Production of vascular endothelial growth factor by murine macrophages: regulation by hypoxia, lactate, and the inducible nitric oxide synthase pathway

Abstract

Murine thioglycolate-induced peritoneal macrophages (MPMs) and the murine RAW264.7 macrophage-like cell line (RAW cells) constitutively produce vascular endothelial growth factor (VEGF). VEGF production is increased under hypoxic conditions or after cell activation with interferon-gamma (IFNgamma) and endotoxin (lipopolysaccharide, LPS). In contrast, tumor necrosis factor-alpha is produced only by IFNgamma/LPS-activated cells. Lactate (25 mmol/L) does not increase VEGF production by these cells. However, hypoxia, lactate, and IFNgamma/LPS-activated MPMs express angiogenic activity, whereas normoxic, nonactivated MPMs do not. Lack of angiogenic activity is not due to an antiangiogenic factor(s) in the medium of these cells. Angiogenic activity produced by hypoxia and lactate-treated MPMs is neutralized by anti-VEGF antibody, which also neutralizes most of the angiogenic activity produced by IFNgamma/LPS-activated MPMs. The inducible nitric oxide synthase inhibitors Ng-nitro-L-arginine-methyl ester (1.5 mmol/L) and aminoguanidine (1 mmol/L) block production of angiogenic activity by MPMs and RAW cells. In RAW cells, Ng-nitro-L-arginine-methyl ester and AG block IFNgamma/LPS-activated, but not constitutive, VEGF production, whereas in MPMs, neither constitutive nor IFNgamma/LPS-activated VEGF synthesis is affected. Synthesis of tumor necrosis factor-alpha is also unaffected. In contrast to normoxic, nonactivated MPMs, inducible nitric oxide synthase-inhibited, IFNgamma/LPS-activated MPMs produce an antiangiogenic factor(s). We conclude that VEGF is a major contributor to macrophage-derived angiogenic activity, and that activation by hypoxia, lactate, or IFNgamma/LPS switches macrophage-derived VEGF from a nonangiogenic to an angiogenic state. This switch may involve a posttranslational modification of VEGF, possibly by the process of ADP-ribosylation. ADP-ribosylation by MPM cytosolic extracts or by cholera toxin switches rVEGF165 from an angiogenic to a nonangiogenic state. In IFNgamma/LPS-activated MPMs, the inducible nitric oxide synthase-dependent pathway also regulates the expression of an antiangiogenic factor(s) that antagonizes the bioactivity of VEGF and provides an additional regulatory pathway controlling the angiogenic phenotype of macrophages.

Figure 1.

Nitrite production by MPMs. Cells were incubated in DMEM-1% FCS with or without sodium lactate (25 mmol/L), IFNγ (100 U/ml) and LPS (100 ng/ml), l-NAME (1.5 mmol/L), or AG (1 mmol/L), as indicated. Media were harvested 18 and 48 hours after challenge with IFNγ/LPS. Results are means ± SD of triplicate determinations in a typical experiment. Similar results were found in at least three separate experiments.

Figure 2.

VEGF production by RAW264.7 cells (A) and MPMs (B). Cells were incubated in DMEM-1% FCS with or without sodium lactate (25 mmol/L), IFNγ (100 U/ml) and LPS (100 ng/ml), l-NAME (1.5 mmol/L), or AG (1 mmol/L), as indicated. Media were harvested 18 and 48 hours after challenge with IFNγ/LPS. Results are means ± SD of triplicate determinations in a typical experiment. Similar results were found in at least three separate experiments.

Figure 3.

Competitive RT-PCR analysis of VEGF mRNA levels in control (nonstimulated) MPMs 24 hours after plating. Varying amounts of total RNA (1 to 200 ng) isolated from MPMs were reverse transcribed and amplified by PCR through 25 cycles in the presence of a VEGF RNA minigene (2.5 pg) that amplifies using the same primers as the native VEGF mRNA, as described in Materials and Methods. The RNA minigene yields an amplified PCR product of 293 bp, and the native VEGF mRNA yields a 362-bp fragment. The amount of total RNA that yields an amplification band of the same intensity as the minigene is determined from these analyses.

Figure 4.

RT-PCR analysis of VEGF isoforms produced by IFNγ/LPS-activated MPMs with or without AG treatment. Total RNA isolated from MPMs was reverse transcribed and amplified by PCR, as described in Materials and Methods. PCR primers were located in exons 3 and 8, resulting in the amplification of three PCR products corresponding to 652, 580, and 448 bp.

Figure 5.

TNFα production by MPMs. Cells were incubated in DMEM-1% FCS with or without sodium lactate (25 mmol/L), IFNγ (100 U/ml) and LPS (100 ng/ml), l-NAME (1.5 mmol/L), or AG (1 mmol/L), as indicated. Media were harvested 8, 24, and 48 hours after challenge with IFNγ/LPS. Results are means ± SD of triplicate determinations in a typical experiment. Similar results were found in at least three separate experiments.

Figure 6.

ADP-ribosylation of rVEGF₁₆₅ by bacterial toxins and by macrophage cytosolic extract. Lane A: rVEGF (500 ng) was incubated with macrophage cytosolic extract (see Materials and Methods) in the presence of [³²P]NAD⁺. The total labeling reaction was analyzed on the 0.1% SDS-15% PAGE gel. Lane B: The rVEGF₁₆₅-macrophage cytosolic extract labeling mixture was immunoprecipitated with anti-VEGF antibody, and the immunoprecipitated VEGF was analyzed by SDS-PAGE. A dominant ³²P-labeled band migrating in the same position as rVEGF₁₆₅ (determined by Western analysis of the same blot) is indicated. Lane C: rVEGF₁₆₅ was incubated with cholera toxin subunit A and [³²P]NAD⁺ as described in Materials and Methods. Lane D: Cholera toxin was incubated with [³²P]NAD⁺ in the absence of rVEGF₁₆₅. Lane E: rVEGF₁₆₅ was incubated with pertussis toxin and [³²P]NAD⁺, as described in Materials and Methods.