(A) MDA-MB-468 cells were infected with lentivirus containing BIKDD or vector alone, and were analyzed by cell numbers, MTT and clonogenicity assays (setting at 100% in vector group). (B) Cells were imaged, and were immunoblotted by N-terminal Bik antibody and cleaved PARP antibody, respectively. Scale bar indicates 10 μm. (C) 200,000 MDA-MB-468 cells were analyzed by flow cytometry 7 days post infection or paclitaxel treatment once at 1.0 μg/ml dose using fluorescent-conjugated CD44 and CD24 antibodies. The data represent the mean of three independent experiments. For mammosphere formation assay, indicated cells were used and the mammospheres was counted under microscope with size ≥100 μm. (D) MDA-MB-468 mammospheres established as above were infected with lentivirus containing either vector alone or BIKDD, and the number of re-formed mammospheres were counted within 7 days. Scale bar indicates 50 μm. (E) BT474 cells eliminated the CD44⁺/CD24⁻ population and mammosphere formation of BCICs. All error bars indicate SD.

(A) MDA-MB-468 cells were infected with lentivirus containing BIKDD or shRNAs against scramble, BCL-2, BCL-xL, MCL-1, or all of them, respectively. Cells from each group were analyzed by flow cytometry 7 days post virus infection. (B) Indicated numbers of cells were subjected to mammosphere culture post infection up to two weeks. (C) The protein expression levels of Bcl-2, Bcl-xL, Mcl-1, and tubulin were detected 7 days after virus infection. All error bars indicate SD. See also Figure S2.

(A) Schematic diagram of engineered claudin-4-based constructs. (B) VISA-claudin4-BikDD selectively inhibited the growth of multiple breast cancer cell lines in vitro. The percentage of growth inhibition compared to CMV-BikDD is presented. (C–D) VISA-claudin4-BikDD greatly suppressed tumor growth and dissemination of 4T1 mouse breast orthotopic synergeic tumors. Mice were injected with 0.75 mg/kg VISA-claudin4-BikDD or CMV-BikDD as indicated by arrows, respectively. (E) Mice survival was monitored after 28 days post tumor transplantation. (F) VISA-claudin4-BikDD at 0.75 mg/kg significantly reduced the tumor growth of MDA-MB-435 human breast cancer orthotopic xenograft in nude mice. (G) Mice survival was monitored after 60 days post tumor transplantation. (H) VISA-claudin4-BikDD is much safer than CMV-BikDD with intravenous treatment of 5 mg/kg plasmid/Liposome complex. All error bars indicate SD. See also Figure S4.

(A) BCICs from MDA-MB-468, MCF7/HER2, MDA-MB-435, BT549 and HBL100 cells isolated by either FACS sorting or mammosphere culture, were transient transfected with VISA-claudin4-Luc, with pRL-TK as internal control. The luciferase signal was measured in each population of unsorted parental cells and BCICs, normalized with DNA concentration of firefly/renilla luciferase to exclude the potential different transfection rate. The percentage of the luciferase signals as compared with unsorted total population (setting at 100%) was presented. (B) The CD44⁺/CD24⁻ population of MDA-MB-468 tumor in the different groups were determined at Day 26 (the day of last drug treatment). The P-values shown were obtained by comparison with the control group. (C) For off-therapy study, we monitored tumor growth in different group after treatment stopped at Day 26 post first drug treatment, and the tumor growth rate (slope) of off-therapy stage (day 26 to day 49) was calculated by exponential trendline in Excel software (D). All error bars indicate SD. See also Figure S6.

(A–B) patient primary breast tumor cells were isolated from radiated-patient primary tumor after surgery, as described in experimental procedures. After BikDD treatment, cells were subjected to FACS analysis. 3,000 cells were used for mammosphere formation, and mammospheres were counted with size at >100 μm. (C) Mouse MMTV-Neu primary tumor cells were isolated from primary tumor in MMTV-Neu transgenic mouse and subjected to FACS analysis using either mouse CD24⁺/CD29⁺ (top) or CD24⁺/CD49f⁺ (bottom) fluorescent-conjugated antibodies after BikDD treatment. (D) For mammosphere formation, 5,000 cells were used and mammospheres were counted with size >100 μm. (E) MMTV-Neu primary mammospheres were treated with either vector alone or BikDD via lentivirus infection, and mammospheres were counted after reformation. Scale bar indicates 50 μm. (F) Indicated numbers of parental MDA-MB-468 cells and mammosphere cells with or without BikDD treatment, were inoculated into the mammary fat pad of NOD/SCID mice, and mice were monitored for the next months for tumorigenesis. All error bars indicate SD. See also Figure S1.

(A) The claudin-4 promoter was selectively activated in breast cancer cell lines. (B) Schematic diagram of engineered claudin-4-based constructs. (C) The promoter activity of engineered VISA-claudin4-Luc was robust and selectively activated in breast cancer cell lines in vitro. (D) The promoter activity of VISA-claudin4-Luc was expressed much longer than that of CMV promoter in breast cell lines. The ratio of VISA-claudin4-Luc/CMV-Luc at day 2 and day 5 was presented in table. (E) The engineered VISA-claudin4-Luc was robust and selectively expressed in tumor in vivo. Mice were subjected to luciferase imaging 48 hours after injection with 2.5 mg/kg DNA plasmid per mice in 100 μl plasmid:liposome complex, 5 mice per group. The photon signals were quantified, and the percentage of the photon signals as compared with the CMV-Luc (setting at 100%) was presented. (F) Mice were then dissected, and tumor tissue and other specified organs were removed out and subjected to ex vivo imaging. All error bars indicate SD. See also Figure S3.

(A) VISA-claudin4-BikDD at 0.5 μg sensitized multiple EGFR⁺/HER2⁺ breast cancer cell lines to 1 μM lapatinib, but not in the EGFR⁻/HER2⁻ MCF7 breast cancer cell line and the MCF10A normal breast cell line. The percentage of the cell viability as compared with itself without lapatinib treatment (setting at 100%) was presented. (B) Cancer cells were transiently transfected with siRNAs against BCL-2, BCL-xL, MCL-1, or all of them, respectively, and were subjected to lapatinib at indicated concentrations for 72 hours, and the OD_570nm was measured 4 hours after MTT staining. (C) VISA-claudin4-BikDD has synergistic effects with lapatinib in multiple human breast xenograft mouse models. Mice were treated with 0.75 mg/kg VISA-claudin4-BikDD-liposome complexes intravenously as indicated by arrows. Lapatinib was orally administrated at 50 mg/kg every other day. All error bars indicate SD. See also Figure S5.