Expression of CXCR4 Predicts Poor Prognosis in Patients with Malignant Melanoma

Malignant melanoma represents about 2% to 3% of all malignancies in the United States and North Europe. The incidence of malignant melanoma has increased steadily during the last decades with a high mortality rate because of the metastatic dissemination of tumor cells and generation of chemoresistance (1). Although surgery can be curative in localized disease, a large number of patients develop distant metastases. Despite multiple clinical trials that have tested a wide variety of anticancer strategies, ranging from surgery to immunotherapy, radiotherapy, and chemotherapy, the average survival rate is of 6 to 10 months. To date, the most important prognostic factors are Breslow's tumor thickness, ulceration, and lymph node status (2). Diagnosis of malignant melanoma at very early stages (i.e., before the malignant melanocytes become invasive) allows for complete surgical excision of the primary tumors. In these cases there is over a 95% success rate at stage I/II and it can even substantially extend the long term survival of patients with lymph node infiltration (stage III; refs. 1, 2).

The migration of tumor cells to a secondary site from their primary location one is a key event in cancer metastasis. Many studies suggest the mechanisms used for homing of leukocytes and hematopoietic progenitors may be appropriated for the dissemination of tumors via the bloodstream and lymphatics. CXCR4 is a seven-domain trans-membrane chemokine receptor predominantly expressed on lymphocytes where it activates chemotaxis. The CXCR4/CXCL12 pathway has been recently involved in stimulating the metastatic process of many different neoplasms (3), where CXCR4 is able to activate a plethora of phenomena such as chemotaxis, invasion, angiogenesis, and proliferation. These phenomena have been shown with different methodologic approaches in breast cancer (4, 5), small cell lung cancer (6), ovarian cancer (7, 8), pancreatic cancer (9), lymphoma (10–12), neuroblastoma (13), glioblastoma (14, 15), renal cell carcinoma (16), thyroid cancer (17–19), rabdomyosarcoma (20), prostate cancer (21), and colorectal cancer (22).

The role of CXCR4 receptor in the biology of malignant melanoma has been previously described. CXCR4 expression was detected in human melanoma cell lines and in patient samples. The authors described expression of CXCR3 and CXCR4 receptors regulating cell motility during invasion as well as cell proliferation and survival (23). Insights into this mechanism showed that CXCL12 promotes melanoma cell invasion through activation of membrane type I–bound matrix metalloproteinase and small GTPase proteins RhoA, Rac1, and CDC42 (24). Interestingly, CXCR4 expression on melanoma cells is increased by transforming growth factor-β1 and specific inhibition with anti-transforming growth factor-β antibodies blocked CXCR4 expression and melanoma cell invasion toward CXCL12 (24). Overexpression of CXCR4 dramatically enhanced the metastatic accumulation of B16 melanoma cells in mice lungs (25). Expression of functional CXCR4 receptor on melanoma cells indicates that it might contribute to cell motility during invasion of lymph nodes as well as to regulation of cell proliferation and survival.

The aim of the present study was to evaluate the expression of CXCR4 in malignant melanoma patients and its possible prognostic value in 71 human primary melanoma samples and relative sentinel lymph nodes (SLN).

Patients were seen at the Clinical Immunology Department of the National Cancer Institute of Naples from 1996 to 2003. The histologic sections were reviewed by two expert pathologists (G.B. and R.F.) to verify the histologic diagnosis of malignant melanoma, before performing immunohistochemistry. Standard therapeutic strategies were applied for these patients, according to SLN status. Patients with soft tissue recurrence or local disease were treated with surgery. Patients with advanced disease were mainly treated with palliative chemotherapy. A few selected patients underwent pulmonary metastasectomy in the case of a solitary lesion or single lobe involvement and adequate disease-free survival (DFS, at least 1 year). Locations of the primary tumor were grouped in four classes: head/neck, trunk, extremities, and feet. Breslow's tumor thickness in the studied samples was selected to be >1 mm. SLN status was recorded. In particular, the presence of metastasis was considered as follows: micrometastasis if their diameter was ≤2 mm, macrometastasis if the diameter was >2 mm.

