Homotypic gap junctional communication associated with metastasis suppression increases with PKA activity and is unaffected by PI3K inhibition

Abstract

Loss of gap junctional intercellular communication (GJIC) between cancer cells is a common characteristic of malignant transformation. This communication is mediated by connexin proteins that make up the functional units of gap junctions. Connexins are highly regulated at the protein level and phosphorylation events play a key role in their trafficking and degradation. The metastasis suppressor breast cancer metastasis suppressor 1 (BRMS1) upregulates GJIC and decreases phosphoinositide-3-kinase (PI3K) signaling. On the basis of these observations, we set out to determine whether there was a link between PI3K and GJIC in tumorigenic and metastatic cell lines. Treatment of cells with the well-known PI3K inhibitor LY294002, and its structural analogue LY303511, which does not inhibit PI3K, increased homotypic GJIC; however, we found the effect to be independent of PI3K/AKT inhibition. We show in multiple cancer cell lines of varying metastatic capability that GJIC can be restored without enforced expression of a connexin gene. In addition, while levels of connexin 43 remained unchanged, its relocalization from the cytosol to the plasma membrane was observed. Both LY294002 and LY303511 increased the activity of protein kinase A (PKA). Moreover, PKA blockade by the small molecule inhibitor H89 decreased the LY294002/LY303511-mediated increase in GJIC. Collectively, our findings show a connection between PKA activity and GJIC mediated by PI3K-independent mechanisms of LY294002 and LY303511. Manipulation of these signaling pathways could prove useful for antimetastatic therapy.

Figure 1

LY294002 increases GJIC. (A) Cell lines treated with LY294002 (10 μM) exhibit higher levels of calcein dye transfer from donor to acceptor cells in a fluorescence dye transfer assay. (B) Seven cancer cell lines exhibited a significant increase in calcein spread. Data shown are represented as fold increase in dye transfer compared to non-treated (NT) cells. Arrows highlight donor cells. [* = p<0.05, ** = p<0.01]

Figure 2

Inhibition of AKT is not responsible for increases in GJIC. (A) Nanomolar concentrations of Wortmannin were used to inhibit AKT phosphorylation at Ser 473. (B) Quantified results of calcein dye transfer assays showing no increases in GJIC using Wortmannin based phospho-AKT inhibition, while rather slight decreases were observed. [* = p<0.05, ** = p<0.01]

Figure 3

Direct inhibition of AKT does not increase GJIC. (A) The AKT inhibitor, AKTinbVII, was used to decrease phospho-AKT levels directly (10 μM), without a corresponding increase in GJIC (B). Similar to AKT inhibition with Wortmannin, slight decreases in GJIC were observed.

Figure 4

Both LY303511 and LY294002 increase GJIC. LY303511 treatment (10 μM) increases calcein transfer in MDA-MB-231, MDA-MB-435 and C8161.9 cells (A), despite not potently inhibiting PI3K (B). [* = p<0.05, ** = p<0.01]

Figure 5

Cx43 localizes to the plasma membrane after LY294002 or LY303511 treatment. (A) Representative immunoblot analysis from three independent experiments of Cx43 in C8161.9 and MDA-MB-231 6 hr post treatment with 10μM LY294002 and LY303511. (B) Quantification of Cx43 immunoblot analyses (n=3) [** = p<0.01]. (C) Exposure of C8161.9 or MDA-MB-231 cells to LY294002 (10 μM) or LY303511 (10 μM) for 6 hr resulted in substantially more plasma membrane localization of Cx43 than untreated (NT) cells. Arrows highlight Cx43 plaques between cells.

Figure 6

PKA activity is increased by LY294002 and LY303511 and required for enhanced GJIC. (A) Immunofluorescence analysis of CREB phosphorylated at Ser 133 in MDA-MB-231 revealed an increase in phosphorylation after 2 hr treatment with LY294002 (10 μM) and LY303511 (10 μM). (B) Analysis of phosphorylation of CREB at Ser 133 in nuclear extracts from MDA-MB-231 treated with LY294002 (10 μM) showed phosphorylation of CREB within 15 min. Treatment with the PKA agonist 8-BR-cAMP (0.1 mM, 0.5 mM, 1.0 mM) increased GJIC in a dose dependent manner in MDA-MB-231 (C), while inhibition of PKA activity by H89 (10 μM) when used in co-treatments with LY294002 (10 μM) inhibited GJIC compared to LY294002 alone (D, calcein) (E, quantified). [*=p<0.05]. Pretreatment of cells with the adenylate cyclase inhibitors 2′5′-dideoxyadenosine (F) or SQ 22,536 (G) (10 μM,75 μM,100 μM) did not reduce the ability of LY294002 (10 μM) to increase GJIC.