Macrophage Inflammatory Protein-3A Is a Novel Serum Marker for ...

Imaging, Diagnosis, Prognosis

Macrophage Inflammatory Protein-3A Is a Novel Serum Marker for Nasopharyngeal Carcinoma Detection and Prediction of Treatment Outcomes Kai-Ping Chang,1,7 Sheng-Po Hao,1 Jui-Hung Chang,6 Chih-Ching Wu,6 Ngan-Ming Tsang,2 Yun-Shien Lee,3 Chen-Lung Hsu,4 Shir-Hwa Ueng,5 Shiau-Chin Liu,1 Yu-Lun Liu,1 Pei-Cih Wei,3 Yin Liang,6 Yu-Sun Chang,6 and Jau-Song Yu6,8

Abstract

Purpose: We herein examine whether macrophage inflammatory protein-3a (MIP-3a) is a biomarker for nasopharyngeal carcinoma (NPC) and whether it is involved in modulating NPC cell functions. Experimental Design: The study population comprises 275 NPC patients and 250 controls. MIP-3a levels in tissues and sera were examined by immunohistochemistry and ELISA, respectively. EBV DNA load and EBV viral capsid antigen IgA were measured by quantitative real-time PCR and immunofluorescence assay, respectively. Effects of MIP-3a on NPC cell motility were investigated byTranswell migration/invasion assays and RNA interference. Results: MIP-3a was overexpressed in NPC tumor cells. Serum MIP-3a levels were significantly higher in untreated patients, recurrent patients and patients with distant metastases versus nonNPC controls, patients with complete remission, and long-term disease-free patients. In the prospective cohort, serum MIP-3a levels were significantly higher in untreated NPC patients with advanced tumor-node-metastasis stage versus early stage and also correlated with EBV DNA load. Measurement of MIP-3a, EBV DNA, and viral capsid antigen IgA levels in serial serum/plasma samples from treated patients at 6-month intervals revealed a high association between MIP3a level, EBV DNA load, and disease status. Among 155 consecutive NPC patients, subjects with pretreated MIP-3a serum levels over 65 pg/mL had worse prognoses for overall survival and distant metastasis-free survival in univariate and multivariate analysis. Additionally, cell functional assays showed that MIP-3a contributed to migration and invasion of NPC cells, which could be effectively inhibited by MIP-3a knockdown. Conclusions: MIP-3a may be a novel biomarker and prognosticator for NPC and is involved in migration and invasion of NPC cells.

Authors’ Affiliations: Departments of 1Otolaryngology-Head and Neck Surgery, 2 Radiation Oncology, 3Genomic Medicine Research Core Laboratory, 4Internal Medicine, Division of Hematology-Oncology, and 5 Pathology, Chang Gung Memorial Hospital at Lin-Kou; 6Molecular Medicine Research Center; 7Graduate Institute of Clinical Medical Sciences; and 8Department of Cell and Molecular Biology, Chang Gung University,Taiwan Received 1/10/08; revised 4/2/08; accepted 4/9/08. Grant support: National Science Council of Taiwan grants NSC94-2314-B-182A188, NSC-94-2314-B-182A-162, NSC94-2745-B-182-003-URD, NSC95-2320B-182-049-MY2, and NSC96-2314-B-182A-109-MY3 and Chang Gung University and Chang Gung Memorial Hospital grants CMRPG360211, CTRP1001, and CMRP140041. 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: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). Requests for reprints: Jau-Song Yu, Department of Cell and Molecular Biology, Chang Gung University, No. 259 Wen-Hwa 1st Road, Kwei-Shan, Taoyuan, Taiwan. Phone: 886-3-2118800, ext. 5171; Fax: 886-3-2118891; E-mail: yusong@ mail.cgu.edu.tw. F 2008 American Association for Cancer Research. doi:10.1158/1078-0432.CCR-08-0090

www.aacrjournals.org

Nasopharyngeal carcinoma (NPC), a malignant neoplasm of the head and neck, is rare in western countries but common in Southeast Asia, including Taiwan. NPC is radiosensitive and traditionally treated by radiotherapy; with concurrent chemoradiotherapy, the 5-year overall survival (OS) and disease-free survival for advanced NPC recently exceeded 70% (1, 2). Although many patients with advanced NPC have a good probability of cure with concurrent chemoradiotherapy, tumor stage at initial presentation is still a major prognosticator of patient survival (3, 4). Unfortunately, most NPC patients in endemic areas present with advanced stages at diagnosis, owing to the insidious clinical course of NPC and the relatively inaccessible anatomical site of the nasopharynx. Thus, searching for new tumor markers for NPC is still merited. Because EBV genomes are present in almost every NPC tumor cell, irrespective of histologic differentiation and geographic distribution (5 – 7), various EBV-derived/related factors have been used as NPC tumor markers. The EBV-specific viral capsid antigen (VCA), IgA, shows good sensitivity but has a high falsepositive rate for primary screening and a poor specificity for

6979

Clin Cancer Res 2008;14(21) November 1, 2008


may be a useful supplement to EBV VCA IgA and EBV DNA load for primary screening or post-treatment monitoring of NPC. Finally, we found that NPC tumor cell lines could express and secrete MIP-3a, which promoted the migration and invasion of NPC cells.

