Serum levels of the chemokine CCL2 are elevated in malignant pleural mesothelioma patients

Background Malignant pleural mesothelioma (MPM) is a debilitating disease of the pleural cavity. It is primarily associated with previous inhalation of asbestos fibers. These fibers initiate an oxidant coupled inflammatory response. Repeated exposure to asbestos fibers results in a prolonged inflammatory response and cycles of tissue damage and repair. The inflammation-associated cycles of tissue damage and repair are intimately involved in the development of asbestos-associated cancers. Macrophages are a key component of asbestos-associated inflammation and play essential roles in the etiology of a variety of cancers. Macrophages are also a source of C-C motif chemokine ligand 2 (CCL2), and a variety of tumor-types express CCL2. High levels of CCL2 are present in the pleural effusions of mesothelioma patients, however, CCL2 has not been examined in the serum of mesothelioma patients. Methods The present study was carried out with 50 MPM patients and 356 subjects who were possibly exposed to asbestos but did not have disease symptoms and 41 healthy volunteers without a history of exposure to asbestos. The levels of CCL2 in the serum of the study participants was determined using ELISA. Results Levels of CCL2 were significantly elevated in the serum of patients with advanced MPM. Conclusions Our findings are consistent with the premise that the CCL2/CCR2 axis and myeloid-derived cells play an important role in MPM and disease progression. Therapies are being developed that target CCL2/CCR2 and tumor resident myeloid cells, and clinical trials are being pursued that use these therapies as part of the treatment regimen. The results of trials with patients with a similar serum CCL2 pattern as MPM patients will have important implications for the treatment of MPM.


Background
A causal association between exposure to at least some types of asbestos and lung carcinomas and malignant pleural mesothelioma (MPM) has been long recognized [1], and in 2012 the WHO/International Agency for Research on Cancer (IARC, Lyon) classified all forms of asbestos (chrysotile, crocidolite, amosite, tremolite, actinolite, and anthophyllite) as carcinogenic to humans [2]. The 2014 updated Helsinki Criteria notes that while the use of asbestos is banned in many industrialized countries, the global production of asbestos remains at over two million metric tons a year, with an estimated 125 million people being exposed to asbestos in the workplace [3]. Furthermore, workers engaged in cleaning debris at sites of natural disasters and workers involved in demolition work may be exposed to asbestos. For example, asbestos-related disease is predicted to be significant in workers engaged in debris cleaning operations after the Great Hanshin Earthquake that occurred in Japan in 1995. Worldwide, asbestos exposure results in an estimated 255,000 deaths annually, with a significant fraction (over 30,000 in 2016) of these deaths due to mesothelioma [4]. In Japan, the number of patients that die of MPM is currently 1,500 a year (Vital Statistics, Ministry of Health Labour and Welfare, Japan, 2015) and the incidence of MPM is predicted to remain relatively high in the coming years due to past exposure to asbestos.
Macrophages are considered to be essential constituents of many types of solid tumors [5,6], and mesotheliomas are heavily infiltrated by macrophages [7][8][9][10]. The subtypes of macrophages within a tumor is heterogeneous [11]; in general however, tumor development is associated with the presence of macrophages with M2-like characteristics, particularly in patients with a poor prognosis [8, [12][13][14]. M2-like macrophages function in resolution of inflammation and protection and repair of damaged tissue [15][16][17][18]. One of the basic functions of M2-like macrophages that is associated with tissue protection and repair is immunosuppression [11], and tumors have generally been found to contain macrophages with immunosuppressive characteristics [5,[19][20][21][22].
Another important myeloid cell population that is associated with tumors are myeloid-derived suppressor cells, and there is almost universal agreement that accumulation of myeloid cells with MDSC-like phenotypes in the blood or tumor correlates with disease progression, poor prognosis, poor response to therapy, and decreased overall survival [23][24][25][26][27][28][29]. MDSCs are associated with tumor progression in mouse models of mesothelioma [30][31][32], and MDSCs are believed to be associated with mesotheliomas in human patients [33,34].
C-C motif chemokine ligand 2 (CCL2), also known as monocyte chemotactic protein- Whether the disparate findings of the studies cited above are due to differences in tumor stage, CCL2 being associated with a tumorigenic response in some cases and to a tumoricidal response in others, differing immune suppression mechanisms in different tumor types or the patient cohorts studied, or to some other factor is not known. It is clear, however, that the role of CCL2 in tumorigenesis is likely to be affected by tumor-specific factors. The current study was undertaken to investigate serum CCL2 levels in mesothelioma patients. We found that serum CCL2 levels were increased in mesothelioma patients and that this increase was dependent on advancing mesothelioma stage. Healthy subjects possibly exposed to asbestos (356 subjects; 33 females and 323 males; age 68.7 ± 8.3 years; Range 35 -96 years): Serum samples were collected from patients who visited or were hospitalized in the Japan Labour Health and Welfare Organization Asahi Rosai Hospital and the Saiseikai Chuwa Hospital. All of the enrolled subjects possibly exposed to asbestos had certified documents issued by the Japanese Ministry of Health, Labour and Welfare for the compensation of medical care. These subjects had no detectable asbestos-associated disease. Since the hospital records of patients not suffering from mesothelioma were not available to us, it is not known whether any of the subjects in this group had a health condition or treatment that would increase their serum CCL2 levels, for example see patient 356 (Table S1). However, while there was a tendency for this group to have higher serum CCL2 levels compared to the healthy, unexposed volunteers, the difference between these groups was not statistically significant. All participants were provided written informed consent before inclusion in the study. Serum samples were then obtained, coded, and stored in aliquots at -80°C until use.

