Role of glutathione metabolism in host defense against Borrelia burgdorferi infection

Significance

Inflammation plays a crucial role in the pathogenesis of Lyme disease, caused by the spirochete Borrelia burgdorferi. Intracellular metabolism is increasingly being recognized as a major determinant of inflammation. In this study, we investigated how B. burgdorferi affects host cell metabolism by analyzing the intracellular metabolome in vitro, as well as the circulating metabolome in patients with early-onset Lyme disease. We identify glutathione metabolism as the most important target of B. burgdorferi infection and discover that this pathway is essential for cytokine production, likely through glutathionylation. These findings not only provide more insight into the pathogenesis of Lyme disease but also underline how host–pathogen interactions in metabolism can play crucial roles in host defense against pathogens.

Abstract

Pathogen-induced changes in host cell metabolism are known to be important for the immune response. In this study, we investigated how infection with the Lyme disease-causing bacterium Borrelia burgdorferi (Bb) affects host metabolic pathways and how these metabolic pathways may impact host defense. First, metabolome analysis was performed on human primary monocytes from healthy volunteers, stimulated for 24 h with Bb at low multiplicity of infection (MOI). Pathway analysis indicated that glutathione (GSH) metabolism was the pathway most significantly affected by Bb. Specifically, intracellular levels of GSH increased on average 10-fold in response to Bb exposure. Furthermore, these changes were found to be specific, as they were not seen during stimulation with other pathogens. Next, metabolome analysis was performed on serum samples from patients with early-onset Lyme disease in comparison with patients with other infections. Supporting the in vitro analysis, we identified a cluster of GSH-related metabolites, the γ-glutamyl amino acids, specifically altered in patients with Lyme disease, and not in other infections. Lastly, we performed in vitro experiments to validate the role for GSH metabolism in host response against Bb. We found that the GSH pathway is essential for Bb-induced cytokine production and identified glutathionylation as a potential mediating mechanism. Taken together, these data indicate a central role for the GSH pathway in the host response to Bb. GSH metabolism and glutathionylation may therefore be important factors in the pathogenesis of Lyme disease and potentially other inflammatory diseases as well.

Lyme disease, caused by Borrelia burgdorferi (Bb) sensu lato, is the most common vector-borne disease in the Northern hemisphere (1, 2), transmitted by ticks. Lyme disease most often presents locally with a migrating skin rash called erythema migrans (EM) but, if left untreated, can give rise to inflammatory complications in the joints (3), heart (4), or nervous system (5). In most cases, Lyme disease can be effectively treated by antibiotics, yet a small percentage of patients experience persisting symptoms even after extensive antibiotic treatment (6, 7).

Interestingly, Bb is not known to produce toxic factors (8). The majority of Lyme disease symptoms are therefore attributed to the host’s immune response against the pathogen. In addition, it is hypothesized that persistent symptoms after treatment are not due to continuous infection, but rather due to an aberrant inflammatory response (7, 9, 10). Together, this suggests a crucial role for the host immune response in the initiation and outcome of the infection.

An upcoming topic in the study of the immune system is immunometabolism, which investigates the impact of cellular metabolism on immune cell function. This is of particular interest in the case of Bb as the spirochete is known to have very limited metabolic capabilities (11, 12). This might cause the spirochete to induce specific changes in host cell metabolism. Supporting this, we recently showed that Bb induces a switch in central glucose metabolism in host mononuclear cells which was crucial for cytokine production (13).

In the present study, we aimed to further explore the metabolic pathways induced by Bb and analyze their role in immune cell function. To achieve this, we performed metabolomic analysis of primary human monocytes stimulated with Bb or other inflammatory stimuli. Identified pathways were then further validated using in vitro intervention experiments to elucidate their role in the inflammatory response. Lastly, we examined the relevant metabolites in serum samples from acute Lyme disease patients.

Results

Primary Human Monocytes Exposed to Bb Display Altered Glutathione Metabolism.

To determine which metabolic pathways were affected by Bb infection, metabolome analysis was performed on primary human monocytes stimulated with Bb or medium control for 24 h. Pathway analysis was performed to identify specific metabolic pathways altered by Bb exposure (Table S1). As seen in Fig. 1A, the pathways most significantly affected by Bb were glutathione (GSH) metabolism, arachidonic acid metabolism, and pyrimidine metabolism. When analyzing individual metabolites, eight compounds related to GSH metabolism were found among the top 25 most significantly affected metabolites (Fig. 1B, with references to Fig. S1). Most noteworthy, Bb stimulation induced a dramatic increase in reduced GSH levels (Fig. 1C), while only modestly increasing oxidized glutathione (GSSG) (Fig. 1D). This indicates a shift in the GSH/GSSG ratio, suggesting a more antioxidative state. Other metabolites significantly increased by Bb stimulation included polyamines, nucleotide metabolites, and phospholipid metabolites. Supporting the previously demonstrated role of glucose metabolism in Bb infection (13), lactate was among the most significantly affected metabolites.

