Toll-like receptor 3 regulates angiogenesis and apoptosis in prostate cancer cell lines through hypoxia-inducible factor 1 alpha

Abstract

Toll-like receptors (TLRs) recognize microbial/viral-derived components that trigger innate immune response and conflicting data implicate TLR agonists in cancer, either as protumor or antitumor agents. We previously demonstrated that TLR3 activation mediated by its agonist poly(I:C) induces antitumor signaling, leading to apoptosis of prostate cancer cells LNCaP and PC3 with much more efficiency in the former than in the second more aggressive line. The transcription factor hypoxia-inducible factor 1 (HIF-1) regulates several cellular processes, including apoptosis, in response to hypoxia and to other stimuli also in normoxic conditions. Here we describe a novel protumor machinery triggered by TLR3 activation in PC3 cells consisting of increased expression of the specific I.3 isoform of HIF-1 alpha and nuclear accumulation of HIF-1 complex in normoxia, resulting in reduced apoptosis and in secretion of functional vascular endothelial growth factor (VEGF). Moreover, we report that, in the less aggressive LNCaP cells, TLR3 activation fails to induce nuclear accumulation of HIF-1 alpha. However, the transfection of I.3 isoform of hif-1 alpha in LNCaP cells allows poly(I:C)-induced HIF-1 activation, resulting in apoptosis protection and VEGF secretion. Altogether, our findings demonstrate that differences in the basal level of HIF-1 alpha expression in different prostate cancer cell lines underlie their differential response to TLR3 activation, suggesting a correlation between different stages of malignancy, hypoxic gene expression, and beneficial responsiveness to TLR agonists.

Figure 1

Poly(I:C) treatment induces time- and dose-dependent HIF-1α nuclear accumulation in PC3 cells Western blot analysis. (A) Cells were treated with poly(I:C) (25 µg/ml) for 16 hours, and whole-cell extracts were analyzed for HIF-1α. (B) Cells were treated with the indicated doses of poly(I:C) for 6 hours, and nuclear extracts were analyzed for HIF-1α. (C) Nuclear extracts from cells treated with 5 µg/ml poly(I:C) for the indicated times. β-Actin was used as control for equalamounts (30 µg/lane) of protein loaded. Data shown are typical of three separate experiments with similar results.

Figure 2

Poly(I:C) stimulation specifically induces the I.3 isoform of HIF-1α in PC3 cells. (A) Cells untreated or treated with 5 µg/ml poly(I:C) for the indicated times were analyzed for hif-1α mRNA levels by qRT-PCR. Each point represents the mean of triplicate samples from three independent experiments with SD as error bars, *P ≤ .05. (B) Cells were treated with 5 µg/ml poly(I:C) for 5 hours 45 minutes and then with CHX (2 µg/ml) or Act D (2 µg/ml) for 15 minutes (upper panel). Cells were treated with 5 µg/ml poly(I:C) for 20 hours 30 minutes with or without 5 µg/ml MG-132 for the last 3 hours in the presence or absence of CHX (2 µg/ml) for 30 minutes (lower panel). Treatment with 100 µM CoCl₂ for 3 hours was used as hypoxic stimulus. HIF-1α protein was detected by Western blot on whole-cell lysates. β-Actin was used as loading control. (C) qRT-PCR with specific primers was used to evaluate the different hif-1α isoforms induced after poly(I:C) stimulation. Data shown in panel B are typical of three separate experiments with similar results. The histogramin panel C represents the mean of triplicate samples from three independent experiments with SD as error bars, *P ≤ .05, **P = .01.

