Regulation of iNOS function and cellular redox state by macrophage Gch1 reveals specific requirements for tetrahydrobiopterin in NRF2 activation

Abstract

Inducible nitric oxide synthase (iNOS) is a key enzyme in the macrophage inflammatory response, which is the source of nitric oxide (NO) that is potently induced in response to proinflammatory stimuli. However, the specific role of NO production, as distinct from iNOS induction, in macrophage inflammatory responses remains unproven. We have generated a novel mouse model with conditional deletion of Gch1, encoding GTP cyclohydrolase 1 (GTPCH), an essential enzyme in the biosynthesis of tetrahydrobiopterin (BH4) that is a required cofactor for iNOS NO production. Mice with a floxed Gch1 allele (Gch1(fl/fl)) were crossed with Tie2cre transgenic mice, causing Gch1 deletion in leukocytes (Gch1(fl/fl)Tie2cre). Macrophages from Gch1(fl/fl)Tie2cre mice lacked GTPCH protein and de novo biopterin biosynthesis. When activated with LPS and IFNγ, macrophages from Gch1(fl/fl)Tie2cre mice induced iNOS protein in a manner indistinguishable from wild-type controls, but produced no detectable NO, as judged by L-citrulline production, EPR spin trapping of NO, and by nitrite accumulation. Incubation of Gch1(fl/fl)Tie2cre macrophages with dihydroethidium revealed significantly increased production of superoxide in the presence of iNOS expression, and an iNOS-independent, BH4-dependent increase in other ROS species. Normal BH4 levels, nitric oxide production, and cellular redox state were restored by sepiapterin, a precursor of BH4 production by the salvage pathway, demonstrating that the effects of BH4 deficiency were reversible. Gch1(fl/fl)Tie2cre macrophages showed only minor alterations in cytokine production and normal cell migration, and minimal changes in basal gene expression. However, gene expression analysis after iNOS induction identified 78 genes that were altered between wild-type and Gch1(fl/fl)Tie2cre macrophages. Pathway analysis identified decreased NRF2 activation, with reduced induction of archetypal NRF2 genes (gclm, prdx1, gsta3, nqo1, and catalase) in BH4-deficient Gch1(fl/fl)Tie2cre macrophages. These findings identify BH4-dependent iNOS regulation and NO generation as specific requirements for NRF2-dependent responses in macrophage inflammatory activation.

Keywords: As3MT; BH4; GCH1; GTPCH; Macrophage; NOS2; Nitric oxide; Tetrahydrobiopterin; iNOS.

Graphical abstract

Fig. 1

(A) Schematic showing the targeting of the mouse Gch1 locus indicating the position of the loxP sites flanking exons 2 and 3, which encode the active site of the GTPCH enzyme. Arrows indicate the position of primers that produce the 1030-bp product from the floxed allele (solid green arrows) and the 1392-bp product following excision of the floxed DNA (solid and open green arrows). (B) Flow cytometry dot plots show expression of CD11b and F4/80 on the cell surface of wild-type (WT) and Gch1^fl/flTie2cre bone marrow-derived macrophages, indicating similar differentiation in both genotypes after 7 days of culture of bone marrow cells in macrophage differentiation media. (C) Genomic PCR shows the presence of the floxed allele in WT cells and the near complete excision of the floxed allele in Gch1^fl/flTie2cre bone marrow-derived macrophages. (D) Measurement of BH4 in the cells shows a near complete loss of BH4 in Gch1^fl/flTie2cre cells. (n=3, P<0.05).

Fig. 2

(A) Western blotting shows a lack of expression of GTPCH protein in Gch1^fl/flTie2cre macrophages, compared to wild-type (WT) macrophages that show increased expression following activation with LPS (100 ng/ml) and IFNγ (10 ng/ml) for 24 h. Both genotypes of cells showed similar induction of iNOS protein following activation. Equal protein loading was demonstrated by detection of β-tubulin, n=3 independent cell preparations. (B) Gene expression analysis of cells 24 h after stimulation showed a significant induction of Gch1 RNA; in contrast Gch1 RNA was undetectable in Gch1^fl/flTie2cre cells (n=3, P<0.05). (C) BH4 and BH2 were upregulated in wild-type macrophages following activation. All forms of biopterin were significantly reduced to near undetectable levels in Gch1^fl/flTie2cre cells. (D) Treatment of macrophages with 5 μM Sepiapterin for 24 h significantly increased BH4 content in Gch1^fl/flTie2cre cells, to levels that were not significantly different from wild-type cells (n=4, P<0.05). *P<0.05, difference by genotype; § P<0.05, difference from control. Two-way ANOVA.

