Cystathionine γ-lyase exacerbates Helicobacter pylori immunopathogenesis by promoting macrophage metabolic remodeling and activation

Monocyte-derived macrophages are a key component of the innate immune response, providing a wide range of functions, including response to pathogens, antigen presentation, immune regulation, and wound repair (1–3). In response to bacterial pathogens such as Helicobacter pylori, macrophages exhibit an M1-like phenotype that is characterized by expression of genes encoding chemokines and cytokines that orchestrate the recruitment and activation of other innate cells as well as a robust adaptive T cell response that perpetuates inflammation (4–10). Proinflammatory macrophages also express enzymes that produce effector molecules such as NO and ROS, which help limit the pathophysiology of infection (8, 9, 11–13). However, these events can contribute to dysregulation of cellular homeostasis and tissue damage. A unique feature of macrophage biology is the ability to maintain plasticity to regulate the proinflammatory response and/or repair damage in responding to environmental cues and metabolic remodeling (14). Therefore, understanding the molecular mechanisms that dictate and regulate macrophage activation has therapeutic potential that is still largely unexplored.

Cystathionine γ-lyase (CTH; also known as CSE) breaks down cystathionine into cysteine, α-ketobutyrate, and ammonia as the last step of the mammalian reverse transsulfuration pathway (RTP) (15, 16). Cystathionine is generated by cystathionine β-synthase (CBS) from homocysteine and serine (15). Homocysteine is directly downstream of the demethylation of S-adenosylmethionine (SAM) to S-adenosylhomocysteine, an important step in DNA, histone, and protein methylation (Figure 1) (17). Another major role of SAM, a product of methionine metabolism, is the biosynthesis of the polyamines spermidine and spermine by spermidine synthase and spermine synthase, respectively (18). Both enzymes require the decarboxylated form of SAM (dcSAM), which is synthesized by the enzyme S-adenosylmethionine decarboxylase (SAMDC), to donate an amino-propyl group (Figure 1) (18).

Figure 1

SAM metabolism. Schematic of key metabolites and enzymes (red) downstream of SAM. cysta, cystathionine; HCY, homocysteine; SAH, S-adenosyl-l-homocysteine.

H. pylori is a Gram-negative bacterium that is estimated to colonize approximately 4.4 billion people, making it 1 of the most common bacterial pathogens in humans (19). H. pylori is the strongest risk factor for development of gastric adenocarcinoma, the fourth leading cause of cancer mortality in the most recent 2020 GLOBOCAN data (20), with 89% of all gastric cancer cases associated with evidence of H. pylori colonization (21, 22). Although complex regimens of antibiotic treatment are available, the prevalence of antibiotic-resistant strains is on the rise (23–25). In addition, the benefit of H. pylori eradication in cancer prevention relies on low rates of reinfection and sustained negative H. pylori colonization status (26, 27). These issues are compounded by the lack of clinical and biological markers to predict patients at risk for precancerous lesions (28). H. pylori elicits a vigorous, yet generally ineffective, innate and adaptive mucosal immune response, resulting in a chronically active inflammatory state that contributes to the progression from gastritis to adenocarcinoma (29–32). Thus, new strategies are needed to target the host immune response to reduce antibiotic use and circumvent the risk of antibiotic resistance.

We recently reported that H. pylori upregulates expression of CTH in mouse and human gastric macrophages (Gmacs) and reduces spermidine and spermine synthesis by limiting SAM availability in macrophages in vitro (33). We now present evidence that genetic deletion of Cth results in reduced gastritis, suppression of both innate and adaptive immune markers, and downregulation of metabolic pathways, including those involved in cellular respiration with H. pylori or with classical macrophage stimuli. These findings were independent of changes in polyamine metabolism. Additionally, there were changes in adaptive immune responses in the gastric mucosa with Cth deletion, but effects on T cells appeared to be independent of any effects of CTH within T cells. Taken together, our data suggest that modulation of CTH activity affects the innate immune response of macrophages and pathogenic gastric inflammation, thus representing a potential target for therapeutic intervention in H. pylori pathogenesis.

Deletion of Cth reduces gastritis in acute models of H. pylori pathogenesis. We determined that CTH expression was localized to Gmacs in H. pylori–infected human stomach tissues and persists at an increased level during progression along the histological cascade toward carcinogenesis (33). Thus, we used Cth-deficient mice (34) to determine the role of CTH in vivo during H. pylori infection. First, we confirmed that Cth mRNA expression is eliminated in the gastric tissues of Cth^–/– mice with and without H. pylori infection (Figure 2A). Immunofluorescence of infected gastric tissues from WT mice showed colocalization of CTH with CD68⁺ Gmacs (Figure 2B); deletion of Cth led to loss of H. pylori–stimulated CTH protein in CD68⁺ Gmacs (Figure 2B). Then, to assess the effect of CTH in vivo, we used well-established models of H. pylori infection: C57BL/6 WT and Cth^–/– mice were infected with H. pylori strain PMSS1 for 4 or 8 weeks (12, 35). Animals lacking Cth demonstrated significantly decreased histologic gastritis at 4 weeks (Supplemental Figure 1, A and B; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.155338DS1) and 8 weeks (Figure 2, C and D), and no difference in histology in uninfected mice (Supplemental Figure 1A). Consistent with our previous studies in other mutant mice (11, 12, 36, 37), decreased levels of gastritis were associated with increased levels of gastric colonization by H. pylori in Cth^–/– mice (Figure 2, E and F, and Supplemental Figure 1C). Expression of the genes encoding for (a) the proinflammatory marker TNF-α, (b) the antiinflammatory marker arginase 1 (ARGI1), and (c) the Th1 cytokine IFN-γ, were all significantly downregulated in Cth^–/– mice infected with H. pylori PMSS1 for 4 weeks when compared with infected WT mice (Supplemental Figure 1, D and E).

