Fig 1.
Genetic analysis of human breast cancer gene expression data associate GRK3 with tumor, basal subtype, and metastasis.
Microarray datasets previously analyzed in Harrell, et al., were reanalyzed to (A) compare CXCR4 and GRK expression levels between normal breast tissue and tumor tissue, (B) determine the significance of CXCR4, GRK3, and GRK5 expression levels in human breast cancer subtypes, and (C) examine the association of GRK3 expression in metastasis to liver and lymph nodes. Microarray data in (C) was grouped based on relative level of GRK3 expression (low, medium, and high). (D) TCGA database was used to compare CXCR4 and GRK expression levels between normal breast tissue and tumor tissue, as in (A). (E) TCGA database was used to determine changes in CXCR4, GRK3, and GRK5 expression levels in human breast cancer subtypes as in (B). Statistical significance determined by Students t-test (A,D), ANOVA (B,E), and Log-Rank test (C).
Fig 2.
CXCR4:GRK3 ratio correlates with the invasiveness of human breast cancer lines.
Human breast cancer lines were analyzed by quantitative real time PCR to determine the mRNA expression levels of GRK3 and CXCR4. Transcript copy number was determined using the standard curve method. Data shown are the average of two to four independent experiments. Error bars represent the SEM.
Fig 3.
Alterations in GRK3 affect migratory responses of human breast cancer lines to CXCL12.
GRK3 expression was altered by overexpression or shRNA silencing in human breast cancer lines. (A) Chemotaxis toward media or 50 nM CXCL12 of MDA-MB-231 transiently transfected with GRK3 or control plasmid was assessed by a real-time modified Transwell assay. Transfection efficiency was routinely 40%. Results shown are the mean of 5 independent experiments. (B) MDA-MB-231 invasion through Matrigel was analyzed after 24 hours and staining with Calcien-AM. The chemoinvasion index is defined as the ratio of relative fluorescence of cells migrated toward CXCL12 over media control. Results shown are the mean of 3 independent experiments. (C) Chemotaxis toward media or 500nM CXCL12 of stably transduced MDA-MB-468 cells was assessed by a real-time modified Transwell assay. Results shown are the mean of 4 independent experiments. (D) MDA-MB-468 invasion through Matrigel was analyzed using a 96-well invasion assay. Results shown are the mean of 3 independent experiments. Error bars for all data represent the SEM. Statistical significance determined by analysis of covariance (ANCOVA) linear regression model (A and C) or by a two-tailed t-test (B and D): * p < 0.05; ** p < 0.01; *** p < 0.001; N.S. not significant.
Fig 4.
GRK3 regulates CXCL12-specific CXCR4 internalization and β-arrestin recruitment.
(A) Control cells (empty plasmid) and GRK3-overexpressed MDA-MB-231 cells were treated for the indicated times with 100 nM CXCL12 at 37°C. Surface expression of CXCR4 on the surface of cells was determined by flow cytometry. Data shown are the mean of three experiments normalized to the zero time point. Error bars represent SEM. Statistical analysis was performed using a two-tailed t-test. **p < 0.01. (B) Using a TANGO arrestin-recruitment assay, HTLA cells were transfected with either CXCR4 alone or CXCR4 plus GRK3 as detailed in the Materials and Methods. Cells were plated in a 384 well plate and stimulated with CXCL12 at the indicated Molar concentrations. Luminescence was measured 24 hours post-stimulation. Error bars represent +/- SEM (n = 3). (C) TANGO results testing the CXCR4 antagonist AMD-3100. As in (B), HTLA cells were transfected with CXCR4 and GRK3 plasmids and stimulated with 10−7 M CXCL12 (one concentration point above EC50) following pre-treatment with AMD3100 at the indicated Molar concentrations. (D) TANGO results testing the CXCR4 antagonist MSX-122. HTLA cells were transfected with CXCR4 and GRK3 plasmids and stimulated with 10−7 M CXCL12 following pre-treatment with MSX-122 at the indicated Molar concentrations.
Fig 5.
Increasing the CXCR4:GRK3 ratio in 66cl4-luc mammary tumor cells increases in vivo metastasis.
