Effect of lapatinib on signaling in normal human mammary epithelial cells (HMECs). Lapatinib was obtained from GlaxoSmithKline (Research Triangle Park, NC). HMECs (1 × 10⁶ cells) from Clonetics (San Diego, CA) were treated with lapatinib (100 nM), dimethyl sulfoxide (0.1%), and epidermal growth factor from R&D Systems (Minneapolis, MN) (final concentration of 20 nM) and harvested at 0, 5, 10, 15, 30, and 60 minutes later. Protein lysates were generated and subjected to immunoblot analysis as previously described (24). A) Expression of total and activated tyrosine kinase receptors of the Her family. B) Expression of total and activated mitogen-activated protein kinases. The following primary antibodies were purchased from Cell Signaling Technology, Inc: phospho-epidermal growth factor receptor (p-EGFR) (rabbit polyclonal, 1:2000 dilution), total epidermal growth factor receptor (rabbit polyclonal, 1:2000 dilution), phospho-ErbB2 (rabbit polyclonal, 1:2000 dilution), phospho-ErbB3 (rabbit polyclonal, 1:1000 dilution), total ErbB4 (rabbit polyclonal, 1:1000 dilution), phospho-extracellular signal-regulated kinase I/II (rabbit polyclonal, 1:2000 dilution), total extracellular signal-regulated kinase I/II (rabbit polyclonal, 1:1000 dilution), phospho-Akt (rabbit polyclonal, 1:1000 dilution), total Akt (rabbit polyclonal, 1:2000 dilution), phospho-c-Jun NH₂-terminal kinase (rabbit polyclonal, 1:1000 dilution), total c-Jun NH₂-terminal kinase (JNK) (rabbit polyclonal, 1:1000 dilution), phospho-p38 (rabbit polyclonal, 1:1000 dilution), and total p38 (rabbit polyclonal, 1:1000 dilution). Other primary antibodies used were as follows: total ErbB2 (rabbit polyclonal, 1:2000 dilution, NeoMarkers/Labvision), total ErbB3 (rabbit polyclonal, 1:1000 dilution, Santa Cruz Biotech, Santa Cruz, CA), phospho-ErbB4 (rabbit polyclonal, 1:1000 dilution, Orbigen, San Diego, CA), and anti–β-actin (mouse monoclonal, 1:8000 dilution, Sigma-Aldrich, St Louis, MO). Secondary antibodies against mouse or rabbit were purchased from Amersham (Piscataway, NJ) and used at a 1:4000 dilution.

Tumor-free interval in MMTV-erbB2 mice treated with lapatinib or vehicle. FVB strain female MMTV-erbB2 mice (3 months old) were housed in the institutional animal facilities, fed a controlled diet of AIN-76A Purified Diet (Harlan Teklad, Madison, WI), and randomly assigned to receive treatment with vehicle (hydroxylpropyl methylcellulose, Methocel K15M Premium, DOW Chemical Company) or lapatinib (at 30 or 75 mg/kg body weight; GW572016, GlaxoSmithKline) for 6 d/wk starting at 3 months of age (n = 16 for each of the lapatinib-treated groups and n = 17 for the vehicle group). Lapatinib was suspended in hydroxypropyl methylcellulose and administered by twice-daily oral gavage using a 20-gauge gavage needle in a volume of 0.1 mL. Tumor development was monitored twice weekly, and tumor growth was measured using calipers. Tumor volume was calculated by multiplying the square of the width (w) by the length (l) and dividing by two (V = w²l/2). The tumor-free interval was defined from the start of treatment to the first appearance of a palpable mammary tumor at least 100 mm³ in size. Tumor-free interval curves were estimated by the Kaplan–Meier method and compared between vehicle and lapatinib treatment (low and high doses) using the generalized Wilcoxon test (two-sided). The tick mark in the low-dose lapatinib arm represents a censored mouse that died at day 207 without a tumor. Error bars show 95% confidence intervals at 300 days.

Effect of lapatinib on the development of premalignant lesions and on biomarker expression in mammary glands. MMTV-erbB2 mice were randomly assigned to receive treatment 6 d/wk with vehicle or lapatinib (75 mg/kg body weight) for 5 months beginning at 3 months of age (n = 20 mice per group). The mice were killed by CO₂ asphyxiation at 8 months of age, and their mammary glands were harvested. Mammary gland samples were fixed overnight in 4% phosphate-buffered formaldehyde and then embedded in paraffin for histological examination and biomarker measurement. A) Histology: Mammary tissue sections (4 μm thick) were stained with hematoxylin–eosin. Representative fields containing normal mammary gland, hyperplasia, mammary intraepithelial neoplasia (MIN), and invasive mammary tumors are shown. B) Premalignant lesions and microscopic invasive cancers. One #3 mammary gland was removed from each mouse and processed for histological assessment, and the number and types of premalignant lesions in one randomly chosen section of each mammary gland were assessed microscopically. Mice were categorized according to the most advanced lesion observed for each mouse. The data were analyzed using Fisher exact test (two-sided). C) Proliferation of mammary gland epithelial cells from MMTV-erbB2 mice treated with lapatinib or vehicle. Ki67 expression in mammary glands was examined by immunohistochemical staining using an anti-Ki67 antibody (rabbit monoclonal, undiluted, NeoMarkers). Data were generated from 20 mice from each experimental group. Representative fields are shown (left panels). The percentage of Ki67-positive cells was quantified by counting at least 1000 cells per one section per mouse (right panel). Error bars correspond to 95% confidence intervals. P value (two-sided) was calculated using the Wilcoxon rank sum test. D) Expression of cyclin D1, epiregulin, and p27 mRNAs in mouse mammary glands. The #4 mammary glands were collected from the MMTV-erbB2 mice treated with vehicle or lapatinib for 5 months. Enriched mammary epithelial cell pellets were collected, and total RNA was extracted from one #4 mammary gland from 15 of the 20 mice from each group (the #4 mammary glands from the remaining five mice were used for whole-mount preparations). These 15 mammary glands were used to prepare three pooled samples of five #4 mammary glands each. RNA was then extracted from these three pools of mammary glands as previously described (26). The relative mean mRNA expression level vs the level of cyclophilin mRNA from these three pooled samples as measured by quantitative real-time reverse transcription–polymerase chain reaction is plotted. Error bars correspond to 95% confidence intervals. The P values were from two-sided Student t tests. E) Immunoblot analysis of phospho-epidermal growth factor receptor (p-EGFR), phospho-ErbB2, and p27 in mouse mammary epithelial cells. The total cell lysate samples described in panel D (from 15 of the 20 mice per group) were used in this analysis. Antibodies against p-EGFR, phospho-ErbB2, and p27 were used (described in the legend to Figure 1). The blots were densitometrically quantified, and the mean values for the ratios of p-EGFR to actin, phospho-ErbB2 to total ErbB2, and p27 to actin were plotted. Error bars correspond to 95% confidence intervals. P values (two-sided) were obtained using Student t test.