Immunohistochemistry. Two serial 5-μm sections of formalin-fixed, paraffin-embedded cutaneous melanoma samples were stained one by standard H&E and the other by the biotin-streptavidin-peroxidase method (YLEM). Deparaffinized sections were microwaved in 1 mmol/L EDTA (pH 8.0) for two cycles of 5 minutes each to unmask epitopes. After treatment with 1% hydrogen peroxidase for 30 minutes to block endogenous peroxidases, the sections were incubated with monoclonal antibodies (anti-CXCR4, clone 44716, R&D systems, Minneapolis, MN) for 2 hours at room temperature. The sections were then incubated with biotin-labeled secondary antibody (1:30) for 30 minutes and with streptavidin-peroxidase (1:30) for 10 minutes. Slides were stained for 10 minutes with AEC chromogen (DAKO, Milan, Italy) and then counterstained with hematoxylin, washed, and mounted in waterish medium. The dilutions of the monoclonal antibody, biotin-labeled secondary antibody, and streptavidin-peroxidase were made with PBS (pH 7.4) containing 5% bovine serum albumin. All series included positive controls of well-characterized sections (melanomas and breast cancer). Negative controls were obtained by substituting the primary antibody with a mouse myeloma protein of the same subclass, at the same concentration as the monoclonal antibody. All controls gave satisfactory results. In all cases, normal epithelial cells were negative for CXCR4 expression. Neoplastic cells with CXCR4 cytoplasmatic and/or membrane immunohistochemical expression were considered positive cells. Macrophage positivity was used as adequate internal positive control for each case, to validate technical procedure. Staining was categorized into four semiquantitative classes based on the rate of stained (positive) tumor cells: absence of staining, <10% positive cells (low), 10% to 50% positive cells (moderate), and >50% positive cells (high). Topographical expression of CXCR4 was also evaluated. Prevalent expression in radial or vertical component was considered when expression in such component was twofold compared with the other. Moreover, CXCR4 expression was evaluated in SLN. In particular, cases were considered positive when >10% of metastatic cell showed cytoplasmatic and/or membrane CXCR4 expression. Slides were evaluated by two blinded observers (G.B. and R.F.); discordant cases were discussed and concordance was then achieved.

Statistical Analysis. Correlations between CXCR4 expression, baseline patient features, and tumors were studied by contingency tables and the χ² test. DFS was defined as the time elapsed from the date of the initial diagnosis to the appearance of local relapse or distant metastasis or death, whichever occurred first. Overall survival (OS) was defined as the time elapsed from the date of the initial diagnosis to death or to the date of the last available information on vital status. Kaplan-Meier product limit method was applied to draw DFS and OS curves. Univariate analysis was done with the log-rank test. Cox proportional-hazards regression was used to analyze the effect of several risk factors on DFS and OS. The probability of the end point (hazard) was estimated with a coefficient: exp(b). The quantity exp(b) can be interpreted as the instantaneous relative risk of an event, at any time, for an individual with the risk factor present compared with an individual with the risk factor absent, given both individuals are the same on all other covariates. Risk factors (covariates) were considered dichotomous (male versus female, age < 70 versus age ≥ 70, ulceration versus no ulceration, SLN involved versus SLN not involved, Breslow ≤ 2.00 mm versus >2.00 mm, CXCR4 absent versus CXCR4 expressed). Ninety-five percent confidence intervals (95% CI) of hazard ratios are also reported.

Patient's Features. From 1996 to 2003, 71 specimens from surgically resected primary cutaneous malignant melanomas with Breslow tumor thickness >1 mm were tested for the CXCR4 expression. All resections and SLN biopsies were done by the same surgical team (Surgical Oncology B, National Cancer Institute of Naples). Three patients referred to our center only at the moment of the diagnosis and there was no adequate follow-up to allow time-to-event analysis. Characteristics of all patients and primitive melanoma are summarized in Table 1. Overall, 31 of 71 tumors (43.6%) expressed CXCR4. Median age was 53 years; 14 patients ages ≥70 years. Genders were equally represented. Melanomas located in the trunk or extremities accounted for 61.9% and 29.5% of the cases, respectively. Twenty-nine patients (40.8%) had Breslow tumor thickness >1.00 and ≤ 2.00 mm, 26 (36.6%) patients >2.00 and <4.00 mm, and 16 patients >4.00 mm. Clark level was II in one case (1.4%), III in 24 cases (33.8%), and IV in 46 cases (64.7%). Ulceration was present in 33 cases (46.5%) and nodular appearance in 32 lesions (45.1%). Twenty-five SLNs (35.3%) were infiltrated by malignant melanoma; in particular, 12 were micrometastasis. No correlation was found between CXCR4 expression and clinicopathologic characteristics.

