Higher expression of chemokine receptor CXCR7 is linked to early and metastatic recurrence in pathological stage I nonsmall cell lung cancer

Authors


Abstract

BACKGROUND:

The authors elucidated particular chemokine receptors that are expressed on lung cancer cells, as well as the clinical significance of the expression of these chemokine receptors in completely resected nonsmall cell lung cancer (NSCLC).

METHODS:

The authors examined gene expression of chemokine receptors (CCR1-11, CXCR1-7, XCR1, and CX3CR1) in 11 cell lines of lung cancer, and gene expression of CXCR3, CXCR4, and CXCR7 (CXCR3/4/7) in surgical specimens of 127 patients who underwent complete resection for their NSCLC between May 2001 and December 2002, using quantitative real-time reverse transcriptase–polymerase chain reaction (PCR). Mutation detection analysis of the EGFR genes using the PCR single-strand conformational polymorphism method was evaluated in patients with pathological (p-) stage I adenocarcinoma.

RESULTS:

Substantial expression of CXCR3/4/7 mRNA was observed in all NSCLC cell lines examined. In p-stage I NSCLC, CXCR4 and CXCR7 expression values in patients with postoperative metastatic recurrence (Rec-Distant) were significantly higher than in those without recurrences (P = .003 and P = .007, respectively). In addition, the 5-year disease-free survival (DFS) rate of high CXCR7-expressing patients (63.2%) was significantly lower than that of low CXCR7-expressing patients (84.8%) (P = .033). The EGFR mutation was significantly more frequent in patients with higher CXCR7 expression (14 of 21 patients) than in those with lower CXCR7 expression (12 of 32 patients) (P = .038). A multivariate analysis confirmed that high CXCR7 expression was an independent and significant factor predicting a poor DFS in p-stage I NSCLC patients (P = .041).

CONCLUSIONS:

Higher expression of CXCR7 is associated with Rec-Distant and poor DFS in patients with p-stage I NSCLC. Cancer 2009. © 2009 American Cancer Society.

Primary lung cancer is the leading cause of cancer deaths in most industrialized countries. Although some advances in surgery, chemotherapy, and radiotherapy have been made recently,1, 2 the prognosis remains unsatisfactory. The most important reason for the poor prognosis is the high incidence of nodal and/or distant metastases documented in primary lung cancer patients. To improve the prognosis, it is essential to determine the biological characteristics of nonsmall cell lung cancer (NSCLC) and find novel therapeutic targets.

Chemokines are a superfamily of small, cytokine-like proteins that bind to and activate a family of chemokine receptors. Over 50 chemokines have been identified, and these chemokines are divided into 4 families (CXC, CX3C, CC, and C) on the basis of the positions of 4 conserved cysteine residues. Chemokine receptors are 7 transmembrane G-coupled proteins.3-5 To date, 20 chemokine receptors (CCR1-11, CXCR1-7, XCR1, and CX3CR1) have been identified. Important roles of chemokines and their receptors have been demonstrated in inflammation, infection, tissue injury, allergy, cardiovascular diseases, and malignant tumors.6 A complex network of chemokines and their receptors influence the regulation of tumor progression and metastasis.5, 7, 8 Furthermore, chemokine receptors have been demonstrated to play several key roles in cancer metastasis. Indeed, chemokine receptors may potentially facilitate tumor dissemination at each of the key steps of metastasis, including adherence of tumor cells to endothelium, extravasation from blood vessels, metastatic colonization, angiogenesis, proliferation, and protection from the host response via activation of key survival pathways such as PI-3K and Akt.9 Recent studies have clearly demonstrated the importance of chemokine receptor expression in metastasis to specific organs (eg, lymph nodes, bone marrow, liver, and lungs) by breast cancer,10 melanoma,11 gastric carcinoma,12 and lung cancer13 cells. In particular, the chemokine CXCL12 (stromal cell-derived factor-1; also known as SDF-1) is thought to act through its receptor, CXCR4, to promote the growth of primary tumors and progression to metastatic disease in lung cancer.14 In addition, recent data have suggested that CXCR7 (also known as RDC1), recently deorphanized as a chemokine receptor that binds chemokines CXCL11 (interferon-inducible T-cell chemoattractant; also known as I-TAC) and CXCL12, has key functions in promoting tumor development and progression.15

Although chemokine receptor overexpression in tumor tissues may suggest the development of diagnostic agents and therapy targeted at chemokine receptor–overexpressing tumors, few detailed clinical studies have been undertaken to evaluate the clinical significance of chemokine receptor status in NSCLC.13, 16, 17 Particularly, no comprehensive study with regard to all known chemokine receptors has been documented in NSCLC. Thus, in this study, we comprehensively analyzed the expression of all known chemokine receptors in lung cancer cell lines, and assessed chemokine receptor (especially for CXCR3, CXCR4, and CXCR7) expression in tumor tissues as well as in tumor cell lines in correlation with clinical outcomes of resected NSCLC.

