-exposed group (n = 94)MM group (n = 22)
Age (y) 63.0 ± 7.8 61.3 ± 7.4 68.7 ± 7.9 
Sex (M/F) 33/21 94/0 18/4 
Smoking history    
    Nonsmokers 41 (80%) 34 (36%) 6 (27%) 
    Ex-smokers 5 (9%) 49 (52%) 11 (50%) 
    Smokers 8 (15%) 11 (12%) 5 (23%) 

Subjects with Asbestos Exposure

To recruit asbestos-exposed subjects, 500 invitation letters were sent to subjects who worked or were working at the shipbuilding industry. From November 2004 to September 2005, 94 male subjects (response rate, 19%; mean age, 61.3 ± 7.4 years) with history of asbestos exposure were enrolled at the Institute of Occupational Medicine, Polytechnic University of Marche (Ancona, Italy). The subjects had been exposed to asbestos dust on average for more than 20 years. Smokers (12%), ex-smokers (52%), and nonsmokers (36%) were examined. Each subject underwent lung function analysis, chest radiography, and high-resolution computed tomography. Evidence of asbestos-related diseases (fibrosis and pleural plaques) was found in 28 (24%) subjects.

Subjects with No Exposure to Asbestos

The control group consisted of 54 aged-matched subjects (mean age, 63.0 ± 7.8 years; 33 males and 21 females) recruited from November 2004 to January 2007 (response rate, 70%). The subjects were undergoing screening radiography for chemoprevention at the Pneumology Clinic of the University Hospital of Ancona (Ancona, Italy). None of them had ever been exposed to asbestos as documented by their occupational histories. All subjects had normal chest radiographs.

Subjects with MM

Blood samples were collected from 22 patients (mean age, 68.7 ± 7.9; 18 males and 4 females) diagnosed for MM, who were recruited, from November 2004 to January 2007, at the Oncology Clinic of the University Hospital of Ancona, with a response rate of 90%. Exclusion criteria were the presence or suspicion of any infectious disease and previous radical surgery, radiotherapy, as well as chemotherapy for MM. Pathologic diagnosis was done on pleural biopsies obtained by thoracoscopy or thoracotomy. Tumors were classified as epithelial in 11, mixed in 5, and sarcomatoid in 6 patients.

WBC 8OHdG Analysis

The levels of 8OHdG were determined in WBCs using the fluorometric OxyDNA assay kit (Calbiochem) according to the manufacturer's instructions. Whole blood (7 mL) collected into EDTA tubes was immediately centrifuged at 1,500 × g (20°C, 15 min). The buffy coat was removed, placed in a 15 mL Falcon tube, and resuspended in 4 mL PBS. The suspension was then layered onto 4 mL Lympholyte-H (Cedarlane) and centrifuged at 1,000 × g (20°C, 30 min). The resulting cloudy layer was collected and placed in a 15 mL Falcon tube filled with PBS and centrifuged at 230 × g (20°C, 5 min). After removing the supernatant, the pellet was resuspended in 500 μL PBS/500 μL of 2% paraformaldehyde and incubated on ice for 15 min. Two washes in PBS were then made, and after centrifugation, the cells were resuspended in 50 μL of the blocking solution for 1 h at 37°C. The blocking solution was removed and 100 μL of the fluorescent probe for 8OHdG were added and the samples were incubated for 1 h. After washing, the cells were resuspended in the fluorescence-activated cell sorting buffer and analyzed by flow cytometry (FACSCalibur, BD PharMingen Italy). The results are expressed as mean fluorescence intensity (arbitrary units).

SMRP Assay

Plasma levels of SMRPs were determined using a sandwich-type ELISA assay (Mesomark, Schering) according to the manufacturer's instructions and the results are expressed in nmol/L. Briefly, 100 μL of standard and plasma samples (1:100 dilution) were added to a 96-well microtiter plate coated with specific antibodies against SMRPs and then incubated at room temperature for 60 min. After washing, the plate was incubated with a secondary horseradish peroxidase–conjugated antibody. The detection process involved addition of 100 μL of the substrate (3,3′,5,5′-tetramethylbenzidine) to all wells and the absorbance was read at 405 nm using an ELISA plate reader (Sunrise, Tecan). Concentrations of SMRPs were extrapolated from the standard curve and expressed in nmol/L.

