Nicastrin and Notch4 drive endocrine therapy resistance and epithelial to mesenchymal transition in MCF7 breast cancer cells

Abstract

Introduction: Resistance to anti-estrogen therapies is a major cause of disease relapse and mortality in estrogen receptor alpha (ERα)-positive breast cancers. Tamoxifen or estrogen withdrawal increases the dependence of breast cancer cells on Notch signalling. Here, we investigated the contribution of Nicastrin and Notch signalling in endocrine-resistant breast cancer cells.

Methods: We used two models of endocrine therapies resistant (ETR) breast cancer: tamoxifen-resistant (TamR) and long-term estrogen-deprived (LTED) MCF7 cells. We evaluated the migratory and invasive capacity of these cells by Transwell assays. Expression of epithelial to mesenchymal transition (EMT) regulators as well as Notch receptors and targets were evaluated by real-time PCR and western blot analysis. Moreover, we tested in vitro anti-Nicastrin monoclonal antibodies (mAbs) and gamma secretase inhibitors (GSIs) as potential EMT reversal therapeutic agents. Finally, we generated stable Nicastrin overexpessing MCF7 cells and evaluated their EMT features and response to tamoxifen.

Results: We found that ETR cells acquired an epithelial to mesenchymal transition (EMT) phenotype and displayed increased levels of Nicastrin and Notch targets. Interestingly, we detected higher level of Notch4 but lower levels of Notch1 and Notch2 suggesting a switch to signalling through different Notch receptors after acquisition of resistance. Anti-Nicastrin monoclonal antibodies and the GSI PF03084014 were effective in blocking the Nicastrin/Notch4 axis and partially inhibiting the EMT process. As a result of this, cell migration and invasion were attenuated and the stem cell-like population was significantly reduced. Genetic silencing of Nicastrin and Notch4 led to equivalent effects. Finally, stable overexpression of Nicastrin was sufficient to make MCF7 unresponsive to tamoxifen by Notch4 activation.

Conclusions: ETR cells express high levels of Nicastrin and Notch4, whose activation ultimately drives invasive behaviour. Anti-Nicastrin mAbs and GSI PF03084014 attenuate expression of EMT molecules reducing cellular invasiveness. Nicastrin overexpression per se induces tamoxifen resistance linked to acquisition of EMT phenotype. Our finding suggest that targeting Nicastrin and/or Notch4 warrants further clinical evaluation as valid therapeutic strategies in endocrine-resistant breast cancer.

Figure 1

Tamoxifen-resistant cells (TAM-R) are resistant to tamoxifen (4-OH-TAM), phenotypically distinct, more invasive and migratory compared to wild-type MCF7. (A) MCF7 and TAM-R cells treated with vehicle (EtOH) or 10^-7 M 4-OH-TAM were plated (3 × 10³/well) in 96-well plates and allowed to adhere. One plate was fixed and annotated as Day 0. A sulforhodamine B (SRB) assay was performed every two days until Day 6. The experiment was repeated three times and each time six technical replicates were used. (B) Cells were stained with Crystal Violet and 10X images were taken with a bright-field microscope when cells were 50% confluent (bar represents 400 μm). (C) Boyden chamber-based assay was used to determine the cells invasive or migratory capacity. Cells were allowed to invade or migrate for 72 or 18 hrs respectively before the insert was fixed, cut, and mounted in Mowiol infused with DAPI. 4X images were taken with a fluorescent microscope (bar represents 1,000 μm). The results are representative of three biological and two technical replicates. (D) Quantification of microRNA (mRNA) levels of epithelial to mesenchymal transition (EMT) markers or Notch genes (G) analysed by qRT-PCR. Fold change is shown in TAM-R compared to MCF7 cells, normalised to GAPDH. Results represent three biological as well as three technical replicates of each. (Bars represent standard deviation (SD) ^*P <0.05 ^**P <0.01, ^***P <0.0001, t Student, two-tails). (E) Western blot validation for representative EMT markers and Notch proteins (H). ActinB was used as loading control. (F) Representative images showing E-cadherin expression in MCF7 and TAM-R cells at cell-cell contact.

Figure 2

Nicastrin (NCST) monoclonal antibodies (mAbs) and gamma secretase inhibitor (GSI) Pfizer can reverse epithelial to mesenchymal transition (EMT) process. (A) Binding of anti-NCST mAbs to endogenous cells surface Nicastrin of tamoxifen-resistant (TAM-R) and MCF7 cells. Non-permeabilised cells were incubated with 50 μg/ml of mAbs1/2, followed by incubation with the secondary anti-rat FITC antibody. Rat immunoglobulin G (IgG) was used as control. Binding was assessed by FACS. (B) TAM-R cells were pre-incubated for 30 minutes with 50 μg/ml of mAb1/2, or 10 μM GSIPF (PF03084014). For invasion assay, cells were seeded in the upper compartment on top of matrigel-coated membrane and allowed to invade for 72 hrs. For motility assay, matrigel was excluded. Pre-treated cells were seeded on 6-well plates for 54 hrs, then harvested and counted. A total of 50,000 were transferred to the upper compartment for 18 hrs. The results are representative of three biological and two technical replicates. (C, D) qRT-PCR showing mAbs and GSIPF treatment modulate EMT and Notch-responsive genes. Cells were treated as in B and analysed 72 hrs later. microRNA (mRNA) levels are represented as fold induction normalised to GAPDH and compared to control. Results represent three biological as well as three technical replicates of each. (E) Representative western blot showing downregulation of Nicastrin and Notch proteins following mAbs and GSIPF treatments. IQGAP1 was chosen as EMT representative protein. Quantitation represents the average from three biological experiments (bars represent standard deviation (SD)).