Fig. 3

(A) Representative EPR spectra from wild-type (WT) and Gch1^fl/flTie2cre macrophages following activation with LPS (100 ng/ml) and IFNγ (10 ng/ml). The characteristic triple-peaked trace associated with NO-Fe(DETC)₂ signal was readily detectable in the wild-type cells, but absent in the Gch1^fl/flTie2cre cells. Quantification of the EPR signal indicated nitric oxide production only by activated wild-type cells (n=3). (B) Representative HPLC traces showing arginine to citrulline conversion by wild-type macrophages, which was absent in Gch1^fl/flTie2cre cells. Quantification of the NMA-inhibitable arginine to citrulline conversion indicated significantly reduced NOS enzyme activity in the Gch1^fl/flTie2cre macrophages. (C) Nitrite accumulation in the cell culture media over 24 h of cell stimulation was measured using the Griess assay. No nitrite production was detected in the absence of stimulation with LPS/IFNγ, but after stimulation was readily detected in wild-type supernatants, but was virtually undetectable levels in the Gch1^fl/flTie2cre cells (n=6, P<0.05). (D) The production of nitrite was significantly reduced by treatment with L-NAME (2 mM) over the 24 h stimulation in wild-type cells, and was completely blocked by incubation with iNOS inhibitor 1400W (10 nM). Production of nitrite by Gch1^fl/flTie2cre cells was significantly enhanced by coincubation of the cells with 5 μM sepiapterin during the 24 h cell stimulation. (E) Macrophage lysates from activated cells were run under nonreducing conditions to assess iNOS monomer and dimer formation. Total iNOS and GTPCH content was assessed from duplicate reduced lysate samples under standard blotting conditions. The identity of the bands obtained were confirmed using control lysates from wild-type nonactivated (u/t BMDM) and activated macrophage lysates (LPS/IFNγ BMDM) and activated iNOS^−/− lysates (iNOS^−/− BMDM).

Fig. 4

Following activation with LPS (100 ng/ml) and IFNγ (10 ng/ml) cells or control cells were pretreated with or without PEG-SOD (100 U/ml) for 30 min prior to incubation with dihydroethidium for 30 min. The production of 2-hydroxyethidium and ethidium was quantified using HPLC with fluorescence detection. The specificity of the 2-hydroxyethidium signal to superoxide was confirmed by quantification of the PEG-SOD-inhibitable signal. (A) In the absence of cell stimulation no significant difference in superoxide was detected between wild-type (WT) and Gch1^fl/flTie2cre cells. (B) Following activation Gch1^fl/flTie2cre cells produced significantly more superoxide than wild-type cells. The superoxide signals detected were not inhibitable by L-NAME (2 mM). (C) Representative HPLC trace showing the alterations in dihydroethidium oxidation in wild-type and Gch1^fl/flTie2cre cells. Notably Gch1^fl/flTie2cre cell pellets were bright red in color (ethidium) following the experiment. (D) Quantification of the ethidium signal indicated a significant increase in Gch1^fl/flTie2cre cells under both basal and stimulated conditions (E). The elevated ethidium signal was not significantly inhibited by PEG-SOD (100 U/ml) or L-NAME (2 mM). (F) Superoxide burst was induced by treatment with 2 μM PMA in the presence of dihydroethidium. The PEG-SOD-inhibitable production of superoxide was not significantly difference between genotypes. (G) Induction of the superoxide burst did not alter the elevated production of ethidium by the Gch1^fl/flTie2cre cells. (H) Pretreatment of the cells with Sepiapterin for 24 h or vitamin C for 30 min significantly reduced the elevated ethidium signal detected in the Gch1^fl/flTie2cre cells (n=3–6 per group, P<0.05). *P<0.05, difference by genotype; § P<0.05, difference from control. Two-way ANOVA.

Fig. 5

(A) Supernatants were collected from wild-type (WT) and Gch1^fl/flTie2cre macrophages 24 h after stimulation with LPS (100 ng/ml) and IFNγ (10 ng/ml). Proteome Profiler arrays, cytokine detection antibodies prespotted onto nitrocellulose filters in duplicate, were used to profile cytokine secretion. The intensity of the cytokine signal was quantified using densitometry for n=4 biological replicates per genotype. (B) Representative cytokine arrays show no significant differences between genotypes. (C) Detection of PGE2 by enzyme immunoassay indicated no alteration in production by Gch1^fl/flTie2cre cells. (D) Detection of TNFα and IL-6 (E) by ELISA indicated induction of both cytokines by stimulation, with a small but significant reduction in cytokine production by the Gch1^fl/flTie2cre cells (cytokine detection n=6 per genotype, P<0.05). (E) Representative trace from the RTCA-DP software showing macrophage migration toward chemoattractant or buffer control. Quantification of the response in the xCELLigence assay showed that migration toward either CCL5 or C5a was not significantly altered in Gch1^fl/flTie2cre macrophages (n=3 per genotype).

Fig. 6

(A) Heat map showing gene expression of genes significantly regulated by >1.5-fold passing a fdr P<0.05 across a time course of gene expression. All genes significantly regulated by genotype in at least one time point are shown, ranked by fold difference between wild-type (WT) and Gch1^fl/flTie2cre macrophages at 24 h (n=3 per genotype per time point, paired data). Data shown are the average normalized log2 expression value for each gene normalized to the expression across all samples. Green represents a positive change and red represents a negative change. (B) Confirmation of gene array result by quantitative RT-PCR of RNA from cells with or without stimulation with LPS (100 ng/ml) and IFNγ (10 ng/ml) for 24 h. (C) Confirmation of decreased expression of NRF2 target genes 24 h following stimulation as above. (RT-PCR studies n=6 per genotype, two-way ANOVA *P<0.05, ** P<0.01, ***P<0.001 for difference by genotype).