Previously, we have reported that the mutations in the EGFR gene were found in approximately half of our Japanese adenocarcinoma patients, and that the EGFR gene mutations may play an important role in the etiology of adenocarcinoma in nonsmokers.18 Moreover, our study,19 TRIBUTE,20 and the IDEAL/INTACT21 study recently showed that patients with EGFR gene mutations had a better prognosis than those without EGFR gene mutations. These data suggested that adenocarcinoma with EGFR gene mutations should be classified into a different category from nonadenocarcinoma or carcinoma without EGFR gene mutations. Thus, in this study, we analyzed the correlation between EGFR gene mutations and chemokine receptors in adenocarcinoma patients.

MATERIALS AND METHODS

Cell Lines

Human cell lines used in this study were as follows: 1 lung adenocarcinoma cell line (A549) from Riken Cell Bank (Tsukuba, Japan); 5 lung adenocarcinoma cell lines (H1650, H1975, H441, H522, H23), 1 lung squamous cell carcinoma cell line (H520), 2 lung large cell carcinoma cell lines (H1299, H460), and 1 lung small cell carcinoma cell line (H146) from American Type Culture Collection (Rockville, Md); and 1 lung adenocarcinoma cell line (PC-9) from Immuno-Biological Laboratories (Gunma, Japan). All cell lines were cultured in RPMI1640 (Nacalai Tesque, Kyoto, Japan) supplemented with 10% fetal calf serum.

Patients and Tissue Samples

A total of 127 consecutive patients with NSCLC who underwent resection at the Department of Thoracic Surgery, Kyoto University Hospital between May 2001 and December 2002 were included in this study (Table 1). Clinical data for patients involved were obtained from inpatient and outpatient medical records, chest x-ray films, whole-body computed tomography films, bone scanning data, and operation records. Regarding smoking status, we classified patients into ≤20 or >20 pack-years. One hundred twelve patients underwent complete resection of the lobe or segment in which the tumor resided and received hilar and mediastinal lymph node dissections. Seven patients received partial lung resection and lymph node sampling. Eight patients underwent incomplete resection. No patient with pathological (p-) stage I disease was exposed to chemotherapy or radiotherapy before his or her tumor was resected. All tissue samples were immediately snap-frozen in liquid nitrogen and stored at −80°C until use.

Table 1. mRNA Expression in Tumor Tissues According to Patients' Characteristics
 No.CXCR3CXCR4CXCR7
MeanSEPMeanSEPMeanSEP
  1. SE indicates standard error; sq-ad, squamous-adeno; w-m, well-moderate; p-, pathological; pT, pathological tumor; pN, pathological nodal; F-L, free-local.

All1270.0130.002 0.2120.026 0.0230.006 
Sex          
 Men860.0140.002.6380.2040.030.6680.0270.009.308
 Women410.0120.002 0.2280.049 0.0130.002 
Age, y          
 Lower (≤67)650.0130.002.7090.2020.034.7060.0130.002.111
 Higher (>67)620.0140.003 0.2220.040 0.0330.013 
Histology          
 Squamous380.0080.002.2570.2270.060.8740.0500.020.044
 Adeno760.0160.003 0.2110.029 0.0110.002(sq-ad .009)
 Large50.0140.009 0.2360.177 0.0060.004 
 Others80.0110.003 0.1340.044 0.0110.003 
Differentiation          
 Well320.0130.002.9840.2210.052.9770.0090.002.267
 Moderate650.0130.003 0.2070.038 0.0220.003(w-m .010)
 Poor300.0140.005 0.2130.047 0.0380.025 
p-Stage          
 I790.0140.002.5950.2060.030.7370.0170.002.264
 II160.0070.002 0.1590.048 0.0110.003 
 III280.0160.006 0.2570.076 0.0460.028 
 IV40.0170.010 0.2370.043 0.0170.010 
pT          
 1620.0130.002.9210.2150.035.5790.0250.012.937
 2450.0130.003 0.2220.048 0.0190.003 
 3110.0160.012 0.1030.053 0.0160.006 
 490.0090.005 0.2740.113 0.0310.017 
pN          
 0860.0130.002.2800.1930.028.5630.0170.002.418
 1170.0090.002 0.2820.070 0.0210.009 
 2230.0200.007 0.2420.081 0.0450.033 
 310.000 0.012 0.001 
Smoking, pack-years          
 Lower (≤20)480.0120.002.4420.1950.039.6060.0110.002.136
 Higher (>20)790.0140.003 0.2220.034 0.0300.010 
Postoperative recurrence region          
 Free710.0120.002.2420.1680.020.2290.0150.002.112
 Distant ± local260.0150.005 0.2630.079(F-L .043)0.0190.005(F-L .037)
 Local only220.0090.002 0.2980.093 0.0560.035 
 Residual80.0260.016 0.1990.061 0.0150.006 

For histological examination, samples were fixed in 10% (v/v) formalin and then embedded in paraffin. A 4-μm section was prepared from each sample and stained with hematoxylin and eosin.