Fig. 1.

Metabolome analysis of primary monocytes stimulated with Bb versus RPMI. (A) Scatter plot of KEGG metabolic pathways in primary human monocytes (n = 5) affected by Bb stimulation, showing log P value of the enrichment analysis (y axis and visualized by node color) and pathway impact, taking into account the importance of the affected metabolites within a pathway (x axis and visualized by node radius; range 0 to 1, where 1 is maximal impact). (B) Heat map depicting the top 25 most significantly affected metabolites after Bb stimulation, where red indicates an increase and blue indicates a decrease. Numbers represent references to Fig. S1. #, GSH derivative, formed upon oxidative stress of GSH. (C and D) Raw data of (C) reduced glutathione (GSH) and (D) oxidized glutathione (GSSG) levels in primary monocytes stimulated with Bb or control (RPMI). Box plot indicates median ± min/max values. Max *P > 0.05, **P > 0.01.

Next, we compared fold changes (FCs) relative to medium control (RPMI) in metabolite levels in monocytes exposed to Bb with cells exposed to the TLR4 ligand LPS and the TLR2 ligand Pam3Cys. Fig. 2A shows the top 50 most differentially regulated metabolites between the stimuli. Interestingly, one cluster of eight metabolites (indicated in red) was identified, which were up-regulated (FC > 1) by Bb, while being down-regulated (FC < 1) by LPS stimulation and Pam3Cys stimulation. From this cluster, three metabolites could be linked to GSH metabolism (see Fig. S1 for reference): homocysteine, pyridoxal phosphate (PLP), and flavin adenine dinucleotide (FAD). The other metabolites in this cluster were adenosine monophosphate (AMP), inosinic acid (or inosine monophosphate) (IMP), tetradecanoylcarnitine A, beta-alanine, and hypotaurine. GSH was also slightly increased in LPS- and Pam3Cys-stimulated samples. However, LPS and Pam3Cys induced a twofold increase in GSH levels, while a 10-fold increase was seen after Bb stimulation (Fig. 2 B and C). An overview of all metabolites and enzymes involved in GSH metabolism is given in Fig. S1. All together, these data indicate that GSH metabolism is significantly and specifically influenced by Bb stimulation.

Fig. 2.

Metabolome analysis of primary monocytes stimulated with Bb versus LPS and P3cys. (A) Heat map depicting top 50 most differentially regulated metabolites in primary monocytes stimulated with Bb or the TLR ligands LPS or P3cys (all n = 5), based on fold changes in metabolite levels relative to unstimulated controls. Numbers represent references to Fig. S1. (B and C) Fold changes (relative to unstimulated controls) in levels of GSH (B) and GSSG (C) in primary monocytes stimulated with Bb compared with LPS and Pam3cys. Box plot indicates median ± min/max.

Altered GSH Metabolism Affects Bb-Induced Cytokine Production, but Not Reactive Oxygen Species Production.

To investigate whether altered GSH metabolism plays a role in immune cell function in response to Bb, we performed several in vitro validation experiments. First, we determined whether the increase in GSH levels affects the capacity of host cells to generate reactive oxygen species (ROS). To investigate this, human peripheral blood mononuclear cells (PBMCs) were exposed for 24 h to Bb and then stimulated with serum-opsonized zymosan (SOZ), a potent inducer of ROS. Indeed, exposure to Bb strongly decreased SOZ-induced ROS production (Fig. 3A). However, lowering GSH concentrations using buthionine sulfoximine (BSO) or diethyl maleate (DEM) did not reverse the effect of Bb stimulation on ROS production (Fig. 3 B and C), indicating that this effect was independent of GSH levels.

Fig. 3.