Human Angiogenesis and MMP Arrays

Human angiogenesis (PDGFβ, HGF, bFGF, and VEGFβ) and human MMP (MMP2, MMP9, TIMP1, and TIMP2) arrays were analyzed by multiplex sandwich ELISA (SearchLight, Pierce Biotechnology) according to the manufacturer's instructions. Each well of the microplate was prespotted with target protein-specific antibodies. These antibodies capture the specific target protein in the standard and plasma samples added to the plate (50 μL of 1:5 diluted plasma). Unbound proteins were washed away and biotinylated detecting antibodies were added. After washing, antibody streptavidin-horseradish peroxidase was used for detection. Each sample was tested in duplicate and the results are expressed in ng/mL.

Statistical Analysis

All data are presented as mean ± SD and within a minimum-maximum range. Comparisons between groups were done using Mann-Whitney U test for unpaired samples and Kruskal-Wallis analysis for multiple comparisons, and the rank correlation coefficient according to Spearman was used. Multiple regression analysis was used to estimate the influence of independent variables such as age, smoking, fibrotic changes and pleural plaques, as well as duration of exposure to asbestos on the markers studied (dependent variable). Receiver operating characteristic (ROC) curves were plotted to quantify the marker performance. ROC curves correlate the sensitivity of a diagnostic test within the entire range of the possible false-positive rate. The area under the ROC curve (AUC) indicates the average sensitivity of a marker over the entire ROC curve. The best statistical “cutoff” was calculated by minimizing the distance between the point with specificity = 1 and sensitivity = 1 and the intercept on the ROC curve. AUC values are reported with their 95% confidence intervals. The usage of biomarker combinations in MM diagnosis was assayed with logistic regression. This method was applied to specify a probability, which depends on several factors. Statistical calculations were done using the Statistical Package for the Social Sciences statistical package version 12.0F (SPSS). Statistical differences of at least P < 0.05 were considered statistically significant.

Concentrations of Biomarkers in the Groups

The plasma biomarker concentrations detected in the three groups are summarized in Table 2. High levels of 8OHdG were observed in WBCs of the asbestos-exposed group and MM group compared with subjects without exposure to asbestos (healthy controls). Plasma SMRP concentrations of asbestos-exposed subjects were not significantly different from the age-matched subjects. Conversely, patients with MM showed high levels of SMRPs when compared with both the asbestos-exposed group and the control group. As listed in Table 2, the mean plasma levels of PDGFβ (P < 0.001), HGF (P < 0.0001), bFGF (P < 0.0001), and VEGFβ (P < 0.0001) were significantly increased in the asbestos-exposed group and more in the MM group compared with the control group. The levels of MMP2, MMP9, TIMP1, and TIMP2 as well as the MMPs to TIMPs ratio were not different between groups (data not shown). None of the biomarkers was influenced by sex, age, smoking habits, and the presence or absence of pleural plaques and lung fibrosis (data not shown).

Table 2.

