Anterior gradient-2 plays a critical role in breast cancer cell growth and survival by modulating cyclin D1, estrogen receptor-alpha and survivin

Abstract

Introduction: Anterior-gradient 2 (AGR2) is an estrogen-responsive secreted protein. Its upregulation has been well documented in a number of cancers, particularly breast cancer, for which mixed data exist on the prognostic implications of AGR2 expression. Although emerging evidence indicates that AGR2 is associated with poor prognosis, its function and impact on cancer-relevant pathways have not been elucidated in breast cancer.

Methods: To investigate the biologic role of AGR2 in breast cancer, AGR2 was transiently knocked down, by using siRNA, in T47 D and ZR-75-1 (estrogen receptor-alpha (ER)-positive) and MDA-MB-231 and SK-BR-3 (ER-negative) human breast cancer cell lines. The impact of silencing AGR2 was evaluated in both anchorage-dependent and anchorage-independent growth (soft agar, spheroid) assays. Cell-cycle profiles in ER-positive cell lines were determined with BrdU incorporation, and cell death was measured with Annexin V, JC-1, and F7-26 staining. After transiently silencing AGR2 or stimulating with recombinant AGR2, modulation of key regulators of growth and survival pathways was assessed with Western blot. Combination studies of AGR2 knockdown with the antiestrogens tamoxifen and fulvestrant were carried out and assessed at the level of anchorage-dependent growth inhibition and target modulation (cyclin D1, ER).

Results: AGR2 knockdown inhibited growth in anchorage-dependent and anchorage-independent assays, with a more-pronounced effect in ER-positive cell lines. Cyclin D1 levels and BrdU incorporation were reduced with AGR2 knockdown. Conversely, cyclin D1 was induced with recombinant AGR2. AGR2 knockdown induced cell death in ZR-75-1 and T47 D cells, and also downregulated survivin and c-Myc. Evidence of AGR2-ER crosstalk was demonstrated by a reduction of ER at the protein level after transiently silencing AGR2. AGR2 knockdown in combination with fulvestrant or tamoxifen did not preclude the efficacy of the antiestrogens, but enhanced it. In addition, p-Src, implicated in tamoxifen resistance, was downregulated with AGR2 knockdown.

Conclusions: Transiently silencing AGR2 in ER-positive breast cancer cell lines inhibited cell growth and cell-cycle progression and induced cell death. Breast cancer drivers (ER and cyclin D1) as well as cancer-signaling nodes (pSrc, c-Myc, and survivin) were demonstrated to be downstream of AGR2. Collectively, the data presented support the utility of anti-AGR2 therapy in ER-positive breast cancers because of its impact on cancer-relevant pathways.

Figure 1

siRNA-mediated AGR2 knockdown affects anchorage-dependent and anchorage-independent growth in breast cancer cell lines. T47 D, ZR-75-1, MDA-MB-231, and SK-BR-3 cells were transfected with negative control siRNA (iNC), AGR2 siRNA (iAGR2), or untransfected (UT). KSP (DKSP) and its corresponding control (DNC) were used as transfection controls. Results are expressed as a ratio of untransfected cells (±SD), n = 3. (a) Detection of endogenous AGR2 in breast cancer cell line supernatants by IP-Western and whole-cell lysates by Western. AGR2 knockdown was confirmed in lysates 72 hours after transfection. β-Actin served as a loading control. (b) The impact of iAGR2 on anchorage-dependent growth was evaluated at 96 hours after transfection by using the Cell Titer Glo assay. Anchorage-independent growth assays were also used: (i) soft agar colony formation assay (c), with Alamar blue as a readout; (ii) spheroid assay (d), in which lysed spheroid LDH levels were representative of total cell number after 8 days; corresponding spheroid images were also captured. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 2

AGR2 knockdown reduces cell proliferation in ER-positive breast cancer cells. Cell-cycle profiles were analyzed with BrdU incorporation. Cells were pulse-labeled with 10 μM BrdU 48 hours after transfection and analyzed for BrdU incorporation with FACS. Cells were gated on Sub G₁, G₀/G₁, S, and G₂/M populations. (a) T47 D, (b) ZR-75-1, (c) MDA-MB-231, and (d) SK-BR-3 cells.