Effect of inhibitors of GSH metabolism on ROS production and cytokine production. (A) Area under the curve (AUC) of ROS-induced luminescence on serum-opsonized zymosan (SOZ)-stimulated PBMCs (n = 11) pretreated for 24 h with Bb or control (RPMI). (B and C) AUC of ROS-induced luminescence on SOZ-stimulated PBMCs (n = 8 and n = 6, respectively) pretreated with Bb in the presence of different doses of the GSH synthesis inhibitor BSO (B) (n = 8) or the GSH-depleting agent DEM (C) (n = 6). (D–F) Production of IL-1β (Upper) and TNFα (Lower) by PBMCs after 24 h stimulation with Bb in the presence of different doses (in µM) of (D) BSO (n = 14), (E) DEM (n = 14), or (F) N-acetylcysteine (NAC, n = 15). Bar graphs represent mean ± SEM. *P > 0.05, **P > 0.01, ***P < 0.001.

Next, we determined whether altered GSH metabolism affects Bb-induced cytokine production. PBMCs were pretreated with modulators of GSH metabolism and then stimulated for 24 h with Bb. All compounds were checked for cytotoxicity, and, with the exception of the high DEM concentration, no signs of cytotoxicity were seen (Fig. S2). First, GSH biosynthesis was targeted by inhibiting γ-glutamylcysteine synthase (GCS) using BSO. As shown in Fig. 3D, BSO showed disparate effects on cytokine production, with low concentration decreasing IL-1β production, while increasing TNFα. Depleting GSH with low-concentration DEM gave similar results to GCS inhibition, decreasing IL-1β production and increasing TNFα production (Fig. 3E). High concentrations of DEM strongly decreased the production of both cytokines although this may be confounded by effects on cell viability (Fig. S2). The more potent effect of DEM on cytokine production was reflected by GSH levels as DEM led to an earlier and stronger decrease in intracellular GSH levels than BSO (Fig. S3). Furthermore, DEM may have additional effects by reacting with other thiols and/or activating the transcription factor NRF2 (14). All together, these data suggest that modulation of GSH metabolism strongly affects cytokine production although the effect is dose-dependent.

Surprisingly, increasing GSH synthesis by the addition of the precursor N-acetyl-cysteine (NAC) also strongly decreased production of both cytokines, as seen from Fig. 3F. This suggests that both inhibition and inducing GSH synthesis can decrease cytokine production.

GSH Metabolism Affects Cytokine Production Through Different Mechanisms.

To unravel how modulation of GSH metabolism affects Bb-induced cytokine production, we first analyzed the effect of modulators of GSH metabolism on mRNA transcription. For BSO and NAC, the highest concentrations were used to detect the maximal effect. For DEM, the lower concentration was selected to rule out cytotoxic effects. As seen in Fig. 4A, a substantial increase in mRNA levels of IL1B was seen after 4 h and 24 h of stimulation with Bb. However, these mRNA levels were unaffected by treatment with the GSH modulators.

Fig. 4.

Effect of modulators of GSH metabolism on cytokine transcription and translation. (A and B) mRNA expression of IL1B (A) and TNF (B) in PBMCs from healthy volunteers (n = 6) pretreated for 1 h with DEM, BSO, or NAC and stimulated for 4 h and 24 h with Bb. Box plot indicates median ± min/max. (C and D) Measurement of (total) IL-1β (C) and TNFα (D) protein levels in cellular lysates of PBMCs from healthy volunteers (n = 6) pretreated for 1 h with DEM, BSO, or NAC and stimulated for 24 h with Bb. (E–H) Protein levels of (total) IL-1β and TNFα in cellular lysates (intracellular) (G and H) or cell-free supernatants (secreted) (E and F) of PBMCs from healthy volunteers (n = 6) pretreated for 1 h with the glutathionylation inducer 2-AAPA and stimulated for 24 h with Bb. Vehicle controls were included as needed (indicated by +Veh). Bar graphs represent mean ± SEM. *P > 0.05.

For TNF, a significant increase in mRNA expression was seen after 4 h, but not 24 h of stimulation with Bb (Fig. 4B), suggesting that induction of TNF mRNA is rapid but short-lasting. Depletion of GSH by DEM was able to mildly potentiate mRNA levels, yet the other inhibitors showed no significant effect.

Taken together, these data indicate that the induction of IL1B and TNF mRNA by Bb is largely unaltered by modulation of GSH metabolism. This suggests that the previously shown effects on cytokine secretion mostly take place posttranscriptionally.

To further elucidate this, we measured the intracellular levels of (pro)IL-1β and TNFα to determine whether mRNA was effectively being translated into protein. As shown in Fig. 4C, stimulation with Bb drastically increased levels of intracellular (pro)IL-1β. However, similar to the transcriptome data, intracellular concentrations were not significantly altered by the addition of GSH modulators. This suggests that the effects of GSH modulation on IL-1β secretion occur posttranslationally.

