Association of CXCR4 and CCR7 chemokine receptor expression and lymph node metastasis in human cervical cancer

Abstract

Background: The chemokine receptors CXCR4 and CCR7 have been suggested to play an important role in cancer invasion and metastasis. The expression of these receptors in human cervical cancer, however, has seldom been characterized.

Patients and methods: We investigated the expression of CXCR4 and CCR7 in cervical cancer specimens and determined the association between their expression and the clinicopathological features observed, including patient outcome.

Results: CXCR4 expression was significantly higher in elderly patients (P = 0.025); it was also significantly increased in patients with cancers displaying large tumor size (P = 0.010), deep stromal invasion (P = 0.0004), lymph–vascular space involvement (P = 0.0002), or lymph node metastasis (P < 0.0001). CCR7 expression was significantly higher in cases of squamous cell carcinomas (P = 0.010) and in patients with cancers showing large tumor size (P < 0.0001), deep stromal invasion (P < 0.0001), vaginal invasion (P = 0.047), lymph–vascular space involvement (P = 0.012), or lymph node metastasis (P < 0.0001). Logistic regression analysis revealed that deep stromal invasion (P = 0.017) and CXCR4 (P = 0.016) and CCR7 (P = 0.022) expression were independent factors that influenced pelvic lymph node metastasis. The disease-free survival and overall survival (OS) rates of patients exhibiting both CXCR4 and CCR7 expression were significantly reduced (P < 0.0001). In addition, the expression of both CXCR4 and CCR7 was an independent prognostic factor for OS (95% confidence interval = 1.03–17.86; P = 0.046).

Conclusions: CXCR4 and CCR7 expression may be associated with lymph node metastasis; moreover, the expression of these receptors can serve as an indicator of poor prognosis in patients with cervical cancer.

cervical cancer, CCR7, CXCR4, lymph node metastasis, prognosis

introduction

Tumor cell invasion and subsequent dissemination via the bloodstream and lymph vessels are critical steps in the progression of malignant tumors, including cervical cancer. The mechanisms involved in lymph node metastasis, however, are not fully understood. In a previous study, we demonstrated that the expression of vascular endothelial growth factor-C is involved in the promotion of lymph node metastasis in cervical cancer [1]. The determination of such a predictive molecular marker that is correlated with lymph node metastasis might be important in elucidating the mechanism of metastasis.

Chemokines belong to the small-molecule chemoattractive cytokine family and are categorized into four groups (CXC, CC, CX3C, and C) on the basis of the characteristic presence of four conserved N-terminal cysteines [2–4]. Chemokines mediate their chemical effect on target cells through G-protein-coupled heptahelical receptors, which are structurally characterized by seven transmembrane-spanning domains and are involved in the attraction of mononuclear and polymorphonuclear leukocytes. Recent data indicate that chemokine receptors may direct lymphatic and hematogenous spread and may additionally influence the sites of metastatic growth of different tumors [5]. Different cancers express varying combinations of CC and CXC chemokine receptors. The chemokine receptor CXCR4 appears to be expressed by a majority of cancer types [6]. CXCR4 was initially reported to regulate the homing of lymphocytes to inflammatory tissues [7]. The CXCR4/CXCL12 pathway is involved in stimulating the metastatic process of many different neoplasms, in which CXCR4 activates various phenomena, such as chemotaxis, invasion, angiogenesis, and proliferation [8]. Müller et al. [9] reported that neutralizing the interaction of CXCR4 and CXCL12 in vivo significantly impairs the metastasis of breast cancer cells to regional lymph nodes and the lungs. CCR7—the receptor for the two major chemokines CCL19 and CCL21—is expressed on naive T cells, memory T cells, B cells, and mature dendritic cells and is considered to play an important role in lymphocyte cell trafficking and homing to lymph nodes [10, 11]. It has been reported that CCR7 expression specifically enhances the metastatic ability of B16 melanoma cells through a lymph-mediated but not a blood-borne pathway [12].

