Figure 1.
HeyA8 Ovarian cell line expresses low levels of cell surface CXCR4.
Jurkat and HeyA8 cells were cultured under normal growth conditions. For flow cytometry analysis, 250,000 cells per conditions were used. For phospho-ERK assay, 100,000 cells per conditions were used. (A) Jurkat cells showed high expression levels of surface CXCR4 (solid line) when compared to control IgG (dotted line) whereas HeyA8 ovarian cell line showed low surface CXCR4 expression. (B) ELISA based phospho-ERK levels when normalized to total ERK showed increased % phospho-ERK levels in Jurkat cells when stimulated with either PMA (250 ng/mL) or SDF-1α (10 ng/mL). Phospho-ERK levels when normalized to total ERK showed no change in pERK levels in HeyA8 cells when stimulated with either PMA (250 ng/mL) or SDF-1α (10 ng/mL). (***, P <0.05 vs (-), 1-way ANOVA). Each bar graph is representative of at least 2 experiments. (C) 10 µg of protein was used for western blot analysis of breast cancer cell line MDA-MB-231 and MCF-7 where whole cell lysates and tumor extracts showed detectable CXCR4 protein levels.
Figure 2.
3D spheroid culture condition enhances CXCR4 surface expression.
Cell lines were cultured as 3D spheroids as described in materials and methods and dissociated using accutase cell dissociating buffer for 10 minutes for flow cytometry analysis. 250,000 cells were used for the flow cytometry analysis. (A) HeyA8, (B) ES-2, and (C) FaDu cells showed low levels of detectable surface CXCR4 when compared to control human IgG under normal adherent culturing conditions. When grown under 3D spheroids, all of the cell lines showed a dramatic shift in CXCR4 surface expression. (D) Tet-inducible shRNA knockdown of CXCR4 (shCXCR4) in HeyA8 cells under 3D spheroid condition showed decreased CXCR4 surface expression (Grey line) indicating the specificity of increased CXCR4 surface expression under 3D spheroids. Flow cytometry of CXCR4 expressions in 3D spheroids were pooled representative cells isolated from at least 10 individual spheroids. Each flow graph is representative of 3 (ES2 and FaDu) or 4 (HeyA8) separate experiments.
Figure 3.
Phospho-ERK level is decreased in 3D spheroids.
HeyA8 cells were cultured as 3D spheroids with different number of cells per spheroid. (A) At 5000 cells per spheroid, there was a broad shift in CXCR4 surface expression. At higher cell numbers per spheroid, the expression levels of surface CXCR4 increased. (B) Corresponding normalized pERK levels showed a strong correlation between increasing CXCR4 surface expression with decreasing phospho-ERK levels. (*, P <0.05 vs Adh, 1-way ANOVA) (***, P <0.0001 vs Adh, 1-way ANOVA). (C) At 100,000 cells per spheroid, HeyA8 cells treated with SDF-1α and PMA showed activation of phospho-ERK levels. (**, P <0.005 vs Adh, 1-way ANOVA). Each bar graph is representative of 3 experiments.
Figure 4.
Phospho-ERK inhibition increases CXCR4 surface expression.
HeyA8 cells were cultured under normal adherent conditions. 100,000 HeyA8 cells were treated with MEK1/2 inhibitor U0126 and LY294002 at 10 µM for 48 hours. Cells were harvested and analyzed by flow cytometry. (A) CXCR4 expression was increased when treated with MEK1/2 inhibitor U0126 as compared to DMSO treated (solid lines). Similar to DMSO treatment, there was no detectable surface CXCR4 expression when treated with PI3K inhibitor LY294002. (*, P <0.05 vs Adh, 1-way ANOVA) (***, P<0.0001 vs Adh, 1-way ANOVA). (B) 200,000 HeyA8 cells were plated onto two 6 well plates and treated with U0126. One plate was harvested for CXCR4 expression and the other for phospho-ERK assay. The phospho-ERK assay showed decreasing levels with increasing U0126. Flow cytometry analysis showed a dose dependent increase in CXCR4 expression. (C) MDA-MB-231 and MCF-7 cells either grown as 3D spheroids or treated with U0126 showed increase CXCR4 expression similarly to HeyA8 cells. Results are representative of 3 experiments.
Figure 5.
3D spheroids enhance NFAT3 and CXCR4 expression.
