P-glycoprotein and tumor progression.

Development of resistance to chemotherapy is a hallmark of many advanced human cancers. The molecular basis of clinical resistance is not well understood. Over the past 15 years, a number of different molecular mechanisms of resistance to anticancer drugs have been identified in many cell lines. One mechanism involves the increased expression of the plasma membrane P-glycoprotein (Pgp), which is thought to function as an adenosine triphosphate-dependent efflux pump that reduces the cellular accumulation of a wide range of chemotherapeutic drugs (12)Studies (1-3) of biopsy samples from patients have demonstrated that elevated levels of Pgp can be detected in tumors of every histologic type. In addition, numerous studies have been conducted to determine whether or not Pgp expression is associated with clinical outcome. The conclusions of such studies have varied; however, the basis for these differences in conclusions may be attributed in part to differences in the techniques used for detecting Pgp in tumor samples. Nevertheless, several studies (3-7) on leukemias, lymphomas, and some childhood solid tumors have demonstrated a strong association between the detection of Pgp in tumor samples and poor response to chemotherapy. The observed association between Pgp expression and poor outcome in the clinical studies raises the question of how changes in Pgp expression are regulated in malignant cells. At present, mechanisms involved in the regulation of Pgp expression in human cancers and in normal human tissues are unknown. In this issue of the Journal, Schneider et al. (8) address the question of whether or not Pgp expression was associated with p53 mutation and HER-2/neu expression in a number of human gynecological tumors. The rationale for their study was based on the hypothesis that, since previous experiments have shown that c-Ha-ras and mutant p53 appear to activate promoter activity of the human MDR1 (Pgp) gene using assays of expression of reporter gene p'asmid constructs transfected into NIH 3T3 cells, it is possible that genes associated with oncogenic development may also activate concomitantly Pgp in human cancers. Such an association may provide an explanation for the frequent observation of elevated Pgp expression in many advanced cancers that have not been treated with chemotherapy. Although the question posed by the report by Schneider et al. is an important one, finding a definitive answer is far from simple. It has been reported that p53 appears to be involved in the regulation of a wide variety of cellular processes (9,10). As far as its ability to regulate transcription is concerned, the central core region of wild-type p53 protein has sequence-specific DNA-binding activity, while the N-terminus of p53, when fused to the DNA-binding domain of yeast protein Gal4, can function as a transcriptional activator. As a result of these two activities, wild-type p53 protein can stimulate the expression of reporter genes that contain a p53-binding sequence in their promoters. In contrast, a large number of cellular and viral promoters that lack a p53-binding sequence, such as the MDR1 promoter, are repressed by wild-type p53 protein (77). It is interesting that p53 missense mutants have been shown to activate a number of promoters, including those present in the MDR1 and PCNA genes as well as in the long terminal repeat (LTR) of human immunodeficiency virus (12-16). Activation by mutant p53 is not dependent on the presence of a consensus p53-binding sequence. Indeed, the mechanisms by which mutant p53 can stimulate the expression of genes that do not contain p53-binding sites is not yet known. All of these studies have relied on transient expression-reporter gene assays, and it is pertinent to consider the physiological relevance of such assays in which constitutively overexpressed p53 protein acts on plasmid-borne extrachromosomal copies of a promoter-reporter gene. It has been demonstrated, for example, that certain promoters that are inhibited by p53 protein in transient transfection assays are not inhibited in their native state, implying that regulation of genes by p53 is subject to the natural chromatin state of the endogenous gene. In light of these observations and of the enormous difficulty in extrapolating findings from transient expression-reporter gene assays to endogenous gene regulation, the results presented by Schneider et al. (8) are not entirely surprising and argue against the hypothesis that MDR1 gene expression is regulated by p53 protein. Schneider et al. found no association between p53 gene status and MDR1 expression in mammary, endometrial, and cervical rumors. Moreover, el Rouby et al. (17) reported a similar