Levels of plasma biomarkers

Control group (n = 54)Asbestos-exposed group (n = 118)MM group (n = 22)
SMRPs (nmol/L) 1.1 ± 1.5 1.7 ± 6.0 10.9 ± 16.2*, 
    Min-max 0.2-7.8 0.2-44.1 1.0-57.6 
8-OHdG (AU) 5.1 ± 2.1 8.4 ± 3.8 9.2 ± 4.7 
    Min-max 2.1-10.0 2.7-26.4 3.3-17.8 
MMP2 (ng/mL) 601.5 ± 31.8 475.7 ± 24.5 487.7 ± 26.4 
    Min-max 233-1,690 149-1,120 112-1,848 
MMP9 (ng/mL) 244.0 ± 243.5 163.2 ± 132.0 220.6 ± 250.7 
    Min-max 12-1,022 22-582 10-1,008 
TIMP1 (ng/mL) 318.7 ± 250.2 221.5 ± 103.4 243.5 ± 60.6 
    Min-max 91-1,004 53-655 161-300 
TIMP2 (ng/mL) 276.3 ± 431.0 203.0 ± 255.1 114.5 ± 49.6 
    Min-max 68-1,765 38-1,479 39-200 
PDGFβ (ng/mL) 19.1 ± 19.0 21.3 ± 21.4 47.4 ± 30.1*, 
    Min-max 0.5-52 0.3-81 3.0-106 
HGF (ng/mL) 5.0 ± 3.5 7.4 ± 5.1 15.0 ± 8.5*, 
    Min-max 1.4-13.7 1.4-21.3 2.2-33.0 
bFGF (ng/mL) 0.6 ± 0.4 0.8 ± 0.4 1.4 ± 0.8*, 
    Min-max 0.1-1.5 0.1-2.1 0.3-3.2 
VEGFβ (ng/mL) 0.5 ± 0.2 0.7 ± 0.3 1.7 ± 1.7*, 
    Min-max 0.1-1.0 0.2-2.2 0.1-6.0 
Control group (n = 54)Asbestos-exposed group (n = 118)MM group (n = 22)
SMRPs (nmol/L) 1.1 ± 1.5 1.7 ± 6.0 10.9 ± 16.2*, 
    Min-max 0.2-7.8 0.2-44.1 1.0-57.6 
8-OHdG (AU) 5.1 ± 2.1 8.4 ± 3.8 9.2 ± 4.7 
    Min-max 2.1-10.0 2.7-26.4 3.3-17.8 
MMP2 (ng/mL) 601.5 ± 31.8 475.7 ± 24.5 487.7 ± 26.4 
    Min-max 233-1,690 149-1,120 112-1,848 
MMP9 (ng/mL) 244.0 ± 243.5 163.2 ± 132.0 220.6 ± 250.7 
    Min-max 12-1,022 22-582 10-1,008 
TIMP1 (ng/mL) 318.7 ± 250.2 221.5 ± 103.4 243.5 ± 60.6 
    Min-max 91-1,004 53-655 161-300 
TIMP2 (ng/mL) 276.3 ± 431.0 203.0 ± 255.1 114.5 ± 49.6 
    Min-max 68-1,765 38-1,479 39-200 
PDGFβ (ng/mL) 19.1 ± 19.0 21.3 ± 21.4 47.4 ± 30.1*, 
    Min-max 0.5-52 0.3-81 3.0-106 
HGF (ng/mL) 5.0 ± 3.5 7.4 ± 5.1 15.0 ± 8.5*, 
    Min-max 1.4-13.7 1.4-21.3 2.2-33.0 
bFGF (ng/mL) 0.6 ± 0.4 0.8 ± 0.4 1.4 ± 0.8*, 
    Min-max 0.1-1.5 0.1-2.1 0.3-3.2 
VEGFβ (ng/mL) 0.5 ± 0.2 0.7 ± 0.3 1.7 ± 1.7*, 
    Min-max 0.1-1.0 0.2-2.2 0.1-6.0 

NOTE: Values are presented as mean ± SD and within a minimum-maximum range. Statistical differences among the groups were calculated by the nonparametric Kruskal-Wallis test.

*

P = 0.00005, MM group versus asbestos-exposed group.

P = 0.00005, MM group versus control group.

P = 0.00005, asbestos-exposed group versus control group.

Biomarker Correlations

Table 3 presents the Spearman correlation coefficients among the analyzed biomarkers. A positive correlation was observed between MMP2 and MMP9 and between their inhibitors TIMP1 and TIMP2. The angiogenic factors (PDGFβ, HGF, bFGF, and VEGFβ) positively correlated with each other. MMP2 and the inhibitor TIMP2 negatively correlated with the PDGFβ, HGF, and bFGF levels. Only a negative correlation was found between MMP9 and PDGFβ. Notably, no correlations were found between 8OHdG and any of the markers tested. The SMRPs positively correlated with TIMP1 and with all angiogenic factors evaluated.

Table 3.

Correlation coefficients according to Spearman between biomarkers

SMRPs8OHdGMMP2MMP9TIMP1TIMP2PDGFβHGFbFGFVEGFβ
SMRPs 1.00 −0.06 −0.20 −0.02 0.25* −0.13 0.47 0.42 0.28* 0.23* 
8OHdG  1.00 −0.17 0.15 −0.08 0.03 −0.01 −0.06 −0.02 0.05 
MMP2   1.00 0.44 0.51 0.76 −0.39 −0.45 −0.44 0.16 
MMP9    1.00 0.56 0.45 −0.26* −0.12 −0.20 0.04 
TIMP1     1.00 0.62 −0.01 0.00 −0.08 0.09 
TIMP2      1.00 −0.23* −0.29* −0.27* −0.06 
PDGFβ       1.00 0.81 0.78 0.46 
HGF        1.00 0.90 0.65 
bFGF         1.00 0.76 
VEGFβ          1.00 
SMRPs8OHdGMMP2MMP9TIMP1TIMP2PDGFβHGFbFGFVEGFβ
SMRPs 1.00 −0.06 −0.20 −0.02 0.25* −0.13 0.47 0.42 0.28* 0.23* 
8OHdG  1.00 −0.17 0.15 −0.08 0.03 −0.01 −0.06 −0.02 0.05 
MMP2   1.00 0.44 0.51 0.76 −0.39 −0.45 −0.44 0.16 
MMP9    1.00 0.56 0.45 −0.26* −0.12 −0.20 0.04 
TIMP1     1.00 0.62 −0.01 0.00 −0.08 0.09 
TIMP2      1.00 −0.23* −0.29* −0.27* −0.06 
PDGFβ       1.00 0.81 0.78 0.46 
HGF        1.00 0.90 0.65 
bFGF         1.00 0.76 
VEGFβ          1.00 

