Manipulating Glucose Metabolism during Different Stages of Viral Pathogenesis Can Have either Detrimental or Beneficial Effects Free

Virus infections cause tissue damage in several ways, one of which is to induce an inflammatory reaction orchestrated by T cells that respond to viral Ags. One such example is the blinding immuno-inflammatory reaction called stromal keratitis (SK), which occurs in the cornea of the eye following infection with HSV (1, 2). In such reactions, the proinflammatory effector T cells may be more tissue damaging if regulatory components of immunity, such as certain cytokines or cells with regulatory functions, are deficient (3–6). Thus, one aim of therapy with these usually chronic tissue-damaging lesions is to shift the balance of different components involved in the immune response to the infection. Few, if any, effective therapies are readily available to achieve this objective. However, recent studies in the field of cellular metabolism have drawn attention to the fact that nutrient uptake and utilization may differ among cell types involved in immune responses (7–9). Moreover, it has become evident that manipulating metabolic pathways represents a potential means of rebalancing immune responses, and this approach is being mainly explored in the cancer and autoimmunity fields where the imbalance largely involves different subsets of T cells (10–14).

Application of the metabolic reprogramming approach has focused on manipulating glucose and fatty acid metabolism, which can show major differences between immune cells involved in reactions (15). However, few, if any, studies so far have focused on infectious diseases, but this topic is highly relevant because many chronic tissue-damaging infections are not subject to control by effective vaccines, or by readily acceptable (or affordable) means of therapy. In fact, targeting metabolic events represents a logical approach to pathogen control because many cause major changes in metabolism not only in the cells they infect, but also impact on the function of distant uninfected organs such as the liver, kidney, cardiovascular system, and even the brain (16). Some of the general physiological consequences of systemic infections has been highlighted by recent studies (16, 17). However, the general topic of how virus infections, particularly those that cause local infections, influence physiological responses is still poorly understood. Our present studies record some metabolic consequences of local infections in the eye with HSV.

Our results show that ocular HSV infection in mice led to increased fed and fasted blood glucose levels at the time when the virus no longer persists in ocular tissues. In addition, CD4 T cells from infected mice showed increased glucose uptake both at the corneal lesion site and in the draining lymph nodes. The CD4 T cells from HSV-infected animals were highly metabolically active and displayed increased glucose uptake in vitro compared with T cells from naive animals. In vitro experiments also indicated that the effector function of inflammatory T cells was dependent on glucose concentration. Moreover, inhibition of glucose uptake by 2-deoxy-glucose (2DG) limited the differentiation of effector T cells in vitro. In contrast, regulatory T cells (Treg) were unaffected by 2DG in vitro. Finally, and of potential therapeutic relevance, in vivo administration of 2DG resulted in diminished SK lesions, a consequence of reduced effector T cell responses. Taken together, we show that local infection with HSV results in changes in glucose homeostasis causing increased blood glucose levels, which may act to stimulate the generation and sustenance of inflammatory CD4 effector T cells, which, in the special environment of the eye, can result in damaging consequences. Although changes in blood glucose levels were not evident during the acute phase of ocular infection, therapy with 2DG during that phase resulted in death from herpes encephalitis in many animals. Possible explanations for these findings are discussed.

Female C57BL/6 mice were purchased from Harlan Sprague-Dawley (Indianapolis, IN), BALB/c DO11.10 RAG2^−/− mice were purchased from Taconic and kept in a pathogen-free facility where food, water, bedding, and instruments were autoclaved. All the animals were housed in American Association of Laboratory Animal Care–approved facilities at the University of Tennessee, Knoxville, TN. All investigations followed guidelines of the Institutional Animal Care and Use Committee, and adhered to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research. HSV-l RE strain was used in all procedures. The virus was grown in Vero cell monolayers (American Type Culture Collection, Manassas, VA), titrated, and stored in aliquots at −80°C until used.

