Knock-down of amphiregulin inhibits cellular invasion in inflammatory breast cancer

Abstract

We have previously shown that SUM-149 human breast cancer cells require an amphiregulin (AREG) autocrine loop for cell proliferation. We also demonstrated that AREG can increase epidermal growth factor receptor (EGFR) stability and promote EGFR localization to the plasma membrane. In the present studies we successfully knocked-down AREG expression in SUM-149 cells by lentiviral infection of AREG shRNA. In the absence of AREG expression, SUM-149 cell growth was slowed, but not completely inhibited. Furthermore, cells infected with AREG shRNA constructs showed an increase in EGFR protein expression by Western blot. Immunofluorescence and confocal microscopy showed that following AREG knock-down, EGFR continued to localize to the cell surface. Soft agar assays demonstrated that AREG knock-down cells retain anchorage-independent growth capacity. Additionally mammosphere forming assays and Adefluor staining analysis showed that knock-down of AREG expression did not affect the expression of stem cell phenotypes. However, following AREG knock-down, SUM-149 cells demonstrated a dramatic decrease in their ability to invade a Matrigel matrix. Consistent with this observation, microarray analysis comparing cells infected with a non-silencing vector to the AREG knock-down cells, identified genes associated with the invasive phenotype such as RHOB and DKK1, and networks associated with cell motility such as integrin-linked kinase signaling, and focal adhesion kinase signaling. AREG was also found to modulate WNT and Notch signaling in these cells. Thus, AREG functions in regulating the invasive phenotype, and we propose that this regulation may be through altered signaling that occurs when AREG activates plasma membrane localized EGFR.

Figure 1. AREG Knock-down in SUM-149 cells

SUM-149 cells were stably infected with various AREG shRNA lentiviral expression constructs. Levels of AREG mRNA expression in cells infected with each individual construct and the combined sh2, sh3, sh4 constructs (shPld) was measured by real-time RT-PCR (A&B). SUM-149 mRNA expression was set to one. Secreted AREG was measured by ELISA (B). Conditioned media was harvested after 24 hours from the SUM-149, SUM-149sh4, SUM-149shPld, and SUM-149shNS cells. The cells were counted and the concentration of AREG was normalized to cell number. AREG protein expression in the whole cell was measured by western blot (D).

Figure 2. AREG knock-down decreases the rate of proliferation in SUM-149 cells

(A) Infected cells were treated with 1ug/ml puromycin 12 hours post-infection (Day1) and counted at days 4, 7, and 10. B) Cells at low passage following shRNA infection were seeded at 1×10⁴ cells per well and counted at approximately 24 hour intervals for 94 hours, and doubling times were calculated. Error bars represent the standard deviation of three replicates in the experiment.

Figure 3. Continued proliferation of AREG knock-down cells is EGFR dependent

(A) Western blot showing EGFR protein levels and (B) phosphorylation at specific EGFR tyrosine residues during normal growth of MCF10A, SUM-149, SUM-149sh4, SUM-149shPld, and SUM-149shNS cells. (C) SUM-149, SUM-149sh4, SUM-149shPld, and SUM-149shNS cells were seeded at 1×10⁴ cells per well. The first bar in each group represents the number of cells on day one. The next four bars show the number of cells after seven days in their normal growth media or with everyday exposure to 1 μg/ml EGFR Ab, 0.5 μmol/L Iressa, or 1 μg/ml AREG Ab. Error bars represent the standard deviation of three replicates in the experiment.

Figure 4. The role of EGF family ligands in growth of AREG knock-down cells

(A) Relative mRNA expression by real-time RT-PCR of AREG, EREG, EGF, TGF-a, BTC, HB-EGF, and EPGN in the SUM-149, SUM-149sh4, SUM-149shPld, and SUM-149shNS cells. (B) Relative mRNA expression by real-time RT-PCR of AREG, EREG, TGF-a, BTC, HB-EGF, and EPGN in the SUM-149 and SUM-149 cells treated with 0.5 μmol/L Iressa for one or 24 hours. (C) Bars show growth of early passage cells (p18) Cells were seeded at 1×10⁴ cells per well. The first bar in each group represents the number of cells on day one. The next four bars show the number of cells after seven days of growing in their normal growth media or with everyday exposure to 1ug/ml IgG mouse Ab, 1ug/ml IgG goat Ab, 1 μg/ml AREG Ab, 1 μg/ml EGF Ab, 1ug/ml EPGN Ab, 1ug/ml EGFR Ab or 0.5 μmol/L Iressa. Error bars represent the standard deviation of three replicates in the experiment. (E) Relative mRNA expression by real-time RT-PCR of AREG in late passage knock-down cells (>p18) with SUM-149 parental cells normalized to one.

Figure 5. Effect of AREG knock-down on soft agar colony formation (A&B) and mammosphere growth (C)

The bars represent the number of colonies that grew in soft agar after three weeks (A) or four weeks (B) of growth divided by the number of cells seeded (1×10⁵). Cells were counted with GelCount, automated mammalian cell counter. Error bars represent the standard deviation of three replicates in the experiment. (C) Cells were seeded at 1×10⁴ cells in ultra-low attachment plates in mammosphere media with (IHE) or with out (IH) 10 ng/ml of exogenous EGF. The bars represent the number of spheres counted by visualization with a microscope. Error bars represent the standard deviation of three replicates in the experiment.

Figure 6. AREG knock-down inhibits cellular invasion

(A) MCF10A, SUM-149, SUM-149sh4, SUM-149shPld, and SUM-149shNS cells were seeded in serum free media on a Matrigel matrix and incubated at 37°C for 24 hours with serum containing media as the chemoattractant. The bars represent the average number of cells that invaded the matrix divided by the number of cells seeded (2.5×10⁴ cells). Error bars represent the standard deviation of three replicates in the experiment. (B) Photos represent one field of view after staining invading cells.

Figure 7. Validation of increased expression of RHOB and DKK1 in AREG knock-down cells

(A)Western blot showing RhoB protein expression in parental and knock-down cells. βActin is used as a loading control. Relative mRNA expression by real-time RT-PCR of (B) RHOB and (C) DKK1 mRNA expression with SUM-149 or SUM-149shNS parental cells normalized to one.