Figure 2

Deletion of Cth reduces gastritis in an acute model of H. pylori pathogenesis. WT and Cth^–/– mice were infected or not with H. pylori PMSS1 for 8 weeks. (A) CTH mRNA expression in gastric tissues of WT and Cth^–/– mice infected or not with H. pylori; n = 3–7 mice per genotype. (B) Representative immunofluorescence images of CTH (green) colocalized (yellow) with the macrophage marker CD68 (red) from WT and Cth^–/– mice infected or not with H. pylori; n = 3 mice per genotype. DAPI (blue). (C) H&E images from infected mice; arrowheads highlight inflammatory cells. (D) Histologic gastritis scores. Each symbol is a different mouse; data are pooled from 2 independent experiments. (E) H. pylori colonization in gastric tissues from D. (F) Correlation between gastritis in D and H. pylori colonization in E. All values are reported as mean ± SEM. Statistical analyses, where shown, are: (A and D) 1-way ANOVA with Kruskal-Wallis test, followed by a Mann-Whitney U test; (E) Student’s t test; (F) correlation and significance determined by Pearson’s product-moment correlation test. *P < 0.05, **P < 0.01, ****P < 0.0001. (B and C) Scale bars: 50 μm. Ctrl, control.

CTH mediates induction of immune response–specific gene sets. Because H. pylori infection produces a chronic inflammatory state in humans, we used an established mouse model of chronic H. pylori infection to assess differential gene expression, in which mice were infected with H. pylori strain SS1 for 16 weeks (11, 12). Similar to the acute H. pylori infection models, Cth^–/– mice exhibited significantly decreased histologic gastritis (Figure 3, A and B) and increased H. pylori colonization (Figure 3C). Furthermore, the expression of the genes encoding for the proinflammatory genes TNF-α, chemokine (C-X-C motif) ligand 1 (CXCL1; the murine homolog of the neutrophil chemokine IL-8), IL-12β, and for the Th17 marker IL-17 and the Th1 marker IFN-γ were downregulated in Cth^–/– mice chronically infected with H. pylori SS1 for 16 weeks compared with infected WT animals (Figure 3D). Similarly, the transcript of the antiinflammatory marker Il10 was also higher in WT mice versus Cth^–/– mice during H. pylori infection (Figure 3E). CTH expression was not induced in the Ly6G⁺ neutrophils (Supplemental Figure 2) or CD3⁺ lymphocytes (Supplemental Figure 3A) found in the gastric tissues of infected WT mice. Although WT CD4⁺ splenocytes express increased levels of Cth mRNA with costimulation of CD3 and CD28 (Supplemental Figure 3B), Cth^–/– splenocytes did not exhibit differences in cytokine or T cell transcription factor expression or proliferation compared with WT splenocytes (Supplemental Figure 3, C–E). Thus, CTH is not upregulated in either gastric neutrophils or T cells with H. pylori infection, and CTH does not have a primary effect on T cell function. However, our data demonstrate that CTH expression contributes to the gastric immune response to H. pylori and affects the macrophage innate response and, hence, the adaptive T cell response.

Figure 3

Deletion of Cth reduces gastritis in a chronic model of H. pylori pathogenesis. WT and Cth^–/– mice were infected or not with H. pylori SS1 for 16 weeks. (A) H&E images from infected mice; arrowheads highlight inflammatory cells. (B) Histologic gastritis scores. Each symbol is a different mouse; data pooled from 2 independent experiments. (C) H. pylori colonization in gastric tissues from B. (Tissues from 1 of the experiments was used for RNA-Seq). mRNA expression of (D) proinflammatory and (E) antiinflammatory markers in gastric tissues; n = 5–6 H. pylori–infected mice per genotype. (B–E) All values are reported as mean ± SEM. Statistical analyses, where shown: (B, D, and E) 1-way ANOVA with Kruskal-Wallis test, followed by a Mann-Whitney U test; (C) Student’s t test. *P < 0.05. (A) Scale bars: 50 μm. Ctrl, control.

