| The signal transducer and activator of transcription 3 is frequently constitutively activated in many types of human cancer. Activation of STAT3 depends on phosphorylation of tyrosine 705 (pY705). Activated STAT3 forms a dimer through the reciprocal interation of pY705-SH2 domain, and this dimer form STAT3 can be easily translocated into nuclear, where STAT3 binds to specific DNA and regulates the expression of its down stream gene. Those proteins upregulated by STAT3 are usually of promoting cell proliferation or apoptosis resistence, such as cyclin D1, Bcl-xL and survivin. The constitutively activated STAT3 causes oncogenesis, promotes tumor development and induces anticancer-drug resistence. So STAT3 is an attractive target for cancer therapy. This study focused on exploring a novel strategy for virtual screening small molecules in disrupting STAT3 dimer to repress the STAT3 pathway. According to known STAT3 inhibitors, we built a QSAR model to predict compounds capacity of inhibiting STAT3, which was based on X-score. This QSAR shown a well prediction in a training set and some test sets, with a conventional correlation coefficient r2>0.5 between the predicted values and experimental values. Approximately 420,000 compounds in Chembridge data base were screened using DOCK and Autodock vina. Then selected compounds were brought to bioassay. As a result, compound ICBDS 020 could inhibit luciferase activity depending on STAT3, and repressed breast cancer proliferation. This study displays a novel idea in virtual screening STAT3 inhibitors, and may be potentially suitable for designing or screening inhibitors of other targets. Compound ICBDS 020 is also potentially a lead compound for therapy of cancers with constitutively activated STAT3. |