Thus, the chemokine receptors CXCR4 and CCR7 may be important in lymph node metastasis; however, their expression in human cervical cancer has not been characterized. Therefore, we investigated the expression of CXCR4 and CCR7 in cervical cancer specimens by using immunohistochemistry. In addition, the association between CXCR4 and CCR7 expression and the clinicopathological features of cervical cancer, including patient prognosis, was determined.

patients and methods

patients and tissue samples

In this study, we examined 174 patients presenting with the International Federation of Gynecology and Obstetrics (FIGO) stages IB–IIB cervical cancers. Each of these patients underwent radical hysterectomy and pelvic lymphadenectomy at the Department of Obstetrics and Gynecology of the Okayama University Medical School. Tumor specimens were obtained at the time of surgery and were immediately fixed in 10% neutral-buffered formalin and embedded in paraffin. Informed consent was obtained from each patient before sample collection. Histological cell typing was conducted according to the World Health Organisation classifications as follows: 102 were classified as squamous cell carcinomas, 49 as adenocarcinomas, and 23 as adenosquamous carcinomas. Clinical staging was assessed based on the FIGO staging system as follows: 92 were allocated to stage IB, four to stage IIA, and 78 to stage IIB. The median age at the time of surgery was 46 years (range 25–67 years). Patients with lymph node metastasis, parametrial involvement, deep stromal invasion, or marked lymph–vascular space involvement were treated with an adjuvant external whole-pelvic irradiation, combination chemotherapy, or chemoradiation. Disease-free survival (DFS) and overall survival (OS) rates were defined as the interval from the initial surgery to clinically or radiologically proven recurrence and death, respectively. The end date of the follow-up study for conducting the analysis was 31 March 2006, and the median duration of the follow-up was 56.5 months (range 1–131 months).

immunohistochemistry and staining evaluation

Sections of 4-μm thickness were obtained from several representative areas of each tumor specimen and were mounted on to glass slides for immunostaining according to the labeled streptavidin biotin procedure of the Dako LSAB kit (Dako North American Inc., Carpinteria, CA). Briefly, after the slides were dewaxed in xylene and rehydrated in an alcohol series, antigen retrieval was carried out in a microwave oven in 10 mM citric acid buffer (pH 6.0) for 3 × 10 min. The sections were then incubated with 0.3% hydrogen peroxide to block endogenous peroxidase activity, followed by incubation with normal horse serum for 5 min at room temperature. Immunostaining was then carried out by incubation with anti-CXCR4 mouse mAb (R&D systems Inc., Minneapolis, MN) and anti-CCR7 rabbit mAb (Epitomics Inc., Burlingame, CA) for 2 h at room temperature. The sections were subsequently incubated for 20 min with biotinylated anti-mouse and anti-rabbit immunoglobulin, followed by incubation with peroxidase-conjugated streptavidin for 20 min and 0.05% 3,3′-diaminobenzidine tetrahydrochloride solution (Wako Pure Chemical Industries Ltd., Osaka, Japan) containing hydrogen peroxide for 10 min. Finally, the slides were counterstained with Mayer's hematoxylin and mounted in an aqueous mounting medium. At each step, the slides were washed carefully in phosphate-buffered saline (pH 7.4).

The specimen was considered as positive if the distinct cytoplasmic staining was observed in >50% cancer cells. In our study of endometrial cancer, CXCR4 and CCR7 protein levels were significantly correlated to messenger RNA levels on the basis of these criteria. Microscopic analyses were conducted independently by two of the authors who had no prior knowledge of the clinical data. The final evaluations of a few ambiguous cases were carried out using a conference microscope.

statistical analyses

The association between the variables was tested using the chi-square test or a stepwise logistic regression analysis. The survival rates were calculated by the Kaplan–Meier method and the differences between the survival curves were examined by the log-rank test. Factors found to be significant were then selected for a stepwise Cox multivariate proportional hazard model in order to determine their prognostic values. These analyses were carried out using the Stat-View 5.0 software (Abacus Concepts Inc., Berkley, CA). Probability values of <0.05 were considered statistically significant.

results

CXCR4 protein expression in cervical cancer tissue

Figure 1A, B, and C shows the representative immunostaining profiles of CXCR4 in cervical cancer specimens. In normal cervical tissues, we found no distinct CXCR4 staining. Of the 174 tumor samples examined, 110 (63%) were positively stained and 64 (37%) were negatively stained. The association between CXCR4 expression and clinicopathological factors is shown in Table 1. CXCR4 expression was significantly higher in elderly patients (P = 0.025) and in FIGO stage II tumors (P = 0.024); it was also significantly elevated in patients with cancers showing large tumor size (P = 0.010), deep stromal invasion (P = 0.0004), lymph–vascular space involvement (P = 0.0002), or lymph node metastasis (P < 0.0001). Furthermore, no significant association was observed between CXCR4 immunoreactivity and histological type, vaginal invasion, parametrial invasion, or ovarian metastasis.