CXCR4 Expression in Primary Melanoma: Pattern of Expression. Out of 71 melanoma specimens tested for the CXCR4 expression 31 (43.7%) stained positive for CXCR4. The positive cases were divided into three score classes according to their staining: low in 15 cases (48%), moderate in 10 (32%), and high in 6 (19%). Figure 1 shows representative positive staining of human melanoma (A-F) and sentinel lymph nodes (G and H). In all melanoma specimens, normal skin tissue resulted in negative staining. Melanoma samples showed mainly a subepidermal expression (Fig. 1A and B; high and low expression group, respectively). A small subgroup of 10 of 31 positive cases expressed CXCR4 also in the vertical growth phase (Fig. 1C and D; high and low expression group, respectively). Different cellular localization of CXCR4 were detected; the staining was cytoplasmic (Fig. 1E) and/or membrane (Fig. 1F). The staining mainly involved cytoplasms. In a few cases, CXCR4 was exclusively detected at the membrane level (8 of 31, 25%). CXCR4 expression was detected in involved lymph nodes in both diffuse metastasis (Fig. 1G) and micrometastatis (Fig. 1H).

CXCR4 Expression: Correlation with Outcome. At the time of this analysis, after a median follow-up for alive patients of 38 months, 25 patients progressed, and 19 died. Univariate analysis of prognostic factors for DFS and OS is summarized in Table 2. SLN metastases, presence of ulceration and CXCR4 expression had a significant prognostic value for DFS and OS. The predictive ability of ulceration and CXCR4 expression for DFS, and of SLN metastases and CXCR4 expression for OS were confirmed by a multivariate analysis Table 3) adjusted by age, sex, Breslow, SLN status, presence of ulceration, and CXCR4 expression. Graphic pattern of Kaplan-Meier estimated curves (Fig. 2A and B) suggest that prognosis is unfavorable for patients expressing CXCR4. For CXCR4-positive tumors median DFS and OS were 22 and 35 months, respectively; and hazard ratios of relapse and death as compared with patients with CXCR4-negative tumors of 2.5 (95% CI, 1.2-6.1) and 3.1 (95% CI, 1.1-7.2).

Lymph node spreading represents the most common way of metastasis in malignant melanoma. Expression of CXCR4 in SLN biopsies was done in 56 of 71 primary melanoma. As described, 25 SLNs were infiltrated by melanoma. CXCR4 expression was detected in 10 of 17 (58.8%) evaluable and melanoma-positive SLNs.

In the present study, we investigated the potential role of CXCR4 as a prognostic factor in human melanoma. CXCR4 expression was detected in 31 of 71 primary cutaneous melanoma selected to have a Breslow > 1 and in 10 of 17 evaluable melanoma involved lymph nodes. Interestingly, the expression of CXCR4 was able to predict the prognosis and this ability holds true both for DFS (P = 0.0154) and OS (P = 0.0009). To the best of our knowledge, this is the first report demonstrating the prognostic role for CXCR4 in malignant melanoma.

It has been postulated that organ specific metastasis might be governed, in part, by interactions between chemokine receptors on cancer cells with metastatic potential and chemokine gradients in target organs (26). Melanoma cell lines express functional chemokine receptors CCR7, CXCR3, CCR10, and CXCR4 which are able to activate cell motility during invasion, cell proliferation, and survival (23–25). CXCR4 staining mainly involved cytoplasms. In a few cases, CXCR4 was exclusively detected at the membrane level (8 of 31, 25%). It is noteworthy that four of eight patients showing a specific membrane staining died at the observation time point (38 months of follow-up). However, we will not approach any conclusions since the number of examined events is too small. The cytoplasmatic detection of the CXCR4 receptor was previously described in cancers of the breast (29), lung (30), colon (31–33), prostate (34), pancreas (9), and in CD34⁺ hematopoietic cells (35). It might represent a functional status of the receptor because the binding to the specific ligand induce receptor internalization (36) and specific signals are required in order to express CXCR4 on cell surface (22).

