Cell culture

The breast cancer cell lines BT474 (Cat# TCHu143) and MDA-MB-231 (Cat# TCHu227) were obtained from National Collection of Authenticated Cell Cultures (Shanghai, China). BT474 cells were maintained in RPMI-1640 (Gibco, Grand Island, NY, USA, Cat# C11875500BT) complemented with 10% fetal bovine serum (FBS) (SuperCulture, Shenzhen, China, Cat# 60211031). MDA-MB-231 cells were maintained in DMEM (Gibco, Cat# C11995500BT) complemented with 10% FBS. All medium contained 100 U/ml penicillin and 100 μg/ml streptomycin (SparkJade, Shandong, China, Cat# CM0004). Cells were grown in a 5% CO2 incubator at 37 °C. All cells were routinely tested for mycoplasma. When treated or transfected, cells were randomly allocated into different treatment groups.

Generation of stable cell lines

mCasExpress system contains two plasmids, pCDH-FRT-STOP-FRT-ZsGreen-puro plasmid and pCW57-Lyn11-NES-DEVD-flpO-hygro15. To generate stable CDH12 knockdown and control cell lines, shCDH12 (TGGATTAGCCGGAACAACAATTCAAGAGATTGTTGTTCCGGCTAATCCTTTTTT) or shNC (TGTTCTCCGAACGTGTCACGTCAATTCAAGAGAACGTGACACGTTCGGAGAATTTTTT) was inserted into pLKO.1 vector. For stable lines overexpressing CDH12, pLVX-IRES-BSD-CDH12 were generated by sub-cloning the coding sequence of CDH12 into the pLVX-IRES-BSD vector, and the empty pLVX-IRES-BSD was used as a control. Each of these plasmids was transfected into HEK293T cells together with pCMV-dR8.2 dvpr (Addgene# 8455) and pCMV-VSV-G (Addgene# 8454) using Lipofectamine 2000 (Invitrogen, New York, USA, Cat# 11668030). The supernatant was harvested and filtered with a 0.45 μm filter at 48 h post transfection and applied to breast cancer cells overnight in the presence of 10 μg/mL polybrene (Solarbio, Cat# H8761). The infected breast cancer cells were then selected for 7–10 days in growth medium containing 2 μg/mL puromycin (Solarbio, Beijing, China, Cat# P8230), 200 μg/mL hygromycin (Solarbio, Cat# H8081) or 10 μg/mL Blasticidin S (Solarbio, Cat# B9300).

siRNA transfection

For the knockdown experiments, cells were transfected with siCDH12 (sense: GCAAGCCACUUUACACCAUTT, antisense: AUGGUGUAAAGUGGCUUGCTT) or siNC (sense: UUCUCCGAACGUGUCACGUTT, antisense: ACGUGACACGUUCGGAGAATT)25, siCREB (Santa Cruz Biotechnology, Cat# sc-29281), or siERK (Santa Cruz Biotechnology, Cat# sc-29307), using the Lipofectamine 2000 transfection reagent according to the manufacturer’s instructions.

Isolation of the ZsGreen+, the ZsGreen−, and the control populations

BT474Cas and MDA-MB-231Cas cells were treated with 1 μg/mL DOX (Sangon Biotech, Shanghai, China, Cat# A600889) for 6 h. Then BT474Cas cells were treated with 20 ng/mL ADR (MedChemExpress, Shanghai, China, Cat# HY-15142A) or 1 μg/mL CDDP (MedChemExpress, Cat# HY-17394) plus 1 μg/mL DOX. MDA-MB-231Cas cells were treated with 30 ng/mL ADR or 3 μg/mL CDDP plus 1 μg/mL DOX. 24 h later, the treatment medium was replaced with fresh growth medium to allow cells to recovery. After 120 h recovery, the ZsGreen+ cells and ZsGreen− cells were sorted on Beckman MoFlo (Beckman Coulter, Indianapolis, IN, US). To isolate control populations, cells were treated with 1 μg/mL DOX for 30 h, then grew in growth medium for 120 h and subjected to sorting on Beckman MoFlo.

Colony formation assay

Cells were seeded at 1 × 103 cells/well in six-well plates. After two weeks, the colonies were fixed with 4% paraformaldehyde and stained with 0.2% crystal violet (Solarbio, Cat# G1065). The colonies were counted and imaged on Olympus TH4–200 microscope (Olympus, Tokyo, Japan).

EdU incorporation assay

EdU incorporation and staining were performed using EdU-555 kit (Beyotime, Shanghai, China, Cat# C0075) according to the manufacturer’s instructions. The percentage of EdU+ cells were measured using CytoFLEX S (Beckman, USA).

Migration and invasion assay

Cell migration and invasion assays were performed using 24-well BD Falcon 8.0 μm transwell inserts (Falcon, BD Biosciences, MA, USA, Cat# 353097). For migration assay, 3 × 105 BT474 cells or 5 × 104 MDA-MB-231 cells were seeded into the top chamber of the insert. For invasion assay, 3 × 105 BT474 cells or 5 × 104 MDA-MB-231 cells were seeded into the top chamber of the insert coated with Matrigel (BD Biosciences, Billerica, MA, USA, Cat# 354234). The upper chamber was filled with serum-free medium, and the lower chamber was filled with the medium containing 10% FBS. After 48 h (for BT474 cells) or 12 h (for MDA-MB-231 cells) incubation, the migrated and invaded cells were fixed with 4% paraformaldehyde (Solarbio, Cat# P1110) and stained with 0.5% crystal violet. The migrated and invaded cells were imaged using an inverted microscope (IX73, Olympus, Tokyo, Japan), and the migration or invasion capacity was determined by the percentage of area covered by the migrated or invaded cells, which was quantified using Image J software (National Institute of Health, USA).