NOTE: Correlation coefficients were determined according to Spearman test. The levels of 8OHdG were expressed as arbitrary units, SMRPs as nmol/L, and MMP2, MMP9, TIMP1, TIMP2, PDGFβ, HGF, bFGF, and VEGFβ as ng/mL. Correlations with P < 0.05 were considered statistically significant.

Abbreviation: AU, arbitrary units.

*

P < 0.05.

P < 0.01.

Diagnostic Validity of the Single and Combined Markers

The ROC curves were generated to analyze the diagnostic values of individual markers (Figs. 1 and 2). SMRPs represent a marker with the highest AUC, allowing to discriminate between patients with MM and both the control subjects (AUC = 0.920 ± 0.030; P = 0.0001) and the asbestos-exposed subjects (AUC = 0.927 ± 0.022; P = 0.0001). An AUC curve that did not reach statistical significance was observed by comparing the asbestos-exposed subjects with the control subjects (AUC = 0.459 ± 0.042; P = 0.502). The WBC 8OHdG level was found to be appropriated to evaluate the asbestos exposure. ROC analyses comparing the subjects with those without asbestos exposure showed an AUC of 0.775 ± 0.037 (P = 0.001). The AUC for discriminating between patients affected by MM and the control age-matched subjects was 0.788 ± 0.090 (P = 0.004). An AUC not statistically significant was found between asbestos-exposed subjects and patients with MM (AUC = 0.556 ± 0.110; P = 0.536; Fig. 1).

Figure 1.

ROC curves for 8OHdG and SMRPs. The AUCs were determined for 8OHdG and SMRPs, discriminating age-matched control subjects (Ctrl) from asbestos-exposed subjects (Exp), asbestos-exposed subjects from MM patients, and age-matched control subjects from MM patients. Differences with P < 0.05 were considered statistically significant.

Figure 1.

ROC curves for 8OHdG and SMRPs. The AUCs were determined for 8OHdG and SMRPs, discriminating age-matched control subjects (Ctrl) from asbestos-exposed subjects (Exp), asbestos-exposed subjects from MM patients, and age-matched control subjects from MM patients. Differences with P < 0.05 were considered statistically significant.

Close modal
Figure 2.

ROC curves for PDGFβ, HGF, bFGF, and VEGFβ. The AUCs were determined for PDGFβ, HGF, bFGF, and VEGFβ, discriminating age-matched control subjects (Ctrl) from asbestos-exposed subjects (Exp), asbestos-exposed subjects from MM patients, and age-matched control subjects from MM patients. Differences with P < 0.05 were considered statistically significant.

Figure 2.

ROC curves for PDGFβ, HGF, bFGF, and VEGFβ. The AUCs were determined for PDGFβ, HGF, bFGF, and VEGFβ, discriminating age-matched control subjects (Ctrl) from asbestos-exposed subjects (Exp), asbestos-exposed subjects from MM patients, and age-matched control subjects from MM patients. Differences with P < 0.05 were considered statistically significant.

Close modal

No diagnostic value was observed for MMP2, MMP9, TIMP1, and TIMP2. The AUC values were not statistically significant to differentiate the three groups (data not shown). As observed for SMRPs, the PDGFβ levels distinguished MM patients from both the control subjects (AUC = 0.783 ± 0.065; P = 0.001) and the asbestos-exposed subjects (AUC = 0.765 ± 0.061; P = 0.001) but not the asbestos-exposed subjects from the control group (AUC = 0.534 ± 0.071; P = 0.632). Conversely, HGF, bFGF, and VEGFβ significantly discriminated the asbestos-exposed subjects from the controls and the MM patients and the latter from the controls (Fig. 2). In addition to the diagnostic performance defined by the AUCs of the ROC analyzed, the sensitivity and specificity results of selected markers were calculated at defined cutoffs. Table 4 presents the diagnostic sensitivity and specificity (90% limits) of the SMRPs, 8OHdG, PDGFβ, HGF, bFGF, and VEGFβ to distinguish between the healthy persons and the asbestos-exposed subjects, the asbestos-exposed subjects and the MM patients, and the healthy persons and the MM patients.