Assay method
Enzyme-linked immune-absorbent assay (ELISA) kits (CCL2: DCP00, R&D systems, Minneapolis, USA) were used for measuring CCL2, following the manufacturer's instructions. The minimum detectable level of human CCL2 ranged between 0.57 and 10.0 pg/ml for these ELISA kits. All samples had measured CCL2 levels above the minimum detectable levels.

Statistics
In Table 1, patient gender, age, and serum CCL2 are presented as mean ± SD. In Tables 2-5, Analysis of Variance (ANOVA) was used to calculate the estimated marginal means and standard error. Fisher's exact test was used to test the significance of the differences of the nominal data (the data pertaining to gender).
The Kruskal-Wallis (one-way ANOVA) test was used to test the significance of the differences in patient age. Analysis of covariance (ANCOVA) was used to compare the estimated marginal means of serum CCL2 levels adjusted for the covariates of age and gender. The homogeneity of the variance of the serum CCL2 levels was tested using Welch's test. The significance of the differences between the means was tested using the Bonferroni test when the variance was homogenous and Tamhane's T2 test when the variance was not homogenous. p-values were determined using pairwise comparison tests (pairwise comparisons are shown in Additional file 3, Tables S3 -S8). p-values < 0.05 were considered statistically significant. All statistical analyses were carried out with statistical software package SPSS 24.0 (SPSS, Chicago, IL, USA).

Results
A summary of the gender, age, and serum CCL2 levels of the study subjects is shown in Table 1. Individual CCL2 levels are shown in Additional file 1, Table S1.
The pairwise comparisons of the groups is shown in Additional file 3, Tables S3 and S4. The mean CCL2 level in the serum of the mesothelioma patients is significantly elevated compared to the Possibly Exposed (no apparent disease) group, and this increase is dependent on the stage of the disease.  Tables 2 and 3 show the unadjusted serum CCL2 means and 95% confidence intervals and the estimated CCL2 means and 95% confidence intervals when the data is adjusted based on the covariates of age and gender. In Table 2, the data was adjusted using the Unexposed (no apparent disease), Possibly Exposed (no apparent disease), and Mesothelioma (all patients) groups. The pairwise comparisons of these groups is shown in Additional file 3, Table S5. In Table 3, the data was adjusted using the Unexposed (no apparent disease), Possibly Exposed (no apparent disease), and Mesothelioma stages 1-4 groups. The pairwise comparisons of these groups is shown in Additional file 3, Table S6. After adjusting the data, the estimated mean CCL2 level in the serum of the mesothelioma patients is significantly elevated compared to the Possibly Exposed (no apparent disease) group, and this increase is dependent on the stage of the disease.   Tables 4 and 5 show the results when these two patients are removed from data analysis. Table 4 shows he unadjusted serum CCL2 means and 95% confidence intervals and the estimated CCL2 means and 95% confidence intervals when the data is adjusted based on the covariates of age and gender using the Unexposed (no apparent disease), Possibly Exposed (no apparent disease), and Mesothelioma (all patients) groups. The pairwise comparisons of these groups is shown in Additional file 3, Table S7. Table 5 shows he unadjusted serum CCL2 means and 95% confidence intervals and the estimated CCL2 means and 95% confidence intervals when the data is adjusted based on the covariates of age and gender using the Unexposed (no apparent disease), Possibly Exposed (no apparent disease), and Mesothelioma stages 1-4 groups. The pairwise comparisons of these groups is shown in Additional file 3, Table S8. After removal of patients 31 and 50 from the data analysis, CCL2 levels in the mesothelioma patients are still significantly higher than the CCL2 levels in the Unexposed (no apparent disease) group and this increase is dependent on the stage of the disease. a Different from the Unexposed (no apparent disease) and the Possibly Exposed (no apparent disease) groups at p < 0.01 b Different from the Possibly Exposed (no apparent disease) groups at p < 0.05.

