The 10A, 10AT, 10ATG3B and 10CA1a cell lines were obtained from Dr Novak and Dr Miller at the Karmanos Cancer Institute (Detroit, MI, USA). Benign 10A cells, the progenitor line of the 10A series, are spontaneously immortalized breast epithelial cells obtained from a female patient with fibrocystic breast disease (3). The 10A cells were transfected with a mutated T₂₄ Ha-ras gene to generate premalignant 10AT cells (17,18). The premalignant 10ATG3B cell line was generated from 10AT cells that underwent this process of transplantation in nude/beige mice and re-establishment in culture three times (2). The 10ATG3B cells were more tumorigenic than 10AT cells. The fully malignant 10CA1a cell line was generated from a xenograft obtained from sequential passages by trocar transplantation (19). The 10CA1a cell line led to rapidly growing tumors with 100% efficacy.

The cells were cultured in Dulbecco’s modified Eagle’s medium/F-12 medium supplemented with 10 mg/ml of human insulin, 20 ng/ml of epidermal growth factor (all from Invitrogen, Carlsbad, CA, USA), 0.5 mg/ml of hydrocortisone (Sigma-Aldrich, St. Louis, MO, USA), 5% horse serum, 100 U/ml of penicillin and 100 mg/ml of streptomycin (both from Invitrogen). The cells were maintained in a humidified environment of 5% CO₂/95% air at 37°C as previously described (4,5). The 80% confluent cells were filtered in 60 mm² tissue culture dishes.

Antibodies against retinoblastoma (Rb), its phosphoproteins, poly(ADP-ribose)polymerase(PARP),E2F1,Bax, cyclin A, phospho-Akt (Ser473 and Thr308), phospho-p70S6K (Thr421/Ser424 and Thr389), phospho-S6RP, cyclin-dependent kinase (cdk) 2, p21^Waf1/Cip1 and p27^Kip1 were obtained from Cell Signaling Technology (Beverly, MA, USA). p53 antibody was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA) and cyclin D3 antibody was purchased from BD Transduction (San Jose, CA, USA). Anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody was obtained from Calbiochem (EMD Biosciences Inc., San Diego, CA, USA).

The effect of paclitaxel (Sigma-Aldrich) on cell proliferation was tested in cells plated on flat-bottomed 24-well plates at a density of 1–2×10⁴ cells/well in 1-ml medium and determined on the basis of growth characteristics. Triplicate wells were treated with varying concentrations of paclitaxel for 24, 48 and 72 h following overnight incubation. The relative percentage of metabolically active cells to untreated controls was then determined on the basis of the mitochondrial conversion of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide to formalin. The ability of cells to form formalin by active mitochondrial respiration was detected using a 96-well format plate reader by measuring the absorbance at a wavelength of 550 nm (A_{550 nm}). The percentage of metabolically active cells was compared with the percentage of control cells growing in the absence of paclitaxel in the same culture plate. The percentage of inhibition was calculated by the formula: % inhibition = (OD_control - OD_test/OD_control) ×100 (20).

Each cell of the MCF10AT cell lineage (1×10⁴/well) was treated with 0, 1, 2, 5 and 10 nM paclitaxel for 24 h in 96-well plates containing medium with 1% (v/v) horse serum in a humidified environment of 5% CO₂/95% air at 37°C. Cytoplasmic DNA fragments, which are an indicator of apoptosis, were measured with a DNA cell death detection ELISA^PLUS kit (Roche Diagnostics GmBH, Mannheim, Germany) according to the manufacturer’s instructions.

The cells were treated with 10 nM paclitaxel for 24 and 48 h, harvested in 0.25% trypsin/0.1% EDTA (in phosphate-buffered solution), centrifuged for 5 min at 500 × g and 4°C, washed and fixed with 75% ethanol and then resuspended in 1 ml of propidium iodine staining solution (i.e., 50 μg/ml of propidium iodide, 100 U/ml of ribonuclease A and 0.1% glucose) for 1 h. Apoptotic cells were quantified on the FACSCalibur flow cytometer and identified using Cell Quest software (Becton-Dickinson, San Jose, CA, USA). Red fluorescence, measured at 585/542 nm, indicative of propidium iodide uptake by damaged cells, was measured by logarithmic amplification and electronic compensation for spectral overlap (21).

The cells were treated with 10 nM paclitaxel for 24 and/or 48 h, and subsequently lysed. Protein concentrations were determined as previously described (22). Protein samples (20–40 μg of protein/lane) from three dishes of cells were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on a gel (7.5–15%) and transferred to nitrocellulose. Blots were incubated with the appropriate diluted primary antibody in 5% bovine serum albumin (BSA), Tris-buffered saline (TBS) and 0.1% Tween-20 (TBS-T) at 4°C with gentle agitation, overnight, followed by incubation with a secondary antibody conjugated to horseradish peroxidase for 2 h at room temperature for immunodetection. Protein expression was detected by enhanced chemiluminescence (Amersham Life Science, Piscataway, NJ, USA) on Kodak X-OMAT film (Sigma). Exposed film was scanned and the band density was quantified by Kodak 1D Image Analysis (Eastman Kodak Company, Rochester, NY, USA). GAPDH was used as an internal standard.

