Inflammatory stimuli from macrophages and cancer cells synergistically promote tumor growth and angiogenesis

Abstract

The focus of the present study was whether and how infiltrating macrophages play a role in angiogenesis and the growth of cancer cells in response to the inflammatory cytokine interleukin (IL)-1beta. Lewis lung carcinoma cells overexpressing IL-1beta grew faster and induced greater neovascularization than a low IL-1beta-expressing counterpart in vivo. When macrophages were depleted using clodronate liposomes, both neovascularization and tumor growth were reduced in the IL-1beta-expressing tumors. Co-cultivation of IL-1beta-expressing cancer cells with macrophages synergistically augmented neovascularization and the migration of vascular endothelial cells. In these co-cultures, production of the angiogenic factors vascular endothelial growth factor-A and IL-8, monocyte chemoattractant protein-1, and matrix metalloproteinase-9 were increased markedly. The production of these factors, induced by IL-1beta-stimulated lung cancer cells, was blocked by a nuclear factor (NF)-kappaB inhibitor, and also by the knockdown of p65 (NF-kappaB) and c-Jun using small interference RNA, suggesting involvement of the transcription factors NF-kappaB and AP-1. These results demonstrated that macrophages recruited into tumors by monocyte chemoattractant protein-1 and other chemokines could play a critical role in promoting tumor growth and angiogenesis, through interactions with cancer cells mediated by inflammatory stimuli.

Figure 1

Effect of clodronate liposomes (Cl₂MDP‐LIP) on Lewis lung carcinoma (LLC) tumors in vivo. (a) Inhibition of LLC/neo and LLC/IL‐1β tumor growth by clodronate liposomes. On day 0, 5 × 10⁵ LLC cells were implanted s.c. into each C57BL/6 mouse. Clodronate‐containing or control liposomes containing phosphate‐buffered saline (PBS‐LIP) were injected every 3 days from day 7. The data represent mean tumor volumes ± SD (n = 5 mice). For tumor volumes on day 20, *P = 0.012, **P < 0.001. (b) Representative hematoxylin–eosin images of LLC/neo and LLC/IL‐1β tumor sections (magnification: ×200). (c) Representative images of LLC tumor sections stained for F4/80 (magnification: ×200). (d) Numbers of infiltrating macrophages in LLC tumors, counted in four microscope fields (magnification: ×200). Data represent means ± SD. *P < 0.01, NS, not significant. (e) Representative images of LLC tumor sections stained for CD31 (magnification: ×200). (f) CD31‐positive microvessels quantified by morphometric analysis in LLC tumors in four microscope fields (magnification: ×200). Data represent means ± SD. *P < 0.05, **P < 0.001.

Figure 2

Angiogenic activity induced by coculturing Lewis lung carcinoma (LLC)/neo or LLC/IL‐1β cells with U937 cells. (a) Migration of human umbilical vein endothelial cells (HUVEC) when LLC/neo, LLC/IL‐1β, U937, or co‐cultures of LLC and U937 cells were grown in the outer chamber for 24 h in 0.5% fetal bovine serum, before seeding 1.5 × 10⁵ HUVEC per well in the inner chamber, on polycarbonate filters precoated with human plasma fibronectin. After 8 h incubation at 37°C, HUVEC that had migrated to the lower surface of the filter were counted in four fields per chamber at a magnification of ×200. Data represent means ± SD (n = 4 chambers). *P < 0.01, *P < 0.001. (b) Angiogenesis induced by LLC, U937, and combinations of LLC and U937 cells in the murine dorsal air‐sac assay. Chambers containing 1 × 10⁶ LLC or 5 × 10⁵ U937 cells or combinations of LLC and U937 cells were implanted s.c. On day 5 after the implantation, the chambers were removed, and photographs of the implantation sites were taken. (c) Quantitative analysis of newly formed vessels induced by LLC tumors in combination with macrophages on day 5. The newly formed vessels greater than 3 mm in length were counted. Data represent means ± SD (n = 4 mice). *P < 0.01, **P < 0.001.