Table 4.

Diagnostic sensitivities and specificities of MMPs, TIMPs, and PDGFβ, HGF, bFGF, and VEGFβ to distinguish between healthy persons and asbestos-exposed subjects, asbestos-exposed subjects and MM cancer patients, and healthy persons and MM cancer patients

MarkerCtrl vs Exp
Exp vs MM
Ctrl vs MM
ng/mLSensitivity (%)Specificity (%)ng/mLSensitivity (%)Specificity (%)ng/mLSensitivity (%)Specificity (%)
SMRPs (nmol/L) 0.2 90 1.0 90 79 1.0 90 78 
 1.9 90 1.7 72 90 1.9 68 90 
8OHdG (AU) 4.4 90 48 3.9 90 3.9 90 37 
 7.9 43 90 12.7 20 90 7.9 60 90 
PDGFβ (ng/mL) 0.8 90 18 4.3 90 38 3.3 90 45 
 45.1 14 90 50.2 45 90 45.1 45 90 
HGF (ng/mL) 1.8 90 21 4.0 90 36 4.2 90 48 
 9.3 32 90 14.7 35 90 9.3 75 90 
bFGF (ng/mL) 0.3 90 31 0.5 90 34 0.5 90 52 
 1.1 30 90 1.4 45 90 1.1 50 90 
VEGFβ (ng/mL) 0.3 90 34 0.3 90 0.3 90 17 
 0.7 34 90 1.0 60 90 0.7 70 90 
MarkerCtrl vs Exp
Exp vs MM
Ctrl vs MM
ng/mLSensitivity (%)Specificity (%)ng/mLSensitivity (%)Specificity (%)ng/mLSensitivity (%)Specificity (%)
SMRPs (nmol/L) 0.2 90 1.0 90 79 1.0 90 78 
 1.9 90 1.7 72 90 1.9 68 90 
8OHdG (AU) 4.4 90 48 3.9 90 3.9 90 37 
 7.9 43 90 12.7 20 90 7.9 60 90 
PDGFβ (ng/mL) 0.8 90 18 4.3 90 38 3.3 90 45 
 45.1 14 90 50.2 45 90 45.1 45 90 
HGF (ng/mL) 1.8 90 21 4.0 90 36 4.2 90 48 
 9.3 32 90 14.7 35 90 9.3 75 90 
bFGF (ng/mL) 0.3 90 31 0.5 90 34 0.5 90 52 
 1.1 30 90 1.4 45 90 1.1 50 90 
VEGFβ (ng/mL) 0.3 90 34 0.3 90 0.3 90 17 
 0.7 34 90 1.0 60 90 0.7 70 90 

NOTE: Data result from ROC analysis done with 54 control subjects, 94 asbestos-exposed subjects, and 22 MM patients. The cutoffs correspond to the values at 90% sensitivity and specificity as indicated.

Abbreviations: Ctrl, control subjects; Exp, asbestos-exposed subjects.

To evaluate whether a marker combination may increase the predictive value for early detection of MM, the logistic regression analysis was done. No increase in the accuracy of determining the disease was observed by the Wald test. The SMRPs alone highly discriminate the MM patients from the healthy controls and the asbestos-exposed subjects. Additionally, the logistic regression analysis revealed that VEGFβ can increase the discriminative power of 8OHdG in determining the subjects with the risk to develop the disease (control group versus asbestos-exposed group; P = 0.0001, Wald test). The probability of the risk to develop the disease can be calculated by the formula P = exp (−6.6 + 1.698 X1 + 0.001358 X2) / 1 + exp (−6.6 + 1.698 X1 + 0.001358 X2), with X1 = 8OHdG WBC level and X2 = VEGFβ plasma level. Accordingly, the AUC increases from 0.775 ± 0.037 for 8OHdG and 0.714 ± 0.062 for VEGFβ to 0.925 ± 0.035 for the 8OHdG/VEGFβ combination. The increase in sensitivity and specificity is reflected by the ROC curves in Fig. 3A. For better visualization, dot plots are shown for the SMRPs and the 8OHdG/VEGFβ combination (Fig. 3B).