Total cellular RNA was isolated from cells with TRIzol reagent (Invitrogen) and purified with an RNeasy mini kit (Qiagen, Valencia, CA, USA). Single-stranded oligo(dT)-primed cDNA was generated from 2 μg total RNA using RNA reverse transcriptase (Applied Biosystems, Foster City, CA, USA). The quantitative reverse transcribed PCR (RT-qPCR) reaction was prepared using the Power SYBR-Green Master Mix and RT-qPCR was performed with the StepOnePlus Real-Time PCR system (both from Applied Biosystems). The semi-quantitative RT-PCR reaction was prepared using GoTaq^® PCR Master Mix (Promega, Madison, WI, USA). Each sample had a final volume of 15 μl containing ~100 ng of cDNA. Primers used for analysis of human MDR1, MRP1 and BCRP were: MDR1 (forward, 5′-AAG CCA CGT CAG CTC TGG ATA-3′ and reverse, 5′-CGG CCT TCT CTG GCT TTG T-3′, 73 bp); MRP1 (forward, 5′-AAG AAA ACA GGG AAG CAG CA-3′ and reverse, 5′-GCT CTC TGG GTT TGA AGT CG-3′, 150 bp); BCRP (forward, 5′-CCC GCG ACA GTT TCC AAT GA-3′ and reverse, 5′-GGC GTT GAG ACC AGG TTT CA-3′, 171 bp); and GAPDH (forward, 5′-GTC AAC GGA TTT GGT CGT ATT-3′ and reverse, 5′-AGT CTT CTG GGT GGC AGT GAT-3′). GAPDH was used as an internal standard. The amplification reaction was carried out with 2 μl of cDNA product for 40 cycles; each cycle consisted of 95°C for 30 sec, 60°C for 30 sec, and 72°C for 30 sec, followed by a final 1-min elongation at 72°C. Relative mRNA levels were assessed using the 2^−ΔΔCt method or 1.5% agarose gel electrophoresis.

Data were presented as means ± standard deviation. Statistical significance (P<0.05) between groups was determined by analysis of variance (ANOVA), followed by the Tukey-Kramer multiple comparison analysis (4,5).

Paclitaxel differentially inhibited cell proliferation of 10A, 10AT, 10ATG3B and 10CA1a cells in a time-dependent (24, 48 and 72 h) and concentration-dependent (0–10 nM) manner as determined by the MTT assay (Fig. 1). Paclitaxel treatment at 10 nM for 48 h inhibited 10A, 10AT and 10ATG3B cell proliferation by 62.3, 46.2 and 42.3%, respectively. However, paclitaxel inhibited 10CA1a cell proliferation by 29.2% at 48 h, showing that the 10CA1a tumor cells were less sensitive to paclitaxel than the benign (10A) and premalignant (10AT and 10ATG3B) cells (P<0.01, Fig. 1). The same pattern of inhibition by paclitaxel was also observed at 24 and 72 h (Fig. 1).

We determined whether paclitaxel induced apoptosis and cell cycle arrest of the MCF10AT cell series. Paclitaxel differentially arrested the cell-cycle progression of the MCF10AT cell series at the subG₁/G₁ phase (Fig. 2). The treatment of MCF10AT cell lineage with 10 nM paclitaxel for 24 h induced the apoptosis of cells in sub-G₁ phase by 23.6, 26.1, 25.2 and 8.96% in 10A, 10AT, 10ATG3B and 10CA1a, respectively, showing less sensitivity in 10CA1a cells. However, treatment with 10 nM paclitaxel for 48 h resulted in the progression to G₁/S-phase arrest. The percentage of cells in the G₁ phase increased by 27.6 and 26.7% in premalignant 10AT and 10ATG3B cells as compared to the control at 24 h, while the corresponding values in the S phase decreased by 11.3 and 19.8% (Fig. 2). G₁ arrest by paclitaxel was prominent in the 10AT and 10ATG3B premalignant cells at 48 h and 20.2 and 10.2% apoptosis was consistently detected in 10A and 10CA1a cells, respectively, at 48 h (Fig. 2). These results indicated that paclitaxel inhibited cell proliferation through cell-cycle arrest from the subG₁ to G₁ phase at 24 h and from G₁ to S phase at 48 h in the MCF10AT cell lineage, which was accompanied by a decrease in S-phase progression. Additionally, the 10CA1a tumor cells were more resistant to paclitaxel-mediated cell cycle arrest (Fig. 2).