Pathological staging was determined using the current tumor node metastasis classification system (International Union Against Cancer, 1997). The histological type and the grade of cell differentiation in these patients were determined using the pathological diagnosis by 2 pathologists at the Kyoto University Hospital Laboratory of Anatomic Pathology, and then confirmed by 1 pathologist (T.M.) specializing in lung pathology according to classification by the World Health Organization classification system 2004.

Postoperative locoregional recurrence (Rec-Local) was defined as radiographic or pathologic documentation of recurrence within the primary tumor bed, ipsilateral hilum, or mediastinum. Postoperative metastatic recurrence (Rec-Distant) was defined as clinical or radiographic evidence of M1 disease according to the 2002 American Joint Committee on Cancer definition.22

A written informed consent to perform genetic analyses was obtained from each patient before surgery, and the study itself was approved by the Ethics Committee of Kyoto University Graduate School and Faculty of Medicine.

RNA Isolation and Quantification of Chemokine Receptor Expression With Real-Time Reverse Transcription–Polymerase Chain Reaction

Total RNA was isolated from frozen tissues using an RNeasy Mini Kit (QIAGEN, Tokyo, Japan), following the manufacturer's protocol.23 To digest contaminated DNA, the RNA extract was incubated with RNase-free DNase I (Amersham Pharmacia Biotech, Uppsala, Sweden) during RNA isolation. Reverse transcription (RT) of total RNA was performed using Ready-To-Go You-Prime First-Strand Beads (Amersham Biosciences, Piscataway, NJ) and random hexamers (Amersham Pharmacia Biotech, Piscataway, NJ). To quantify expression of the chemokine receptor genes, real-time RT–polymerase chain reaction (PCR) was performed using the Light Cycler thermal cycler system (Roche Diagnostics, Tokyo, Japan) following the manufacturer's protocol. The sense and antisense primers used for quantitative amplification of chemokine receptor mRNAs (CCR1-11, CXCR1-7, XCR1, and CX3CR1) are shown in Table 2, and those used for amplification of glyceraldehyde-3-phosphate dehydrogenase (GAPDH), used as an internal control, were as follows:

  • sense, 5′-ACAACAGCCTCAAGATCATCAG-3′

  • antisense, 5′-TCTTCTGGGTGGCAGTGATG-3′

Table 2. Primer Sequences for Quantitative Real-Time RT-PCR Analysis and Amplicon Sizes
Gene NameForward PrimerReverse PrimerAmplicon (bp)NCBI Accession No.
  1. RT-PCR indicates reverse transcription–polymerase chain reaction; NCBI, National Center for Biotechnology Information.

CCR1CACGGACAAAGTCCCTTGGCAAAGGCCCTCTCGTTCAC115NM_001295
CCR2AGACAAGCCACAAGCTG/AACGTGACTTCTTCACCGCTCTC110NM_000647
CCR3TATCCGGGCAAGAACTTATCGGCCCACGTCATCATAGTAGG113NM_001837
CCR4GAGCCGCTTTCAGAAAAGCGCTTGGGGATACTTTCATACAG133NM_005508
CCR5GACATCCGTTCCCCTACAAGAGATGAACACCAGTGAGTAGAG148NM_000579
CCR6GATGGTCATCACATTGGTGAGAGCCCCTTCAGCTCACAG115NM_004367
CCR7GCACAGCCTTCCTGTGTGCGTCCGTGACCTCATCTTG112NM_001838
CCR8TGAGCTGCACTCACATGAGTAGTAGTAGTCGGTCACTGTTG126NM_005201
CCR9TGTCCCAGGGAGAGTTGCTGTGGATTCAGAGCCATAGTC104NM_006641
CCR10GCCACAGAGCAGGTTTCCACGGTCAGGGAGACACTG122NM_016602
CCR11GACAATGCTAGGTGCATTCCGCATCTTCATGAGTGTCCTTG133NM_016557
CXCR1CTGCAGCTCCTACTGTTGGTTCATCTGCAGGTGGCATG130NM_000634
CXCR2AACCTTGAGGCACAGTGAAGTCAAAGCTGTCACTCTCCATG139NM_001557
CXCR3CATGGTCCTTGAGGTGAGTGGTACAGCACGAGTCACTCTC97NM_001504
CXCR4CGTCAGTGAGGCAGATGACTGCAATAGCAGGACAGGATG114NM_003467
CXCR5TCCTGCTGGTCTTCATCTTGGTGGACAATGGCCAGGTAG148NM_001716
CXCR6GCTCATCTCTGGAACAAACTGTGCTGAACTGCAGGAAGTC131NM_006564
CXCR7AGTTGTTGCAAAGTGCTCAGTCTGAGTAGTCGAAGAGATGC132NM_020311
XCR1GACGCATGTAAAGAGGTGTAGAAGACCCAGGCCTGGTTC126NM_005283
CX3CR1CAGATCCAGAGGTTCCCTTGTGACGGAGTAGAATATGGACAG142NM_001337