To evaluate global changes in the transcriptome of Gmacs, we isolated F4/80⁺ cells (Supplemental Figure 4A) from the gastric tissues of WT and Cth^–/– mice infected with H. pylori SS1 for 16 weeks and performed RNA-Seq. Hierarchical clustering and heatmap analysis highlighted differential transcript abundance between genotype and infection status (Supplemental Figure 4B). Principal component analysis demonstrated a distinct distribution of transcript profiles from WT and Cth^–/– Gmacs that was dependent on H. pylori infection (Supplemental Figure 4C). Overall, there were 24,781 transcripts identified that were present in 2 or more samples. Of the total transcripts, 17,246 were known mRNAs and 7535 were unknown. Using a FDR of 0.1, we identified differentially expressed genes (DEGs), downregulated and upregulated, between Gmacs from infected Cth^–/– and WT mice (Figure 4A). The complete list of DEGs, including Ensemble transcript identifiers, official gene symbols, fold changes, and P values, is provided in Supplemental Data Set 1. Gene ontology–term gene set enrichment analysis of the significant DEGs between Gmacs from infected Cth^–/– versus WT mice evidenced suppression of gene sets associated with immunity and response to stimuli (Figure 4B and Supplemental Data Set 2). Notably, gene-level expression analysis of the gene ontology term “immune system process” revealed consistently reduced expression in Gmacs from infected Cth^–/– mice compared with WT mice (Figure 4, C and D).

Figure 4

CTH mediates induction of immune response–specific gene sets. Transcriptomic analysis using RNA-Seq of F4/80⁺ enriched gastric cells (Gmacs); n = 5 individual infected mice per genotype. (A) Volcano plot of DEGs in Cth^–/– Gmacs compared with WT Gmacs from infected mice (fold change [FC] > 2; FDR < 0.05). Immune-associated genes are enlarged. Gray dots, not significant; green dots, log₂ fold change > 2; blue dots, adjusted (adj) P < 0.05; red dots, log₂ fold change > 2 and adjusted (adj) P < 0.05. (B) Ridge plot displaying gene set enrichment analysis of DEGs in infected Cth^–/– compared with infected WT Gmacs (log₂ fold change > 2; FDR < 0.05). (C) Enrichment score plot of the gene set “immune system process” identified in A. (D) Heatmap showing the gene-level expression of genes within the immune system process gene set shown in B. Ext, external; ES, ES, enrichment score.

We next determined how the response to infection differs between WT and Cth-deficient mice when compared with their respective uninfected controls. We performed differential gene expression analysis to compare the response to infection between genotypes. There were 37 downregulated genes shared by WT and Cth^–/– Gmacs and 145 genes that were induced in both WT and Cth^–/– cells during infection (Supplemental Figure 5A). Immune-related genes upregulated by infection in WT Gmacs included Irgm1, Irgm2, Ifit1, and Ido1, in addition to upregulated Cth expression (Figure 5A). In contrast, immune-related genes were either not induced or were downregulated by infection, such as Ifit2, Cybb, and Fcna, in Cth^–/– Gmacs (Figure 5B). Pathway analysis of significant DEGs using the Database for Annotation, Visualization, and Integrated Discovery Bioinformatics Resource (https://david.ncifcrf.gov/) revealed 4 downregulated pathways from the Kyoto Encyclopedia of Genes and Genome Pathways database in the transcriptome of Cth^–/– mice compared with WT mice (Supplemental Figure 5B and Supplemental Data Set 3). We also assessed function-associated keywords and identified 26 downregulated keywords (Supplemental Figure 5C). Downregulated pathways and keywords in cells from infected Cth^–/– mice compared with WT mice included metabolic pathways and the following downregulated keywords: “immunity,” “S-adenosyl-l-methionine,” and “methyltransferase” (Figure 5C and Supplemental Figure 5, B and C). Downregulation of these pathways suggests that Gmacs from Cth^–/– mice have altered immune function and impaired metabolic activation compared with WT mice. Furthermore, the downregulation of genes associated with SAM and methyltransferases is consistent with our hypothesis that CTH activity can modulate the reactions upstream of the RTP.

Figure 5

Identification of H. pylori–regulated genes in Gmacs. (A and B) The volcano plots show differential gene signatures in Gmacs from infected mice compared with uninfected mice (fold change [FC] > 2; FDR < 0.05). Gray dots, not significant; green dots, log₂ fold change > 2; blue dots, adjusted (adj) P < 0.05; red dots, log₂ fold change > 2 and adjusted P < 0.05. Genes of interest are enlarged. (A) Gmacs from H. pylori–infected WT mice compared with uninfected WT mice. (B) Gmacs from H. pylori–infected Cth^–/– mice compared with uninfected Cth^–/– mice. (C) Heatmaps displaying significantly altered pathways of DEGs in Cth^–/– Gmacs compared with WT Gmacs of infected mice when compared with uninfected controls (interaction P < 0.01). Fold change is infected divided by control for each genotype.

CTH promotes the metabolism of SAM through the RTP during H. pylori infection. To relate the transcriptional changes to metabolic signatures, we next performed an untargeted global metabolomic analysis of the gastric tissues from infected WT and Cth^–/– mice. Overall, there were 1300 positively charged metabolites and 734 negatively charged metabolites significantly affected by CTH in infected mice (Supplemental Figure 6, A and B). Pathway analysis evidenced metabolic pathways involving SAM metabolism; namely “(S)-reticuline biosynthesis,” “spermidine biosynthesis,” “spermine biosynthesis,” and “l-serine degradation” were significantly affected by deletion of Cth in infected mice (Figure 6A and Supplemental Data Set 4). We also found that pathways involved in energy production, such as the TCA cycle and glycolysis, were affected by the deletion of Cth (Figure 6A).