The apoptotic effect of paclitaxel was confirmed by the measurement of cytoplasmic DNA fragments, the hallmark of apoptosis (Fig. 3). The cells treated with 0, 1, 2, 5 and 10 nM paclitaxel for 24 h showed that DNA fragments increased in a concentration-dependent manner with less sensitivity in fully malignant 10CA1a cells (Fig. 3). Treatment of 10A, 10AT, 10ATG3B and 10CA1a cells with 10 nM paclitaxel resulted in DNA fragmentation levels of 300, 210, 270 and 181%, respectively, as compared to each control, while the 10CA1a cells were significantly less sensitive to paclitaxel than the benign and premalignant cells (P<0.05).

Immunoblot analysis results of subG₁/G₁ checkpoint protein expression at 24 h, particularly p53 and PARP are shown in Fig. 4. Paclitaxel increased p53 protein expression by 87, 102, 812 and 84% in 10A, 10AT, 10ATG3B and 10CA1a cells, respectively, and induced cleaved PARP fragmentation. Cleaved PARP fragmentation was minimally detected at 48 h except in 10A cells. These results correlated well with hte paclitaxel-mediated induction of apoptosis with the decreased sensitivity in 10CA1a cells (Fig. 2).

Cell cycle-dependent phosphorylation by cdk inhibited Rb target binding, thus allowing cell-cycle progression (23). The levels of Rb protein were significantly reduced in each of the cell lines at 24 and 48 h post-paclitaxel treatment (Fig. 4). The levels of phospho-Rb (Ser780, Ser795 and Ser807/811) were differentially decreased in the four cell lines at 24 and 48 h post-paclitaxel treatment, with less decrease in 10CA1a cells (Fig. 4). Paclitaxel increased the cdk inhibitor, p21^Waf1/Cip1 protein expression in 10A, 10AT and 10ATG3B cells at 24 h by 2.57-, 1.53- and 2.48-fold, respectively, but negligibly in 10CA1a cells at both 24 and 48 h (Fig. 4). p27^Kip1 protein, another cdk inhibitor, was also significantly increased in 10A, 10AT, 10ATG3B and 10CA1a cells by 2.73-, 2.22-, 2.40- and 1.71-fold, respectively, post-paclitaxel treatment for 24 h (Fig. 4). Cyclin D3 levels were minimally changed at 24 h post-paclitaxel; however, 48-h treatment with paclitaxel effectively reduced cyclin D3 levels in all the cell lines by 11–85% (Fig. 4). These results supported, in part, that paclitaxel inhibited cell proliferation by G₁/S cell-cycle arrest in the MCF10AT cell lineage.

Activation of Akt was shown to play a key role in the development of therapeutic resistance (24). The basal protein expression of phospho-Akt (Ser473 and Thr308) was significantly higher in 10CA1a tumor cells (4) and was not inhibited by paclitaxel at 24 h (Fig. 5A). These phenomena applied to the downstream targets of Akt, i.e., phospho-p70S6K (Thr421/Ser424 and Thr389) and phospho-S6RP, where paclitaxel inhibited these phosphoproteins efficiently in 10A, 10AT and 10ATG3B cells but not sufficiently in 10CA1a cells (Fig. 5A).

We also observed other proteins involved in apoptosis and the G₁/S cell-cycle checkpoint, i.e. E2F1, Bax, cyclin A and cdk2. Paclitaxel reduced cdk2 protein levels in 10A, 10AT, 10ATG3B and 10CA1a cells at 24 h by 55.0, 70.3, 68.2 and 18.2%, respectively, compared to the untreated controls (Fig. 5B). Cyclin A, the binding partner of cdk2 for S-phase progression, was inhibited by paclitaxel in the same pattern as cdk2 (Fig. 5B). Transcription factor E2F1 binds preferentially to Rb in a cell cycle-dependent manner and mediates cell proliferation and apoptosis (25). Paclitaxel reduced E2F1 protein levels in the 10A, 10AT, 10ATG3B and 10CA1a cells at 24 h by 52.2, 55.0, 59.9 and 3.9%, respectively, compared to the untreated controls (Fig. 5B). The apoptosis regulator Bax was upregulated by the tumor suppressor protein p53 and was found to be involved in the p53-mediated apoptosis (26). The 10A, 10AT and 10ATG3B cells gradually increased basal Bax protein expression and paclitaxel-induced Bax protein; however, Bax was minimally expressed in 10CA1a cells (Fig. 5B).

We evaluated the causes of chemoresistance to paclitaxel in 10CA1a cells by measuring the basal expression of multidrug resistance-associated genes in the MCF10AT cell lineage using RT-qPCR. MDR1 was significantly decreased (P<0.001) in 10AT, 10ATG3B and 10CA1a cells. However, MRP1 and BCRP gene expression was significantly increased in 10AT and 10CA1a cells compared to benign 10A cells (Fig. 6A). The effects of paclitaxel on the MDR1, MRP1 and BCRP gene expression were not significant in the MCF10AT cell lineage, with only minimal changes observed at 24 h (Fig. 6B).