Translational Relevance NPC, one of the most common cancers in Southeast Asia, is commonly diagnosed late due to its deep location and vague symptoms. It is highly associated with EBV, and EBVderived markers are currently used for NPC diagnosis, but the high false-positive rates are problematic for primary screening in endemic areas. In current study, we found that the serum levels of MIP-3a were significantly elevated in untreated patients, recurrent patients and patients with distant metastases versus non-NPC controls, patients with complete remission, and long-term disease-free patients. More importantly, MIP-3a serum levels were dynamically associated with the disease status of NPC. NPC patients with higher pretreated MIP-3a serum levels had worse prognoses for overall survival and distant metastasis-free survival in univariate and multivariate analyses.When MIP-3a was combined with traditional EBV markers, the screening efficacy of these two-marker panels was better than that of each alone, indicating that addition of MIP-3a levels to the EBV-based screening protocol may improve the efficacy of primary NPC screening. Collectively, MIP-3a may represent a useful serum biomarker and prognosticator for primary screening and post-treated monitoring for NPC.

Materials and Methods

discriminating NPC from other EBV-associated diseases in endemic areas (8, 9). Cell-free EBV DNA has been shown to be a good indicator and prognosticator for NPC primary screening and OS, respectively (10 – 12). However, because quantitative analysis of plasma/serum EBV DNA requires elaborate procedures and specialized equipment, this screening method is usually inaccessible to general practitioners in endemic areas and is therefore mostly used for post-treatment monitoring. Compelling evidence has emerged in recent years suggesting that chemokines play important roles in regulating processes critical to tumor progression, such as proliferation and metastasis (13, 14). Chemokines represent a large family of proteins in humans; four families of chemokines have been described according to the relative position of the conserved cysteine residues: CC, CXC, XC, and CX3C (15). Macrophage inflammatory protein-3a (MIP-3a), encoded by the CCL20 gene, is a CC-chemokine that induces leukocyte migration into inflammation sites and regulates leukocyte trafficking through lymphoid tissues (16). Increased expression of MIP-3a has been reported in several inflammatory conditions (17 – 19) and cancers, including breast adenocarcinoma, hepatocellular carcinoma, and pancreatic ductal cell adenocarcinoma (20), and this chemokine has also been implicated in promoting growth and migration of pancreatic cancer cells (20, 21). However, no previous study has investigated the potential role of MIP-3a in NPC. We hypothesized that MIP-3a might also be overexpressed in NPC and may participate in modulating the pathophysiology of tumor cells. Here, we provide the first evidence that MIP-3a is overexpressed in NPC tumor cells and that its serum level is significantly elevated in NPC patients and tightly associated with the disease status and treatment outcome of NPC. We also compared the efficacy of MIP-3a, EBV VCA IgA, and EBV DNA load as markers for disease diagnosis and found that MIP-3a


