INTRODUCTION: Triple-negative breast cancer (TNBC) high rate of relapse is thought to be due to the presence of tumor-initiating cells (TICs), molecularly defined as being CD44high/CD24-/low. TICs are resilient to chemotherapy and radiation. However, no currently accepted molecular target exists against TNBC and, moreover, TICs. Therefore, we sought the identification of kinase targets that inhibit TNBC growth and eliminate TICs. METHODS: A genome-wide human kinase small interfering RNA (siRNA) library (691 kinases) was screened against the TNBC cell line SUM149 for growth inhibition. Selected siRNAs were then tested on four different breast cancer cell lines to confirm the spectrum of activity. Their effect on the CD44high subpopulation and sorted CD44high/CD24-/low cells of SUM149 also was studied. Further studies were focused on polo-like kinase 1 (PLK1), including its expression in breast cancer cell lines, effect on the CD44high/CD24-/low TIC subpopulation, growth inhibition, mammosphere formation, and apoptosis, as well as the activity of the PLK1 inhibitor, BI 2536. RESULTS: Of the 85 kinases identified in the screen, 28 of them were further silenced by siRNAs on MDA-MB-231 (TNBC), BT474-M1 (ER+/HER2+, a metastatic variant), and HR5 (ER+/HER2+, a trastuzumab-resistant model) cells and showed a broad spectrum of growth inhibition. Importantly, 12 of 28 kinases also reduced the CD44high subpopulation compared with control in SUM149. Further tests of these 12 kinases directly on a sorted CD44high/CD24-/low TIC subpopulation of SUM149 cells confirmed their effect. Blocking PLK1 had the greatest growth inhibition on breast cancer cells and TICs by about 80% to 90% after 72 hours. PLK1 was universally expressed in breast cancer cell lines, representing all of the breast cancer subtypes, and was positively correlated to CD44. The PLK1 inhibitor BI 2536 showed similar effects on growth, mammosphere formation, and apoptosis as did PLK1 siRNAs. Finally, whereas paclitaxel, doxorubicin, and 5-fluorouracil enriched the CD44high/CD24-/low population compared with control in SUM149, subsequent treatment with BI 2536 killed the emergent population, suggesting that it could potentially be used to prevent relapse. CONCLUSION: Inhibiting PLK1 with siRNA or BI 2536 blocked growth of TNBCs including the CD44high/CD24-/low TIC subpopulation and mammosphere formation. Thus, PLK1 could be a potential therapeutic target for the treatment of TNBC as well as other subtypes of breast cancer.