Figure 6

CTH promotes the metabolism of SAM through the RTP during H. pylori infection. (A) Bubble plot of metabolomic analysis of gastric tissues from WT and Cth^–/– mice at 8 weeks p.i. with H. pylori PMSS1 (FDR < 0.05). Pathways involving SAM metabolism or energy production are enlarged. The x-axis and node color represent pathway overlap and the y-axis and node radius represent the P value; n = 8 infected mice per genotype. (B) Abundance of dcSAM, SAM, and cystathionine in the gastric tissues of WT and Cth^–/– mice at 4 weeks p.i. with H. pylori PMSS1; n = 4 uninfected and 6–8 infected mice per genotype. All values are reported as mean ± SEM. Statistical analyses, where shown were conducted with 1-way ANOVA with Newman-Keuls post hoc test. *P < 0.05, **P < 0.01, ***P < 0.001.

We then used targeted metabolomics to verify our findings. Specifically, dcSAM, SAM, and cystathionine levels were measured by liquid chromatography–mass spectrometry in the tissues of WT and Cth^–/– mice infected or not with H. pylori (Figure 6B). Levels of dcSAM remained relatively similar; however, SAM levels were significantly accumulated in naive Cth^–/– mice and significantly decreased in Cth^–/– mice with infection with no change in SAM levels in the WT mice (Figure 6B). Importantly, levels of cystathionine, the substrate for CTH, were markedly increased in the Cth^–/– mice, confirming that CTH activity was lost (Figure 6B).

Because there was an abundance of SAM in naive Cth^–/– gastric tissues, we investigated whether SAM availability directly regulates the macrophage response to H. pylori. We treated bone marrow–derived macrophages (BMmacs) with SAM in the absence or presence of H. pylori for 24 hours. We confirmed that exogenous SAM treatment results in accumulation of SAM and cystathionine in the BMmacs, which decreased with infection (Supplemental Figure 7A). Analysis of mRNA expression revealed that treatment with SAM significantly reduced expression of Tnf and Il12b, while increasing expression of Il1b and Il6, having no effect on Nos2 and the p35 subunit of IL-12 (Il12a), and increasing expression of the prototypical markers of antiinflammatory macrophages, Arg1, and Tgfb1 (Supplemental Figure 7, B–E). To assess the in vivo effect of SAM on H. pylori–induced disease, WT mice were orally gavaged with SAM starting 1 week after infection (p.i.) (Supplemental Figure 7F). There was no change in gastritis score or colonization after SAM treatment (Supplemental Figure 7, G and H).

CTH is downstream of SAM, which is needed for the biosynthesis of dcSAM, and thus for spermidine and spermine production. We have previously shown that polyamines can regulate H. pylori–induced macrophage activation and thus affect disease progression (6, 7, 9, 12, 36). To determine if CTH may affect H. pylori–induced immune activation by a mechanism linked to polyamines, we next measured polyamine levels in the gastric tissues of WT and Cth^–/– mice infected or not with H. pylori. We found that there were no significant differences in putrescine, spermidine, or spermine levels between WT and Cth^–/– mice (Supplemental Figure 8A). This may be explained by the absence of any changes in the mRNA expression levels of enzymes within the polyamine pathway in the gastric tissues (Supplemental Figure 8B). We next directed the consumption of SAM toward the RTP and away from polyamine synthesis using the SAMDC inhibitor, SAM486A (also known as CGP 48664 and commercially as Sardomozide) (38). Inhibition of SAMDC in BMmacs resulted in a significant accumulation of putrescine with and without H. pylori infection (Supplemental Figure 8C). Spermidine levels remained relatively the same independent of treatment or infection, whereas spermine levels significantly decreased in H. pylori–infected cells after SAMDC inhibition (Supplemental Figure 8C). BMmacs treated with SAM486A and infected with H. pylori for 24 hours displayed significantly increased expression of Tnf but decreased expression of Il12b compared with the untreated infected BMmacs (Supplemental Figure 8D). To assess the in vivo effect of SAMDC inhibition on H. pylori–induced disease, C57BL/6 mice were administered SAM486A by i.p. injection, beginning 1 week p.i., every other day for the remainder of the 4 week infection (Supplemental Figure 8E) (39). As expected, SAM486A treatment led to an increase of putrescine levels in the gastric tissues from the mice (Supplemental Figure 8F). In contrast, the levels of spermidine and spermine did not change with either infection or treatment with SAM486A (Supplemental Figure 8F), suggesting another mechanism maintains polyamine levels in the stomach. Mice treated with SAM486A had no change in gastritis scores or H. pylori colonization (Supplemental Figure 8, G and H). In summary, SAM supplementation and SAMDC inhibition had moderate effects on macrophage gene expression in vitro; however, this did not translate to in vivo effects, suggesting that SAM, SAMDC, and polyamines do not mediate the effects of CTH on H. pylori–induced inflammation and that other mechanisms should be identified.