Patients, clinical staging protocol, oncologic treatment, and clinical outcome assessment. The study population was a prospective cohort including 166 consecutively and newly identified NPC patients and 250 control subjects seen in Chang Gung Memorial Hospital at Lin-Kou from March 2001 to November 2007. An additional 109 serum samples were obtained from a retrospective cohort in our tissue bank, including 21 cases with locoregional recurrent NPC, 48 cases with distant metastasis, and 40 long-term NPC survivors (>5 years). The tumornode-metastasis stage was defined according to the 2002 cancer staging system revised by the American Joint Committee on Cancer (22). All NPC patients had been biopsy-proven and had undergone routine checkups including head and neck magnetic resonance imaging, chest X-rays, abdominal ultrasonographies, and bone scans, before treatment and every 6 months after treatment, according to the standard protocol. Fluorodeoxyglucose-positron emission tomography scans were done in all untreated NPC patients to confirm the initial tumor stage. The control subjects were all volunteers undergoing routine health examinations or individuals presenting with otolaryngologic-related, nonneoplastic diseases. Patients and controls with histories of malignant disease were excluded from the study. All subjects signed an institutional review board-approved informed consent before study participation. Blood samples were collected before treatment and during regular follow-ups. The recurrent NPC tumors were detected during regular follow-ups by endoscopy at our otolaryngology’s clinic, and the recurrent tumors were usually biopsied for pathologic proof immediately and blood samples were taken at the same time. For detecting patients with distant metastasis, the systemic workups (e.g., chest X-ray and bone scan) might be done 2 to 3 months before regular outpatient follow-ups when blood samples were usually taken and reports of the systemic workups could be interpreted. The time point of detecting distant metastasis was defined as the time when these systemic workups have been done. The characteristics of all study subjects are given in Supplementary Table S1. All patients enrolled in the prospective cohort had been treated with definitive radiotherapy (cumulative dose of external beam radiotherapy z64.8 Gy). According to our current NPC treatment protocol, all patients whose tumor stages were z3 received additional cisplatinbased concurrent chemoradiotherapy in the Department of Radiation Oncology at Chang Gung Memorial Hospital (1, 2). Patients who were diagnosed with distant metastatic disease at presentation (M1 stage) and/or who had undergone previous treatment at another institute were excluded from the present study. Patients were followed-up at 3-month intervals during the first 3 years after therapy and at 6-month intervals thereafter. Quantitative real-time PCR and immunohistochemistry. Quantitative real-time PCR analysis for MIP-3a was done with a HT 7900 Sequence Detection System (Applied Biosystems) as described in the Supplementary Data. Immunohistochemical analyses were done using an automatic immunohistochemistry staining device according to the manufacturer’s procedures (Bond, Vision Biosystems) using an antiMIP-3a antibody (R&D Systems). The immunohistochemical staining and scoring methods are described in the Supplementary Data. Cell culture. The NPC-TW02 and NPC-TW04 cell lines were used for functional assays. Both cell lines came from the primary nasopharyngeal tumors of untreated NPC patients. TW02 was derived from a keratinizing squamous cell carcinoma (the more common type of NPC in the western countries) of a 36-year-old woman (23). On the other hand, TW04 was derived from an undifferentiated carcinoma, the

6980


Identification of MIP-3a as a Novel Serum Marker for NPC most common histopathologic type of NPC in the endemic area, and obtained from a 36-year-old female with metastatic neck mass (24). Cell culture, preparation of conditioned medium, and Western blotting were done as described in the Supplementary Data. ELISA of MIP-3a. MIP-3a levels in tested samples were determined using the Quantikine ELISA kit for human MIP-3a (R&D Systems) as detailed in the Supplementary Data. Immunofluorescence assay for EBV VCA IgA. Serum EBV titers were determined using an immunofluorescence assay specific for EBV VCA IgA (Meridian Bioscience) according to the manufacturer’s protocol and as reported previously (25). The levels of EBV VCA IgA were determined by titration, and cutoff values were set at 1:40. Quantitation of plasma EBV DNA load. DNA extraction and quantitation of EBV DNA load in plasma were done according to the previously described protocols (11, 12, 25). Cell migration and invasion assays. Migration and invasion of NPC cells were evaluated using a chemotaxis chamber (Corning) and a Cell Invasion Assay Kit (Chemicon), respectively. Detailed procedures for both assays are described in the Supplementary Data. Knockdown of MIP-3a by RNA interference. A 22-nucleotide duplex (5¶-GGATACACAGACCGTATTCTTC-3¶) was designed for short hairpin RNA targeting against MIP-3a in NPC-TW04 cells. The procedures used to generate MIP-3a-silenced stable clones are detailed in the Supplementary Data. Statistical analysis. All statistical data are expressed as mean F SE, except for EBV VCA IgA, which is given as geometric mean titer F SE. The serum target proteins levels and plasma EBV DNA load were

compared using the unpaired Student’s t test. Where differences were identified by F test, Student’s t tests were adjusted for unequal variances (Mann-Whitney’s U test). In comparing the immunohistochemical scores of paired samples from the same subject, Wilcoxon signed rank test was used. Receiver operator characteristic (ROC) curves were constructed by plotting sensitivity versus (1 - specificity), and the areas under the curves (AUC) were analyzed with the Hanley and McNeil method. Kaplan-Meier plots were used for survival analysis, with statistical significance measured by the log-rank test. All statistical tests were two-sided, and P < 0.05 was considered statistically significant. All statistical analyses were done using SPSS version 13.0.