Figure 3

Poly(I:C)-induced HIF-1α accumulation is TLR3-dependent. (A) PC3 cells stably transfected with a plasmid containing the dominant-negative form of TLR3 (PC3-TLR3-DN) or with a plasmid containing only the resistance to puromycin (PC3-PURO) were subjected to total DNA extraction, and the presence of the TLR3-DN plasmid was assessed using PCR with specific primers. Nontransfected PC3 cells and the plasmid used for the transfection (pZERO-hTLR3) were used as negative and positive controls, respectively. β-Actin was used as a control for equal amounts of DNA loaded. (B and C) Both stably transfected cell lines were treated for 6 hours with 5 µg/ml poly(I:C), and the increase of hif-1α mRNA and nuclear protein accumulation were evaluated by qRT-PCR (B) and Western blot analysis (C). β-Actin was used to normalize both qRT-PCR and Western blot. Data shown in panels A and C are typical of three separate experiments with similar results. The histogramin panel B represent the mean of triplicate samples from three independent experiments with SD as error bars, *P = .05.

Figure 4

Poly(I:C) increases the production of active VEGF through HIF-1α. (A and B) vegf mRNA and protein levels of PC3 cells were evaluated after 5 µg/ml poly(I:C) stimulation for the indicated time points using qRT-PCR (A) and ELISA (B), respectively. (C and D) Formation of capillary-like structures. Representative phase-contrast microphotographs of capillary-like structures (C) and quantitative evaluation of their morphogenesis in HUVEC cultures (D). HUVECs were exposed to supernatants of PC3 cells untreated (ctr) or treated with poly(I:C) for 48 hours in the presence or absence of VEGF neutralizing antibodies (anti-VEGF Ab). HUVEC exposed to serum-free medium (DMEM) with or without poly(I:C) for 48 hours were used as negative controls. (E) Cells were transiently transfected with a vector encoding an RNA interfering hif-1α (int. hif-1α) or with a vector encoding a scramble RNA sequence (int. scramble), then treated with 5 µg/ml poly(I:C) for 6 hours and analyzed for total HIF-1α protein level using Western blot. β-Actin was used as a control for equal amounts of protein loaded. (F) Transfected cells were treated with poly(I:C) for 24 hours, and vegf mRNA expression was evaluated by qRT-PCR. The histograms in A, B, D, and F represent the mean of triplicate samples from three independent experiments with SD as error bars, *P ≤. 05, **P ≤. 01. Data shown in panel E are typical of three separate experiments with similar results.

Figure 5

Overexpression of I.3 hif-1α in LNCaP cells determines poly(I:C)-induced VEGF production. (A) LNCaP cells were treated for the indicated times with 25 µg/ml poly(I:C), and Western blot analysis of HIF-1α was performed on nuclear extracts. Cells treated with 100 µM CoCl₂ for 3 hours were used as positive control for HIF-1α accumulation. β-Actin was used as a loading control. (B, C, and D) LNCaP cells were transiently transfected with a plasmid containing the I.3 isoform of HIF-1α (hif-1α I.3) or with a control plasmid (pcDNA3). Cells were then treated with 5 µg/ml poly(I:C) for 6 hours, and HIF-1α protein levels were analyzed using Western blot (B). Analysis of the I.1 and I.3 isoform of hif-1α mRNA was performed by using qRT-PCR (C). ELISA was used to analyze VEGF concentration in the CM of untransfected or transfected cells with the indicated plasmids and then treated or left untreated with poly(I:C) (D), the histogram represents the mean ± SEM of three independent experiments. **P ≤ .01. Data shown in panels A and B are typical experiments repeated three times with similar results. Histograms in panels C and D represent the mean of triplicate samples from three independent experiments with SD as error bars, *P ≤ .05, **P ≤ .01.

Figure 6

Up-regulation of HIF-1α inhibits poly(I:C)-induced apoptosis in LNCaP and PC3 cells. PC3 cells transfected with scramble or hif-1α interfering plasmids, treated or not with 5 µg/ml poly(I:C) for 24 hours, were analyzed for the apoptotic rate using PI staining (A) and caspase-3 activation (B). LNCaP cells overexpressing I.3 isoform of hif-1α were treated with 5 µg/ml poly(I:C) for 24 hours, and PI staining (C) and caspase-3 activation (D) were performed to evaluate the apoptotic rate. Histograms in panels A and C represent the mean of triplicate samples from three independent experiments with SD as error bars, **P ≤ .01. Data shown in panels B and D are typical experiments repeated three times with similar results.