Significantly decreased expression levels of chemokine receptor CXCR4 (A) and matrix metalloproteinase MMP2 (B) were found in primary melanomas of patients with positive sentinel lymph nodes (SLNs) (n = 11) compared with SLN-negative patients (n = 17). ΔΔCT indicates delta-delta threshold cycle. The y-axes represent the amounts of the respective messenger RNA (calculated as 2−ΔΔCT). A superficial spreading melanoma with a tumor thickness of 0.65 mm from a clinically SLN-negative female patient served as the calibrator. The mean values for each group are represented as horizontal lines.
Otto K, Starz H, Becker JC, Schrama D. Overexpression of Matrix Metalloproteinases, Chemokines, and Chemokine Receptors Relevant for Metastasis in Experimental Models Not an Indication of Lymph Node Metastases in Human Melanoma. Arch Dermatol. 2007;143(7):945-955. doi:10.1001/archderm.143.7.947
Morphologic characteristics such as tumor thickness and ulceration are the major accepted indicators of risk for metastatic spread. Since matrix metalloproteinases and chemokines and their receptors are involved in the complex process of metastasis, we tested whether their expression predicts the risk of melanoma progression. To this end, the relative messenger RNA (mRNA) expression of chemokine receptors CCR7 and CXCR4, the chemokine CXCL8, and the matrix metalloproteinases MMP2 and MT1-MMP in primary cutaneous melanomas from 28 patients was correlated with the presence of micrometastases in the respective sentinel lymph nodes.
We gained approval from our local ethics committee, and analyses were performed with the patients’ informed consent. Primary tumors were trimmed to remove most of the surrounding unaffected skin, and total RNA was isolated from frozen tissue sections. After complementary DNA was generated from this RNA, relative expression of CCR7, CXCR4, CXCL8, MMP2, and MT1-MMP was determined by real-time polymerase chain reaction and the comparative delta-delta threshold cycle (ΔΔCT) method where glyceraldehyde-3-phosphate dehydrogenase served as endogenous control. The results were statistically tested after log-normal transformation using the 2-tailed nonpaired t test.
For CXCR4 (P = .01) and MMP2 (P = .04), an inverse relationship between mRNA expression and presence of lymph node metastases was detected (Figure). Primary tumors associated with concurrent lymph node metastasis, while not reaching statistical significance, also showed lower expression of CCR7 (P = .06) and MT1-MMP (P = .09). For CXCL8 (P = .96), no correlation was observed.
The statistically significant inverse correlation of the expression of genes known to promote metastasis (CXCR4 and MMP2) and the presence of lymph node micrometastases was unexpected. For example, Müller et al1 detected an increased expression of CXCR4 and CCR7 in human melanoma cells compared with primary melanocytes. In a murine melanoma model, however, CXCR4 expression did not enhance occurrence of lymph node metastasis.2
Notably, cell signaling is influenced not only by the amount of expressed chemokine receptors or their ligands but also through regulatory mechanisms. Indeed, despite high expression of CXCR4 on germinal center T cells, their migration to CXCL12 was diminished owing to follicular dendritic cell–mediated expression of regulators of G protein signaling 13 and 16.3 In addition, tumor cells are highly flexible, ie, they can change from proteolytic migration to proteolysis-independent movement in an ameboid manner.4 Furthermore, heterogeneity of tissues has to be taken into account; in contrast to tumor cell lines, tumors consist of different proportions of tumor, fibrocytic, vascular, and inflammatory cells. Therefore, the variances in mRNA expression could partly be due to variances in cell composition. Thus, polymerase chain reaction analyses of microdissection samples should reflect the cytokine expression in single cells more accurately.
Nevertheless, since tumor cells and surrounding host cells form a complex environment, the investigation of the whole tumor is in accordance with physiologic conditions. Such an analysis, however, forecloses identification of individual cells characterized by overexpression of a given mRNA, which may be essential for the metastatic potential. In addition, similar RNA levels may result in different protein expression by differences in the protein turnover, ie, translation and degradation. Therefore, our results do not exclude the relevance of the investigated factors for promoting metastasis but demonstrate the unsuitability of global analysis of these transcripts for determining patient prognosis.
Correspondence: Dr Schrama, Department of Dermatology, University of Wuerzburg, Josef-Schneider-Strasse 2, Building D8, 97080 Wuerzburg, Germany (firstname.lastname@example.org).
Financial Disclosure: None reported