CTH suppresses DNA methylation and supports activated macrophage gene expression. We observed that expression of CTH was specifically induced by H. pylori in BMmacs compared with another Gram-negative bacterium, Citrobacter rodentium, or the classical stimuli of M1-like macrophages (e.g., LPS plus IFN-γ), M2-like macrophages (e.g., IL-4), or regulatory macrophage–like macrophages (e.g., IL-10) (Figure 7A).

Figure 7

CTH suppresses DNA methylation and supports M1 macrophage gene expression. (A) Expression of Cth by WT BMmacs 24 hours after challenge; n = 4–10 biological replicates from 3 independent experiments. (B) Representative plot of CD11b and F4/80 expression of WT and Cth^–/– BMmacs 24 hours p.i. with H. pylori. (C) Fold change of metabolite levels in BMmacs 24 hours p.i. with H. pylori compared with control (ctrl); n = 4 biological replicates. (D) Representative plot and (E) MFI quantification of 5mC staining in WT and Cth^–/– BMmacs 6 hours p.i. with H. pylori; n = 4 biological replicates. (F) Expression of proinflammatory macrophage activation markers by WT and Cth^–/– BMmacs 6 hours p.i. with H. pylori; n = 4–10 biological replicates per genotype from 3 independent experiments. (G) NOS2 protein immunoblot and densitometric analysis of WT and Cth^–/– BMmacs 24 hours p.i. with H. pylori; n = 4 biological replicates per genotype. (H) NO₂^– concentration in supernatants from WT and Cth^–/– BMmacs; n = 8 biological replicates from 2 independent experiments. (I) Expression of antiinflammatory macrophage activation markers by WT and Cth^–/– BMmacs 6 hours p.i. with H. pylori; n = 4–10 biological replicates per genotype from 3 independent experiments. (J) Phosphorylated RELA (pRELA) protein immunoblot and densitometric analysis of WT and Cth^–/– BMmacs 30 minutes p.i. with H. pylori; n = 3–4 biological replicates from 2 independent experiments. All values are reported as mean ± SEM. Statistical analyses, where shown: (A) 1-way ANOVA with Dunnett’s test; (C) 1-way ANOVA with Newman-Keuls post hoc test; (E, F, H, and I) 1-way ANOVA with Kruskal-Wallis test, followed by a Mann-Whitney U test; (J) Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001 vs. uninfected; ^§P < 0.05, ^§§P < 0.01 vs. WT.

Currently available inhibitors of CTH are moderately selective but are active against other pyridoxal phosphate–dependent enzymes (40). Therefore, to assess the role of CTH in macrophages, we generated BMmacs from WT and Cth^–/– mice (Figure 7B and Supplemental Figure 9A). We confirmed that CTH was not expressed in Cth^–/– BMmacs (Supplemental Figure 9B) and that there was no change in expression of Cbs in the absence of CTH or after H. pylori infection (Supplemental Figure 9C). Levels of homocysteine, the substrate for transsulfuration, were significantly higher in Cth^–/– BMmacs independent of infection, and cystathionine levels were the same between WT and Cth^–/– BMmacs with or without infection (Figure 7C and Supplemental Figure 9D). SAM levels were significantly higher in WT infected BMmacs compared with the uninfected control, whereas SAM levels were significantly decreased in Cth^–/– BMmacs with and without H. pylori infection (Figure 7C and Supplemental Figure 9D). DNA methylation was significantly increased in infected Cth^–/– BMmacs 6 hours p.i. with H. pylori (Figure 7, D and E). At 24 hours p.i., DNA methylation was reduced to almost uninfected WT levels; however, Cth^–/– BMmacs had significantly higher levels of DNA methylation compared with WT, independent of infection (Supplemental Figure 9, E and F).

We next assessed whether the increase in DNA methylation affected gene expression. Levels of proinflammatory macrophage markers Tnf, Nos2, Cxcl10, Il1b, Il6, and Il12b were significantly reduced in Cth^–/– BMmacs 6 hours p.i. with H. pylori (Figure 7F). The decreased Nos2 mRNA expression in infected Cth^–/– BMmacs correlated with decreased levels of NOS2 protein and NO production (Figure 7, G and H). Expression of the antiinflammatory markers Chil3 and Tgfb1, but not Arg1 and Il10, was decreased in Cth^–/– BMmacs at 6 hours p.i. with H. pylori (Figure 7I). At 24 hours p.i. with H. pylori, expression of Tnf, Arg1, and Tgfb1 was significantly reduced in Cth^–/– BMmacs (Supplemental Figure 9, G and H). These findings support the idea that CTH can regulate macrophage gene expression through DNA methylation.

We have previously shown that H. pylori enhances macrophage gene expression through NF-κB activation (11, 41). We next found that WT and Cth^–/– BMmacs had similar levels of MyD88 expression (Figure 7J), which is directly linked to TLR signaling upstream of NF-κB activation. In contrast, phosphorylation of the NF-κB p65 subunit RELA at Ser-536 that was activated by H. pylori infection in WT BMmacs was significantly decreased in Cth^–/– cells (Figure 7J). NF-κB activation and nuclear translocation are regulated by the phosphorylation of inhibitor of κ light polypeptide gene enhancer in B cells, inhibitor α (NFKBIA, also known as IκBα), and subsequent degradation. There was no difference in the phosphorylation or total levels of NFKBIA (Supplemental Figure 9I). These findings suggest that CTH is important in early signaling cascades in macrophages by enhancing the phosphorylation of NF-κB, independent of NFKBIA degradation.