RNA extraction and quantitative RT-PCR

Total RNA was extracted using TRIzol Reagent (Thermo Fisher Scientific, MA, USA, Cat# 15596026). mRNA was first converted into cDNA using RevertAid Reverse Transcriptase (Thermo Fisher Scientific, MA, USA, Cat# K1691) according to the manufacturer’s instructions. Quantitative RT-PCR (qRT-PCR) was performed using FastStart Essential DNA Green Master (Roche, Basel, Switzerland, Cat# 4368706). β-actin was used as an internal reference. Primers used are listed in Table S1.

Western blot

Protein extraction was performed using RIPA buffer (Beyotime, Cat# KGP702) supplemented with 1 mM PMSF (Keygen, Nanjing, China, Cat# KGP610). Protein samples were quantified using the BCA assay kit (Spark Jade, Cat# EC0001). 40 μg protein was separated in SDS-PAGE and transferred onto PVDF membranes. Membranes were incubated overnight at 4 °C with primary antibodies (1:1000), and then with secondary antibodies (1:5000) at room temperature for 1 h. Protein bands were detected using Clarity Western ECL blotting substrates (Thermo, Cat# 32209) and ChemiDoc Imager (Tanon, 5200). The antibodies used are listed in Table S2.

RNA sequencing

Total RNA of cells was extracted with TRIzol reagents (Thermo) following the manufacturer’s protocol. mRNA was enriched using Oligo(dT) beads, fragmented, reversely transcribed into cDNA, and ligated to Illumina sequencing adapters. The cDNA library was sequenced using Illumina nova 6000 by Novogene Biotechnology Co. (Beijing, China). Genes with fold change more than 2 and false detection rate less than 0.05 were considered differentially expressed. GO enrichment analysis was performed using DAVID (https://david.ncifcrf.gov). Transcription factor enrichment analysis was performed using ChEA3 (https://maayanlab.cloud/chea3/).

Xenograft model

All animal studies were conducted following the protocol approved by the Animal Care and Use Committee of School of Basic Medical Sciences of Shandong University. Six-week-old female BALB/c nude mice (Beijing Vital River Laboratory Animal Technology, Beijing, China, Cat# 401) were randomly allocated into different groups. 5 × 106 BT474 cells or 2 × 106 MDA-MB-231 cells were inoculated into the mammary fat pads of the mice. Tumors were measured with a caliper, and the tumor volume was calculated using the formula V = maximal diameter × perpendicular diameter2/2. When the largest tumor reached the maximum allowed size, mice were sacrificed. The tumors were surgically removed, weighted, photographed. Portions of the tumors were immediately frozen in liquid nitrogen for preparation of RNA and protein samples or fixed in 4% buffered formalin for immunohistochemistry.

Immunohistochemistry

Tumor sections were embedded in paraffin wax. After deparaffinization and rehydration, the sections were subjected to antigen recovery then immersed in 3% H2O2 for 10 min to quench endogenous peroxidase. After being blocked with 10% goat serum at room temperature for 1 h, the sections were incubated with anti-Ki-67 antibody (1:200, Cell Signaling Technology, Danvers, MA, USA, Cat# 9027 T) overnight at 4 °C and second antibody (1:200, Jackson ImmunoResearch, Philadelphia, PA, USA, Cat# 111-035-003) at 37 °C for 1 h. The immunoreactions were visualized using horseradish peroxidase (HRP) conjugated DAB. Negative controls were performed by omitting the primary antibodies. Sections were observed and imaged using Olympus BX51 microscope (Olympus, Japan).

In vivo lung metastasis assay

2 × 106 cells MDA-MB-231 cells were resuspended in 100 μL PBS and injected through the tail vein into BALB/c nude mice. After 6 weeks, the mice were sacrificed, and the lungs were collected, fixed and analyzed.

Targeted bisulfite sequencing assay

Genomic DNA was extracted from cells with QIAGEN kit (QIAGEN, Hilden, Germany) according to the manufacturer’s protocol. DNA was quantified and diluted to 20 ng/μL. Genomic DNA (400 ng) was subjected to sodium bisulfite treatment using EZ DNA Methylation™-GOLD Kit (Zymo Research) according to manufacturer’s protocol. CpG islands adjacent to the promoter region of CDH12 gene were assessed by Genesky BioTech (Shanghai, China). Briefly, CpG islands were selected for measurement according to the following criteria: 0.60 or higher ratio of observed/expected dinucleotides CpG; 50% or higher GC content; 100 bp minimum length. Two regions from CpG islands of CDH12 were selected and sequenced. PCR amplicons of target CpG regions were separated by agarose electrophoresis and purified using QIAquick Gel Extraction kit (QIAGEN, Hilden, Germany, Cat# 28704). The products were sequenced on an Illumina MiSeq benchtop sequencer (Illumina, CA, United States).

ChIP-qPCR assay

ChIP assays were performed as described previously51. The site-specific primers used in qPCR are listed in Table S3. The antibodies used in ChIP assays are listed in Table S4.

Statistical analysis

All data are presented as the mean ± standard error of the mean (SEM). Data were analyzed using GraphPad Prism 5 (GraphPad software Inc., San Diego, CA, USA). One-way Analysis of Variance (ANOVA) with Tukey test was used for comparison of three or more groups. The two-tailed unpaired t-test was used to analyze the difference between two groups. P < 0.05 was considered statistically significant. The assumption of equal variance was validated by F-test. No data were excluded from analysis. The sample sizes were chosen empirically based on the observed effects and previous reports. When collecting and analyzing data of RT-qPCR, immunostaining and xenograft volumes, the investigators were blinded to the group allocation. All the experiments were repeated at least three times and the representatives were shown in the figures.