Corneal infections of C57BL/6 were conducted under deep anesthesia induced by i.p. injection of tribromoethanol (Avertin). Mice were scarified on the cornea with a 27-gauge needle, and a 3 μl drop containing 1 × 10⁴ PFU of HSV-1 was applied to the eye. The eyes were examined on different days postinfection (pi) with a slit-lamp biomicroscope (Kowa Company, Nagoya, Japan), and the clinical severity of keratitis of individually scored mice was recorded as previously described (18). Briefly, the scoring system was as follows: 0, normal cornea; +1, mild corneal haze; +2, moderate corneal opacity or scarring; +3, severe corneal opacity but iris visible; +4, opaque cornea and corneal ulcer; +5, corneal rupture and necrotizing keratitis. The naïve uninfected mice were scarified on the cornea with a 27-gauge needle without the addition of any virus.

The 2DG (Sigma-Aldrich) was dissolved in PBS and administered i.p. at 250 mg/kg twice a day starting from either day 0 or day 5 pi until day 14 pi. The control group either received an equal volume of PBS or was left untreated. The dose of 2DG was based on preliminary dose titrations.

Blood glucose levels were measured from the tail blood at different time points pi using a Bayer Contour glucose meter and compared with naive uninfected animals. For fasting blood glucose levels, animals were fasted for 16 h in a clean cage with water followed by measurement of glucose levels in the blood from the tail. For cytokine measurements, serum was isolated from the blood collected using retro-orbital bleeding. At least 10 mouse cytokines were profiled using a multiplex platform and data were extracted based on cytokine-specific standards by Eve Technologies (Calgary, AB). Five independent serum samples were used from mice at different time points pi. Serum from uninfected mice after 4 d of scarification was used as control.

At day 15 pi, corneas were excised, pooled group wise, and digested with Liberase (Roche Diagnostics, Indianapolis, IN) for 45 min at 37°C in a humidified atmosphere of 5% CO₂. After incubation, the corneas were disrupted by grinding with a syringe plunger on a cell strainer, and a single-cell suspension was made in complete RPMI 1640 medium. The single-cell suspensions obtained from corneal samples were stained for different cell surface molecules for FACS analyses. Draining cervical lymph nodes (DLNs) were obtained from mice sacrificed at 15 days pi and single-cell suspensions were used. All steps were performed at 4°C. Briefly, cells were stained with respective surface fluorochrome-labeled Abs in FACS buffer for 30 min, then stained for intracellular Abs. Finally, the cells were washed three times with FACS buffer and resuspended in 1% paraformaldehyde. The stained samples were acquired with a FACS LSR II (BD Biosciences, San Jose, CA), and the data were analyzed using FlowJo software (Tree Star, Ashland, OR). To determine the number of IFN-γ–producing T cells, intracellular cytokine staining was performed. In brief, corneal cells were either stimulated with PMA (50 ng) or ionomycin (500 ng) for 4 h in the presence of brefeldin A (10 μg/ml) in U-bottom 96-well plates (18). After this period, Live/Dead staining was performed followed by cell surface and intracellular cytokine staining using a Foxp3 intracellular staining kit (eBioscience) in accordance with the manufacturer’s recommendations. Dead cells were gated out using Live/Dead staining. Cells are mentioned as Th1 if they are CD4⁺ IFN-γ⁺ and Treg if they are CD4⁺ Foxp3⁺.

For 2-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino]-2-deoxyglucose (2-NBDG) uptake in vivo, mice were injected i.v. with 100 μg 2-NBDG per mouse diluted in PBS to either naive C57BL/6 animals or day 15 pi animals. Then 15 min following the injection, DLNs were collected and single-cell suspensions were stained as described above and analyzed using flow cytometry (19).

CD4 (RM4-5), CD45 (53-6.7), CD11b (M1/70), Ly6G (1A8), F4/80 (BM8), IFN-γ (XMG1.2), CD25 (PC61), CD44 (IM7), Foxp3 (FJK-16S), anti-CD3 (145-2C11), anti-CD28 (37.51), GolgiPlug (brefeldin A), and anti–pro-IL-1 β (NJTEN3) from either eBioscience or BD Biosciences. Anti-mouse phospho-S6 ribosomal (D57.2.2E) was from Cell Signaling and hGlut1 from R&D Systems. PMA and ionomycin were from Sigma-Aldrich. The Live/Dead staining kit and 2-NBDG were from Life Technologies. Recombinant IL-2, IL-12, IL-6, and TGF-β were from R&D Systems. Glucose-free RPMI 1640 media (Life Technologies) was prepared using dialyzed FBS and glucose (Sigma-Aldrich) was added at concentrations of 0.1–20 mM.