The role of CTH in the macrophage proteome. To further determine the role of CTH in the macrophage response to H. pylori, we analyzed the proteome of WT and Cth^–/– BMmacs. Using a 4-plex tandem mass tag isobaric mass tagging–based approach, we identified 213 proteins in WT BMmacs and 142 proteins in Cth^–/– BMmacs that were upregulated with infection, with 101 of those proteins shared between the 2 genotypes (Figure 8A and Supplemental Data Set 5). In addition, we identified 75 proteins in WT BMmacs and 45 proteins in Cth^–/– BMmacs that were downregulated with infection, with 29 of those proteins shared between the 2 genotypes (Figure 8A and Supplemental Data Set 5). We then assessed the H. pylori–induced proteome of WT BMmacs using Ingenuity Pathway Analysis (Figure 8B and Supplemental Data Set 6). The pathway most significantly activated by infection was “metabolism of reactive oxygen species,”with a Z-score of 3.51 (Figure 8B and Supplemental Data Set 6).

Figure 8

CTH modifies macrophage activation patterns of proteins during H. pylori infection. Proteomic analysis of BMmacs from WT and Cth^–/– mice infected or not with H. pylori for 24 hours; n = 4 biological replicates per genotype. (A) Venn diagram showing commonly downregulated and upregulated proteins. (B) Enrichment analysis of pathways activated by H. pylori infection in WT BMmacs. (C) Enrichment analysis of the pathways in B activated by H. pylori infection in Cth^–/– BMmacs. (FDR < 0.05; Z-score > 2).

We next assessed alterations in the H. pylori–induced proteome in the absence of CTH. Cth^–/– BMmacs infected with H. pylori compared with uninfected Cth^–/– BMmacs exhibited upregulation of only 27 of the 89 pathways (30%) activated by H. pylori infection in WT BMmacs (Figure 8C and Supplemental Data Set 7) and downregulation of only 2 of the 8 pathways (25%) inhibited by H. pylori infection in WT BMmacs (Supplemental Figure 10 and Supplemental Data Sets 6 and 7). Although “synthesis of reactive oxygen species” and “production of reactive oxygen species” were present in both WT and Cth^–/– BMmacs, importantly, the “metabolism of reactive oxygen species” was not present in the pathways induced by H. pylori in Cth^–/– BMmacs (Figure 8, B and C). Thus, CTH has the capacity to regulate protein expression in macrophages and contribute to macrophage activation and functions.

CTH supports mitochondrial respiration while promoting glycolysis in H. pylori–infected macrophages. To elucidate the mechanism by which CTH supports macrophage activation and contributes to H. pylori–induced disease, we investigated the downstream metabolic effects of Cth deletion in macrophages. Because the metabolomic pathway analysis indicated changes in glycolysis and the TCA cycle, we conducted extracellular flux assays on WT and Cth^–/– BMmacs to assess mitochondrial function and cellular respiration. As expected, glycolysis and glycolytic capacity were increased with H. pylori infection in WT BMmacs, whereas Cth^–/– BMmacs had impaired glycolysis and glycolytic capacity in naive cells or those exposed to H. pylori (Figure 9A). WT BMmacs stimulated with LPS plus IFN-γ for 24 hours had decreased oxygen consumption rate (OCR) but increased extracellular acidification rate (ECAR) (Supplemental Figure 11, A and B), whereas WT BMmacs stimulated with IL-4 for 24 hours displayed both increased OCR and increased ECAR (Supplemental Figure 11, C and D). Mitochondrial oxidative phosphorylation was decreased (Figure 9B) and ECAR was increased with H. pylori infection in WT BMmacs (Figure 9C). Cth^–/– BMmacs displayed increased OCR and ECAR 24 hours after challenge with H. pylori (Figure 9, B and C) or with LPS plus IFN-γ (Supplemental Figure 11, A and B) but did not change with IL-4 stimulation (Supplemental Figure 11, C and D). Uninfected Cth^–/– BMmacs exhibited lower basal and maximal respiration than did uninfected WT BMmacs, which did not change with infection, unlike the WT BMmacs (Figure 9D). Elevated ECAR levels with H. pylori infection or LPS plus IFN-γ stimulation confirmed that the Cth^–/– BMmacs were still viable (Figure 9C and Supplemental Figure 11B). The decreased mitochondrial respiration also resulted in decreased mitochondrial ATP production and spare respiratory capacity, an indicator of metabolic plasticity, in the Cth^–/– BMmacs (Figure 9D), suggesting that Cth^–/– BMmacs do not exhibit metabolic activation in response to infection.