Results Overexpression of MIP-3a in tumor cells of NPC tissues. We first used quantitative real-time PCR to examine the mRNA levels of MIP-3a in 14 paired NPC tumor and adjacent normal tissues and found significantly more MIP-3a transcripts in tumor specimens (100.2 F 34.6 versus 16.7 F 9.2; P = 0.0002; Fig. 1A). To distinguish which cell types in the tumor mass expressed MIP-3a, we did immunohistochemical staining of tissue sections from 65 patients in the prospective cohort. MIP3a was highly expressed in the cytoplasm of tumor cells but was relatively absent from the infiltrating lymphocytes (Fig. 1B). Moreover, the paired normal nasopharyngeal epithelium samples showed lower or no expression of MIP-3a (Fig. 1B).

Fig. 1. Overexpression of MIP-3a in NPC tissues. A, box-plot analysis of MIP-3a mRNA transcript levels in the 14 paired pericancerous adjacent normal (AN) and tumor tissues as assessed by quantitative real-time PCR. The glyceraldehyde 3-phosphate dehydrogenase (GADPH) gene was used as an internal control for normalization. B, immunohistochemical staining of MIP-3a in paired pericancerous adjacent normal and tumor tissues from two representative cases. Bar, 100 Am. Brown signals, expression of MIP-3a. Adjacent normal tissues were biopsied from the pericancerous normal epithelium of the same study subjects. Images in the box (left ; 200) were enlarged (right ; 400). C, box-plot analysis of the immunohistochemical staining scores of MIP-3a in 28 paired adjacent normal and tumor tissues.


6981



Statistical analysis of the 28 paired samples available from these 65 patients revealed that MIP-3a expression was significantly higher in tumor cells versus nontumor cells (135.4 F 11.9 versus 22.5 F 7.4; P < 0.0001; Fig. 1C). However, we did not find a strong correlation between MIP-3a expression and T stage, N stage, overall stage, or pathologic classification of NPC (Supplementary Table S2). Correlation between serum MIP-3a levels and disease status. We then tested whether serum MIP-3a levels were elevated in NPC patients. The serum MIP-3a levels in 166 untreated NPC patients from the prospective cohort were 83.6 F 7.7 pg/mL, which was significantly higher than those from 250 controls (21.8 F 0.9 pg/mL; P < 0.0001; Fig. 2A). Among this cohort, 115 NPC patients showed complete remission (CR) 6 months post-treatment; their serum MIP-3a levels at this point were significantly reduced (from 83.6 F 7.7 to 19.0 F 1.3 pg/mL; P < 0.0001; Fig. 2A). To further evaluate if MIP-3a could be used to monitor treatment outcome, we analyzed serum samples from the retrospective cohort containing 101 posttreated NPC patients with various disease statuses. We found that (a) the serum MIP-3a levels in patients with CR were significantly lower than those in patients with recurrence (19.0 F 1.3 versus 86.5 F 23.9 pg/mL; P < 0.0001) or distant metastasis (19.0 F 1.3 versus 230.3 F 26.1 pg/mL; P < 0.0001); (b) the serum MIP-3a levels in recurrent patients were not statistically different from those in untreated patients (P = 0.898); (c) the serum MIP-3a levels in patients showing distant metastasis were significantly higher than those in untreated patients (P < 0.0001) and recurrent patients (P < 0.0001), respectively; and (d) the serum MIP-3a levels were not statistically different between controls and patients with CR (P = 0.080), between controls and long-term survivors (21.8 F 0.9 versus 24.2 F 3.2 pg/mL; P = 0.969), or between patients with CR and long-term survivors (P = 0.243; Fig. 2A). Overall, these findings suggest that serum MIP-3a could be a potential marker for detection of NPC. Serum MIP-3a levels and OS. From 166 patients of the prospective cohort, we did survival analysis on 155 patients undergoing complete standardized treatment and regular follow-up. We constructed ROC curves between events and censors and selected 65 pg/mL MIP-3a as a cutoff value for stratification, because this MIP-3a serum level was equal to the mean F 3 SD of serum levels of control groups (21.8 F 14.3 pg/ mL) and exhibited a minimum P value with bootstrap resampling (10,000 permutations). As shown by Kaplan-Meier plot, the 5-year OS for patient subgroups stratified by serum MIP-3a levels 65 pg/mL were 81.2% and 64.4%, respectively; this difference was statistically significant using a log-rank test (P = 0.008; Fig. 2B). We conducted a multivariate analysis of the serum MIP-3a levels on OS using Cox proportional regression models adjusted by other prognostic factors, such as gender, age at onset, tumor overall stage, and use of chemotherapy. The results indicated that patients with higher serum MIP-3a levels (>65 pg/mL) have higher risk for OS compared with the patients with lower serum MIP-3a levels (