In contrast to the IL-1β data, intracellular TNFα levels were substantially affected by modulation of GSH metabolism. Similar to what was seen for secreted TNFα levels, stimulation with Bb led to increased intracellular TNFα levels, which were potentiated by the addition of DEM and inhibited by the addition of NAC (Fig. 4D). BSO had no significant effect on intracellular TNFα levels. The effect of DEM treatment could at least partially be explained by increased mRNA levels. However, NAC treatment did not significantly affect mRNA levels of TNF, while completely shutting down intracellular levels. This suggests that the inhibitory effect of this compound is due to an effect on the translation of TNF mRNA into protein.

Taken together, these data suggest that modulation of GSH metabolism affects cytokine production through different mechanisms, for a large part taking place at the (post)translational level. Therefore, we hypothesized that the process of glutathionylation may play a role. Glutathionylation is a posttranslational modification in which a GSH molecule binds directly to a protein. To determine whether glutathionylation plays a role in Bb-induced cytokine production, we made use of 2-AAPA, an inhibitor of glutaredoxin-1 (15), which is expected to increase glutathionylation. As shown in Fig. 4 E and F, a high dose 2-AAPA decreased secretion of both IL-1β and TNFα. However, similar to the GSH modulators, intracellular levels of (pro)IL-1β remained unchanged (Fig. 4G), while intracellular levels of TNFα were affected to the same extent as secreted levels (Fig. 4H). This suggests that 2-AAPA modulates cytokine production in a similar fashion as modulators of GSH metabolism, supporting the hypothesis that glutathionylation mediates the effect of GSH on cytokines.

Overall, these data show that modulation of GSH metabolism substantially affects the secretion of IL-1β and TNFα through different mechanisms. For IL-1β, GSH seems to affect activation and/or processing of IL-1β. For TNFα, decreasing GSH levels appeared to induce mRNA transcription, while increasing GSH levels interfered with mRNA translation. In both cases, these alterations may be mediated by altered levels of glutathionylation.

GSH Metabolism Is Altered in Patients with EM.

To further validate our in vitro data, we performed metabolome analysis on serum of patients with EM. First, pathway analysis was performed to compare patients with EM to healthy controls. As shown in Fig. 5A, the aminoacyl-tRNA biosynthesis pathway, involved in mRNA translation, was the most significantly affected pathway in EM patients. Next to this, arachidonic acid metabolism was found to be altered in these patients, in accordance with our monocyte metabolome data. Further supporting our monocyte metabolome data, GSH metabolism was among the top 10 most significantly affected pathways. In addition, cysteine metabolism and methionine metabolism, both upstream from GSH metabolism, were found in the top 10 most affected pathways. Taken together, this suggests that our model of primary human monocytes exposed to Bb correlates very well to the circulating metabolome in vivo.

Fig. 5.

Metabolome analysis on serum samples of acute EM patients compared with other infections. (A) Scatter plot of KEGG metabolic pathways in serum samples from patients with erythema migrans (EM) (n = 10) versus healthy controls (HC) (n = 10), showing enrichment log(P) value (y axis) and pathway impact, determined by topological analysis (x axis). Node color is based on its P value, and the node size is based on pathway impact values. The top 10 most significantly affected metabolic pathways are included for reference. (B) Heat map depicting the top 50 most differentially regulated metabolites, determined by ANOVA, between acute EM patients (n = 10), patients with acute Gram-negative infection (G-neg, n = 5), patients with acute Gram-positive infection (G-pos, n = 5), and healthy controls (HC, n = 10). #, Fig. S1, metabolite 1; ¥, Fig. S1, metabolite 11.

Next, we compared patients with EM and healthy controls to patients with acute bacterial infection. Fig. 5B depicts the top 45 most differentially expressed metabolites between the four groups. Notably, one cluster was seen, indicated in blue, with metabolites significantly increased in patients with Bb infection, while remaining unchanged in patients with other infections. The top five metabolites in this cluster turned out to be γ-glutamyl amino acids (Fig. 5B, indicated by “¥”), breakdown products from γ-glutamyl transpeptidase activity (GGT). This is again supportive of our previous data as GGT is involved in the recycling of oxidized GSH (see Fig. S1 for reference). Interestingly, l-methionine, a possible precursor for GSH, was found to be strongly decreased compared with both healthy controls and patients with other infections (Fig. 5B, indicated by “#”).

