Cyclin A and beta-catenin expression in actinic keratosis, Bowen's disease and invasive squamous cell carcinoma of the skin

Abstract

Actinic keratosis (AK) has been defined as a precancerous lesion or an early phase in the evolution of squamous cell carcinoma (SCC) and histological changes seen in the individual cells of an AK are indistinguishable from those seen in SCC, which invade the dermis. Cyclin A is an increasingly utilized proliferation marker that has functions in both S phase (DNA replication) and initiation of mitosis, whereas alterations of beta-catenin, the molecule involved in cell-cell adhesion and in signalling transduction, could promote invasive and proliferative capacities of malignant tumours. To determine cyclin A and beta-catenin expression pattern in cutaneous SCC and in in situ lesions classified as keratinocytic intraepidermal neoplasia (KIN) and, using traditional terms, as AK and Bowen's disease (BD), and to analyse it in relation to SCC differentiation, diameter and thickness. Immunohistochemical staining was performed on 110 formalin-fixed paraffin-embedded tissue samples with the streptavidin-biotin technique using antibodies to cyclin A and beta-catenin. On histological examination, 53 lesions were diagnosed as AK, 16 as BD and 41 as SCC-11 well differentiated (WD), 16 moderately differentiated (MD) and 14 poorly differentiated (PD). Using KIN classification, 22 lesions were KIN1, 23 were KIN2 and 24 were KIN3. For cyclin A, distribution and labelling index (LI), and for beta-catenin, level of membranous staining and presence of aberrant (nuclear/cytoplasmic) localization were examined. Diffuse cyclin A presence was observed more frequently in BD than in AK (P < 0.0001) or SCC (P = 0.0002), and in SCC-PD compared with SCC-WD (P < 0.0001) or SCC-MD (P = 0.0003). Differences between KIN3 and KIN2, as well as KIN3 and KIN1 lesions, were statistically significant (P < 0.0001), and the same result appeared when KIN1 and KIN2 cases were grouped and compared with those of KIN3 (P < 0.0001). Cyclin A LI was significantly lower (P < 0.05) in AK than in BD or SCC, but no difference between BD and SCC was found, and LI in BD was even higher than in SCC-WD or SCC-MD, while analysis regarding SCC differentiation and KIN classification revealed the same correlation as for the cyclin A distribution. Reduced or absent beta-catenin membranous staining was found in 90 cases (81.8%), more often in SCC than in AK (P = 0.03) or in AK and BD grouped together (P = 0.02). There was no statistical difference between SCCs of various level of differentiation, or between different KIN grades. Diffuse loss of membranous beta-catenin staining showed 36 lesions (32.7%), more frequently SCC than AK (P = 0.003) or AK and BD grouped (P = 0.006), as well as SCC-PD compared with SCC-WD (P = 0.01) and SCC-MD (P = 0.03), whereas all KIN comparisons remained nonsignificant. Aberrant beta-catenin cellular localization demonstrated 28 lesions (25.5%), most often in the basal or peripheral parts and in the lesions with diffuse beta-catenin loss (P = 0.009), but revealed no correlation with the histological type, SCC level of differentiation or KIN grades. Diffuse loss of membranous beta-catenin staining was found to be significantly more frequent in SCC thicker than 4 mm (P = 0.03), while all other comparisons between cyclin A or beta-catenin with the tumour size remained nonsignificant. Cyclin A LI was higher in cases with diffuse loss of membranous staining (P = 0.001) or with aberrant cellular localization of beta-catenin (P = 0.002). Cyclin A LI showed greater difference between AK and BD than between BD and SCC, suggesting that increased proliferation (measured by cyclin A LI) characterizes progression of in situ lesions from AK to BD, whereas reduced beta-catenin expression separates more clearly SCC from the in situ lesions. Diffuse pattern of loss of membranous beta-catenin staining correlated better with the type of lesion, SCC differentiation and tumour size than reduced expression in general or aberrant cellular localization of beta-catenin. KIN classification does not seem to be supported by our findings, except when KIN1 and KIN2 lesions (in situ, partial thickness) are grouped.