The presence of metastasis in the SLN is important for staging and prognosis (1, 2). The metastatic potential of primary melanoma is considerably higher than that of other primary solid tumors when comparing the size of primary lesion. Malignant melanoma metastasizes frequently to regional tumor-draining lymph nodes, preferentially via the lymphatics. Notably, only 8 of 27 melanomas with Breslow between 1 and 2 mm (T2 tumors at tumor-node-metastasis, American Joint Committee on Cancer 2002 staging system) relapsed in lymph nodes; six out of these tumors expressed CXCR4. These observations might suggest that thin melanomas CXCR4-positive cells are more prone to move toward the draining lymph node. Previous reports described a correlation between CXCR4 expression and lymph node metastasis in 79 surgically resected invasive ductal carcinomas (29) and with clinical outcome in 61 patients with completely resected non–small cell lung cancer (30). In melanoma-forced expression of CXCR4 in B16 melanoma cells enhances pulmonary metastasis, but metastasis to lymph nodes, liver, or kidney are not affected (25). Interestingly, we detected CXCR4 expression in 10 of 17 evaluable melanoma involved lymph nodes; only seven involved lymph nodes derived from primary melanoma expressed CXCR4 suggesting a role of other chemokine receptors such as CCR7 (27) and CXCR3 (28) in lymph node metastasis.

Our data show that the higher the expression of CXCR4 the worst the prognosis of patients supporting the inhibition of CXCR4 as a possible therapeutic target in melanoma. Blocking anti-CXCR4 monoclonal antibody (4) or to inhibitory peptide (37) specifically inhibited metastasis to the lungs in breast cancer models. Inhibition of the CXCR4/CXCL12 axis was previously studied since CXCR4 plays an important role in HIV type-1 (HIV-1) infection (38, 39). AMD3100 is a byciclam noncompetitive antagonist of CXCL12 binding to CXCR4. Rubin showed that systemic administration of AMD3100 inhibits intracranial growth of glioblastoma and medulloblastoma xenograft increasing apoptosis and decreasing cell proliferation (40). AMD3100 seemed safe and effective in patients affected by multiple myeloma and NHL aimed to induce a rapid mobilization of CD34⁺ cells (41). Other than AMD3100 several CXCR4 inhibitors are already available (4, 37–42). Recently, a single treatment with a microcapsule containing a CXCR4 antagonist, (4F-benzoyl-TE14011-PLA) was showed to suppress pulmonary metastasis of murine melanoma (42). The expression of the receptor at the protein and at the RNA levels in melanoma cell line derived from patient samples will allow studies on testing of specific CXCR4 inhibitors.

Combating melanoma metastasis remains a major clinical challenge, and thus far results are disappointing. Having a molecular “diagnosis” in melanoma would permit identification of patients at high risk of relapse and death. CXCR4 inhibitors alone or in association with other biological agents targeting enhanced proliferation, increased resistance to apoptosis and increased cell migration may provide therapeutic alternatives and/or support to conventional cytotoxic therapies. Our results are consistent with the hypotesis that the CXCR4/CXCL12 pathway activation could promote tumor progression.

We thank the Melanoma Cooperative Group of Naples: Aprea P., Ascierto P.A., Ayala F., Beneduce G., Bosco L., Botti G., Caracò C., Castello G., Celentano E., Chiofalo M.G., Comella G., Daponte A., De Marco M.R., Graziano F., Leonardi E., Lombardi M.L., Marfella A., Montella M., Mori S., Mozzillo N., Napolitano M., Peluso G.F., Perrone F., Pirozzi G., Satriano S.M.R., Scala S., and Tatangelo F. (National Cancer Institute, G. Pascale, Naples, Italy); Satriano R.A., Vozza A., Ruocco V (2nd University of Naples, Italy); Palmieri G. [Division of Cancer Genetics, ICB-Consiglio Nazionale delle Ricerche, Alghero (SS), Italy]; Cossu A., Tanda F. (University of Sassari, Italy); Dr. M.R. Ventura for data management; and J. Bryce for technical assistance.