Figure 1

Immunohistochemical staining of CXCR4 and CCR7 in cervical cancer specimens. Original magnification ×200. (A) Negative CXCR4 staining in an adenocarcinoma. (B) Positive CXCR4 staining in an adenosquamous cell carcinoma. (C) Positive CXCR4 staining in an adenocarcinoma. (D) Negative CCR7 staining in a squamous cell carcinoma. (E) Positive CCR7 staining in a squamous cell carcinoma. (F) Positive nuclear CCR7 staining in an adenocarcinoma.

Table 1

Association between CXCR4 expression and clinicopathological factors in cervical cancers

Variables	CXCR4 expression		P valuea
	−	+
Age (years)			0.025
<50	47	62
≥50	17	48
FIGO stage			0.024
I	41	51
II	23	59
Histological type			0.446
Adenocarcinoma	20	29
ADSQ	9	14
SCC	35	67
Tumor size (cm)			0.010
≤4	58	82
>4	6	28
Stromal invasion			0.0004
≤2/3	43	43
>2/3	21	67
Vaginal invasion			0.197
Negative	58	92
Positive	6	18
Parametrial invasion			0.377
Negative	52	83
Positive	12	27
LVS involvement			0.0002
Negative	42	40
Positive	22	70
Lymph node metastasis			<0.0001
Negative	61	78
Positive	3	32
Ovarian metastasis			0.621
Negative	63	107
Positive	1	3

a

Chi-square test.

ADSQ, adenosquamous carcinoma; FIGO, International Federation of Gynecology and Obstetrics; LVS, lymph–vascular space; SCC, squamous cell carcinoma.

CCR7 protein expression in cervical cancer tissue

Figure 1D and E shows the characteristic immunostaining profiles of CCR7 in cervical cancer specimens. In normal cervical tissues, we found no distinct CCR7 staining. Of the 174 tumor samples, 102 (59%) were positively stained and 72 (41%) were negatively stained. The association between CCR7 expression and clinicopathological factors is shown in Table 2. CCR7 expression was significantly higher in FIGO stage II cases (P < 0.0001) and squamous cell carcinomas (P = 0.010); it was also significantly increased in patients with cancers showing large tumor size (P < 0.0001), deep stromal invasion (P < 0.0001), vaginal invasion (P = 0.047), lymph–vascular space involvement (P = 0.012), or lymph node metastasis (P < 0.0001). Furthermore, no significant association was observed between CCR7 immunoreactivity and age, parametrial invasion, or ovarian metastasis.

Table 2

Association between CCR7 expression and clinicopathological factors in cervical cancers

Variables	CCR7 expression		P valuea
	−	+
Age (years)			0.847
<50	47	62
≥50	29	36
FIGO stage			<0.0001
I	54	38
II	22	60
Histological type			0.010
Adenocarcinoma	29	20
ADSQ	12	11
SCC	35	67
Tumor size (cm)			<0.0001
≤4	72	68
>4	4	30
Stromal invasion			<0.0001
≤2/3	54	32
>2/3	22	66
Vaginal invasion			0.047
Negative	70	80
Positive	6	18
Parametrial invasion			0.065
Negative	64	71
Positive	12	27
LVS involvement			0.012
Negative	44	38
Positive	32	60
Lymph node metastasis			<0.0001
Negative	72	67
Positive	4	31
Ovarian metastasis			0.446
Negative	75	95
Positive	1	3

a

Chi-square test.

ADSQ, adenosquamous carcinoma; FIGO, International Federation of Gynecology and Obstetrics; LVS, lymph–vascular space; SCC, squamous cell carcinoma.

We noticed nuclear CCR7 staining in 24 samples (14%). Figure 1F shows the representative immunostaining profile of nuclear CCR7 in cervical cancer. The association between nuclear CCR7 expression and clinicopathological factors is shown in Table 3. Nuclear CCR7 expression was significantly lower in patients with cancers showing vascular space involvement (P = 0.039). There was no significant association between nuclear CCR7 immunoreactivity and any other clinicopathological factor.

Table 3

Association between CCR7 expression and clinicopathological factors in cervical cancers

Variables	Nuclear CCR7 expression		P valuea
	−	+
Age (years)			0.638
<50	95	14
≥50	55	10
FIGO stage			0.564
I	78	14
II	72	10
Histological type			0.107
Adenocarcinoma	38	11
ADSQ	20	3
SCC	92	10
Tumor size (cm)			0.349
≤4	119	21
>4	31	3
Stromal invasion			0.168
≤2/3	71	15
>2/3	79	9
Vaginal invasion			0.141
Negative	127	23
Positive	23	1
Parametrial invasion			0.842
Negative	116	19
Positive	34	5
LVS involvement			0.039
Negative	66	16
Positive	84	8
Lymph node metastasis			0.650
Negative	119	20
Positive	31	4
Ovarian metastasis			0.418
Negative	146	24
Positive	4	0

a

Chi-square test.

