Abrogation of cisplatin-induced programmed cell death in human breast cancer cells by epidermal growth factor antisense RNA

Background: Epidermal growth factor receptor (EGF-R) perturbation by receptor ligand(s), e.g., epidermal growth factor (EGF) and transforming growth factor-α (TGF-α), or receptor-specific antibodies accentuates cisplatin-induced toxicity in tumor cells. This sensitization occurs only in tumor cells with high expression of EGF-R but not in those with low expression of EGF-R. Purpose: Therefore, we have studied the role of EGF-R expression on cisplatin-mediated cytotoxicity. Methods: MDA-468 human breast cancer cells were stably transfected with a p-chloramphenicol acetyl transferase (pact[p]-CAT) vector containing a 4.1-kilobase full-length antisense EGF-R complementary DNA. EGF-R content was assessed by ¹²⁵I-EGF binding and EGF-R immunoblot assays. Cisplatin sensitivity was evaluated by (a) colony-forming assay in vitro, (b) xenograft growth in nude mice, (c) cell cycle distribution of propidium iodide-labeled DNA, (d) DNA fragmentation in agarose gels, and (e) terminal deoxynucleotidyl transferase (Tdt) fluorescence in situ. Cisplatin uptake was measured by atomic absorption spectroscopy, and the levels of drug-DNA intrastrand adducts were determined by a dissociation-enhanced fluoroimmunoassay that utilizes an antibody against cisplatin-modified DNA. Results: Selected clones (MDA- 468/AS-EGFR) exhibited more than 90% loss of both ¹²⁵I-EGF binding and receptor content determined by western blot analysis, whereas clones transfected with the vector alone (MDA-468/p-CAT) had EGF-R levels similar to those of the parent cells. By use of a colony-forming assay, the 1-hour IC₅₀ (i.e., the concentration of drug required for 1 hour to achieve 50% cell kill) for cisplatin was 2 μM or less for parental and vector- transfected clones (n = 4), whereas it was 25 μM or more for all MDA- 468/AS-EGFR clones (n = 3). MDA-468/p-CAT clones exhibited internucleosomal DNA fragmentation, enhanced Tdt-end labeling in situ, and G₂ arrest 48 hours after a 1-hour incubation with 3-30 μM cisplatin. Under these conditions, apoptosis and G₂ arrest were undetectable in all MDA-468/AS-EGFR clones. An MDA-468 subline selected after long-term treatment with a TGF-α-Pseudomonas exotoxin A fusion protein 40 lacked EGF binding and also exhibited cisplatin resistance (1-hour IC₅₀: >30 μM) compared with parental cells. This EGF- R-dependent difference in cisplatin response was confirmed in a nude mouse xenograft model by use of high- and low. EGF-R-expressing cell clones. Total intracellular drug accumulation after a 1-hour cisplatin exposure, as measured by atomic absorption spectroscopy, was identical in both groups of cells. Intrastrand drug-DNA adducts, however, were statistically higher in high EGF-R expressors than in low-EGF-R-expressing clones. Conclusions: These data indicate that a critical level of EGF-R signaling, which is amplified in some common human cancers, is necessary for cisplatin-mediated apoptosis in tumor cells and suggest an inhibitory effect of this pathway on the repair of cisplatin-damaged DNA.