A 20-μL sample of reaction mixture containing 10 μL of Quantitect SYBR Green Master Mix (Qiagen K. K., Tokyo, Japan), 0.5 μM of sense and antisense primers, and 0.02 μg of cDNA was prepared. PCR amplification was initiated by preincubation for 15 minutes at 95°C to activate the hot-start–designed DNA polymerase, followed by 40 cycles of the following protocol: denaturation at 94°C for 15 seconds, annealing at 56°C for 15 seconds, and elongation at 72°C for 15 seconds with detection of the fluorescent products.

The quantification data were analyzed with the Light Cycler analysis software version 5.03. The expression levels of each gene were represented as the ratio of each gene expression value to the GAPDH gene expression value of each sample.

Mutation Detection Analysis of the EGFR Genes

Mutation detection analysis of the EGFR genes was evaluated as described previously.18 In brief, genomic DNA was extracted from each frozen tumor sample or each lung cancer cell line, and each exon of the EGFR gene was amplified by the PCR using specific primer sets. Each product was analyzed using the PCR single-strand conformational polymorphism method. If mutational band(s) were detected, the corresponding sample was directly sequenced, and the mutation type was confirmed. Three adenocarcinoma cell lines (H1650, H1975, PC9) displayed an EGFR gene mutation.24, 25

Statistical Analyses

Median values of the series were used to classify each case as having high or low CXCR3, CXCR4, and CXCR7 expression, to compare their relationships with the other biological parameters and clinical follow-up. The median values for CXCR3, CXCR4, and CXCR7 expression were 5.87 × 10−3, 1.12 × 10−1, and 9.06 × 10−3, respectively.

The significance of differences in categorical data was tested using the chi-square test. Differences between continuous variables were examined using the Student t test if the continuous variables were normally distributed, or using the Mann-Whitney U test if not normally distributed. Comparison among 3 continuous variables was performed using analysis of variance. The postoperative survival rate and disease-free survival rate were analyzed using the Kaplan-Meier method with the completely resected 119 patients, and the differences between groups were assessed by the log-rank test.

Multivariate analysis of prognostic factors was performed using the Cox proportional hazard model. Stat View software version 5 (SAS Institute, Cary, NC) was used to perform all statistical calculations. Differences were considered to be statistically significant if the P value was <.05.

RESULTS

CXCR3, CXCR4, and CXCR7 mRNA Are Substantially Expressed in Lung Cancer Cell Lines

Quantitative gene expression analysis of chemokine receptors CCR1-11, CXCR1-7, XCR1, and CX3CR1 revealed substantial expression of CXCR3, CXCR4, and CXCR7 (CXCR3/4/7) mRNAs in 11 human lung cancer cell lines (Fig. 1). All lung cancer cell lines expressed the CXCR3 and CXCR4 genes at various levels, but there was no apparent difference in expression according to the histological types of cell lines. On the contrary, CXCR7 expression was not observed in large (H460, H1299) and small (H146) cell carcinoma cell lines. The mean value of CXCR7 expression in squamous cell carcinoma cell lines was considerably higher than that in adenocarcinoma cell lines (0.0053 vs 0.0013 ± 0.0008, P = .147).

Figure 1.

Expression of chemokine receptors is shown in 11 human lung cancer cell lines. Quantitative reverse transcription–polymerase chain reaction analyses of all known chemokine receptors are shown. Values are expressed as the ratio of the expression value of chemokine receptors to that of glyceraldehyde-3-phosphate dehydrogenase.

Expression of CXCR3, CXCR4, and CXCR7 in Tumor Tissues Obtained From NSCLC Patients

In this study, because we could detect substantial expression of CXCR3/4/7 mRNA by means of in vitro survey of lung cancer cell lines, we quantitatively measured CXCR3/4/7 mRNA expression in primary tumors resected from NSCLC patients.

As shown in Table 1, there was no significant difference in the mean value of CXCR3 expression according to all patient characteristics, such as sex, age, histological type, tumor differentiation, p-stages, pathological tumor factor, pathological nodal factor, smoking status, or postoperative recurrence site.

The mean value of CXCR4 expression in patients with Rec-Local disease was significantly higher than that in patients without postoperative recurrences (Rec-Free) (0.298 vs 0.168, P = .043) (Table 1). There was no significant difference in the mean value of CXCR4 expression according to any other patient characteristics.