Figure 3

AGR2 knockdown induces cell death. ZR-75-1 cells were collected 96 hours after AGR2 knockdown and analyzed for cell death by measuring ssDNA breaks by using the following: (a) F7-26 staining by FACS analysis, and (b) alterations in mitochondrial membrane potential by determining the ratio of JC-1_red to JC-1_green and represented as a ratio of the untransfected control (±SD), n = 3. MG132 and CCCP served as apoptosis and depolarization controls, respectively. Cell death was investigated 96 hours after AGR2 knockdown in T47 D cells by using the following: (c) annexin V (AV) and propidium iodide (PI) and gated on normal (AV^-/PI^-), necrotic (AV^-/PI⁺), early apoptotic (AV⁺/PI^-), and late apoptotic/necrotic (AV⁺/PI⁺) cells; and (d) JC-1 staining, as previously described.

Figure 4

Target modulation of proliferation and survival proteins by AGR2. (a) Lysates 72 hours after transfection with nontargeting control (iNC) or AGR2 (iAGR2) were evaluated with Western blot for modulation of regulators of growth and survival. Note: Some blots may be from different gels run with the same set of samples. (b) ZR-75-1 cells were treated for 6 hours with BSA (5 μg/mL), Novus rhAGR2 (rhAGR2 (N)), or in-house rhAGR2 (rhAGR2(I)), and analyzed with Western blot for cyclin D1 induction. (c) ZR-75-1 cells were plated in eight-well chamber slides and treated with 5 μg/mL of BSA or rhAGR2 (I) for 6 hours. Cells were stained with cyclin D1, and mounting medium containing DAPI was used. Images were taken by using a fluorescent microscope and pseudo-colored in Adobe Photoshop. (d) Quantitation of cyclin D1 immunofluorescence images. The percentage of cells in each bin based on cyclin D1 intensity is represented (Bin 1, weakest staining; Bin 4, brightest staining). Results are expressed as the mean ± SD; n = 4.

Figure 5

Evidence of ER-AGR2 crosstalk. (a) ZR-75-1 cells were treated with vehicle control (DMSO) or E2 (10 nM) for 24 hours analyzed with Western blot. Numbers above bands represent relative AGR2 induction with E2 treatment after quantitation and normalized to β-actin (Image J). (b) Lysates 72 hours after AGR2 knockdown were analyzed with Western blot for ER. **P < 0.01. Note: Some blots may be from different gels run with the same set of samples.

Figure 6

ER-independent activities after AGR2 knockdown. T47 D cells were transfected with iNC or iAGR2, and after 24 hours, cells were treated in combination with 4-hydroxytamoxifen (4-OHT) or ICI 182,780 (ICI) at doses of 100 nM or 1 μM. The growth-inhibitory effects of combination treatments were assessed with Cell Titer Glo 96 hours after transfection (a), or the level of target modulation at 72 hours with Western blot analysis (b). (c) Modulation of p-Src was analyzed 72 hours after AGR2 knockdown. The kinetics of cyclin D1 and ER modulation were determined with Western blot, and densitometric values of cyclin D1 and ER were calculated and normalized to β-actin (Image J) and represented as the mean ± SD, n = 3. *P < 0.05; **P < 0.01; ***P < 0.001, antiestrogen and AGR2 knockdown combination versus AGR2 knockdown alone. Note: Some blots may be from different gels run with the same set of samples.

Figure 7

Impact of rat anti-AGR2 Ab on cell growth and cyclin D1 in T47 D cells. (a) Rat anti-AGR2 Abs were tested for AGR2 specificity by using an ELISA directed against human AGR2 and human AGR3. Species crossreactivity also was assessed by using an ELISA directed against mouse AGR2. (b) After confirming Ab specificity, T47 D cells were treated with an anti-AGR2 Ab (10 μg/mL) for 48 hours or AGR2 siRNA for 72 hours, and cyclin D1 modulation was examined with immunofluorescence. Cells were stained with cyclin D1, and mounting media containing DAPI were used. Images were taken by using a fluorescence microscope and pseudo-colored in Adobe Photoshop. The isotype control Ab used for cyclin D1 staining was an anti-AGR2 Ab of the same isotype but was not shown to modulate cyclin D1 or to have an impact on growth. Cyclin D1 intensity was quantitated by using ImagePro and binned based on intensity, and the percentage of cells in each bin based on cyclin D1 intensity is represented (Bin 1, weakest staining; Bin 4, brightest staining). (c) T47 D, ZR-75-1, and MDA-MB-231 cells were treated for 5 days with 20 μg/mL anti-AGR2 Ab. The relative number of cells was quantitated by using the MTT assay. Results are expressed relative to untreated sample for each cell line.