Figure 3.

ROC curves of 8OHdG and VEGFβ and combination of 8OHdG and VEGFβ as a marker to predict MM. A. The AUCs were determined for 8OHdG and VEGFβ alone and in combination, discriminating age-matched control subjects (Ctrl) from asbestos-exposed subjects (Exp). B. Scatter plot of SMRPs and combination of 8OHdG and VEGFβ, discriminating asbestos-exposed subjects (open circle dots) from age-matched control subjects (close circle dots).

Figure 3.

ROC curves of 8OHdG and VEGFβ and combination of 8OHdG and VEGFβ as a marker to predict MM. A. The AUCs were determined for 8OHdG and VEGFβ alone and in combination, discriminating age-matched control subjects (Ctrl) from asbestos-exposed subjects (Exp). B. Scatter plot of SMRPs and combination of 8OHdG and VEGFβ, discriminating asbestos-exposed subjects (open circle dots) from age-matched control subjects (close circle dots).

Close modal

Biomarker or a combination of several biomarkers that could predict the development of MM or detect the disease in its early stages in the population with high risk would be of paramount importance, particularly given the fact that there is as yet no cure for MM. In the present study, levels of the DNA adduct 8OHdG and factors involved in tumor growth (PDGFβ, HGF, bFGF, and VEGFβ), progression (MMP2, MMP9, TIMP1, and TIMP2), and cell transformation (SMRPs) were measured in healthy subjects, in asbestos-exposed subjects defined as at high risk, and in patients with MM.

8OHdG is an indicator of oxidative DNA damage induced by reactive oxygen species (11). It has been widely used as a biomarker for detecting oxidative stress and oxidative DNA damage in both animal and human studies (25). We found that asbestos-exposed subjects and patients with MM showed comparably high WBC 8OHdG levels differing significantly from those in age-matched controls. A multiple regression analysis revealed that the age, smoking status, and fibrotic changes and pleural plaques were not the most important factors influencing the 8OHdG levels. The increased generation of 8OHdG indicates that high levels of reactive oxygen species are produced in the WBCs of subjects exposed to asbestos. It has been proposed that reactive oxygen species are critical for the development of asbestos-related diseases (7, 8, 10), and oxidative damage to the DNA of WBCs may be induced as a response to increased oxidative stress in the pleural surface of subjects chronically exposed to asbestos (26, 27).

To evaluate whether the 8OHdG levels would be useful in predicting MM in asbestos workers, the ROC analysis was done. We found that the biomarker 8OHdG significantly discriminated the asbestos-exposed subjects from the age-matched controls but not from MM patients (cfFig. 1). It is noteworthy that the 8OHdG levels were not evaluated in the target (mesothelial) cells but in the surrogate cells (WBCs). Thus, analysis of 8OHdG provides information only about the systemic status. The suitability of measuring 8OHdG as a biomarker of oxidative DNA damage depends on a variety of variables, which affect the interpretation of the data. These variables include the steady state between the mature and the newly differentiated or dying WBCs, DNA repair, and cell division and turnover (25). The levels of 8OHdG found in WBCs depend not only on the life span of the cells but also on the recovery of these adducts and individual blood count variability. The value of the 8OHdG levels for predicting cancer on an individual basis is therefore questionable. However, in agreement with other authors (26, 27), our results support the notion that the biomarker 8OHdG detects oxidative DNA damage in humans caused by exposure to asbestos fibers, which are involved in the etiology of MM, but they cannot be used to discriminate between asbestos-exposed individuals with and without MM.

Recently, plasma SMRPs have been proposed as a suitable marker for MM diagnosis (21-23). The release of SMRPs into blood is linked to cell transformation; thus, it could be used as a diagnostic biomarker of MM. In our study, MM patients showed higher plasma SMRP levels relative to the asbestos-exposed and the control subjects. The ROC analysis revealed that the SMRP levels can discriminate MM patients from both the asbestos-exposed and the asbestos-unexposed subjects but do not discriminate the asbestos-exposed individuals from the age-matched controls (cfFig. 1). Thus, the level of SMRPs in plasma can be proposed as a biomarker suitable for diagnosis of existing MM but not to predict the disease.

Differently as found for 8OHdG and SMRPs, the growth factors HGF, bFGF, and VEGFβ can significantly differentiate the high-risk individuals from the healthy controls and the cancer group (cfFig. 2). Such an observation may be important when determining the risk status of individuals with no physical manifestation of the disease.