Discussion
In this study we measured the levels of CCL2 in the serum of 41 healthy volunteers who have not been exposed to asbestos, 356 healthy subjects who have possibly been exposed to asbestos, and 50 mesothelioma patients. The mean CCL2 level in the serum of the mesothelioma patients was significantly elevated compared to both the healthy volunteers who have not been exposed to asbestos and the healthy subjects who have possibly been exposed to asbestos (see Table 1). However, it is known that serum CCL2 levels increase with normal aging [61-63], and analysis of the age of the patients using the Kruskal-Wallis (one-way ANOVA) test showed an age difference between the patients in the Possibly Exposed (no apparent disease) and the mesothelioma groups. Fisher's exact test also showed a gender difference between these groups. Therefore, the data was re-analyzed based on covariates of age (67.97) and gender (1.11). Re-analysis of the data after adjusting for age and gender did not change the conclusions of the study: serum CCL2 was elevated in mesothelioma patients (see Table 2). Mesothelioma patients 31 and 50 (see Supplementary Table S1) had exceptionally high levels of CCL2. After removal of these two patients' data from analysis, serum CCL2 was still elevated in mesothelioma patients (see Supplementary Tables 5 and 6). Therefore, our data indicate that serum CCL2 levels were increased in mesothelioma patients and this increase was not dependent on the age of the patients in the Mesothelioma group or on the presence of the two patients in the Mesothelioma group with exceptionally high levels of serum CCL2. Elevated CCL2 in the serum of mesothelioma patients is in agreement with the high levels of CCL2 present in the pleural effusions of mesothelioma patients reported by Gueugnon et al. [64].
The increase in the serum levels of CCL2 in the mesothelioma patients was dependent on the stage of the disease (see Table 1). Reanalysis of the data adjusting for age and gender also indicated elevated levels of serum CCL2 depended on mesothelioma stage (see Tables 2 and 3). The dependence on mesothelioma stage was still apparent after removal of the two mesothelioma patients with exceptionally high levels of serum CCL2 from data analysis (see Tables   4 and 5). Therefore, as with the increase in the levels of CCL2 in the serum of mesothelioma patients, the dependence of this increase on disease stage was not due to the age of the patients in the Mesothelioma group or on the presence of the two patients in the Mesothelioma group with exceptionally high levels of serum CCL2.
The mean CCL2 level in the serum of the healthy subjects who have possibly been exposed to asbestos was elevated compared to the healthy volunteers who have not been exposed to asbestos. However, as noted above, it is known that serum CCL2 levels increase during normal ageing [61][62][63]. Thus, the levels of CCL2 in the serum in these two groups followed the expected pattern, lower in the healthy unexposed group consisting of primarily younger patients and higher in the healthy possibly exposed group consisting of primarily older patients.
Several studies have reported that increased expression of CCL2 in tumor tissue is associated with advanced tumor stage and worse prognosis: These studies include patients with breast cancer [65][66][67][68], prostate cancer [69,70], gastric cancer [71], colorectal cancer [72,73], esophageal squamous cell carcinoma [74], head and neck squamous cell carcinoma [75], and glial tumors [47]. In agreement with these findings, a number of studies report elevated levels of CCL2 in the serum of cancer patients and/or an association between elevated serum CCL2 and poor prognosis: levels between breast cancer patients and normal donors, but they did find an association between serum CCL2 and poor prognosis.
However, there are also reports that increased expression of CCL2 in tumor tissue is associated with better prognosis: These studies include patients with gastric cancer [59], colorectal cancer [76], liver cancer [77], and non-small cell lung cancer [78].
There are also a number of studies, that report either that serum CCL2 levels in cancer patients are not related to clinical variables or that higher serum CCL2 levels are associated with a better prognosis and/or that lower serum CCL2 levels are associated with worse prognosis. Consequently, therapies that prove effective against other cancers in which the CCL2/CCR2 axis and myeloid-derived cells are associated with disease progression may also prove effective with mesothelioma patients. There is considerable interest in developing therapies that target CCL2/CCR2 and tumor-resident myeloid cells [5,22,[79][80][81][82][83][84][85]. Numerous clinical trials employing these therapies as part of the treatment regimen have been carried out or are currently being pursued [86][87][88][89][90][91][92][93][94].
The success or failure of these trials will have important implications for the treatment of mesothelioma. Another aspect of increased CCL2 in the serum of mesothelioma patients is that it may be possible to use serum CCL2 to monitor a patient's response to treatment [95].