The immunoreactivity of p21 was evaluated in normal mucosa and in adenomas and adenocarcinomas of the human large bowel. In normal mucosa, p21 immunoreactivity was seen in the superficial third of the crypts (maturation compartment) and in surface (terminally differentiated) epithelium, and was mutually exclusive with Ki67/MIB1 reactivity. These data show that p21 expression is inversely related to proliferation and directly related to terminal differentiation. In adenomas, p21‐reactive cells were frequently clustered in the superficial areas and were non‐reactive for MIB1. In adenocarcinomas, p21 staining was heterogeneous: high p21 expression (19 cases) was associated with lower stage and lack of p53 overexpression. p21‐reactive cells were devoid of MIB1 immunoreactivity, but no relationship could be found between p21 and MIB1 labelling indices. p21 heterogeneity may be related to alterations in the p53‐dependent induction pathway: high p21 expression was associated with low to absent p53 reactivity, with presumed normal p53 function; low p21 expression was associated with p53 overexpression, with presumed p53 alteration resulting in loss of function.

Cyclin D1 is a major positive regulator of the G 1 restriction point promoting inactivation of the retinoblastoma protein (RB). The cyclin D1 gene is rearranged, amplified and/or over-expressed in several human neoplasms. In the present series of 64 human breast carcinomas, cyclin D1 amplification (4-to 8-fold) was seen in 24% of cases, and cyclin-D1 immuno-histochemical over-expression was seen in 50% of cases. Amplification and over-expression were statistically associated ; however, divergent result were seen in 30% of cases. Some of these discrepancies may reflect the fact that cyc-lin-D1 expression may be due to mechanisms other than gene amplification. Cyclin-D1 over-expression, but not cyclin-D1 amplification, was associated with positive oestrogen-receptor immunoreactivity. Cyclin-D1 amplification was associated with high RB expression, and 4 cases (7%) with absent RB immunoreactivity showed no cyclin-D1 amplification nor expression. Our data support the hypothesis that cyclin-D1 amplification may be associated with enhanced gene transcription and with high RB expression, that high ER expression may cooperate in maintaining high levels of cyclin-D1 protein, and that loss of RB function, as assessed by the lack of RB immu-noreactivity, may be related to normal cyclin-D1 gene copy number and low cyclin-D1 expression. Int.

The patients with node-positive breast cancer who have a higher likelihood of gaining benefit from adjuvant therapy are those with tumours with cyclin D1 nuclear expression, small size and less than 3 metastatic nodes. Further studies are needed to verify the prognostic value of cyclin D1 in relation to different adjuvant treatments and to deepen the biological pathways that regulate its activation/ suppression in human breast carcinoma.

The calcium-binding protein calbindin-D (28 kD) has been analysed immunohistochemically in different human lymphoid tissues. Combined immunohistochemical staining showed that calbindin-D (28 kD) is expressed by only a proportion of dendritic cells within the light zone of germinal centres, where antigens in the form of immune complexes are trapped and presented to B lymphocytes. All other cell types including macrophages, interdigitating cells, and various lymphocyte populations were negative. The expression of calbindin-D in this functionally relevant subset of follicular dendritic cells could have a role in the regulation of proliferation and selection of memory B-cells by modulating the concentration of calcium ions. Calbindin-D may be a useful marker for analysing in situ the phases of follicular development in different physiological and pathological conditions.