Role and Prognostic Utility of Cutaneous CCL27 Expression in Early Stage Micosys Fungoides

Background: Recruitment of neoplastic T-cells in skin is a critical step in the pathogenesis of Mycosis Fungoides (MF) lesions. The chemokine CCL27 attracts memory CCR10-positive T cells to the skin. Increased CCL27 serum levels and skin epidermal expression have been observed in MF patients compared to normal controls, but the interactions between neoplastic cells and the skin immune system needs to be further elucidated.

Aims: We investigated whether CCL27 expression in MF is related to the density of dendritic cells (DC), CD8+ lymphocytes and CD4+ neoplastic lymphocytes; moreover, whether it influences the clinical response to therapy, it is modifiable by therapy itself and if the changes eventually observed can predict the frequency of recurrences and event-free survival.

Material and methods: CCL27 epidermal immunohistochemical staining was performed in 52 early stage-MF patients, 34 M/18 F, treated by PUVA plus interferon a for 14 months, with a complete follow-up data. Specimens were obtained at diagnosis and at the end of treatment. Formalin-fixed/paraffin-embedded tissue sections were immunostained with anti-CCL27, CD1a, CD8, CD4 antibodies. CCL27 immunoreactivity was graded as + (faint/moderate staining, limited to epidermal basal layer, as in normal skin); ++ (faint/moderate staining in the suprabasal layers; strong staining up to the lower two-thirds of epidermis); +++ (strong staining of full-thickness epidermis). Density of CD1a-positive DC and CD8+ lymphocytes was graded as + (few/isolated cells), ++ (small, ≤5 cell clusters), +++ (larger, >5 cell clusters). Density of CD4+ lymphocytes was graded as + (perivascular infiltrate), ++ (discontinuous subepidermal band), +++ (continuous band). Statistical analysis was also performed (c2 test for tables of contingency; event-free survival curves (EFS) by Kaplan-Meier method).

Results: At diagnosis, CCL27 expression was similar to normal skin in 16 cases and abnormal/suprabasal in 36, with 12 +++ cases. A normal/basal CCL27 expression tended to be correlated with a high DC density and with a low neoplastic infiltrate density (although P=NS). CCL27 at diagnosis was not correlated with clinical response (50 complete remission/CR, 2 partial remission). Treatment induced a significant CCL27 increase (suprabasal in 42 cases with 24 +++ cases; c2 test: P=0.004). Skin CCL27 expression at the end of treatment, but not at diagnosis, was related to recurrence and influenced EFS. During follow-up (median, 92.5 mo.s; range, 43–165), 33 patients relapsed (median EFS, 46 mo.s). A normal/basal CCL27 expression at the end of treatment, was correlated with a lower incidence of disease recurrence (3/9 compared to 30/41 with suprabasal expression; c2 test: P=0.022) and a longer median EFS (111 mo.s vs 39 mo.s with suprabasal expression; log rank test: P=0.031).

Conclusions: Increased CCL27 expression in early-stage MF lesions might be related to a balance between neoplastic cells and immunomodulant dendritic cells. CR induction in almost all patients by treatment is not correlated with a CCL27 reduction, but a normal CCL27 expression after treatment designates a subset of patients (about 20%) with a favourable behaviour. The mechanisms involved in the increased CCL27 expression after therapy in the remaining 80% patients, which have a higher probability to recur, need to be further investigated.