C. Charpin et al., AUTOMATED AND QUANTITATIVE IMMUNOCYTOCHEMICAL ASSAYS OF BCL-2 PROTEININ BREAST CARCINOMAS, British Journal of Cancer, 76(3), 1997, pp. 340-346
Expression of the bcl-2 gene was investigated in 218 human breast carc
inomas by immunohistochemical analysis. Immunodetections were assessed
using (1) frozen sections, (2) documented commercially available mono
clonal antibody (bcl-2/124, Dako), (3) automation of immunoperoxidase
technique (Ventana) and (4) quantitative evaluation of results by imag
e analysis (SAMBA) and statistical analysis of quantitative data (BMDP
software). Bcl-2 protein expression was correlated with current progn
ostic indicators and with molecular markers detected by the same proce
dure as for Bcl-2. It was shown that Bcl-2 expression is not related t
o patients' age, tumour size and type or lymph node status, but an inv
erse relationship was observed between Bcl-2 and tumour grade (P < 0.0
001). An inverse relationship was also observed between Bcl-2 expressi
on and p53 (P < 0.0001), Ki67/MIB1 antigen- (P = 0.0012), and P-gp- (P
= 0.002) positive immunoreactions. In contrast, anti-Bcl-2 positive r
eaction was significantly associated with ER-positive (P < 0.001) and
with ER/PR-positive or ER/PR/pS2-positive immunoreactions (P less than
or equal to 0.005). Bcl-2 expression was independent of CD31 and cath
epsin D expression. Thus, Bcl-2 protein, thought to be antiapoptotic,
exhibits parodoxical expression in human breast carcinomas. It is stro
ngly detected in low-grade tumours (well-differentiated) with low (MIB
1) growth fraction, but is independent of the tumour progression (size
, node status, CD31, and cathepsin D). Bcl-2 acting on apoptosis is re
lated to p53 gene abnormalities in breast carcinomas. Bcl-2 protein ex
pression may also be involved in response to endocrine therapy (associ
ated to ER/PR/pS2 positive immunoreactions) and probably with chemores
istance mechanisms (inverse relationship with P-gp).