At day 2 postocular infection with HSV-1, the corneas were isolated and two corneas were pooled per sample per group. Naive CD4 T cells or cells activated in vitro with or without 2DG were taken with at least 100,000 cells per sample. Total RNA from corneal and isolated T cell populations was isolated using mirVana miRNA isolation kit (Ambion). Liver (≤30 mg) was isolated at day 0, 4, 8, and 15 pi, and total RNA was extracted using RNeasy Fibrous Tissue Mini Kit (Qiagen) as per the manufacturer’s recommendation. cDNA was made with 500 ng of RNA (corneal samples) and entire RNA (isolated T cells) by using oligo(dT) primer and ImProm-II Reverse Transcription system (Promega). Taqman gene expression assays for Glut-1 (SLC2A1), hexokinase 1, hexokinase 2, Il-6ra, Tnfrsf1b, Stat3, Ifngr1, Cyp7a1, Rhoc, Il-1a, Ifnb, Tnfa, and Il-1b were purchased from Applied Biosystems and quantified using a 7500 Fast Real-Time PCR system (Applied Biosystems). The expression levels of different molecules were normalized to β-actin using a Δ cycle threshold calculation. Relative expression between control and experimental groups was calculated using the 2^−ΔΔCt × 1000 formula.

CD4⁺ T cells (total or naive) were purified from single-cell suspensions of pooled DLNs and spleens from HSV-infected or naive C57BL/6 mice using a mouse total or naive CD4⁺ T cell isolation kit according to the manufacturer’s instructions (Miltenyi Biotec, Auburn, CA). The purity was achieved at least to an extent of 95%.

For glycolysis measurements using seahorse extracellular flux analyzer, naive CD4 T cells were used in Treg and Th1 differentiation. For experiments evaluating the effects of glucose concentrations or the effects of 2DG, splenocytes from naive DO11.10 RAG2^−/− mice were used as a precursor population for the induction of Treg and Th1 cells as previously described (18). Briefly, 1 × 10⁶ splenocytes after RBC lysis and several washings or 1 million naive CD4 T cells were cultured in 1 ml glucose-sufficient RPMI 1640 media or glucose-free RPMI 1640 media (dialyzed serum) containing rIL-2 (100 U/ml) and TGF-β (1–5 ng/ml) in the presence or absence of various concentrations of glucose (0.5–20 mM) with plate-bound anti-CD3/CD28 Ab (1 μg/ml) for 5 d at 37°C in a 5% CO₂ incubator. After 5 d, samples were characterized for Foxp3 intracellular staining (eBioscience staining kit), analyzed by flow cytometry. For Th1 differentiation, cells were cultured in the presence of recombinant mouse IL-12 (5–10 ng/ml) and anti–IL-4 (10 μg/ml). After 5 d, samples were restimulated with PMA or ionomycin and analyzed for the production of IFN-γ by intracellular cytokine staining kit (BD Biosciences) using a flow cytometer. The 250 μM 2DG dose was chosen based on the preliminary dose-response experiments and previous reports (20).

Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) values were measured using a Seahorse XF24 metabolic analyzer. Briefly, total CD4 T cells were purified from either infected or naive female B6 mice, or Treg/Th1 cells were differentiated and expanded from naive CD4 T cells in vitro as described above. A total of 1 × 10⁶ cells per well were plated on XF24 plates (Seahorse Bioscience) precoated with 0.5 mg/ml poly-d-lysine (Sigma-Aldrich). Cells were maintained in RPMI 1640 media (Corning) supplemented with 1 mM sodium pyruvate (Sigma-Aldrich) and 10% FBS. Before analyzing, cells were spun down and 530 μl of XF media (with or without glucose) was added to each well, followed by incubation for 30 min in a CO₂-free incubator at 37°C. Seahorse analyzer was then run per the manufacturer’s protocol with oligomycin (1 μM), carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (1 μM), and antimycin A (1 μM) injected through ports A, B, and C respectively for mitochondrial stress test, and glucose (10 mM), oligomycin (1 μM), and 2DG (10 mM) for glycolysis stress test (21).

