G Protein-coupled Receptor Kinase 2 (GRK2) Promotes Breast Tumorigenesis Through a HDAC6-Pin1 Axis

Abstract

In addition to oncogenic drivers, signaling nodes can critically modulate cancer-related cellular networks to strength tumor hallmarks. We identify G-protein-coupled receptor kinase 2 (GRK2) as a relevant player in breast cancer. GRK2 is up-regulated in breast cancer cell lines, in spontaneous tumors in mice, and in a proportion of invasive ductal carcinoma patients. Increased GRK2 functionality promotes the phosphorylation and activation of the Histone Deacetylase 6 (HDAC6) leading to de-acetylation of the Prolyl Isomerase Pin1, a central modulator of tumor progression, thereby enhancing its stability and functional interaction with key mitotic regulators. Interestingly, a correlation between GRK2 expression and Pin1 levels and de-acetylation status is detected in breast cancer patients. Activation of the HDAC6-Pin1 axis underlies the positive effects of GRK2 on promoting growth factor signaling, cellular proliferation and anchorage-independent growth in both luminal and basal breast cancer cells. Enhanced GRK2 levels promote tumor growth in mice, whereas GRK2 down-modulation sensitizes cells to therapeutic drugs and abrogates tumor formation. Our data suggest that GRK2 acts as an important onco-modulator by strengthening the functionality of key players in breast tumorigenesis such as HDAC6 and Pin1.

Keywords: Acetylation; Breast transformation; Cancer; GRK2; HDAC6; Pin1.

Graphical abstract

Fig. 1

Luminal breast tumors oncogenic pathways promote increased GRK2 protein levels in cellular and animal models. (A) Western blot analysis of GRK2 expression levels relative to 184B5 cells in non-transformed (NT), “luminal-like” and “basal-like” transformed breast cancer cells. (B) Molecular features of cell lines used in (A):+, presence; −, absence; null, homozygous deletion. (C) Immunoblot analysis of pSer473- AKT and pan-AKT levels in panel A cell lysates. (d, e) Analysis of GRK2 and HDAC6 expression in luminal breast tumor cells in control or estrogen-depleted conditions (D), or upon estrogen re-stimulation (E). (f) Levels of GRK2 and AKT activation in MDA-MB-468 cells treated with the EGFR-inhibitor AG1478. (G, H). Mammary expression of GRK2 and AKT activation in tumor-bearing HER2-transgenic mice (n = 4, g) or MyrAKT1-transgenic animals (n = 2 per group, h) as compared to normal paired glands or non-transgenic littermates. All data are mean ± SEM, n = 3–4. Actin or GAPDH expression was used as loading control were indicated. *p < 0.05; **p < 0.01; ***p < 0.001.

Fig. 2

GRK2 over-expression potentiates mitogenic signaling pathways in breast cancer cells via HDAC6. (a) Analysis of both AKT and ERK1/2 responses to 100 ng/ml EGF in MCF7-F5luc cells with extra (GRK2 35 clone) or silenced (shGRK2) expression of GRK2. *p < 0.05, **p < 0.01 and ***p < 0.001, and T p < 0.05, TT p < 0.01 comparing pcDNA3 to shGRK2 and GRK2 conditions, respectively. (b–c) Analysis of EGF-triggered ERK1/2 and AKT stimulation upon transfection of GFP-HDAC6-wt or GRK2-phosphodefective HDAC6 constructs in MCF7 cells (b), or stable transfection in MDA-MB-231 cells (c). *p < 0.05, **p < 0.01 and ***p < 0.001 compared to empty vector-bearing cells and T p < 0.05, TT p < 0.01 for comparison of HDAC6-wt to HDAC6 mutant-expressing cells. In all panels data are mean ± SEM (n = 3–4). Representative blots are shown.