ADSQ, adenosquamous carcinoma; FIGO, International Federation of Gynecology and Obstetrics; LVS, lymph–vascular space; SCC, squamous cell carcinoma.

predictors of lymph node metastasis

The predictors of lymph node metastasis were determined by stepwise logistic regression analysis. Pelvic lymph node metastasis was correlated with tumor size, stromal invasion, parametrial invasion, lymph–vascular space involvement, ovarian metastasis, and CXCR4 and CCR7 expression status (Table 4). Multivariate analysis revealed that deep stromal invasion (P = 0.017) and CXCR4 (P = 0.016) and CCR7 (P = 0.022) expression were the independent factors that influenced pelvic lymph node metastasis.

Table 4

Univariate and multivariate analysis with respect to lymph nodes metastasis

Variable	Comparison	Univariate			Multivariate
		Odds ratio	P value	95% confidence interval	Odds ratio	P value	95% confidence interval
Histological cell type	SCC : non-SCC		NS
Tumor size (cm)	≤4 : >4	4.74	0.0002	2.07–10.82		NS
Stromal invasion	≤2/3 : >2/3	15.81	0.0001	4.62–54.15	5.51	0.019	1.32–22.30
Vaginal invasion	Negative : positive		NS
Parametrial invasion	Negative : positive	2.57	0.022	1.15–5.76		NS
LVS involvement	Negative : positive	7.45	0.0001	2.73–20.34		NS
Ovarian metastasis	Negative : positive	12.94	0.029	1.30–128.51		NS
CXCR4	Negative : positive	8.34	0.0007	2.44–28.54	5.77	0.016	1.35–19.41
CCR7	Negative : positive	8.33	0.0001	2.79–24.85	5.35	0.021	1.24–14.04

SCC, squamous cell carcinoma; LVS, lymph–vascular space; NS, not significant.

univariate survival analyses

Figure 2 presents the DFS curves of the 174 patients with cervical cancer on the basis of the CXCR4 expression status. Both the DFS and OS rates of patients exhibiting CXCR4 expression were significantly lower than those of patients lacking CXCR4 expression (P = 0.025 and P = 0.0001, respectively). Figure 3 presents the DFS curves of the 174 patients with cervical cancer on the basis of the CCR7 expression status. Both the DFS and OS rates of patients exhibiting CCR7 expression were significantly lower than those of patients demonstrating the lack of CCR7 expression (P = 0.006 and P = 0.038, respectively). Furthermore, the DFS and OS rates of patients exhibiting both CXCR4 and CCR7 expression were significantly poor (P < 0.0001) (Figure 4). There was no association between nuclear CCR7 expression and patient outcome. The results of the univariate survival analyses of other variables are shown in Table 5.

Figure 2

(A) Disease-free survival and (B) overall survival curves of the 174 cervical cancer patients according to their CXCR4 expression status.

Figure 3

(A) Disease-free survival and (B) overall survival curves of the 174 cervical cancer patients according to their CCR7 expression status.

Figure 4

(A) Disease-free survival and (B) overall survival curves of the 174 patients displaying cervical cancer according to their CXCR4 and CCR7 expression status.

Table 5

DFS and OS analyses of prognostic factors using the log-rank test

Variables	Number	Estimated 5-year DFS (%)	P valuea	Estimated 5-year OS (%)	P valuea
Age(year)			NS		NS
<50	109	87.3		90.5
≥50	65	85.1		91.6
Histology			NS		0.043
SCC	102	87.5		96.2
Non-SCC	72	84.9		83.4
Tumor size(cm)			0.015		0.043
≤4	140	89.9		93.4
>4	34	72.8		81.0
Stromal invasion			0.001		0.003
≤2/3	86	95.2		96.6
>2/3	88	78.2		85.4
Vaginal invasion			NS		NS
Negative	150	86.2		91.5
Positive	24	87.5		87.3
Parametrial invasion			0.002		0.007
Negative	135	90.5		94.4
Positive	39	72.6		79.5
LVS involvement			0.0003		0.0009
Negative	82	97.2		98.8
Positive	92	76.4		83.4
Lymph node metastasis			<0.0001		<0.0001
Negative	139	94.2		95.1
Positive	35	54.2		74.1
Ovarian metastasis			0.004		0.0005
Negative	170	87.3		92.0
Positive	4	50.0		50.0
CXCR4			0.025		0.0001
Negative	64	94.9		97.2
Positive	110	81.1		81.0
CCR7			0.006		0.0378
Negative	76	95.4		96.1
Positive	98	79.4		81.8
Nuclear CCR7			NS		NS
Negative	150	85.1		95.0
Positive	24	89.5		100.0

a

Kaplan–Meier test.