Virus titers were measured in the brain and corneas of HSV-infected mice as described previously (22). Virus titers in all samples were measured using standard plaque assay as described previously (22, 23).

Female C57BL/6 mice were used at 8–16 wk of age. Bone marrow–derived macrophages (BMDMs) were generated as described previously (24) and grown in RPMI 1640 with 10% FCS. The media was supplemented with M-CSF (10 ng/ml). Cells were 98% pure for macrophages (CD45⁺ CD11b⁺ F4/80⁺) and plated out on day 10 at 100,000 cells per well and stimulated with LPS (20 ng/ml) for 24 h.

Statistical significance was determined by either Student t test (comparing two groups) or one-way ANOVA (comparing three or more groups). A p value <0.05 was regarded as a significant difference between groups: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. GraphPad Prism software (GraphPad Software, La Jolla, CA) was used for statistical analysis.

To ascertain if ocular infection with HSV led to changes in glucose metabolism, blood glucose levels were measured in control and infected animals up to 15 d pi, the time when inflammatory responses to the ocular infection were at their peak. Elevated blood glucose levels were not observed in the initial period pi when the replicating virus was present in the eye. However, a moderate increase (∼1.5–2 fold) in blood glucose levels occurred in samples from day 8, a time when ocular inflammatory lesions first become evident in the corneal stroma, as well as at day 15 pi (Fig. 1A). However, at the latter time point the effect was more variable and differences were not significant in some experiments. At day 15, comparisons of blood glucose levels were also made between infected and uninfected animals following a 16 h fasting period. Once again, glucose levels were significantly higher (p < 0.01) in the infected animals compared with uninfected controls in both the fed and fasting states (Fig. 1B).

Previous studies indicated that low-grade inflammation can affect systemic glucose levels by inducing gene expression changes in the liver (25, 26). Accordingly, proinflammatory cytokine levels in the serum were measured at different time points pi. The results indicated that IL-6, TNF-α, and IFN-γ cytokine levels were significantly increased in serum samples of day 4 pi animals compared with uninfected controls (Fig. 1C). Measurement of gene expression changes in the liver at different time points pi indicated a significant increase in the expression of genes in the signaling pathways of proinflammatory cytokines at day 8 pi (∼3 fold), which included Il-6ra, Tnfr2, Ifn-γr1, and Stat3 (Fig. 1D). A recent report suggests that systemic inflammation suppressed cholesterol 7 α-hydroxylase (CYP7A1), the rate-limiting enzyme of the bile acid biosynthesis in the liver (25). This led to the accumulation of intermediate metabolites of the mevalonate pathway, resulting in stabilization of RHOC, a small GTPase induced by inflammation. The outcome was fasting hyperglycemia. We also quantified the expression of both CYP7A1 and RHOC in the liver at day 8 pi, the time when blood glucose was at its peak. Whereas the expression of CYP7A1 was significantly reduced, the expression of RHOC was higher in the liver at day 8 pi (Fig. 1E). These data indicate that ocular infection with HSV-1 results in a change in glucose metabolism likely by effects on the liver mediated by responses to inflammatory cytokines.

As the inflammatory response in the eye to HSV infection is mainly orchestrated by CD4⁺ T cells (1), changes in glucose uptake and glycolytic events were measured in CD4 T cells of animals with SK lesions and compared with CD4 T cells in uninfected animals. The first series of experiments compared glucose uptake in control uninfected and animals with lesions (day 15 pi), following the administration of a fluorescently labeled glucose analog 2-NBDG. Then 1 h later, the DLN were collected and the fraction of CD4⁺ T cells that took up 2-NBDG was measured by flow cytometry. The results indicated that the CD4⁺ T cells from mice with ocular lesions had significantly more cells (∼6-fold) that took up the glucose compared with uninfected controls (Fig. 2A, 2B). Of note, compared with uninfected animals, CD4⁺ T cells from infected mice contained 4-fold more cells with the effector memory phenotype and 7-fold more cells that produced IFN-γ when stimulated in vitro with PMA and ionomycin (Supplemental Fig. 1). The glucose uptake receptor GLUT1, as measured by flow cytometry, was also increased in CD4 T cells at day 15 pi compared with cells from uninfected controls (Fig. 2C).