Fig. 3

GRK2 fosters Pin1 functionality by regulating its acetylation status in a HDAC6-dependent manner. (a) Effect of GRK2 dosage on Pin1 acetylation status. Levels of acetyl-Pin1 in MCF7-F5luc cells stably over-expressing GRK2 or a silencing shRNA-GRK2 construct were determined as detailed in Materials and Methods. *p < 0.05, **p < 0.01 and ***p < 0.001 compared to empty vector, TT p < 0.01 for the indicated comparison (n = 3–4). (b) GRK2 promotes Pin1 de-acetylation in a catalytic-dependent manner. Pin1 acetylation levels were assessed in MCF7tet-GRK2 and MCF7tet-GRK2-K220R cells treated or not with tetracycline for 24 h and stimulated with EGF (100 ng/ml) for the indicated times. *p < 0.05, when comparing EGF-stimulated tetracycline-induced vs non-induced conditions. (c) EGF-induced co-immunoprecipitation of endogenous HDAC6 and Pin1 in tetracycline-induced MCF7tet-GRK2wt cells. **p < 0.01 (or ^#p < 0.05) when compared to unstimulated cells in a two-tail (or one-tail) t-test analysis (n = 3). (d, e). Effect of GRK2 on Pin1 de-acetylation by HDAC6. Pin1 acetylation decays upon transient (d) or stable (e) over-expression of HDAC6-wt, but not of an HDAC6 mutant defective in GRK2 phosphorylation, in MCF7 (d) or MDA-MB-231 (e) cells. (f, g) Prevention of K46-acetylation enhances the binding of Pin1 to mitotic phospho-proteins. Lysates of MCF7 cells in exponential growth or in mitotic arrest by Nocodazol were pulled-down with GST or GST-Pin1 constructs and analyzed by immunoblotting with anti-MPM2 (f) or anti-PLK1 (g) antibodies. Equal loading of GST fusion proteins and mitotic arrest were confirmed by western blot with anti-GST and anti-phospho Histone3 antibodies. In these panels data are mean ± SEM, n = 2–5. *p < 0.05, **p < 0.01 and ***p < 0.001.Representative blots are shown.

Fig. 4

GRK2 endows transformed breast cells with proliferative advantages. (a, b) Growth of parental, shRNA-GRK2, wt or K220R mutant GRK2-infected MCF7 (wt p53) (a), or MDA-MB-231 (mutant p53) (b) cells in normal culture medium was monitored using the xCELLigence technology as described in Materials and Methods. * or T, p < 0.05 and **p < 0.01 when compared to control infected cells. (c–d) HDAC6 underlies the growth-promoting effects of GRK2. (c) xCELLigence-based analysis of growth of non-induced or tetracycline-induced (24 h) GRK2wt-MCF7-Tet cells upon HDAC6 silencing with a pool of siRNAs as detailed in Materials and Methods. *p < 0.05 and ***p < 0.001 (two-tail), when compared to siRNA control-transfected and untreated cells, ^#p < 0.05 and ^##p < 0.01 (one-tail). Proper knockdown of HDAC6 and induction of GRK2 is shown in a representative blot. (d) Stable expression of HDAC6-wt but not of the HDAC6-S1060-1062-1068A mutant fosters growth in MDA-MB-231 cells. (e) Acetylation of Pin1 impairs proliferation of tumor cells. Growth of MCF7 cells co-transfected either with HA-Pin1-wt, HA-Pin1-K64R or HA-Pin1-K46Q and a combination of two specific Pin1 siRNAs or a control siRNA was determined as above. *p < 0.05, ^#p < 0.05, T p < 0.05 in one-tail t-test analysis. Similar expression of Pin1 constructs and knockdown of endogenous Pin1 was confirmed by immunoblot with anti-HA and anti-Pin1 antibodies. A fragment that co-migrates with endogenous Pin1 is detected with a specific Pin1 antibody upon HA-Pin1 transfection. In all panels data are mean ± SEM (n = 3).