Apart from the γ-glutamyl amino acids, several eicosanoid metabolites were found to be significantly elevated in EM patients compared with both healthy controls and patients with other infections.

Together, these findings indicate that strong and specific alterations in GSH metabolism, as well as eicosanoid/arachidonic acid metabolism, are found systemically after Bb infection.

To further support these findings, we made use of publicly available transcriptome data from PBMCs of patients with EM to analyze gene expression levels of GSH-related genes (16). Noteworthy, several genes in GSH metabolism were differentially expressed in patients with Lyme disease in the acute stage (CBS, GSS, GGT1) but also up to 3 wk (MAT2A, GSS, GGT1) and even 6 mo (CBS, GGT1) after diagnosis (Fig. S4). GGT1, the gene encoding for γ-glutamyl transpeptidase (GGT) was the only gene significantly altered in Lyme patients at all time points. However, contrary to the increased levels of γ-glutamyl amino acids found in serum, gene expression of GGT was consistently down-regulated. This suggests a feedback mechanism in which high levels of breakdown products down-regulate gene expression. Nevertheless, these data support the conclusion that GGT activity is altered in patients with Bb infection.

Taken together, we show that GSH metabolism is strongly affected in patients with EM, as seen from altered metabolite levels in serum and persistently altered transcriptional activity in PBMCs.

Discussion

In this study, we have shown that exposure of primary human monocytes to Bb results in significant and specific changes in GSH metabolism. In addition, we show that modulating GSH metabolism significantly affects cytokine production, possibly through glutathionylation. Finally, we provide evidence that GSH metabolism is altered in patients with EM and that these alterations might persist for months after the initial infection.

Previous studies have shown the important role for metabolic pathways in the inflammatory response to pathogens. In this study, we investigated which metabolic pathways are involved in the host response against Bb. Both our metabolome analysis on monocytes and our analysis on serum samples from patients showed altered GSH metabolism after Bb infection. Specifically, intracellular levels of reduced GSH were dramatically increased in Bb-stimulated monocytes.

Considering the antioxidative properties of GSH, it was interesting to find that stimulation with Bb lowered ROS generation in response to a secondary stimulus. Surprisingly, this dampening of the oxidative response appeared independent of GSH levels. Nevertheless, this may have important implications as ROS is an important component of the antimicrobial defense and decreased ROS production might increase susceptibility to other pathogens.

Despite not affecting ROS generation, we found that modulation of GSH metabolism substantially influenced the production of IL-1β and TNFα, two crucial innate cytokines. These findings are in accordance to a recent study by Diotallevi et al. (17), who showed that GSH influences cell signaling and inflammation independent from its antioxidative properties. Interestingly, both inducing GSH synthesis and depleting GSH strongly decreased cytokine production while moderate inhibition increased cytokine production. Multiple studies have previously suggested a role for GSH levels in cytokine production (18–20), yet the underlying mechanism has so far not completely been elucidated.

To provide more insight into how GSH levels affect Bb-induced cytokine production, we examined the effect of GSH modulators on cytokine mRNA levels and intracellular levels. For TNFα, we found that altering GSH levels significantly affected TNFA mRNA levels, in accordance to what was found by Fratelli et al. (21). However, the mRNA levels could not fully account for the effect of the GSH modulators, indicating that part of the effect of GSH metabolism occurs posttranscriptionally. Elaborating on this, we found that modulation of GSH metabolism likely also affects translation of TNF mRNA into protein as intracellular protein levels were affected by the inhibitors in a similar fashion as secreted levels of TNFα. We hypothesize that this effect might be mediated by glutathionylation: the binding of GSH to a protein as a posttranslational modification. Glutathionylation is increasingly being recognized as a modification with important functional consequences (22–24). A recent proteomic analysis on potential targets of glutathionylation found a strong enrichment of proteins involved in RNA processing and translation (25). Alterations in the levels of glutathionylation may therefore affect mRNA translation, thereby influencing protein synthesis. Alternatively, glutathionylation may affect TNFα release through peroxiredoxin-2, as shown by Salzano et al. (26)

Contrary to TNFα, GSH-mediated effects on IL-1β secretion appear to take place at the posttranslational level as intracellular levels of (pro)IL-1β were unaffected by GSH modulation. We believe that glutathionylation may again play a role here as one of the proteins found to be glutathionylated was caspase-1, an enzyme known to be crucial for cleaving pro-IL-1β into active IL-1β and subsequent secretion of the cytokine. Meissner et al. (27) showed that glutathionylation of caspase-1 was an inhibitory modification as blockade of the GSH-binding sites significantly increased IL-1β secretion.