Given that the CD4⁺ T cells from infected animals exhibited enhanced glucose uptake, a second set of experiments measured if these CD4⁺ cells also had changes in glucose metabolism. This was done using seahorse technology by determining the ECAR, which serves as an indicator of lactate production and glycolytic activity (21). CD4⁺ T cells isolated from the DLN of infected animals (day 15 pi) had a higher basal and maximal glycolytic rate (>3-fold) compared with CD4⁺ T cells from uninfected animals (Fig. 2D). Of note, measurements of basal OCR, which serves as an indicator mitochondrial respiration (21), indicated that the ratio of OCR to ECAR was significantly reduced (p < 0.05) in cells from infected animals. This result may indicate a cellular preference for glycolysis over mitochondrial respiration (Fig. 2E). These results imply that upon infection with HSV-1, CD4⁺ T cells increase glucose uptake and glucose utilization (glycolysis), which might be supported by higher glucose levels observed in animals with inflammatory lesions.

Because different T cell subsets participate in SK reactions (1), it was of interest to determine the type of T cell subset in which changes in glucose metabolism were occurring. The focus was on CD4 effector and Tregs because these are the two major subsets that influence the extent of SK lesions (3, 4). The two cell types were isolated from the DLN of day 15 infected mice, separating them based on their expression levels of the IL-2 receptor CD25, with cells that were CD25⁺ taken to be mainly Treg. Thus in separate experiments, we could show that on average ∼93% of the CD25⁺ T cells were Foxp3⁺ on day 15 pi. The CD4⁺CD25⁻ population was considered to be predominantly effector T cells (Supplemental Fig. 2). The uptake of 2-NBDG was significantly higher in the CD25⁻ population (mainly effector T cells) as compared with the CD25⁺ population (mainly Treg) (Fig. 3A).

Because the mTOR pathway is also involved in the control of glucose uptake and glycolysis in T cells (27), mTOR activity was compared in the two subsets. This was done by measuring the phosphorylation of S6 kinase (a component of the mTOR downstream signaling cascade) (28). Higher glucose uptake in the CD25⁻ CD4 effector population was also associated with higher mTOR activity compared with the activity of the CD25⁺ CD4 Treg population (Fig. 3B).

To further determine if glucose requirements differed between the Th1 effector and Treg subsets, experiments were done in vitro to generate Treg and Th1 populations from naive DLN populations. The populations obtained were ∼80% enriched for each subset. After 5 d, levels of glycolysis were compared in the two populations by using a glycolysis stress test that measures extracellular flux analysis (ECAR). As shown in Fig. 3C, the basal ECAR levels were elevated in the Th1 population compared with the Treg population (2.8-fold). In addition, the glycolytic capacity, as measured by an increase in ECAR levels following inhibition of ATP synthase (using oligomycin) (21), was around 3-fold higher in Th1 compared with the Tregs (Fig. 3C). Accordingly, the two major populations of CD4 T cells involved in SK lesions had distinct differences in glucose metabolism, a pattern of events noted in some autoimmune and neoplastic diseases (11, 13).

To begin to approach the question of the potential relevance of raised glucose levels during the inflammatory response to HSV infection, in vitro experiments were done to measure the influence of glucose levels on the efficiency of Th1 and Treg differentiation, cell types critically involved in SK reaction (1, 4). For this, naive CD4 T cells were TCR stimulated under either Treg- or Th1-inducing conditions after which the effect of supplementing cultures with different concentrations of glucose was measured. The levels of glucose evaluated varied from below physiological levels to hyperglycemia (0.5–20 mM). As shown in Fig. 4A, 4B, increasing glucose levels elevated the magnitude of Th1 responses, but had no significant influence on Treg responses, as noted previously (29, 30). Curiously, increasing glucose from physiological levels (∼5 mM) to that observed in vivo in the infected mice at day 8 pi (maximum of 10 mM) led to a significant increase (p < 0.05) in Th1 differentiation (Fig. 4C). The above observation supports the hypothesis that a 2-fold increase in the physiological levels of glucose in the periphery might be relevant in terms of enhancing the generation of Th1 cells in vivo.