Figure 3

Notch4 and Nicastrin knockdown reduce epithelial to mesenchymal transition (EMT) genes expression. (A) qRT-PCR showing Notch4 and Nicastrin knockdown effect on EMT as well as Notch-responsive genes. microRNA (mRNA) levels are normalised to GAPDH. (n = 3 bars represent standard deviation (SD) ^*P <0.05.) (B) Representative images showing E-cadherin expression in MCF7 after control, Notch4 or Nicastrin knockdown. (C) Representative western blot confirming specificity for small interfering RNA (siRNA) targeting Notch4. IQGAP1 was chosen as EMT representative protein (n = 3, bars represent SD ^*P <0.05). (D) Representative western blot confirming Nicastrin siRNA effect on Notch receptors and IQGAP1 (n = 3, bars represent SD ^*P <0.05).

Figure 4

Stem cell content population is increased in tamoxifen-resistant (TAM-R) cells and is affected by anti-Nicastrin (NCST) monoclonal antibodies (mAbs) and gamma secretase inhibitor (GSI) treatment. (A) Percentage of CD44⁺/CD24^- cells assessed by FACS in TAM-R compared to MCF7 cells. The graph shows CD44⁺/CD24^- percentage assessed by FACS in TAM-R cells pre-treated with mAbs or GSIPF as in 2B. All treatments reduce the stem content population. (B) TAM-R cells form more primary and secondary mammospheres. Sphere formation efficacy (SFE) was calculated as the number of spheres formed in 10 days from the original number of single cells seeded and expressed as percentage. Bars represent mean percentage of mammospheres ± standard deviation (SD) from five separate replicates and three separate experiments. A representative image is shown at 10X magnification (^*P <0.05, ^**P <0.01). (C) SFE in secondary mammospheres is greatly reduced in mAb1 and GSIPF-treated cells. A representative image of mammospheres is shown as in B (bars represent SD) ^*P <0.05, ^**P <0.01) (D) MCF7 and TAM-R stem cells were fractionated into cytoplasm, membrane and nucleoplasm. Lysates were immunoblotted and probed for NCST, Notch1, and Notch4. Tubulin was used as cytoplasmic marker. ActinB was used to ensure each fraction was loaded equally.

Figure 5

Stable Nicastrin (NCST) overexpression activates Notch pathways and epithelial to mesenchymal transition (EMT) in MCF7 cells. (A) Representative western blot of NCST, Notch receptors and IQGAP1 levels in control (CT) and NCST overexpressing MCF7 cells. ActinB was used as loading control. (B) qRT-PCR showing Nicastrin overexpression effect on Notch1, Notch4, Notch-responsive genes, and EMT genes (C). microRNA (mRNA) levels were quantified relative to GAPDH. The mean of three different experiments is shown. (D) Representative images showing E-cadherin expression in MCF7 CT and MCF and NCST cells at cell-cell contact. (E) CT and MCF7 NCST cells were stained with Crystal Violet and 10X images were taken with bright-field microscope when cells were 50% confluent (bars represent 400 μM). Boyden chamber-based assay was used to determine the migratory capacity of MCF7 NCST compared to MCF7 CT cells. Cells were allowed to migrate for 18 hrs before the insert was fixed, cut, and mounted in Mowiol infused with DAPI. 4X images were taken with a fluorescent microscope. (Bars represent 1,000 μM). The results are representative of three biological and two technical replicates.

Figure 6

Stable Nicastrin (NCST) overexpression confers 4-OH-TAM resistance in MCF7 cells through Notch4 activation. (A) MCF7 CT and MCF7 NCST cells treated with vehicle (EtOH) or 10^-7 M 4-OH-TAM were plated (3 × 10³/well) in 96-well plates and allowed to adhere. One plate was fixed and annotated as Day 0. A sulforhodamine B (SRB) assay was performed every two days until Day 6. The experiment was repeated three times and each time six technical replicates were used. (B) Representative western blot of estrogen receptor alpha (ERα) protein levels in MCF7 CT and MCF7 NCST cells. ActinB was used as loading control. (C) Quantification of microRNA (mRNA) levels of ERα target genes analysed by qRT-PCR. Fold change is shown in MCF7 NCST compared to MCF7 CT cells, normalised to GAPDH. Results represent three biological as well as three technical replicates of each. (D) Representative western blot confirming Notch4 small interfering RNA (siRNA) in MCF7 NCST cells and the effect on NCST protein. (E) An SRB assay performed as in A following siRNA transfection in MCF7 NCST. The experiment was repeated three times and each time six technical replicates were used.