Using the statistical ROC program, we calculated the different sensitivities and specificities of markers with clinical significance (cfFigs. 1 and 2; Table 4). The best indicator of MM was SMRPs with high sensitivity and specificity. Although lacking such high sensitivity and specificity, the levels of 8OHdG, HGF, bFGF, and VEGFβ alone can distinguish high-risk subjects from healthy persons and MM patients. We found that combination of the “exposure” marker 8OHdG with the growth factor VEGFβ highly increased the sensibility and specificity to discriminate the high-risk populations from the healthy controls (cfFig. 3A and B). We found a significantly increased discriminative capability by the use of a combination of 8OHdG and VEGFβ compared with the use of single variables. Positive correlation between SMRPs and plasma growth factors suggests involvement of growth factors in the development, growth, and progression of MM (cfTable 3). VEGFβ is of particular interest because reduction/blockage of members of the VEGF system has been suggested to be of therapeutic value in MM (28, 29).

Dot plots for marker combinations were used to evaluate the best marker combinations to distinguish the high-risk population from the healthy persons and the MM patients. The combination of SMRPs with 8OHdG and VEGFβ was found to be the best to distinguish the individual groups, suggesting a potential diagnostic indicator for patients in the early stages of MM. The importance of the changes in the levels of 8OHdG and SMRPs as well as growth factors in MM has been previously described in the literature, but here, we have for the first time evaluated their clinical relevance for their potential use for both diagnostic and screening purposes. However, the value of our biomarkers as indicators of both prediction and clinical presentation of MM needs to be validated in prospective studies in larger subject populations in which the exposed subjects will be followed for an adequate period of time. Accordingly, a longitudinal study on our asbestos-exposed group is under way.

In conclusion, the combination of blood biomarkers and radiographic findings could be used to stratify the risk of MM in populations with exposure to asbestos; close surveillance might be indicated in workers with long history of exposure, pleural plaques, fibrosis, and elevated levels of plasma biomarkers. Thus, we propose a novel combination of markers that may be useful for early diagnosis of MM, whereby improving the outcome of the ensuing therapy with novel, promising drugs (17, 30, 31). The importance of our findings may be reconciled with the currently very grim prognosis for mesothelioma patients.

Grant support: Regione Marche (Italy) and Polytechnic University of Marche (Ancona, Italy) research grant. J. Neuzil was supported by Dust Diseases Board of Australia and the Grant Agency of the Academy of Sciences of the Czech Republic.

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.

Note: M. Amati and M. Tomasetti contributed equally to this work.

