Mdr1 gene expression in childhood acute lymphoblastic leukemias and lymphomas: a critical evaluation by four techniques.

Expression of the multidrug resistance gene mdr1 is reported to be an important determinant of responsiveness to therapy and survival in some cancers. Many different methods have been used to evaluate mdr1 expression in these studies. This paper compares four methods for determination of mdr1 expression. We studied the mdr1 gene expression in 36 freshly established cell lines from 28 children with acute lymphoblastic leukemia (16 T-ALL, six BCP-ALL, two B-ALL (L3), two biphenotypic leukemias, two Burkitt's lymphomas). Leukemic specimens were obtained at the time of diagnosis in 16 cases, and after chemotherapy in 20 cases. In all the samples, mdr1 mRNA was measured by slot blotting and reverse transcriptase polymerase chain reaction (rt-PCR), and the presence of the mdr1 product, P-glycoprotein, was detected by immunohistochemistry with the MRK-16 monoclonal antibody. In situ mdr1 RNA hybridization was performed in 30 cases. Complete agreement was noted between all the techniques in 14 cases (39%). Results differed on a single test result in another 39% of the cases. These 78% of cases were considered assessable, and the consensus result was presumed to be correct. By this consensus criterion, immunohistochemistry yields both false negative (11%), and false positive (11%) results. RNA slot blotting has a high (21%) false positive rate. In situ mRNA hybridization and rt-PCR have the highest concordance, 80%. The 28 patients from whom these cell lines were derived appear to represent a very poor prognosis group, since there are only two patients (with Burkitt's lymphoma) who are long-term survivors. Nonetheless, a complete clinical response to therapy was correlated with absence of mdr1 expression in assessable cases (p = 0.04). These four methods of determining mdr1 expression often yield discordant results. Therefore, the use of at least two methods for evaluating mdr1 expression is advisable. Rt-PCR is recommended because of its relative simplicity and specificity. This should be supplemented by a technique (immunohistochemistry or flow cytometry) able to detect heterogeneity of P-glycoprotein expression among cells.