To confirm the role for glutathionylation in TNFα and IL-1β production, we made use of an inhibitor of glutaredoxin-1. This enzyme is involved in reversal of glutathionylation, and inhibition of glutaredoxin-1 is therefore used to increase levels of glutathionylation (15). Importantly, we found that inducing glutathionylation inhibited cytokine production in a similar fashion as GSH modulation, affecting TNFα at the level of protein synthesis and IL-1β at the posttranslational level.

Taken together, these data suggest that modulation of GSH metabolism significantly affects secretion of cytokines, possibly mediated by protein glutathionylation. As Bb exposure leads to a substantial increase in intracellular GSH levels, this might also affect the levels of glutathionylation. In fact, induction of GSH levels by Bb may be a mechanism to down-regulate TNFα production as Bb is known to be a poor inducer of TNFα compared with other pathogenic stimuli (13). This is noteworthy as TNFα is known to be a central player in the pathogenesis of many infections and inflammatory diseases (28). Indeed, inhibiting GSH significantly increased TNFα production while potentiating GSH levels completely shut down TNFα synthesis. In contrast, both inhibition and induction of GSH negatively affected IL-1β production, suggesting that Bb induces GSH levels to an optimal level for IL-1β secretion. Accordingly, Bb is known to be a very potent inducer of IL-1β, and this cytokine is known to be an important driving force in Lyme arthritis (29, 30). Taken together, this suggests that Bb-induced changes in GSH metabolism might play a role in skewing the cytokine profile from TNFα toward IL-1β.

Next to our in vitro data, we found evidence for specific alterations in GSH metabolism in patients with a Bb infection. Metabolome analysis showed increased levels of metabolites related to GSH metabolism in serum of patients with EM compared with both healthy controls and patients with other infections. Supporting this, transcriptome data showed substantial and long-lasting changes in mRNA expression for GSH-related genes in PBMCs from patients with EM compared with healthy controls. Noteworthy, one of the genes most significantly affected in patients with EM, was GGT1, encoding γ-glutamyl transpeptidase (GGT). This corresponded well to our data as we found increased levels of GGT breakdown products, γ-glutamyl amino acids, in serum of patients with EM. Serum GGT levels are regularly measured as a marker of liver function and have also been studied in patients with EM although increased GGT levels were only found in a minority of cases (31). However, standard GGT measurements only account for excreted enzyme while GGT is mainly membrane-bound (32). Therefore, we hypothesize that the elevated levels of γ-glutamyl amino acids in serum of patients with EM are due to increased activity of membrane-bound GGT. Taken together, these findings point toward an important role for GGT in the response to Bb infection.

Our data suggesting an important role for GSH metabolism in Bb infection are supported by several previous studies. Recently, Casselli et al. (33) reported two GST genes (GSTT1 and GSTM1) to be among the most significantly affected genes after Bb exposure in primary human astrocytes. Next to this, a recent genome-wide association study (GWAS) found a genetic variant in MAT2B associated to Bb seropositivity (34). MAT2B encodes for the regulatory subunit of MAT2A, one of the upstream enzymes in GSH metabolism. In our study, we found gene expression of MAT2A to be affected by Bb infection in PBMCs of patients with EM. This suggests that GSH metabolism may also play a role in antibody production against Bb although this will require further investigation.

Taken together, these data show that infection with B. burgdorferi strongly modulates GSH metabolism both in vitro and in patients with EM. As we have shown that GSH metabolism plays a crucial role in B. burgdorferi-induced cytokine production, these findings provide more insight into the pathogenesis of Lyme disease and may help explain the variability in clinical signs and disease outcome.

Methods

Primary Human Monocytes.

PBMCs were isolated from blood donated by healthy male volunteers (n = 5) after written informed consent. Ethical approval was obtained from the committee on research involving human subjects (CMO) Arnhem-Nijmegen (NL32357.091.10).

CD14⁺ monocytes were isolated from PBMCs by positive selection using MACS CD14⁺ magnetic beads (Miltenyi Biotec) according to the manufacturer’s instructions. Cells were resuspended in RPMI (RPMI medium 1640, no glucose; Thermo Fisher Scientific) supplemented with 5.5 mM glucose (Sigma-Aldrich), 1 mM pyruvate (sodium pyruvate; Thermo Fisher Scientific), 10% pooled human serum, 1% Hepes (Sigma), and 1% gentamycin and seeded in six-well plates (3 × 10⁶ cells per well). Cells were left to adhere for 30 min and stimulated as described.