Fig. 5

GRK2 favors the anchorage-independent growth of breast tumor cells in a kinase-dependent manner via regulation of HDAC6 and Pin1. (A) Time-course of GRK2 levels in tetracycline (TET)-induced MCF7tet-GRK2wt or -GRK2-K220R cells. Colony formation by either un-induced or tetracycline-induced MCF7tet-GRK2wt and -GRK2-K220R cells (B–D) or by GRK2- or shRNA-GRK2-infected MDA-MB-231 cells (E) in soft-agar medium was analyzed as described in Materials and Methods. The area of colonies was measured using Image J and the size of colonies was scored as Large, Medium and Small as described in Materials and Methods. Bar-graphs show the distribution (%) of the colonies according to their size (C, E) and the median size (D) of colonies. Data are mean ± SD of 2 duplicate independent experiments. (*p < 0.05, **p < 0.01, ***p < 0.001). (F, G) The GRK2-HDAC6-Pin1 axis fosters malignant growth of luminal and basal breast tumor cells. Soft agar colony formation of MCF7 cells (F) transiently co-transfected with either HA-Pin1-wt, HA-Pin1-K64R or HA-Pin1-K46Q and specific Pin1 siRNAs or a control siRNA, as well as MDA-MB-231 cells (G) stably expressing either HDAC6-wt or the GRK2-deficient phosphorylation mutant HDAC6-S1060-1062-1068A in 96-multiwell plates was monitored by Alamar Blue staining as described in Materials and Methods. Data are mean ± SEM of 3 independent experiments performed by triplicate. *p < 0.05, **p < 0.01, when compared to empty-vector or siRNA-control transfected cells in each condition and ^#p < 0.05, comparing HDAC6-wt with HDAC6 mutant. Representative images of cellular foci are shown.

Fig. 6

GRK2 protein levels are up-regulated in a significant proportion of breast cancer patients and positively correlate with relevant prognostic markers of tumor aggressiveness. GRK2 levels (A) and p-S473 AKT (B) were analyzed by western blot in 27 samples from breast cancer patients with infiltrating ductal carcinoma and 12 controls. Bars indicate median values. Pie charts show group tumor patients based on the degree of change in GRK2 and p-AKT expression levels compared to an arbitrary control [(+), higher than control; (−), lower than control; nc, within the range of normal variation (see Materials and Methods)]. (C, D) Samples of 49 patients were analyzed by immunohistochemistry to detect GRK2 and p-AKT. Samples were scored as positive (moderate or strong staining, +) or negative (none or weak staining, −) for GRK2 and p-AKT levels (C). Samples were stratified by p-AKT levels and the distribution of GRK2 groups plotted in a pie chart. Representative sections of GRK2 and p-AKT staining of two patients are shown in panel d. The distribution of tumor size (E), Ki67 (F) and p53 (G) in the cohort of 27 patients with IDC was plotted according to GRK2 expression in a pie chart. p53 levels were determined by immunostaining and the remaining markers were obtained from clinical data (+, presence; −, absence). (H, I) GRK2 correlation with Pin1 expression and its de-acetylation status. Levels of Pin1 and GRK2 were determined by western blot (+, higher than control; −, lower than control; n.c, within the range of normal variation). Pie chart shows GRK2 distribution according to Pin1 expression (H). Pearson test was performed for statistical analysis (C–H). The acetylation status of Pin1 (I) was determined by immunoprecipitation of Pin1 and subsequent immunoblotting with acetyl-lysine and Pin1 antibodies. Pin1 acetylation ratio was plotted against GRK2 levels and the curve was fitted into a quadratic regression linear model using the BestCurFit program.

Fig. 7

GRK2 modulates in vivo tumor growth in a kinase-dependent manner by regulating proliferative and apoptotic pathways. (a) MCF7tet-GRK2wt or -GRK2-K220R cells pre-treated with tetracycline or vehicle (control) were subcutaneously implanted in doxycycline-treated or un-treated (control) nude mice. Tumor volume was measured and endpoint tumors were immuno-stained for GRK2, Ki67, p53 and cleaved caspase-3 (b) as described in Materials and Methods. (Scale bar: 100 μm). Data are the mean ± SEM from tumor masses of 6–8 mice per group. (c) Tumor growth was monitored (6–10 mice per condition) in nude mice subcutaneously injected with MCF7 cells infected with the indicated adenoviral constructs and verified for proper GRK2 levels by western blot. (d) Immunohistochemical analysis of GRK2, Ki67, p53 and annexin V expression in sections of tumors 8-days post-injection. (Scale bar: 100 μm). (e) Analysis of the orthotopic growth of MCF7-F5luc cells stably expressing GRK2-wt or a silencing shRNA-GRK2 construct injected in the mammary fat pad of nude mice with Xenogen imaging, using luciferin substrate once per week after implantation of cancer cells. Signals from primary breast tumors at 2, 4 and 6 weeks after implantation are shown. Data are mean ± SEM from 3 to 5 mice per condition. In all panels *p < 0.05 or T p < 0.05, **p < 0.01, ***p < 0.001 compared to control condition.