erbB-2 amplification and overexpression have been suggested as potentially useful prognostic markers in bladder cancer. We examined 141 bladder tumor specimens (45 fresh tissue samples and 96 formalin fixed tissue blocks) for erbB-2 amplification using fluorescence in situ hybridization. A dual labeling hybridization using a repetitive pericentromeric probe specific for chromosome 17 and a cosmid probe for the erbB-2 locus was performed to analyze the erbB-2 copy number in relation to chromosome 17 copy number on a cell by cell basis. Amplification (more than twice as many erbB-2 signals as centromere 17 signals per tumor) was found in 10 of 141 tumors. There was considerable heterogeneity in erbB-2 amplification. In a given tumor there was a wide range of erbB-2 copy number in amplified cells. The arrangement of erbB-2 signals in clusters in all amplified cases suggests that erbB-2 amplification occurs intrachromosomally in bladder cancer. Amplification was found only in tumors with aneusomy of chromosome 17 and was more frequent in pT2-T4 tumors than in pTa/T1 tumors. Overexpression was present without amplification in 51 tumors. All tumors with erbB-2 amplification showed erbB-2 overexpression. However, in 5 samples the proportion of cells with amplification was significantly lower than the fraction of cells with overexpression, indicating coexistence of two different mechanisms leading to overexpression in these tumors.

Overexpression of p53 and erbB-2 was studied by immunohistochemistry in formalin-fixed tissue samples of 179 patients with transitional cell carcinoma of the urinary bladder. p53 immunostaining was strongly correlated with tumour stage (P < 0.0001). This was driven by a marked difference in p53 expression between pTa (37% positive) and pT1 (71%) tumours, while there was no difference between pT1 and pT2-4 tumours. Similarly, a strong overall association between p53 expression and grade (P < 0.0001) was driven by a marked difference between grade 1 (28%) and grade 2 tumours (71%), and there was no significant difference between grade 2 and grade 3 tumours. Surprisingly, the frequency of erbB-2 overexpression was higher in pT1 tumours (74%) than in either pTa (49%; P = 0.0265) or pT2-T4 (56%; P = 0.0645) tumours. Both p53 and erbB-2 expression was also associated with metastasis. Metastases were found in 77% of patients with p53 positive primary tumours, but in only 50% of the patients with p53 negative primary tumours (P = 0.022). Metastases were found in 66% of patients with erbB-2 positive primaries, but in only 37% of the erbB-2 negative primaries (P = 0.020). Of 32 patients with positivity for both p53 and erbB-2, 84% developed metastases, as compared to 49% of patients with positivity for either one or neither positive (P = 0.002). We conclude that both p53 and erbB-2 overexpression are associated with early invasion in bladder cancer. Furthermore, p53 and erbB-2 may be important predictors for metastasis.