(A) Balb-c mice with 66cl4-luc control transduced cells visualized by optical imaging at 6 weeks post-implantation demonstrate only primary tumor growth in contrast to mice with 66cl4-luc GRK3 deficient cells (B), which show extensive and distant metastasis. Images shown in C. and D. are representative of two independent experiments. Quantification of total (C) and metastatic (D) tumor in control versus GRK3-deficient tumors as imaged by luciferase activity. Data shown in (C) are n = 10 mice at weeks 0 and 2, n = 9 week 4, n = 7 week 6 (due to disease mortality). Data in (D) are n = 8 for Control, n = 7 for GRK3-deficient. (E) Necropsy of Balb-c mouse with control 66cl4-luc tumor cell implant showing small, encapsulated primary tumor (left panel). Necropsy of animal with GRK3-silenced 66cl4-luc tumor cell implant demonstrating larger, friable primary tumor and prominent neovascularization (right panel). All error bars represent SEM.
Table 1.
GRK3-deficient 66cl4-luc mammary tumors disseminate distant metastasis.
Fig 6.
GRK3 deficient and control 66cl4-luc cells secrete soluble CXCL12, proliferate, and undergo apoptosis similarly.
(A) Viable cell density was measured by colormetric assay (Cell Counting Kit-8, Dojindo) at indicated times and cell number estimated by standard curve (n = 3 + SEM). (B) Cells were plated in triplicate at the indicated number per well and incubated in culture conditions for 48 hours. CXCL12 levels in the supernatants were determined by sandwich ELISA and quantified by standard curve (n = 3; mean + SEM). (C) GRK3-deficient and control 66cl4-luc cells were incubated in adherent versus detached (poly-HEMA) conditions overnight and analyzed for apoptosis by flow cytometry for Annexin V and propidium iodide staining. Data is the mean of 4 independent experiments (error bars = SEM).
Fig 7.
GRK3-deficient 66cl4-luc mammary tumor cells have increased chemotaxis but not invasion.
GRK3 was silenced by lentiviral shRNA in 66cl4-luc cells to examine the effect on migration. 100,000 cells were added per well to the top of a 96-well Fluoroblok Transwell plate, and 12.5nM (100ng/ml) CXCL12 or media was in the lower chamber. Migration was monitored over 4 hours using a Fluoroskan Ascent FL plate reader. (A) Control transduced 66cl4-luc cells do not migrate toward CXCL12 (closed circle) over media control (open square) whereas GRK3 deficient 66cl4-luc cells display directional chemotaxis toward CXCL12 (closed triangle) over media control (open diamond). Statistics significance determined by analysis of covariance (ANCOVA) linear regression model: *** p < 0.001; N.S. not significant. (B) Control and GRK3 deficient 66cl4-luc cells were loaded into the upper chamber of Cultrex invasion plates in triplicate (see Materials and Methods). After addition of media alone or CXCL12 (50 or 100 nM) to the lower chamber, cells were incubated for 24 hours, stained with Calcein, and then counted for invasion into matrix (n = 3 ± SEM).
Fig 8.
Summary model of GRK3 regulation of CXCR4-driven metastasis.
Extracellular ligand stimulation of CXCL12 on G protein-coupled receptor (GPCR) CXCR4 elicits conformational changes of receptor, activation and dissociation of guanine nucleotide binding proteins, and downstream signaling for tumor cell migration (A). Negative regulation of surface receptor expression is mediated by G protein-coupled receptor kinase 3 (GRK3), which phosphorylates the carboxyl terminus of CXCR4 for desensitization, thus prompting β-arrestin recruitment for receptor internalization. The presence of GRK3 limits ligand/receptor signaling by contributing to desensitization and by reducing CXCR4 surface expression (A, inset). Upon GRK3 deficiency, CXCR4 receptor expression is enhanced thus allowing increased opportunities for CXCL12 extracellular ligand/receptor stimulation, signaling, and migration (B). The absence of GRK3 enhances ligand/receptor signaling by prolonging CXCR4 surface expression (B, inset).