Figure 9

CTH contributes to macrophage activation by enhancing mitochondrial function and glycolysis in H. pylori–infected macrophages. (A) Rate of glycolysis and glycolytic capacity of WT and Cth^–/– BMmacs 24 hours p.i. with H. pylori; n = 3 mice per genotype. (B) OCR and (C) ECAR of WT and Cth^–/– BMmacs 24 hours p.i. with H. pylori; n = 5 from 2 independent experiments. Vertical dashed lines indicate the sequential addition of oligomycin (oligo), FCCP, and Rot/AA. (D) Basal respiration, maximal respiration, ATP production, and spare respiratory capacity (SRC) derived from B. (E) Representative plot and (F) MFI quantification of MitoSOX Red staining in WT and Cth^–/– BMmacs 24 hours p.i. with H. pylori; n = 4 biological replicates. (G) Glutathione levels in WT and Cth^–/– BMmacs 24 hours p.i. with H. pylori; n = 3 biological replicates. (H) Cysteine levels in WT and Cth^–/– BMmacs 24 hours p.i. with H. pylori; n = 4 biological replicates. All values are reported as mean ± SEM. Statistical analyses, where shown: 1-way ANOVA with Newman-Keuls post hoc test. *P < 0.05, **P < 0.01, ***P < 0.001. FCCP, carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone; Max, maximum; Rot/AA (also Rot&AA), rotenone and antimycin A.

Because mitochondrial activity contributes to the production of ROS that macrophages generate to combat pathogens, we measured mitochondrial superoxide levels, an indicator of mitochondrial ROS. We found that H. pylori–infected Cth^–/– BMmacs exhibited a significant accumulation of mitochondrial ROS compared with uninfected controls and infected WT cells at 24 hours p.i. (Figure 9, E and F). We next measured the amount of the intracellular antioxidant glutathione (GSH) and found that GSH levels were low in Cth^–/– BMmacs at baseline and remained unchanged with infection, whereas GSH levels decreased with infection in WT BMmacs (Figure 9G). Cysteine, the direct product of CTH activity, is a semi-essential amino acid that is the main precursor for generation of GSH (15, 16). Cysteine levels were significantly elevated in infected WT BMmacs compared with control and infected Cth^–/– BMmacs (Figure 9H), consistent with loss of CTH activity in Cth^–/– BMmacs, and increased CTH activity with H. pylori infection in WT cells. These results suggest that macrophages upregulate CTH expression during H. pylori infection to maintain redox homeostasis.

Taken together, these studies demonstrate that the loss of CTH in macrophages leads to metabolic suppression and impaired immune activation that underlies the decrease in H. pylori–induced gastritis observed in Cth^–/– mice (Supplemental Figure 12).

Macrophages are a fundamental component of the host immune response with the capacity to transition along the spectrum of proinflammatory to antiinflammatory polarization as environmental signals and stimuli change (1, 2). In inflammatory pathologies, such as H. pylori–induced gastritis, overzealous immune responses can lead to more severe disease outcomes, such as gastric cancer (10). Therefore, identification of alternative strategies that target the host to limit chronic inflammatory conditions are needed to prevent disease progression. Our study demonstrates that CTH affects metabolic activation and polarization of macrophages in response to H. pylori infection. Moreover, we used unbiased “omics” approaches to probe the global effects of CTH in the host response to H. pylori and consistently found evidence of a heightened immune response and enhanced cellular respiration in the presence of CTH that contributes to H. pylori–induced disease.

In several previous studies, researchers have investigated the role of CTH in inflammation within the context of H₂S production; however, there is not a consensus and results often depend on exogenous versus endogenous enzymatic H₂S derivation, with the latter being much less understood (42, 43). This variation in findings highlights the need to further characterize the role of CTH in different models of inflammation. Using a genetic model of Cth deletion, here we have shown that Cth^–/– mice have significantly attenuated gastritis in models of acute and chronic H. pylori infection. In this study, we did not measure H₂S, because many bacterial pathogens, including H. pylori (44), produce H₂S, and host cells maintain the ability to produce H₂S in the absence of CTH through CBS or 3-mercaptopyruvate sulfurtransferase activity (15). Instead, we measured homocysteine, cystathionine, and cysteine levels as direct measures of CTH activity. Our work further supports the idea that CTH is a modulator of immune cell function and we have identified, for the first time, to our knowledge, the role of CTH in the stomach in the context of H. pylori infection. In future studies, it would be interesting to examine the interplay between CTH and the gut microbiome.

Our previous findings demonstrated that exogenous overexpression of the human CTH gene in RAW 264.7 cells inhibited markers of macrophage activation (33). However, determining the effect of CTH inhibition on macrophage activation was not feasible in the prior study, because we found that Cth siRNA diminished the number of live H. pylori in infected cells, potentially limiting the macrophage response, and that the CTH inhibitor aminooxyacetic acid directly kills H. pylori (33). In addition, the CTH inhibitor has off-target activity (33, 40). In the present study, genetic KO of Cth led to decreased macrophage activation during H. pylori infection. Transcriptomic analysis of Gmacs revealed that multiple gene sets involved in immune function were downregulated with Cth deletion. This was also reflected in the proteome of macrophages from Cth^–/– versus WT mice infected ex vivo with H. pylori.

