Inorganic phosphate induces cancer cell mediated angiogenesis dependent on forkhead box protein C2 (FOXC2) regulated osteopontin expression

Abstract

Recent studies in both rodents and humans suggest that elevated serum phosphorus, in the context of normal renal function, potentiates, or exacerbates pathologies associates with cardiovascular disease, bone metabolism, and cancer. Our recent microarray studies identified the potent stimulation of pro-angiogenic genes such as forkhead box protein C2 (FOXC2), osteopontin, and Vegfα, among others in response to elevated inorganic phosphate (Pi). Increased angiogenesis and neovascularization are important events in tumor growth and the progression to malignancy and FOXC2 has recently been identified as a potential transcriptional regulator of these processes. In this study we addressed the possibility that a high Pi environment would increase the angiogenic potential of cancer cells through a mechanism requiring FOXC2. Our studies utilized lung and breast cancer cell lines in combination with the human umbilical vascular endothelial cell (HUVEC) vessel formation model to better understand the mechanism(s) by which a high Pi environment might alter cancer progression. Exposure of cancer cells to elevated Pi stimulated expression of FOXC2 and conditioned medium from the Pi-stimulated cancer cells stimulated migration and tube formation in the HUVEC model. Mechanistically, we define the requirement of FOXC2 for Pi-induced osteopontin (OPN) expression and secretion from cancer cells as necessary for the angiogenic response. These studies reveal for the first time that cancer cells grown in a high Pi environment promote migration of endothelial cells and tube formation and in so doing identify a novel potential therapeutic target to reduce tumor progression.

Keywords: FOXC2; angiogenesis; inorganic phosphate; migration; osteopontin.

Figure 1. Pi increases FOXC2, OPN, and Vegfα mRNA levels in A549 and MDA-MB-231 cells

(A) A549 and (B) MDA-MB-231 were cultured in the normal growth medium (10%FBS and 1mM Pi) and Pi added to a final concentration of 1, 3, or 5mM for 6 days (A549) and 3 days (MDA-MB-231). Cells were harvested for total RNA isolation and quantitative real-time PCR. 18S rRNA was used for normalization. *P<0.05 (Student’s t-test). Results expressed as mean±SD (n=3).

Figure 2. Conditioned medium from Pi-treated cancer cells increases HUVEC migration and tube formation

A549 and MDA-MB-231 cells were treated with Pi (1, 3, or 5mM final) for 6 or 3 days respectively, at which point medium was replaced with serum free-phenol free medium for 24h and collected as conditioned medium (CM). (A) CM was used to determine the effects of HUVEC migration using a two chamber porous membrane “Boyden” assay. Cells in the lower chamber were counted after 20h. *P<0.05 compared with 1mM Pi treatment. N=3–5. (B) Cancer cells were treated with Pi (1 or 5mM final) and the medium was replaced with serum free-phenol free, collected, and concentrated by filter centrifugation and added to a HUVEC tube formation assay. The resulting cells were photographed and results quantified by Wimasis GmbH (Munich, Germany) software (C). *P<0.05 (Student’s t-test). Results expressed as mean±SD (n=3).

Figure 3. Knockdown of FOXC2 blocks Pi-induced conditioned medium mediated HUVEC migration and Pi-induced OPN expression

(A) A549 and MDA-MB-231 cells were transfected with two different RNAi’s targeting FOXC2 (RNAi-“1” and RNAi-“3”), or siControl (Ctrl) followed by stimulation with Pi or not (1 and 5mM final). Conditioned medium was collected and used in the two-chamber HUVEC migration assay. Results expressed as mean±SD (n=3–5). * P<0.05, compared with 1mM Pi treatment. (B) A549 and MDA-MB-231 were transfected with two different siRNA’s targeting FOXC2 (1 and 3), or siControl (Ctrl) and left untreated (1mM Pi) or treated with 4mM Pi (5mM final) and medium collected and secreted OPN protein levels measured by ELISA. (C) HCT116 cells were transfected by pcDNA3.1 vector or pcDNA3.1-FOXC2 expression plasmid and the cells treated by 4mM of Pi for 2 days (1 and 5mM final). The harvested samples were analyzed by western blotting. (D) HCT116 cells were co-transfected with a 2 kb OPN promoter reporter and either pcDNA3.1 empty vector plasmid or the pcDNA3.1-FOXC2 expression plasmid. The luciferase activities were measured after 24h. Results expressed as mean±SD (n=4). (E) MDA-MB-231 cells were co-transfected with the OPN reporter and RNAi targeting FOXC2 and treated with 4mM Pi as indicated. Luciferase activity was measured after 24 hours. Results expressed as mean±SD (n=3). *P<0.05, compared with control plasmid and no added Pi (1mM Pi); # P<0.05, compared to control plasmid+Pi (5mM).

Figure 4. Antibody neutralization of OPN blocks Pi-conditioned medium mediated HUVEC migration

(A) A549 and MDA-MB-231 cells were cultured in the growth medium (1mM Pi) and treated with 4mM of Pi (5mM final) for 6 or 3 days respectively. The medium was changed to serum free-phenol free for 24h and collected as conditioned medium (CM). The conditioned medium was incubated with antibodies to OPN or Vegfα for 1 hour prior to addition to the two-chamber HUVEC migration assay. Cells in the lower chamber were counted after 20h. Results expressed as mean±SD (n=3–5). (*P<0.05, compared with 1mM Pi treatment or control; # P<0.05, compared with +Pi treatment; Student’s t-test). (B) A549 CM (untreated) was spiked with recombinant OPN (BSA was used in the control) and two-chamber HUVEC migration assay performed. Results expressed as mean±SD (n=3). (*P<0.05, compared with control (0μg/ml); Student’s t-test).