DFS, disease-free survival; LVS, lymph–vascular space; NS, not significant; OS, overall survival; SCC, squamous cell carcinoma.

multivariate survival analysis

We examined the prognostic value of both CXCR4 and CCR7 expression using the Cox proportional hazard model. Multivariate analysis showed that lymph node metastasis was the only independent prognostic factor for DFS [95% confidence interval (CI) 1.20–10.87; P = 0.022]. In addition, non-squamous cell carcinoma was the strongest independent prognostic factor for OS, followed by both CXCR4 and CCR7 expression (95% CI 1.09–10.20, P = 0.035, and 95% CI 1.03–17.86, P = 0.046, respectively).

discussion

This is the first study analyzing CXCR4 and CCR7 protein expressions in a large series of human cervical cancer specimens. We demonstrated that both CXCR4 and CCR7 expression is significantly higher in patients with cancers displaying large tumor sizes, deep stromal invasion, lymph–vascular space involvement, or lymph node metastasis. Furthermore, CXCR4 expression was significantly higher in elderly patients, while CCR7 expression was significantly greater in patients with cancers exhibiting vaginal invasion. The stepwise logistic regression analysis revealed that both CXCR4 and CCR7 expression influenced pelvic lymph node metastasis independently. This finding indicates that both CXCR4 and CCR7 expression is up-regulated in cases with positive nodes and may contribute to the lymph node metastasis occurring in cervical cancers. The present data are consistent with previous reports that describe a positive correlation between CXCR4 expression and lymph node metastasis in cases of non-small-cell lung cancer (NSCLC), nasopharyngeal cancer, colorectal cancer, and esophageal cancer [13–16]. Positive correlation has also been reported between CCR7 expression and lymph node metastasis in cases of NSCLC, breast cancer, gastric cancer, colorectal cancer, esophageal cancer, prostate cancer, and oral and oropharyngeal squamous cell carcinoma [13, 17–22]. Although the role of chemokines and their receptors in human cancers is complex, the chemokine receptors CXCR4 and/or CCR7 may have a critical role in determining lymph node metastasis in these kinds of tumors.

In concordance with previous studies conducted on other cancer types, the staining for CXCR4 and CCR7 was predominantly cytoplasmic. This might represent a functional status of the receptor because binding to a specific ligand induces receptor internalization [23]. The necessity of internalization for chemotaxis and signaling remains controversial. Endosomes are gaining considerable attention as scaffolds for signaling complexes. The assembly of signaling complexes on intracellular endosomal membranes indicates that the intracellular trafficking itinerary of chemokine receptors may have important implications for signaling [24]. We also showed that nuclear CCR7 expression was observed in some cases, although the nuclear staining of CXCR4 was not observed in any case. Nuclear CCR7 expression was exclusively observed in patients with cancers showing negative lymph–vascular space involvement. Cabioglu et al. [17] reported that the nuclear expression of CCR7 and CXCR4 was significantly higher in lymph node-negative breast cancers. They speculated that the different localization patterns of chemokine receptors, i.e. nuclear versus cytoplasmic, appear to have a different biological significance for the metastatic potential of cancer cells. Further studies are needed to elucidate the role of the nuclear localization of these chemokine receptors.

Thus far, no studies have reported the clinical outcome of cases with CXCR4 and CCR7 expression in cervical cancer. Our study is the first to demonstrate that the expression of CXCR4 and CCR7 is associated with reduced survival outcomes in patients with cervical cancer. Furthermore, the multivariate analysis in the present study showed that the expression of both CXCR4 and CCR7 is an independent prognostic factor for the OS of patients. An increased expression of CXCR4 has been reported to correlate with a poor prognosis in esophageal and nasopharyngeal carcinomas [16, 25], whereas increased expression of CCR7 has been reported to be correlated with a poor prognosis in gastric, colorectal, esophageal, prostate, and oral and oropharyngeal squamous cell carcinomas [18–22]. These studies indicate that CXCR4 and CCR7 may be promising prognostic factors for these types of cancer.

In conclusion, the expression of CXCR4 and CCR7 may be associated with lymph node metastasis in cervical cancer. Our findings also provide evidence that both CXCR4 and CCR7 expression can serve as an indicator of poor prognosis in patients with cervical cancer.

We would like to Chisae Kodera for her help with histochemistry and in the preparation of sections.