Peripheral Blood Mononuclear Cells.

PBMCs were isolated from buffy coats from healthy volunteers obtained from the Sanquin blood bank after informed consent. All human experiments were conducted according to the principles of the Declaration of Helsinki. The study was approved by the Arnhem-Nijmegen ethical review board. Briefly, blood was diluted with sterile PBS (1:1), and a density centrifugation was applied over Ficoll-Paque (Pharmacia Biotech). Next, the interphase containing the PBMCs was collected and washed with ice-cold PBS, and cells were resuspended in medium (RPMI 1640, without glucose, without glutamine; MP Biomedicals) supplemented with 5.5 mM d-glucose (Sigma-Aldrich), 0.2 mM glutamine (glutaMAX; Thermo Fisher Scientific), and 0.1 mM pyruvate, 1% Hepes, and 1% gentamycin.

B. burgdorferi Spirochetes.

B. burgdorferi, ATCC strain 35210 [American Type Culture Collection (ATCC)] was cultured at 24 °C in Barbour–Stoenner–Kelley (BSK)-H medium (Sigma-Aldrich) supplemented with 6% rabbit serum until spirochete growth commenced. Cells were then grown at 34 °C to late logarithmic phase, at which point the spirochetes were checked for motility by dark-field microscopy and harvested. Spirochetes were quantified using a Petroff–Hauser counting chamber, washed with PBS, and stored at −80 °C.

Serum Sample Patients.

Serum samples were obtained from patients at the University Hospital of Infectious Diseases, Cluj-Napoca, Romania. The study protocol was approved by the local medical ethics committee of the University Hospital of Infectious Diseases, Cluj-Napoca (2013/01). Written informed consent was obtained from all participants. Patients with EM (n = 10) were clinically diagnosed by an infectious disease specialist and confirmed by an independent Lyme disease expert. Serum samples were taken before onset of treatment. For comparison, patients with acute bacterial sepsis (n = 10) from the same center were included, as well as healthy controls (n = 10).

Metabolome Analysis.

For the primary monocyte analysis, cells were stimulated for 24 h with B. burgdorferi [multiplicity of infection (MOI) = 0.05] or medium control. After incubation, cell-free supernatants were collected, and cells were scraped and spun down, and dry pellets were snap frozen and stored at −80 °C. Serum samples were stored at −20 °C before analysis.

Metabolomic analysis was performed by Metabolon Inc. In short, proteins were precipitated with methanol, and the resulting extract was divided into five fractions: two for analysis by two separate reverse phase (RP)/ultra-performance liquid chromatography (UPLC)-tandem mass spectrometry (MS/MS) methods with positive ion mode electrospray ionization (ESI), one for analysis by RP/UPLC-MS/MS with negative ion mode ESI, one for analysis by hydrophilic interaction chromatography (HILIC)/UPLC-MS/MS with negative ion mode ESI, and one sample reserved for backup. Briefly, samples were placed on a TurboVap (Zymark) to remove the organic solvent. All methods utilized Waters ACQUITY ultra-performance liquid chromatography (UPLC) and a Thermo Scientific Q-Exactive high resolution/accurate mass spectrometer interfaced with a heated electrospray ionization (HESI-II) source and Orbitrap mass analyzer operated at 35,000 mass resolution. The sample extract was dried and then reconstituted in solvents compatible to each of the four methods. One aliquot was analyzed using acidic positive ion conditions, chromatographically optimized for more hydrophilic compounds. The second aliquot was also analyzed using acidic positive ion conditions; however, it was chromatographically optimized for more hydrophobic compounds. The third aliquot was analyzed using basic negative ion optimized conditions using a separate dedicated C18 column. The fourth aliquot was analyzed via negative ionization following elution from a HILIC column (Waters UPLC BEH Amide 2.1 × 150 mm, 1.7 µm) using a gradient consisting of water and acetonitrile with 10 mM ammonium formate, pH 10.8. Raw data were extracted, peak-identified, and quality control processed using Metabolon’s hardware and software. Compounds were identified by comparison with library entries of purified standards or recurrent unknown entities. Peaks were quantified using area-under-the-curve values, which were rescaled to the median and normalized by Bradford protein concentration. Missing values were imputed with the minimum.

Stimulation Experiments.