Figure 3

Expression of angiogenic factors in lung cancer cells and macrophages in culture. Production of (a) vascular endothelial growth factor (VEGF)‐A and (b) monocyte chemoattractant protein (MCP)‐1 by Lewis lung carcinoma (LLC)/neo and LLC/interleukin (IL)‐1β cells was determined with conditioned medium from cell cultures by enzyme‐linked immunosorbent assay (ELISA). **P < 0.01 (n = 3). (c) Production of angiogenic factors was measured using angiogenesis antibody arrays in conditioned medium from U937 cells cultured with or without 1 ng/mL IL‐1β. Human (d) VEGF‐A, (e) IL‐8, (f) and MCP‐1 expression measured by ELISA from LLC/neo, LLC/IL‐1β, and U937 cells, and co‐cultures of LLC and U937 cells, collected after 48 h in medium with 0.5% fetal bovine serum. Data represent means ± SD (n = 4). NS, not significant, *P < 0.05, **P < 0.001.

Figure 4

Effects of a nuclear factor‐κB inhibitor on interleukin (IL)‐1β‐induced monocyte chemoattractant protein (MCP)‐1, vascular endothelial growth factor (VEGF)‐A, IL‐8, matrix metalloproteinase (MMP)‐2, and MMP‐9 production by U937 cells. (a) ELISA was used to measure MCP‐1, IL‐8, and VEGF‐A production by IL‐1β‐treated U937 cells in the absence or presence of the NF‐κB inhibitor dehydroxymethylepoxiquinomicin. Data represent means ± SD (n = 4). **P < 0.01, *P < 0.05. (b) Gelatin zymography showing three bands with MMP activity corresponding to 92 (proMMP‐9), 72 (proMMP‐2), and 62 kDa (MMP‐2).

Figure 5

Transcription factors in interleukin (IL)‐1β‐stimulated mouse and human lung cancer cells, and nuclear factor (NF)‐κB activation by IL‐1β in human macrophages and lung cancer cells. (a) Expression profiles of transcription factors expressed by mouse Lewis lung carcinoma (LLC)/neo and LLC/IL‐1β cells using transcription arrays. A relative increase in transcription factors in LLC/IL‐1β cells was seen when normalized to LLC/neo cells. (b) Expression profiles of transcription factors expressed by human lung B203L cells in the absence or presence of 1 ng/mL IL‐1β. A relative increase in transcription factors in IL‐1β‐treated cells was seen when normalized to untreated cells. (c) Effect of IL‐1β on binding to the NF‐κB consensus fragment in human U937 cells and the human lung cancer cell lines A549 and B203L. After incubation for the times indicated with 1 ng/mL IL‐1β, nuclear extracts were incubated with ³²P‐labeled NF‐κB consensus fragment and subjected to gel electrophoresis with increasing molar excesses of unlabeled AP‐1, Sp1, or NF‐κB DNA fragments. The position of the retarded NF‐κB protein complex is indicated on the left.

Figure 6

Effect of nuclear factor (NF)‐κB/p65 and AP‐1/c‐Jun small interference RNA (siRNA) on interleukin (IL)‐1β‐induced IL‐8 expression of vascular endothelial growth factor (VEGF)‐A, IL‐8, and monocyte chemoattractant protein (MCP)‐1. (a) Two human lung cancer cell lines, A549 and B203L, were treated with 50 nM NF‐κB/p65 siRNA, 50 nM AP‐1/Jun siRNA, or 50 nM control siRNA for 48 h. Total cellular lysates (for NF‐κB) or nuclear fractions (for c‐Jun) were analyzed by western blot analysis. (b) After treatment with siRNA, cells were incubated for a further 24 h with IL‐1β. Conditioned medium was then collected, and VEGF‐A, IL‐8, and MCP‐1 were measured by ELISA. Data are means ± SD (n = 3). **P < 0.001 and *P < 0.05, compared with control siRNA. None, lipofectamine alone; C, control siRNA; Jun, c‐Jun siRNA.