The mean value of CXCR7 expression in squamous cell carcinoma patients was significantly higher than that in adenocarcinoma patients (0.050 vs 0.011, P = .009). With regard to squamous cell carcinoma and adenocarcinoma, cell line and patient sample data are likely to be concordant. The mean value of CXCR7 expression in large cell carcinoma patients (n = 6) was 0.006 ± 0.004 (Table 1). Although our in vitro data showed that CXCR7 expression was not observed in large cell carcinoma cell lines, CXCR7 expression was observed in large cell carcinoma patients. But the value in large cell carcinoma patients was somewhat lower than that in squamous cell carcinoma or adenocarcinoma patients (0.006, 0.050, 0.011, respectively) (Table 1). The reasons this discrepancy existed between clinical data and cell line data remain to be elucidated. We did not assess CXCR7 expression in small cell carcinoma patients, because the samples from small cell carcinoma patients were not included in this study. The mean value of CXCR7 expression appeared to be increased along with differentiation of tumor cells (well vs moderate, P = .010). The mean value of CXCR7 expression in patients with Rec-Local disease was significantly higher than that in patients with Rec-Free disease (0.056 vs 0.015, P = .037) (Table 1). There was no significant difference in the mean value of CXCR7 expression according to any other patient characteristics.

Higher Expression of CXCR4 and CXCR7 mRNAs Were Observed in Patients With Postoperative Metastatic Recurrence in p-Stage I Disease

To evaluate the metastatic potential of CXCR3/4/7 expression, patients were divided into 3 groups according to the postoperative recurrence site.

In patients with p-stage I disease, CXCR4 and CXCR7 expression values in patients with Rec-Distant disease were significantly higher than in those with Rec-Free disease (P = .003 and P = .007, respectively) (P = .002 and P = .017 for all groups, respectively) (Fig. 2a). There was no significant difference in the mean value of CXCR3/4/7 expression between any other postoperative recurrence sites (CXCR3, P = .310 for all groups).

Figure 2.

CXCR3, CXCR4, and CXCR7 mRNA expression is shown in completely resected pathological (p-) stage I-III nonsmall cell lung cancer according to the recurrence site. In patients with p-stage I disease, CXCR4 and CXCR7 expression values in patients with postoperative metastatic recurrence were significantly higher than in those without postoperative recurrences (P = .003 and P = .007, respectively) (P = .002 and P = .017 for all groups, respectively).

In patients with p-stage II-III disease, however, there was no trend between the mean value of CXCR3/4/7 expression and the postoperative recurrence site. The CXCR3 expression values in patients with Rec-Local disease were significantly higher than in those with Rec-Free disease (P = .049) (P = .514, for all groups) (Fig. 2b). The CXCR4 expression values in patients with Rec-Local disease were significantly higher than in those with Rec-Distant disease (P = .034) (P = .028, for all groups) (Fig. 2b). There was no significant difference in the mean value of CXCR3/4/7 expression between any other postoperative recurrence sites (CXCR7, P = .297 for all groups).

EGFR Gene Mutation Is Expressed More Frequently in Tumors With Higher Expression of CXCR7 mRNA

As in our previous studies,18 mutation detection analysis of the EGFR genes was evaluated in 53 patients with p-stage I adenocarcinoma. Our results showed that EGFR gene mutations existed in 14 of 21 patients (66.7%) with higher CXCR7 expression, but in only 12 of 32 patients (37.5%) with lower CXCR7 expression (Fig. 3a). The EGFR mutations were significantly more frequent in patients with higher CXCR7 expression than in those with lower CXCR7 expression (P = .038). In addition, our in vitro data showed that adenocarcinoma cell lines with EGFR gene mutations had a tendency toward high expression values of CXCR7 (Fig. 3b). As a result, high expression values of CXCR7 were observed in 2 (H1975, PC9) of 3 adenocarcinoma cell lines (66.7%) with EGFR gene mutation, but in none of 4 adenocarcinoma cell lines (0.0%) with wild-type EGFR gene.

Figure 3.

(a) Mutation detection analysis of the EGFR genes in patients with pathological (p-) stage I adenocarcinoma is shown. EGFR mutation was significantly more frequent in patients with higher CXCR7 expression than in those with lower CXCR7 expression (P = .038). (b) High expression values of CXCR7 were observed in 2 of 3 cell lines with EGFR gene mutations, but in none of the 4 cell lines with wild-type EGFR gene.

CXCR7 Was a Significant Disease-free Prognostic Factor in Patients With p-Stage I Disease

To evaluate the prognostic significance of CXCR3/4/7 expression status, patients were divided into those with lower CXCR3/4/7 expression and those with higher CXCR3/4/7 expression according to the median value as a cutoff value.

For all patients, CXCR3/4/7 expression status provides no prognostic value (overall survival, P = .438, P = .942, P = .707, respectively) (disease free survival, P = .376, P = .735, P = .753, respectively) (Table 3). A subset analysis according to p-stage showed that CXCR7 was a significant disease-free prognostic factor in patients with p-stage I disease (5-year disease-free survival rate, 84.8% for patients with lower CXCR7 expression and 63.2% for patients with higher CXCR7 expression; P = .033) (Fig. 4b, Table 3b). Conversely, in the 5-year overall survival rate, there was no significant difference between patients with lower CXCR7 expression and higher CXCR7 expression (74.1% vs 67.3%) (P = .243) (Fig. 4a, Table 3a).