For measurements of cytokines and metabolic parameters, cells were seeded in duplicate in round-bottom 96-well plates (5 × 10⁵ cells per well). Cells were pretreated with one of the following inhibitors: 3-deazaneplanocin A (Cayman Chemical), dl-buthionine-sulfoximine (Sigma), di-ethyl maleate (Sigma), N-acetyl cysteine (Sigma), mercaptosuccinic acid (Sigma), OU749 (Cayman Chemical), 2-AAPA hydrate (Sigma), or vehicle control [i.e., RPMI or dimethyl sulfoxide (DMSO) (WAK-Chemie Medical GmbH)] for 1 h and then stimulated with B. burgdorferi for 24 h, unless otherwise indicated. After incubation, plates were spun down, and cell-free supernatants were collected and stored at −20 °C until assayed. For intracellular cytokine measurements, cells were lysed in 0.5% Triton-X (20 µL per well), and lysates were stored at −20 °C until assayed.

Measurement of Reactive Oxygen Species.

Generation of reactive oxygen species by PBMCs was measured using a luminol-based chemiluminescent assay. Briefly, PBMCs were collected after stimulation and added in sixfold to a white 96-well plate (1 × 10⁵ cells per well). ROS production was induced in four wells by the addition of 3 mg/mL serum-opsonized zymosan in HBSS; the two remaining wells served as controls. Luminol (10 mM), which was added to all wells, is oxidized by ROS to produce the luminescent intermediate luminophore. Luminescence, correlating to total ROS production (both intra- and extracellular), was measured continuously at 425 nM for 1 h using an Infinite 200 PRO microplate reader (Tecan).

Cytokine Measurements.

Cytokine concentrations in cell culture supernatants were measured by sandwich ELISA using commercial kits specific for IL-1β and TNFα (R&D Systems) according to the manufacturer’s instructions. Cell lysates were spun down before measuring to remove insoluble material. Absorbance was measured using an Infinite 200 PRO microplate reader (Tecan).

mRNA Isolation and RT-PCR.

After stimulation, cells were lysed and homogenized in TRIzol (Thermo Fischer), and RNA was isolated according to the manufacturer’s instructions. Isolated RNA was checked for purity and transcribed using an iScript cDNA Synthesis Kit (Bio-Rad). For quantitative polymerase chain reaction (qPCR), Power Sybr Green PCR Master Mix (Applied Biosystems) was used with a 7300 Real-time PCR system (Applied Biosystems). Primers used were as follows: B2M (housekeeping gene) [forward (Fw): ATGAGTATGCCTGCCGTGTG, reverse (Rv): CCAAATGCGGCATCTTCAAAC], IL1B (Fw: GCCCTAAACAGATGAAGTGCTC, Rv: GAACCAGCATCTTCCTCAG), and TNF (Fw: GAGGCCAAGCCCTGGTATG, Rv: CGGGCCGATTGATCTCAGC).

Transcriptome Analysis.

Previously published RNA sequencing data of PBMCs from patients (n = 29) with EM and healthy controls (n = 13) (16) were obtained from the publicly available National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database (accession number GSE63085). Expression of selected genes was compared between EM patients at different time points and healthy controls by Kruskal–Wallis one-way ANOVA with Dunn’s post hoc test using R Software for Statistical Computing, version 3.2.4.

Analysis of Metabolome Data.

Pathway analysis and statistical analysis were performed using Metaboanalyst 3.0 (33) on 332 metabolites (monocytes) or 638 metabolites (serum samples) with available Human Metabolome Database (HMDB) identifiers. For pathway analysis, metabolites were mapped to Kyoto Encyclopedia of Genes and Genomes (KEGG) metabolic pathways, and quantitative pathway enrichment and pathway topology analysis were performed. For comparison of Bb-induced metabolic changes to metabolic changes induced by other TLR-ligands, a parallel dataset of primary monocytes stimulated with LPS and Pam3Cys was used (34). Samples were acquired in an identical manner, and sample analysis was performed simultaneously. For all conditions, fold changes in metabolite levels relative to untreated control (RPMI) were calculated. Further statistical details can be found in the appropriate figure legends.

Statistics.

Statistics for measurements of cytokines and metabolic parameters were performed using GraphPad Prism version 5.03 for Windows (GraphPad Software). Data represent mean ± SEM of n different donors. Unless otherwise stated, means were compared using the nonparametric Wilcoxon matched-pairs signed ranks test, with two-tailed significance level set as P > 0.05. Further statistical details can be found in the appropriate figure legends.

Data Availability.

The metabolome data in this paper have been deposited in the publicly available MetaboLights database (https://www.ebi.ac.uk/metabolights/, accession no. MTBLS625).