Tumour-associated lymphangiogenesis in conjunctival malignant melanoma

Abstract

Background: To evaluate whether tumour-associated lymphangiogenesis, that is the formation of new lymphatic vessels (LVs) induced by a tumour, occurs in and around conjunctival malignant melanoma (MM).

Methods: Clinical files and conjunctival specimens of 20 patients with histologically diagnosed conjunctival MM were analysed. Sections were stained with LYVE-1 and podoplanin antibodies as specific lymphatic endothelial markers and Ki67 as proliferation marker. The tumour area and the area covered by LV (LVA), LV number (LVN) and LV density (LVD) were measured within the tumour and in the peritumoural area in digital images of the specimen. The LV results were correlated with the histopathological characteristics, tumour location, recurrence rate, mitomycin C therapy and presence of metastases.

Results: LVs were detected in all specimens within the tumour and peritumourally. Significantly more Ki67(+) proliferating lymphatic endothelial cells were detected in the tumour and in the peritumoural tissue up to 300 microm compared with the surrounding normal conjunctiva (>300 microm distance). There was a slightly positive correlation between the tumour size and the LVN and LVA in the 50 microm zone adjacent to the tumour. We did not find any significant correlations between LVs and histopathological and clinical characteristics (location, shape, relapses, metastases), possibly due to the small sample sizes. Non-limbal tumours with involvement of tarsus or fornix showed a tendency towards a higher LVD compared with limbal tumours.

Conclusion: Conjunctival MMs display tumour-associated LV within and around the tumour. The MM seems to induce lymphangiogenesis not only in the tumour, but also in its proximity.

Figure 1

Tumour-associated lymphatic vessels (LVs) in malignant melanomas (MMs) of the conjunctiva. (A) Representative image of LV staining with LYVE-1 antibody as specific marker for lymphatic endothelium: LYVE-1 positive peritumoural LVs. (B) Representative image of podoplanin stained LVs in the tumour adjacent conjunctiva. (C) Intratumoural LYVE-1 positive LVs (magnification ×100/×200). The area of primary acquired melanosis is marked with an asterisk. Arrows denote the LYVE-1/podoplanin stained lymphatic vessels. Note that erythrocyte-filled blood vessels are not stained with these lymphatic endothelial specific markers. LU, lymphatic vessel lumen.

Figure 2

Tumour-induced lymphangiogenesis. Significantly more proliferating lymphatic vessels (LVs) were found intratumourally and next to the tumour than in distant conjunctiva (>300 μm). Representative images of conjunctival malignant melanoma specimen stained with LYVE-1 and proliferation marker Ki67. Ki67 positive cells are marked (arrowhead). (A) Representative image of Ki67 positivity in tumour-associated lymphatic endothelial cells; red: Ki-67; brown: LYVE-1 (magnification ×1000). (B) Significantly more proliferating LVs were found intratumourally as well as in the directly adjacent tumour environment than in the more distant conjunctiva. A paired t test was performed for each zone separately with the following results: (1) tumour versus tumour adjacent conjunctiva (50 μm zone) p = 0.063 (not significant); (2) tumour versus mid-peripheral zone (50–200 μm) p = 0.021; (3) tumour versus peripheral zone (200–300 μm) p = 0.031; (4) tumour versus conjunctiva >300 μm from the tumour border p = 0.002; n = 20, median is marked; circles mark three outliers. The results support the hypothesis of tumour-associated active lymphangiogenesis in the proximity of the tumour.

Figure 3

Influence of tumour cross-sectional area on lymphatic vessels (LVs). Analyses of the LV parameters area covered by LV (LVA), LV number (LVN) and LV density (LVD) were performed in the tumour and in the tumour environment (50 μm zone, 50–200 μm zone, 200–300 μm zone). Intratumoural LVN, LVA and LVD were not positively correlated to the tumour cross-sectional area (A). There was a slightly inverse correlation for the LVD in the tumour with the tumour cross-sectional area (r = −0.257, p = 0.237). The LVN in the 50 μm area directly adjacent to the tumour was positively correlated to the tumour cross-sectional area (Pearson correlation coefficient r = 0.64; p = 0.002), demonstrated as a scatter plot (B).

Figure 4

Effect of topical mitomycin C on lymphatic vessels (LVs): scatter plot of the measured LV density (LVD) in the tumour (0), in the adjacent 50 μm zone (50), in the 50–200 μm zone (200) and in the 200–300 μm zone (300) before and after mitomycin C therapy. LVD before mitomycin C therapy is marked as a ring, and LVD after mitomycin C therapy is marked as an asterisk. Sections of four patients were analysed, in one patient there was no surrounding conjunctiva represented on the specimen, and only the tumour area was analysed. Because of the limited number of patients, descriptive analyses were performed.

Figure 5

Analysis of lymphatic vessels (LVs) in relation to tumour location. Descriptive scatter plot: in four of seven patients with non-limbal malignant melanomas (MMs) with involvement of the fornix or tarsus, the LV density (LVD) was higher than in all limbal/epibulbar tumour specimens (n = 7). Seven patients had limbal/epibulbar MMs, seven patients had an MM at the fornix or tarsus, and six patients had a disseminated MM. The analyses of LV parameters (LV number (LVN), area covered by LV and LVD) did not show any statistically significant results in these small sample sizes.