To further measure the influence of glucose levels during the inflammatory response to HSV infection, experiments were done in vitro to measure the effect of the molecule 2DG, which inhibits glucose utilization and hence glycolysis. Naive CD4 T cells were activated in the presence of 2DG for 72 h followed by measurement of basal glycolysis using extracellular flux analysis. The data indicated a 3-fold reduction in basal ECAR levels, suggesting a decrease in basal glycolysis in cells activated in the presence of 2DG when compared with control-activated cells (Fig. 5A). Of note, no difference in cell death was observed at the concentration of 2DG used (data not shown). Collectively, the above results support the hypothesis that activated T cells upregulate glycolysis, which can be inhibited by 2DG.

Additional experiments were done to measure the expression of key genes involved in the glycolytic pathway both in the presence or absence of 2DG. Naive CD4 T cells were activated in vitro by stimulating them with anti-CD3/CD28 for 24 h and the effect of 2DG (250 μM) on gene expression changes was recorded. Compared to naive CD4 T cells, activated CD4 T cells expressed higher levels of several genes involved in glycolysis. These included hexokinase 1 (2-fold), hexokinase 2 (20-fold), and Glut1 (4-fold) (Fig. 5). However, when CD4 T cells were activated in the presence of 2DG, the genes involved in glycolysis remained at the levels or even below those observed in nonactivated control cells (Fig. 5B). The presence of 2DG in the cultures also switched off the raised mTOR activity, as measured by levels of phosphorylated S6 kinase using flow cytometry (Fig. 5C).

To measure the effect of inhibition of glycolysis on differentiation of the two CD4 T cell subsets of interest (Th1 and Treg), naive CD4⁺ T cells were cultured in Treg- or Th1-differentiating conditions for 5 d in the presence or absence of 2DG (250 μM). Although the differentiation of Treg was unaffected at the dose of 2DG used, a 20-fold reduction in the numbers of Th1 cells induced was observed in the presence of 2DG (Fig. 6). These results support the above findings that effector T cells use glycolysis to a higher extent than Treg. The results may also mean that 2DG therapy could be useful against SK by inhibiting T effectors but leaving Treg function intact. This notion is tested in the next section.

To measure the therapeutic potential of 2DG against SK, HSV-infected animals were given daily administrations of either 2DG or PBS (control) starting at day 5 pi. This is the time point when there is at best minimal replicating virus detectable in the infected corneas and early inflammatory reactions start to become evident. Animals were examined at day 15 pi to record and compare the severity of SK lesions. The results were clear cut with animals receiving 2DG therapy showing significantly reduced SK lesion severity (p < 0.01) compared with PBS-treated control animals (Fig. 7A). At day 15 pi, around 12% of 2DG treated animals showed a lesion score of ≥3 compared with 50% in control treated animals. At the termination of experiments on day 15 pi, pools of corneas were collected and processed to identify their cellular composition by FACS analysis. There was a reduction in the number of total CD4 T cells (∼10 fold) infiltrating the corneas of 2DG treated animals compared with control animals (Fig. 7B).

In parallel experiments of similar design, pools of corneas, DLNs, and spleens were collected at 15 d pi from 2DG-treated and control animals. Single-cell suspensions were stimulated in vitro with PMA and ionomycin to activate T cells and to record the numbers of cells that were either IFN-γ producers or expressed the transcription factor Foxp3. In the corneas, the number of CD4 T cells expressing IFN-γ was reduced ∼8-fold in 2DG compared with untreated controls (Fig. 7C). In the DLNs and spleen, 2DG treatment resulted in reduced frequency and numbers of IFN-γ–producing CD4 T cells by ∼2-fold. Although the frequency of Tregs in the DLN remained the same, their number decreased significantly. This might be explained by reduced inflammatory responses in treated animals. However, the number and frequency of Treg in the spleens of 2DG-treated animals remained unchanged (Fig. 7D).

Taken together, our results indicate that daily administration of 2DG starting at day 5 pi significantly diminished HSV-1 induced immunopathology along with a reduction in effector T cell numbers in both the cornea and lymphoid organs, which could in part contribute to reduced lesion severity.