Because CTH has been mainly studied as a producer of H₂S, there is a gap in knowledge about how CTH may influence the metabolic pathways directly upstream of its activity. Our analysis of untargeted metabolomics identified SAM metabolism and cellular respiration as pathways regulated by CTH. SAM supplementation has been shown to reduce airway inflammation (45), and M1-like macrophages exhibit reduced DNA methylation (46, 47). However, S-adenosylhomocysteine, the product of methyl donation by SAM, can bind to methyltransferases and inhibit methylation reactions (48, 49). Although SAM supplementation suppressed proinflammatory and enhanced antiinflammatory gene expression in vitro, there was no protection from H. pylori gastritis in vivo. The polyamine pathway is directly linked to SAM metabolism. Based on the increased gastritis in mice with myeloid-specific deletion of ornithine decarboxylase during H. pylori infection (12), we anticipated that inhibition of SAMDC would result in the accumulation of putrescine and decreased inflammation. However, we observed that SAMDC inhibition increased putrescine levels in the stomach of mice without affecting gastritis, suggesting that putrescine accumulation alone is not sufficient to regulate gastritis. Similarly, we hypothesized that deletion of CTH might free up SAM for the metabolism of putrescine into spermidine and increase inflammation. However, loss of CTH did not alter polyamine levels sufficiently enough to affect gastritis.

There is a growing body of literature describing the metabolic needs of immune cells and, in particular, the role of metabolism in macrophage differentiation and activation (50, 51). During macrophage stimulation, polarizing macrophages remodel their metabolism to meet energetic needs. Proinflammatory macrophages repress oxidative phosphorylation in favor of glycolysis to aid the production of cytokines and ROS (52). In contrast, antiinflammatory M2 macrophages enhance mitochondrial oxidative phosphorylation while still consuming glucose (53). We found that macrophages lacking CTH have impaired ability to metabolically adapt to external stimuli. ROS and nitrogen species are important effector molecules produced by macrophages to combat pathogens and regulate polarization (54); however, high levels can be detrimental and cause cellular dysfunction, including mitochondrial impairment (55). Consistent with these concepts, our findings show that loss of CTH results in depleted intracellular cysteine and GSH and contributes to high levels of mitochondrial superoxide and reduced mitochondrial function. Beyond the altered gene expression of macrophage cytokines, our data highlight an unexpected role of CTH in macrophage metabolism and respiration.

Last, in the context of H. pylori infection, regulation of TNF-α transcription depends on alternative activation of NF-κB through the phosphorylation of the p65 subunit RELA at Ser-536 that does not involve NFKBIA degradation (56). We found that deletion of Cth led to decreased TNF-α expression during infection with H. pylori, in vitro and in vivo, and with LPS plus IFN-γ stimulation. The impaired cytokine production may be linked to the observed decrease in phosphorylation of RELA at Ser-536 in the infected Cth^–/– macrophages, especially because there were no differences in phospho- or total NFKBIA levels between the infected cells. Because many kinases are involved in the phosphorylation of p65, future studies using various receptor and signaling cascade inhibitors may enhance the understanding of the mechanism connecting CTH with the early macrophage responses.

In summary, our findings outline the role of CTH in regulating the macrophage-facilitated response to H. pylori. Increased expression of CTH by macrophages contributes to inflammation in the gastric tissues of mice and markers of proinflammatory macrophages ex vivo in the context of H. pylori infection. Using genetic deletion of Cth, we determined that CTH can affect macrophage function via enhanced gene and protein expression through NF-κB phosphorylation and suppression of DNA methylation and maintenance of mitochondrial function through redox homeostasis. These activities, in turn, can affect the adaptive immune response in T cells, though Cth deletion does not appear to affect T cell function in a cell-autonomous manner. Because of the limitations of current chemical inhibitors of CTH, exploring the full spectrum of pathways related to CTH is essential in the discovery of therapeutic targets that can help reduce the burden of chronic gastritis and risk of gastric cancer in humans.

See Supplemental Methods for detailed experimental procedures.

Materials and reagents. See Supplemental Table 1 for information regarding reagents and kits used in this study. See Supplemental Table 2 for information regarding software and analysis packages used in this study. See Supplemental Table 3 for information regarding Abs used in this study. See Supplemental Table 4 for information regarding RT-PCR primers used in this study

Statistics. All the data shown represent the mean ± SEM. A minimum of 3 biological replicates were used for in vitro studies. Statistical analysis was performed in GraphPad Prism 9.2 (GraphPad Software), and significance was set at P < 0.05. Where data were normally distributed, 2-tailed Student’s t test and 1-way ANOVA with the Newman-Keuls or Dunnett’s post hoc test were used to determine significant differences between 2 groups or multiple test groups, respectively. Where data were not normally distributed, a 1-way ANOVA with the Kruskal-Wallis test, followed by a Mann-Whitney U test, was performed, unless otherwise noted. Detailed statistical analysis for RNA-Seq, metabolomics, and proteomics are included in Supplemental Methods.

Study approval. Mice were used under the protocols M1900034, V1800106, and V2000018 approved by the IACUC at Vanderbilt University and Institutional Biosafety Committee and the Research and Development Committee of the Veterans Affairs Tennessee Valley Healthcare System. Procedures were performed in accordance with institutional policies, AAALAC guidelines, the American Veterinary Medical Association Guidelines on Euthanasia, NIH regulations (Guide for the Care and Use of Laboratory Animals; National Academies Press, 2011), and the US Animal Welfare Act (1966).