1
Robinson BW, Musk AW, Lake RA. Malignant mesothelioma.
Lancet
2005
;
366
:
397
–408.
2
Martino D, Pass HI. Integration of multimodality approaches in the management of malignant pleural mesothelioma.
Clin Lung Cancer
2004
;
5
:
290
–8.
3
Pass HI, Vogelzang N, Hahn S, Carbone M. Malignant pleural mesothelioma.
Curr Probl Cancer
2004
;
28
:
93
–174.
4
Roggli VL. Environmental asbestos contamination: what are the risks?
Chest
2007
;
131
:
336
–8.
5
Kamp DW, Graceffa P, Pryor WA, Weitzman SA. The role of free radicals in asbestos-induced disease.
Free Radic Biol Med
1992
;
12
:
293
–315.
6
Kahlos K, Pitkanen S, Hassinen I, Linnainmaa K, Kinnula VL. Generation of reactive oxygen species by human mesothelioma cells.
Br J Cancer
1999
;
80
:
25
–31.
7
Kamp DW, Israbian VA, Preusen SE, Zang CX, Weitzman SA. Asbestos causes DNA strand breaks in cultured pulmonary epithelial cells: role of iron-catalyzed free radicals.
Am J Physiol
1995
;
268
:
471
–80.
8
Takeuchi T, Morimoto K. Crocidolite asbestos increased 8-hydroxydeoxyguanosine levels in cellular DNA of a human promyelocytic leukaemia cell line, HL-60.
Carcinogenesis
1994
;
15
:
635
–9.
9
Moriya M. Single-stranded shuttle phagemid for mutagenesis studies in mammalian cells: 8-oxoguanine in DNA induced targeted G·C→T·A transversions in simian kidney cells.
Proc Natl Acad Sci U S A
1993
;
90
:
1122
–6.
10
Unfried K, Schurkes C, Abel J. Distinct spectrum of mutations induced by crocidolite asbestos: clue for 8-hydroxydeoxyguanosine-dependent mutagenesis in vivo.
Cancer Res
2002
;
62
:
99
–104.
11
Loft S, Paulsen HE. Cancer risk and oxidative DNA damage in man.
J Mol Med
1996
;
74
:
297
–312.
12
Antony VB, Hott JW, Godbey SW, Holm K. Angiogenesis in malignant mesotheliomas: role of mesothelial cell-derived IL-8.
Chest
1996
;
109
:
21
–2S.
13
Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis.
Cell
1996
;
86
:
353
–64.
14
Harvey P, Clark IM, Jaurand MC, Warn RM, Edwards DR. Hepatocyte growth factor/scatter factor enhances the invasion of mesothelioma cell lines and the expression of matrix metalloproteinases.
Br J Cancer
2000
;
83
:
1147
–53.
15
Ferrara N. Molecular and biological properties of vascular endothelial growth factor.
J Mol Med
1999
;
77
:
527
–43.
16
Filiberti R, Marroni P, Neri M, et al. Serum PDGF-AB in pleural mesothelioma.
Tumour Biol
2005
;
26
:
221
–6.
17
Stapelberg M, Gellert N, Swettenham E, et al. α-Tocopheryl succinate inhibits malignant mesothelioma by disrupting the FGF autocrine loop: the role of oxidative stress.
J Biol Chem
2005
;
280
:
25369
–76.
18
Liu Z, Klominek J. Regulation of matrix metalloprotease activity in malignant mesothelioma cell lines by growth factors.
Thorax
2003
;
58
:
198
–203.
19
Chen WT. Membrane proteases: roles in tissue remodelling and tumour invasion.
Curr Opin Cell Biol
1992
;
4
:
802
–9.
20
Zhiwen L, Klominek J. Inhibition of proliferation, and matrix metalloproteinase production in malignant mesothelioma cells by tyrosine kinase inhibitors.
Neoplasia
2004
;
6
:
705
–12.
21
Robinson BWS, Creaney J, Lake R, et al. Mesothelin-family proteins and diagnosis of mesothelioma.
Lancet
2003
;
362
:
1612
–6.
22
Scherpereel A, Grigoriu B, Conti M, et al. Soluble mesothelin-related peptides in the diagnosis of malignant pleural mesothelioma.
Am J Respir Crit Care Med
2006
;
173
:
1155
–60.
23
Chang K, Pastan I. Molecular cloning of mesothelin, a differentiation antigen present on mesothelium, mesotheliomas, and ovarian cancers.
Proc Natl Acad Sci U S A
1996
;
93
:
136
–40.
24
Watanabe H, Okada G, Ohtsubo K, et al. Expression of mesothelin mRNA in pure pancreatic juice from patients with pancreatic carcinoma, intraductal papillary mucinous neoplasm of the pancreas, and chronic pancreatitis.
Pancreas
2005
;
30
:
349
–54.
25
Halliwell B. Can oxidative DNA damage be used as a biomarker of cancer risk in human? Problems, resolutions and preliminary results from nutritional supplementation studies.
Free Radic Res
1998
;
29
:
469
–86.
26
Dusinska M, Collins A, Kazimirova A, et al. Genotoxic effects of asbestos in humans.
Mutat Res
2004
;
553
:
91
–102.
27
Marczynski B, Kraus T, Rozynek P, Raithel HJ, Baur X. Association between 8-hydroxy-2′-deoxyguanosine levels in DNA of workers highly exposed to asbestos and their clinical data, occupational and non-occupational confounding factors, and cancer.
Mutat Res
2000
;
468
:
203
–12.
28
Masood R, Kundra A, Zhu S, et al. Malignant mesothelioma growth inhibition by agents that target the VEGF and VEGF-C autocrine loops.
Int J Cancer
2003
;
104
:
603
–10.
29
Lee AY, Raz DJ, He B, Jablons DM. Update on the molecular biology of malignant mesothelioma.
Cancer
2007
;
109
:
1454
–61.
30
Tomasetti M, Gellert N, Procopio A, Neuzil J. A vitamin E analogue suppresses malignant mesothelioma in a pre-clinical model: a prototype of a future drug against a fatal neoplastic disease?
Int J Cancer
2004
;
109
:
641
–2.
31
Tsiouris A, Wallesby RK. Malignant pleural mesothelioma: current concepts in treatment.
Nat Clin Pract Oncol
2007
;
4
:
344
–52.