We conducted the first phase 2 and pharmacologic study to evaluate the combination of novobiocin (a coumeromycin antibiotic that has been shown to augment alkylating agent cytotoxicity in experimental models) and high-dose cyclophosphamide and thiotepa followed by autologous marrow support in women with chemosensitive advanced breast cancer. Its aims were (1) to determine progression-free survival (PFS) and overall survival (OS), (2) to evaluate the pharmacokinetics of cyclophosphamide and thiotepa, and (3) to measure the ability of novobiocin to reverse alkylator drug resistance in vitro. Forty-one women with chemotherapy-responsive advanced breast cancer received cyclophosphamide (4 g/m2) for peripheral blood stem cell mobilization (treatment 1) followed by high-dose cyclophosphamide (1.5 g/m2 per day for 4 days), thiotepa (200 mg/m2 per day for 4 days), and novobiocin (4 g/day orally for 7 days) (treatment 2) and autologous marrow support. The median PFS was 10 months (range, 0.2-70.6 months) and OS, 21.5 months (range, 0.2-70.6 months). There was no statistically significant relationship between PFS or OS and area-under-the-curve values of cyclophosphamide, thiotepa, or 4-hydroxycyclophosphamide. Patient plasma samples (n = 12) obtained during novobiocin therapy were able to reverse alkylator drug resistance in an in vitro colony-forming assay. Correlative laboratory studies in an in vitro model system demonstrated that patient plasma after novobiocin treatment resulted in the magnitude of resistance reversal that had been predicted by prior preclinical experiments. Clinically, however, this activity of novobiocin did not translate into a substantial increase in PFS or OS compared with historical controls treated with high-dose alkylator therapy alone.