PCR array of transcription factors was performed according to manufacturer's protocol. The 3D spheroids have 100,000 cells per spheroid. (A) A list of transcriptional factors with the corresponding fold change in mRNA expression in cultured 3D spheroids relative to adherent HeyA8 cells. (B) NFAT3 mRNA showed an average of ∼ 20 fold increase in the 3D spheroid. (*, P <0.05 vs Adh, 1-way ANOVA). (C) Corresponding to the increase in NFAT3 in 3D spheroids, real-time PCR showed a 30-50-fold increase in CXCR4 mRNA in 3D spheroids when compared to adherent HeyA8 cells. (*, P <0.05 vs Adh, 1-way ANOVA). (D) Adherent and 3D spheroids western blot analysis of CXCR4 protein in HeyA8, MCF-7, and MDA-MB-231 showed higher CXCR4 levels with 3D spheroids than with adherent cells. (E) siRNA targeting NFAT3 decreased the number of CXCR4 positive cells compared to siCtr as determined by flow cytometry. Western blot analysis showed decreased NFAT3 following siNFAT3 vs siCtr and CXCR4 expression also decreased following siNFAT3. Flow cytometry of CXCR4 expression in 3D spheroids were pooled representative cells isolated from at least 10 individual spheroids. Real-time PCR results were normalized to actin. (*, P <0.05 vs siCtr, 1-way ANOVA). The PCR array results are the average of 2 separate experiments, the NFAT3 mRNA, CXCR4 mRNA, siNFAT3, and western blot experiments are representative of 3 separate experiments.
Figure 6.
NFAT3 regulates CXCR4 expression.
HeyA8 cells were cultured as a monolayer (adherent) and as 3D spheroids as described in materials and methods. HeyA8 adherent cells were treated with calcineurin-NFAT inhibitor cyclosporin A (1 µM) andFK506 (3 µM) for 72 hours. (A) Flow cytometry analysis of HeyA8 adherent cells showed basal CXCR4 surface protein expression was attenuated with the treatment of cyclosporin A as compared to no treatment group (Basal CXCR4). (***, P <0.0001 vs. CXCR4, 1-way ANOVA). (B) HeyA8 adherent cells treated with FK506 showed basal CXCR4 surface expression decreasing similar to cyclosporine A treated cells. (***, P <0.0001 vs. CXCR4, 1-way ANOVA). (C) HeyA8 3D spheroids were treated with Cyclosporin A (1 µM) for 7 days. Flow cytometry showed decreasing levels of surface CXCR4 protein expression treated with cyclosporin A. (**, P <0.05 vs. 3D spheroid, 1-way ANOVA). (D) Picture of HeyA8 3D spheroids from Fig. 6C. Cyclosporin A treated 3D spheroids compared to 3D spheroid control showed cells dissociated from the 3D spheroids and attaching to the bottom as adherent monolayer. (E)(F) Flow cytometry analysis of adherent HeyA8 and MCF-7 cells treated with ionomycin showed significant increase in Basal CXCR4 expression (**, P<0.05 vs. CXCR4, 1-way ANOVA) compared to no treatment (Basal CXCR4). Fig. 6 results are representative of 3 experiments.
Figure 7.
ERK inhibition and 3D spheroids increased NFAT transactivation.
HeyA8 cells plated at 50,000 cells in a 24 well plate for transfections. 500 ng of NFAT-Luc/Renilla mixture (40:1) were transfected per well and incubated overnight. U0126 were treated for 16-24 hours and 3D spheroids were formed for 48 hours prior to luciferase assay. (A) U0126 treated adherent HeyA8 cells showed slight significant increase of NFAT transactivation at the 20 µm but not at 10 µm. (B) 3D spheroid with +NFAT-Luc showed a greater increase in NFAT transactivation compared to adherent HeyA8 cells with +/−NFAT-Luc and 3D spheroid with no NFAT-Luc. (**, P<0.05 vs no treatment control; ****, P<0.05 vs. adherent +NFAT-Luc, 1-way ANOVA). The reporter assays are representative of 3 experiments.
Figure 8.
NFAT3 protein binds CXCR4 promoter.
HeyA8 cells were cultured as monolayer and as 3D spheroids (100,000 cells per spheroid) as described in materials and methods. (A) CXCR4 gene sequence according to GeneBank accession number AF005058. Regions with putative transcriptional factor binding consensus sequences are underlined and in bold. The block arrows indicate real-time PCR primers used for the chromatin immunoprecipitation analysis. (B) Chromatin immunoprecipitation of adherent HeyA8 cells and 3D spheroid showed significant (p <0.005) increase in relative abundance of NFAT3 and Pol II on the CXCR4 promoter. (C) Similar to 3D spheroids, adherent cells treated with 20 µM of U0126 showed significant (p <0.05) increases in relative abundance of NFAT3 and PolII on the CXCR4 promoter. Average values were derived from 2 independent IPs with qPCRs for each IP performed in duplicate (n = 4).
Figure 9.
Model for ERK-NFAT mediated regulation of CXCR4 expression.
Our findings suggest that the enhanced phospho-ERK levels in adherent cells prevent NFAT3 mediated CXCR4 up-regulation. The increase in phospho-ERK levels may directly or indirectly inhibit NFAT3 activity but when phospho-ERK levels were decreased through 3D spheroid culturing or treatment with MEK inhibitor U0126, activated NFAT3 transcriptional factor localized to the nucleus and occupied the CXCR4 promoter to activate CXCR4 expression.