Transcriptional and Post-Translational Upregulation of HER3 (ErbB3) Counteracts Antitumor Effect of HER2 Tyrosine Kinase Inhibitors.

Abstract

Abstract We hypothesized that sustained inhibition of HER3 and its output to PI3K/Akt is required for the optimal antitumor effect of HER2 inhibitors. Therefore, we examined the temporal effect of the HER2 tyrosine kinase inhibitor (TKI) lapatinib (lap) on feedback upregulation of active HER3 in HER2-overexpressing breast cancer cells. A time course with lap-treated cells showed 3 to 5-fold upregulation of HER3 RNA and protein, beginning at 4 h and increasing through 48 h. P-Tyr immunoblot of HER3 immunoprecipitates revealed recovery of HER3 phosphorylation at and beyond 13 h of treatment. Site-specific antibodies revealed HER3 phosphorylation at Y1197 and Y1289, two of the six p85 binding sites in HER3. Recovery of P-HER3 correlated temporally with recovery of T308 P-Akt. The upregulation of HER3 RNA upon treatment with lap suggested that inhibition of active HER2 and PI3K/Akt derepresses the transcription factor FoxO3a. Putative FoxO3a binding sites were identified within the 5' flanking region upstream of the HER3 transcription start site. Transfection with FoxO3a siRNA reduced basal and lap-induced HER3 RNA levels 2 to 5-fold compared to control cells. Conversely, overexpression of FoxO3a increased HER3 RNA 2.5-fold, which could be further enhanced by lap treatment. In addition to these transcriptional mechanisms, the recovery of P-HER3 upon lap-induced inhibition of HER2 suggested engagement of another tyrosine kinase transactivating HER3 and/or that HER2 had been incompletely inhibited by the TKI. However, IGF-IR, Src, and MET TKIs did not inhibit the recovery of P-HER3. On the other hand, the addition of trastuzumab (tz) to lap-treated cells prevented recovery of P-HER3, suggesting that disruption of a ligand-independent HER2-HER3 interaction was involved in partial maintenance of HER3 phosphorylation.The upregulation of HER3 RNA and partial maintenance of P-HER3 and P-Akt suggested that combined inhibition of HER2 and HER3 will synergistically inhibit tumor cell viability. Transfection with HER3 siRNA sensitized HER2+ breast cancer cells to each lap and tz as assessed by Apo-BrdU (apoptosis) and 3D-Matrigel growth assays. Further, treatment with AMG-888, a HER3 monoclonal antibody (AMGEN-U3), sensitized cells to each lap and tz. Ongoing studies include the treatment of BT474 xenografts in athymic mice with lap ± AMG-888 using [18F]-FDG-PET as a non-invasive imaging biomarker to predict treatment outcome. Finally, we examined HER3 levels by immunohistochemistry in sections from tumor blocks of patients enrolled in a neoadjuvant trial where lap was given alone during the first 6 weeks of therapy. The percent and intensity of tumor cell staining was calculated as a histoscore (Human Pathol. 26:291, 1995). On week 2 of therapy, HER3 levels increased 135% above pre-therapy levels (n=8; p=0.03, Mann-Whitney). These data suggest that upon inhibition of the HER2 tyrosine kinase, HER2+ breast cancers 1) upregulate HER3 by transcriptional mechanisms and partially maintain HER3 function by post-translational mechanisms; 2) this compensatory phosphorylation of HER3 partially maintains PI3K/Akt; and 3) inhibition of HER3 sensitizes HER2-dependent breast cancer cells to HER2 inhibitors. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 63.