Figure 4.

CXCR7 mRNA expression and postoperative 5-year (a) overall survival and (b) disease-free survival are shown in completely resected pathological stage I nonsmall cell lung cancer. In comparison according to the status of CXCR7 expression, patients who had high expression of CXCR7 showed a significantly less favorable disease-free survival rate compared with patients who had low expression of CXCR7 (P = .033). Pts indicates patients.

Table 3. CXCR3, CXCR4, and CXCR7 mRNA Expression in Completely Resected p-Stage I-III NSCLC
 CXCR3PCXCR4PCXCR7P
% Higher% Lower% Higher% Lower% Higher% Lower
  1. Each figure shows the 5-year (4-year) overall survival rate of each patient group.

  2. p- indicates pathological; NSCLC non–small-cell lung cancer.

a. Postoperative 5-Year Overall Survival Rate
  All65.554.0.43861.558.8.94260.359.3.707
  p-Stage I80.059.5.22477.165.0.31567.374.1.243
 Histologic type         
  Squamous66.745.5.69666.745.5.57755.6(50.0).687
  Adeno88.967.3.19582.477.9.84781.078.0.712
 Age, y         
  Lower (≤67)88.966.7.35888.271.8.29682.472.2.924
  Higher (>67)72.752.4.31368.558.1.57657.374.0.225
 Sex         
  Men69.244.1.25762.655.4.75959.357.4.795
  Women100.082.4.10594.484.6.34581.3100.0
 Smoking, pack-years         
  Lower (≤20)100.088.9.141100.088.2.11994.494.7.532
  Higher (>20)61.933.6.32853.346.7.81948.150.2.587
 Differentiation         
  Well100.080.0.250100.081.8.08181.8100.0
  Moderate76.556.8.42370.962.0.72462.673.3.476
  Poor33.350.0.36233.350.0.56760.028.6.345
 p-Stage II-III33.342.9.73627.847.6.24040.037.5.901
b. Postoperative 5-Year Disease-free Survival Rate
  All61.153.2.37659.055.5.73556.058.3.753
  p-Stage I78.467.3.24178.367.6.28763.284.8.033
 Histologic type         
  Squamous85.745.5.06385.745.5.06359.3(100.0)
  Adeno80.576.6.67778.479.2.86066.086.9.044
 Age, y         
  Lower (≤67)83.377.8.67882.478.9.78376.584.2.563
  Higher (>67)73.157.0.22174.355.6.21453.484.6.036
 Sex         
  Men78.253.8.06178.058.3.13063.571.8.539
  Women78.682.4.80677.884.6.60362.5100.0
 Smoking, pack-years         
  Lower (≤20)84.288.9.64485.088.2.73877.894.7.122
  Higher (>20)72.046.3.06269.950.0.17251.671.2.227
 Differentiation         
  Well88.280.0.50693.872.7.12963.6100.0
  Moderate79.461.0.26673.965.0.59363.680.0.262
  Poor44.466.7.61444.466.7.67360.053.6.896
 p-Stage II-III23.726.7.62817.931.8.25635.718.6.252

We further analyzed the postoperative survival according to the expression status of CXCR3/4/7 in patients with p-stage I disease among specific subsets. As a rule, the 5-year disease-free survival rate of patients with lower CXCR7 expression seemed somewhat better than that of patients with higher CXCR7 expression, and disease-free prognostic significance was revealed in the older and adenocarcinoma patients (84.6% vs 53.4%, P = .036; and 86.9% vs 66.0%, P = .044, respectively) (Table 3b).

Finally, in patients with p-stage II-III disease, there was no significant difference in the 5-year overall survival rate and disease-free survival rate (Table 3).

Multivariate Analysis of Prognostic Factors

A multivariate analysis confirmed that higher CXCR7 expression was an independent and significant disease-free prognostic predictor in patients with p-stage I disease (P = .041; hazard ratio, 3.923; 95% confidence interval, 1.058-14.539), but not a significant factor for overall prognosis (P = .829) (Table 4). Smoking status was also a significant disease-free prognostic factor (P = .016). Age and smoking status were also significant overall prognostic factors (P = .048 and P = .019, respectively) (Table 4).

Table 4. Multivariate Analysis of Prognostic Factors of the Survival of the Patients With Completely Resected p-Stage I NSCLC
Prognostic FactorsNo.βDisease-free SurvivalPβOverall SurvivalP
Relative Hazard (95% CI)Relative Hazard (95% CI)
  1. p- indicates pathological; NSCLC non–small-cell lung cancer; CI, confidence interval; pT, pathological tumor.