Although we found no evidence for hyperglycemia in the acute phase of HSV infection, experiments were done to measure the effects of 2DG therapy at the time when the virus was actively replicating in the infected cornea. Animals were ocularly infected with HSV-1 and were either treated daily with 2DG or PBS control, starting from the day of infection. Under the infection conditions used, HSV infection of untreated animals failed to cause detectable illness or signs of encephalitis. However, around 40–50% (in three separate experiments) of 2DG-treated animals developed encephalitis and most had to be terminated by day 10 pi (Fig. 8A). By day 8 pi, affected animals became lethargic, lost weight, showed ruffled fur and hunched appearance, along with signs of incoordination. Brains were collected from encephalitic 2DG-treated animals to quantify the levels of virus present. High virus levels of HSV were detectable (>4 log) in the brain homogenates in all animals that showed signs of encephalitis by day 10 pi (Fig. 8B). These animals also had detectable virus in ocular swabs at day 6 pi, which is 1 d beyond the time when the virus is regularly present in untreated infected animals. Additionally, the levels of virus in ocular swabs were around 10-fold higher than in untreated mice (Fig. 8C). Of note, the virus could not be detected in the brains at day 10 pi or in the ocular tissue at day 6 pi in the control animals when infected with the same dose of virus that caused encephalitis in the 2DG-treated animals.

The reduction in antiviral inflammatory cells or mediators could explain the increased viral burden in the animals treated with 2DG. To test this possibility, pools of corneas were isolated from animals treated with 2DG or PBS controls at day 2 pi and were evaluated for the abundance of macrophages and neutrophils. In addition, the effector function of macrophages and neutrophils was measured by the expression of pro–IL-1β using flow cytometry. Although the number and frequency of pro–IL-1β–expressing neutrophils remained unchanged with 2DG treatment (Supplemental Fig. 3), the frequency and the number of pro–IL-1β–expressing macrophages in the cornea were significantly reduced in 2DG-treated animals (>3-fold) (Fig. 8D). This might account for the failure to stop the spread of virus to the CNS. To assess if molecules involved in antiviral responses were affected upon 2DG treatment, the expression of IFN-β, IL-1α, and TNF-α genes was determined in the corneas of both control and 2DG-treated animals at day 2 pi. The results indicated no significant differences in the expression of these molecules (Fig. 8E). Taken together, the increase in viral load in the cornea as well as the failure to stop the spread of virus to the brain could be in part explained by reduction in the innate effector function of macrophages, which also require glucose for their function as demonstrated in the next section.

Macrophages are reported to play an important role in early stages of infection by controlling HSV-1 replication and dissemination within the trigeminal ganglion (31). Hence, effects of 2DG on macrophage activation were measured using LPS stimulation of BMDMs. BMDMs were stimulated with LPS in the presence or absence of 2DG (250 μM) for 24 h, followed by measurement of pro–IL-1β expression using flow cytometry. The results indicated that stimulation of macrophages with LPS in the presence of 2DG resulted in a 2-fold reduction in both the frequency of macrophages expressing pro–IL-1β and expression of pro–IL-1β per cell basis (mean fluorescence intensity [MFI]) as compared with LPS stimulation alone (Fig. 8F), supporting previous reports (32). In conclusion, inhibition of glucose utilization in macrophages inhibited the effector function of macrophages, which might partially explain the high viral burden in the animals treated with 2DG.

Inflammatory reactions are likely to be prolonged and cause excessive tissue damage when the activity of the principal orchestrators, usually either CD4 or CD8 T cells, are not constrained by inhibitory molecules or by cells that function as regulators (3–6). This is seen in the viral infection model we have used in this report, wherein CD4 T cell–driven inflammatory reactions in the eye cause the chronic vision-impairing lesion SK. The challenge with SK is to understand the events that result in lesions and to find therapies that limit the severity and duration of ocular damage. In the present report, we make the unexpected finding that blood glucose levels change significantly during the course of infection. Whereas levels remained normal during the early phase of infection when the virus was actively replicating in the cornea, they increased around 2-fold during the time when inflammatory responses to the virus was occurring. We could show that glucose levels influenced the extent of induction of the inflammatory T cell subset in vitro that mainly drives SK lesions, but not Treg. Additionally, if glucose utilization was limited in vivo as a consequence of therapy in the inflammatory phase with the drug 2DG, lesions were diminished compared with untreated infected controls. In addition, lesions in 2DG-treated animals contained less proinflammatory effectors. Glucose metabolism also influenced the acute phase of infection when the replicating virus was present in the eye. Thus, therapy with 2DG to limit glucose utilization caused mice to become susceptible to the lethal effects of HSV infection, with the virus spreading to the brain causing encephalitis. Taken together, our results indicate that glucose metabolism changed during the course of HSV infection and that modulating glucose levels can influence the outcome of infection, being detrimental or beneficial according to the stage of viral pathogenesis.

