ErbB-2
The erbB-2 gene encodes for a 185-kDa receptor tyrosine kinase with homology to the epidermal growth factor receptor erbB-1, erbB-3, and erbB-4. The natural ligand for erbB-2 has not been identified, but erbB-2 phosphorylation may occur by transmodulation when other members of the erbB kinase family react with their ligands (e.g., epidermal growth factor; Akiyama et al., 1988; Stern and Kamps, 1988; Johnson et al., 1993) and neu differentiation factor (Culouscou et al., 1993; Plowman et al., 1993; Carraway et al., 1994; Kita et al., 1994; SUwkowski et al., 1994; Tzahar et al., 1994). This transmodulation occurs through heterodimerization of erbB-2 with the other kinases.

Although overexpression of erbB-2 is associated with poor prognosis in human breast cancer, erbB-2 is rarely detected in benign lesions (De Potter et al, 1989; Lizard-Nacol et al., 1995; Millikan et al., 1995). ErbB-2 is not overexpressed in either lobular CIS (Gusterson et al., 1988b; Ramachandra et al., 1990; Porter et al., 1991) or invasive lobular carcinoma (Porter et al., 1991; Barbareschi et al., 1992).

However, overexpression of erbB-2 is frequent in DCIS and invasive ductal carcinoma. The precipitous rise in expression at the stage of DCIS suggests that overex-pression of erbB-2 may be sufficient to produce this step in progression.

Although erbB2 is detected in DCIS more frequently than in invasive cancer (Gusterson etal., 1988a, b; De Potter etal., 1989; Kobayashietal., 1992), studies of mixed lesions containing both in situ and invasive components invariably report that the two components express erbB-2 coordinately (Ramachandra et al., 1990; Liu et al., 1992; Iglehart et al., 1995).

Thus, DCIS in which erbB-2 is not amplified gives rise to invasive carcinoma in which erbB-2 is not amplified and DCIS in which erbB-2 is amplified gives rise to invasive carcinoma in which erbB-2 is amplified. Since erbB-2 is not amplified in the majority of invasive breast cancers, these data suggest that DCIS in which erbB-2 is not amplified is more likely to progress to invasive carcinoma than is erbB-2-positive DCIS. The latter conclusion is somewhat counterintuitive because erbB-2 expression by invasive breast cancer is associated with poor prognosis (Slamonetal., 1987,1989; Paiketal., 1990; Gullick et al., 1991; Gustersonet al, 1992; Sauer et al., 1992). If erbB-2-expressing DCIS had a poor prognosis as indicated by more rapid/frequent progression to invasive cancer, the incidence of erbB-2 expression might be expected to be higher in invasive cancer than in DCIS.

Both the incidence of erbB-2 amplification and the propensity of DCIS lesions to progress to invasive cancer vary with the histotype of DCIS. Comedo DCIS is more frequently erbB-2-positive than is cribriforming DCIS (Bartkova et al., 1990; Ramachandra et al., 1990). Ramachandra and colleagues found that 19% of pure cribriforming DCIS and 55% of mixed lesions containing both cribriforming in situ and invasive components expressed erbB-2. On the other hand, 62% of pure comedo DCIS lesions were erbB-2-expressing compared to only 49% of mixed lesions (Ramachandra et al., 1990).

These results indicate that erbB-2-positive cribriforming DCIS are more likely to progress to invasive cancers than are erbB-2-negatrvQ lesions, whereas erfoB-2-negative comedo DCIS are more likely to progress than are erbB-2-positive comedo DCIS. Once again, this suggests complementation groups, a requirement for coexpression of particular combinations of genotypic and phenotypic characteristics in order for cancer to develop and progress.