Age, y       
 Lower (≤67)36 1.232 1.048
 Higher (>67)43.6401.897 (0.664-5.422) 1.0442.841 (1.008-8.008) 
Sex       
 Men47 1.336 1.688
 Women32.7322.080 (0.468-9.238) .3581.431 (0.249-8.209) 
Smoking status, pack-years       
 Lower (≤20)37 1.016 1.019
 Higher (>20)421.8506.357 (1.409-28.68) 2.2019.036 (1.440-56.70) 
Histology       
 Adeno53 1.641 1.966
 Squamous21−.5520.576 (0.143-2.319) −.0980.907 (0.223-3.679) 
 Others5−.7800.458 (0.046-4.519) .1651.179 (0.152-9.138) 
Differentiation       
 Well27 1.313 1.227
 Moderate40−.1860.830 (0.202-3.419) .5171.678 (0.371-7.593) 
 Poor121.0882.968 (0.511-17.25) 1.4414.225 (0.786-22.69) 
pT       
 T146 1.266 1.237
 T233.6081.837 (0.629-5.364) .6051.830 (0.671-4.991) 
CXCR3       
 Lower (≤0.00587)38 1.297 1.436
 Higher (>0.00587)41−.6210.538 (0.167-1.728) −.4260.653 (0.223-1.910) 
CXCR4       
 Lower (≤0.11204)39 1.332 1.394
 Higher (>0.11204)40−.4930.611 (0.226-1.654) −.4300.651 (0.242-1.748) 
CXCR7       
 Lower (≤0.00960)35 1.041 1.829
 Higher (>0.00960)441.3673.923 (1.058-14.54) .1401.150 (0.323-4.094) 

DISCUSSION

In previous clinical studies, it has been reported that chemokine receptor status in human cancers significantly correlated to prognosis and/or metastases in a variety of malignant tumors such as T-cell leukemia (CCR4),26 hepatocellular carcinoma (CCR6),27 gastric carcinoma (CCR7),12 renal cell carcinoma (CXCR3),28 ovarian cancer (CXCR4),29 osteosarcoma (CXCR4),30 colorectal cancer (CCR7 and CXCR4),31, 32 and malignant melanoma (CXCR3 and CXCR4).33, 34 However, no clinical study on CXCR3 and CXCR7 expression, and only 1 clinical study on CXCR4 expression (which used immunohistochemistry), have been reported with regard to NSCLC.17 In this work, we report the first comprehensive analyses of chemokine receptor expression in lung cancer cell lines, and correlations of chemokine receptor expression with prognosis, especially for CXCR7, in tumors resected from NSCLC patients.

Many recent studies examining the role of chemokines in cancer have focused on CXCL12 and its receptor CXCR4, and how they affect metastatic and prognostic processes for patients.9, 13 Heretofore, it has generally been believed that CXCL12 regulates these processes via a single cell surface receptor known as CXCR4. Recently, however, Burns and colleagues35 characterized an alternate receptor CXCR7 that binds with high affinity to CXCL12 and to a second chemokine CXCL11. Interestingly, CXCL11 is 1 of the ligands of CXCR3 that we examined in this study. This membrane-associated CXCR7 is expressed on many tumor cell lines, activated endothelial cells, and fetal liver cells, but on few other cell types. Miao and colleagues15 reported that CXCR7 promotes the growth of breast and lung cancer. Moreover, Burns and colleagues35 showed that expression of CXCR7 provides cells with a growth and survival advantage as well as increased adhesion properties. In addition, they demonstrated that CXCR7 facilitated angiogenesis and that blockade of CXCR7 inhibited tumor growth in mouse models. In our study, we demonstrated that, in patients with p-stage I NSCLC, CXCR4 and CXCR7 expression values in patients with Rec-Distant disease were significantly higher than in those with Rec-Free disease. Conversely, in patients with p-stage II-III NSCLC, CXCR4 expression values in patients with Rec-Local disease were significantly higher than in those with Rec-Distant disease. In addition, CXCR7 expression values in patients with Rec-Local disease were somewhat higher than in those with Rec-Free and Rec-Distant disease. In patients with p-stage II-III NSCLC, patients who were exposed to preoperative chemotherapy or radiotherapy were included. Therefore, we performed the subset analysis between the preoperative therapy group and the preoperative nontherapy group, but no significant difference was found (data not shown). Our results suggest that the overexpression of CXCR4 and CXCR7 may be related to the distant metastatic processes of NSCLC. CXCR4 and CXCR7 may be candidates for molecular targeting treatment to interfere with postoperative distant metastasis in patients with p-stage I NSCLC.

