Characterization of the Microenvironment in Positive and Negative Sentinel Lymph Nodes from Melanoma Patients

Abstract

Melanomas are aggressive skin tumors characterized by high metastatic potential. Our previous results indicate that Natural Killer (NK) cells may control growth of melanoma. The main defect of blood NK cells was a decreased expression of activating NCR1/NKp46 receptor and a positive correlation of NKp46 expression with disease outcome in stage IV melanoma patients was found. In addition, in stage III melanoma patients, we identified a new subset of mature NK cells in macro-metastatic Lymph nodes (LN). In the present studies, we evaluated the numbers of NK cells infiltrating primary cutaneous melanoma and analyzed immune cell subsets in a series of sentinel lymph nodes (SLN). First, we show that NKp46+ NK cells infiltrate primary cutaneous melanoma. Their numbers were related to age of patients and not to Breslow thickness. Then, a series of patients with tumor-negative or -positive sentinel lymph nodes matched for Breslow thickness of the cutaneous melanoma was constituted. We investigated the distribution of macrophages (CD68), endothelial cells, NK cells, granzyme B positive (GrzB+) cells and CD8+ T cells in the SLN. Negative SLN (SLN-) were characterized by frequent adipose involution and follicular hyperplasia compared to positive SLN (SLN+). High densities of macrophages and endothelial cells (CD34), prominent in SLN+, infiltrate SLN and may reflect a tumor favorable microenvironment. Few but similar numbers of NK and GrzB+ cells were found in SLN- and SLN+: NK cells and GrzB+ cells were not correlated. Numerous CD8+ T cells infiltrated SLN with a trend for higher numbers in SLN-. Moreover, CD8+ T cells and GrzB+ cells correlated in SLN- not in SLN+. We also observed that the numbers of CD8+ T cells negatively correlated with endothelial cells in SLN-. The numbers of NK, GrzB+ or CD8+ T cells had no significant impact on overall survival. However, we found that the 5 year-relapse rate was higher in SLN with higher numbers of NK cells.

Fig 1. Immunohistological staining and counting of NKp46⁺ NK cells in primary cutaneous melanomas.

(A) Representative staining of NK cells with anti-NKp46 mAb. (B) Enumeration of NK cell numbers infiltrating cutaneous melanoma (left panel) and paired analysis of intra and peritumoral NK cells (right panel). (C) Correlation of NK cell numbers and age (left panel) or gender (right panel) of the patients. (D) Correlations of NK cell numbers with Breslow thickness (left panel), or with presence of tumor ulceration (right panel). (E) Correlation of NK cell numbers and 5 years-relapse rates.

Fig 2. Immunohistological staining and density evaluation of macrophages (CD68) and endothelial cells (CD34) in SLN.

Due to the high numbers of labelled cells, staining was evaluated as score: (+) moderate, (++): high, (+++): intense staining. Representative staining (X20 and X40) with (A) anti-CD68 and (B) anti-CD34 in SLN (P: Peritumor area; T: Tumor area). Percentages of CD68 and CD34 density staining in SLN^- (C) and SLN⁺ (D) in peritumor (Peri) (left panel) and intratumor (Intra) areas (right panel).

Fig 3. Immunohistological staining and enumeration of cytotoxic cells infiltrating SLN from melanoma patients.

Representative staining and counts of immune cells in SLN^- and SLN⁺: (A) NKp46⁺ NK cells, (B) GrzB⁺ cells, (C) CD8⁺ T cells. (D) Correlation analyses of NKp46⁺ cells, GrzB cells and CD8⁺ T cell numbers between intra- and peritumor area in SLN⁺. (E, F) Correlation analyses between numbers of cytotoxic immune cells (NKp46⁺ cells, GrzB⁺ cells and CD8⁺ T cells) in the SLN⁺ (E) and in the SLN^- (F).

Fig 4. Correlation analyses between numbers of cytotoxic immune cells (GrzB⁺, NKp46⁺, CD8⁺ T cells) and the densities of (A) macrophages or (B) endothelial cells in SLN^- and SLN⁺.

Fig 5. Association between numbers of cytotoxic (A) NK cells, (B) GrzB⁺ cells, CD8⁺ T cells (C) and 5-year-relapse rate of the patients.