Potential Clinical Value of p53 Analysis in Colorectal Cancer

abstract & commentary

Synopsis: In patients with colorectal carcinomas, examination of tumor cells for p53 gene expression by immunohistochemical techniques employing antibodies has been proposed to offer prognostic value. In the current report, sequence analysis of the complete coding region of tumor cell p53 genes was completed in 190 samples from 189 Swedish patients that were treated between the years 1988 and 1992. In this series, the presence of p53 mutations correlated well with cancer-specific survival, whereas the immunohistochemistry results did not. Thus, this technically more laborious method may prove to be of greater clinical value, particularly if new methods are developed to simplify the technique and reduce the expense.

Source: Kressner U, et al. J Clin Oncol 1999; 17:593-599.

The importance of mutations in certain genes to the development and progression of colorectal carcinomas has been established.1 The suppressor gene p53, located on the short arm of chromosome 17, encodes a 53-kd nuclear phophoprotein that regulates the cell cycle.2 Mutations in this gene have been associated with about half of all human malignancies including colorectal carcinoma.1 Other studies have demonstrated that overexpression of p53, detected by immunohistochemical methods, is found in a large portion of cases (from 30-70%) and this overexpression has been interpreted as a surrogate marker for p53 mutation because the half-life of the nonmutated physiologic protein is short and it is usually not detectable. In several,3,4 but not all5 studies, overexpression of p53 has correlated with poorer prognosis in patients with colorectal cancer.

In the report by Kressner and collegues, sequencing of the entire coding region of the p53 gene was undertaken in 191 frozen tumor samples from 189 patients obtained nearly a decade ago in Sweden. To do this, RNA was extracted from the frozen samples and synthesized into cDNA by reverse transcription. p53 was amplified by the polymerase chain reaction and sequenced using an automated laser fluorescence sequencer. From the same tissue samples, p53 overexpression was evaluated by immunohistochemistry using a monoclonal antibody to the p53 protein (DO-7).

Mutations were detected in 99 samples from the 189 patients (52%). Survival time was clearly shorter for patients who had p53 mutations when compared to those that did not. Mutations outside the evolutionarily conserved regions of the gene were associated with an even worse prognosis. Multivariate analysis indicated that the presence of a p53 mutation was an independent prognostic factor. In contrast, there was no significant relationship between overexpression of p53 protein (by immunohistochemistry) and cancer-specific survival.


There has been some controversy about the value of determining abnormalities of p53 gene expression in samples of colon tumors. Although some have demonstrated a correlation with survival,3,4 others have not.5 An explanation for these conflicting data may be found in this report. Not all p53 mutations will result in overexpression of the p53 protein. In this series, of the 188 cases in which the comparison was possible, there was a concordance of immunohistochemistry and presence or absence of mutations in 74% of the subjects. Twenty-eight of the tumors that were positive by cDNA sequencing were negative by immunohistochemistry. This relates to the fact that certain mutations resulted in "stop" codons, and others in deletions that resulted in decreased expression. Interestingly, 20 (22%) of the tumors that displayed the wild-type gene on cDNA sequencing were positive by immunohistochemistry. This would suggest that in some cases, p53 gene expression was influenced by factors other than mutations in the coding regions.

The important finding was the correlation of mutations and survival in a series in which the immunohistochemistry failed to demonstrate the same. This would suggest that immunohistochemistry, although technically much easier and less expensive, is not as good a marker of p53 mutation as hoped. Certainly the techniques used by Kressner et al would be problematic and currently not feasible in the typical clinical setting. However, new techniques for mutation detection are being developed and these may ultimately be applied to simplify the search for p53 mutations in clinical samples from colon cancer patients.

Other than offering prognostic information, p53 analysis may offer important treatment guidance. Fluorouracil, commonly used to treat patients with colorectal cancer, inhibits DNA synthesis by inhibition of thymidylate synthase and also fragments DNA directly. Damaged DNA may result in the tumor cell arresting in G1 or it may stimulate cell death by apoptosis. This response may rely on a normal p53 protein. Thus, drug-induced cytotoxicity may not occur in tumors with dysfunctional p53. Certainly, this concern could be tested by examining chemotherapy-induced responses in patients for whom the presence or absence of tumor cell p53 mutation is established. Inasmuch as a large portion (30-70%) of colorectal cancer patients has tumor cells with these mutations, chemotherapy that relies on a functional p53 gene to produce cytotoxicity could be avoided.


1. Fearon E, et al. Cell 1990;61:759-767.

2. Harris CC, et al. N Engl J Med 1993;329:1318-1327.

3. Goh HS, et al. Cancer Res 1995;55:5217-5221.

4. Smith DR, et al. Br J Cancer 1996;74:216-223.

5. Dix B, et al. Int J Cancer 1994;59:547-551.