In addition, we found that high CXCR7 expression in resected NSCLC tissues was a significant factor in predicting a poor disease-free survival, especially in patients with p-stage I NSCLC. In our data, a multivariate analysis confirmed that higher CXCR7 expression was an independent and significant factor in predicting a poor disease-free prognosis (P = .033), but not a significant factor for overall survival (P = .243). These results suggest that higher expression of CXCR7 implies a poor disease-free prognosis in patients with NSCLC, although a significant difference exists only for patients with p-stage I NSCLC, not for patients with p-stage II-III NSCLC. In patients with p-stage II-III, several patients were exposed to postoperative chemotherapy or radiotherapy. Although the reason for these results was unclear, postoperative therapy might influence postoperative overall survival, but not disease-free survival. Thus, 1 of the limits of this study is the heterogeneity of the samples with p-stage II-III disease, which includes both samples exposed to preoperative therapy and samples not exposed to preoperative therapy. In addition, we demonstrated that higher expression of CXCR7 also correlated to a higher incidence of EGFR gene mutations. Choi and colleagues showed that mutations within the tyrosine kinase domain of EGFR caused significant EGF-independent tyrosine phosphorylation of EGFR, and induced constant EGF-independent activation of EGFR.36 Moreover, Phillips and colleagues showed that a combination of low oxygen condition (many solid tumors were under this condition) and EGFR activation by EGF in the NSCLC cells dramatically enhanced CXCR4 expression and promoted metastasis.37 Although no detailed experimental and clinical studies with regard to CXCR7 have been undertaken, constant activation of EGFR by the mutations within the tyrosine kinase domain might induce high expression of CXCR7 as well as CXCR4. Recently, we reported that adenocarcinoma with EGFR gene mutations have a better prognosis than in those without mutations.19 Moreover, molecular analysis of the tumors of patients that participated in the TRIBUT20 or IDEAL/INTAC21 study showed that those with an EGFR mutated tumor had a better prognosis. This might be 1 of the reasons that CXCR7 is an independent and significant disease-free prognostic factor only for disease-free prognosis, but not for overall prognosis. In brief, although postoperative early and metastatic recurrences are induced by higher expression levels of CXCR7, a higher incidence of EGFR gene mutations results in a good overall prognosis. Although the clinical relevance is difficult to assess, to establish the reliable clinical significance of CXCR7 status, large-scale and prospective clinical studies should be conducted. In addition, it seems appropriate to perform pilot studies in other types of tumors.

A recent study of current genomic data using cDNA microarrays reported that the gene expression of CCR7 appeared to be associated with poor prognosis in stage I-II NSCLC.38 However, in that report, the CXCR3/4/7 genes were not included in microarray data sets. Heretofore, several cDNA microarrays studies regarding NSCLC have been reported, but all those studies did not evaluate the prognostic roles regarding the CXCR3/4/7 genes.39-43 Therefore, in the future, we propose to assess genomic data using cDNA microarrays and including CXCR3/4/7 genes in the large-scale and prospective clinical studies. Furthermore, we only assessed gene expression of CXCR3/4/7 in this study. To clarify the exact clinical significance as well as the exact biological function, we must assess the protein expression. Previously, Miao and colleagues showed that in murine lung and breast carcinoma cells, RNA interference of CXCR7 reduces growth of experimental lung metastases.15 Wang and colleagues showed that in prostate cancer cells, CXCR7 expression increases with increasing tumor grade using the quantitative mRNA analysis, and also showed the significant expression of CXCR7 protein using fluorescence-activated cell sorting analysis. These existing data suggested that mRNA expression of CXCR7 and its protein expression are likely to be concordant.44 With regard to CXCR3 and CXCR4, some studies showed the relation between mRNA and protein expression in melanoma, NSCLC, colorectal cancer, and breast cancer cells.14, 32, 45, 46 These studies also suggested that mRNA expression of CXCR3 and CXCR4 and its protein expression are likely to be concordant. As well as these studies, we also propose to assess protein expression of CXCR3/4/7 in future studies.

In conclusion, our results demonstrate that higher expression levels of CXCR7 appear to be associated with postoperative metastatic recurrence and poor disease-free prognosis in patients with p-stage I NSCLC. Our results also demonstrate that higher expression levels of CXCR7 correlates to a higher incidence of EGFR gene mutations; therefore, the CXCR7 expression may not be a significant factor for overall survival. Studies of CXCR7 may lead to the development of diagnostic agents and targeted therapy for patients with p-stage I NSCLC.

Acknowledgements

We thank Drs. Kazumasa Takenaka, Shinya Ito, Masatsugu Nakagawa, and Eiji Ogawa (Department of Thoracic Surgery, Kyoto University Hospital) for their help in collecting clinical samples and for their technical assistance. We thank Drs. Daisuke Harada, Hirokazu Kotani, Yoshiaki Ueda, and Toshiaki Manabe (Laboratory of Anatomic Pathology, Kyoto University Hospital) for pathological diagnosis of tumors included in this study. We also thank Miss Seiko Sakai for helpful assistance in preparation of the manuscript.

Conflict of Interest Disclosures

No funding for this project was received from any sponsors.

Ancillary