Few reports have noted any changes in blood glucose levels in response to virus infections except in experimental situations where infection can set off diabetes mellitus (33, 34). Moreover, in the situation we described, the mild hyperglycemia was absent during the time when the virus was actively replicating in the eye or elsewhere in the body. Instead, hyperglycemia only occurred during the inflammatory reaction to the virus, which in the case of HSV ocular infection becomes the vision-impairing lesion SK. Moreover, SK is a local lesion in the eye accompanied by an inflammatory response in the innervating trigeminal ganglion (35). Two relevant questions emerge: what accounts for the mild hyperglycemia, and does the event have any consequence?

With regard to causation, the likeliest explanation could be the production of factors from the inflammatory site that caused the liver to change its level of glucose production. Candidates could include hormones, cytokines, or the release of lipid mediators from inflammatory cells (36–39). In a complex model of sustained inflammation induced by LPS, elevated blood glucose levels were observed and were shown to be mediated by TNF-α (25). This cytokine acted on the liver to suppress the expression of the bile acid biosynthesis enzyme CYP7A1, which normally influences glucose production by acting on the hepatic mevalonate pathway, resulting in the inhibition of insulin signaling (25). In our system, elements of the virus with TLR components would likely not be involved because the virus is absent in the bloodstream especially at the time of raised glucose levels. However, we did observe increased blood levels of some cytokines, which included TNF-α, IL-6, and IFN-γ at day 4 pi, as well as an increase in the expression of their receptors in the liver at day 8 pi. Moreover, we could show that animals with high blood glucose levels had reduced levels of CYP7A1, which could mean that the mechanism inducing the hyperglycemia was similar to that described by the Okin and Medzhitov (25). Additional experiments are underway to further evaluate potential mechanisms that could cause raised glucose levels in our system, which differs from the systemic model used by the Okin and Medzhitov (25) in being only a local inflammatory lesion in a small organ, the eye.

The second question about the hyperglycemia observed was the issue of its potential relevance. Two lines of indirect evidence implied that the effect could have relevance. Firstly, the change of glucose levels was ∼2-fold; whether this change in glucose concentration was potentially meaningful was tested in an in vitro induction system. In this system, the levels of CD4 Th1 effector cells generated from naive precursors were compared using media with controlled concentrations of glucose. Curiously, increasing the glucose concentrations from physiological levels to the levels observed in infected animals significantly increased the number of effector Th1 cells induced, and had no effect on the levels of Treg induction, confirming the previous observations (29, 30). This could mean that the 2-fold increase in blood glucose observed in vivo might also serve to enhance Th1 differentiation, but further studies are needed to verify this possibility.

Another indirect approach also indicated that the change in glucose levels might help drive the inflammatory T cell response. Accordingly, when glucose utilization was suppressed by treating mice with 2DG from the time of onset of ocular inflammation, SK lesions were significantly reduced in magnitude. Along with this, the composition of the cellular constituents involved in ocular lesions was markedly changed with reduced numbers of CD4 effectors. Other reports have also shown that 2DG therapy (13, 14) can control both autoinflammatory lesions and graft-versus-host disease inflammatory reactions. However, in these inflammatory models, therapy with 2DG alone was usually noneffective and additional antimetabolite therapies such as metformin (13) and 6-diazo-5-oxo-l-norleucine (14) were needed to be used in combination to counteract lesions. The fact that 2DG alone was a highly effective therapy against SK could relate to the less chronic nature of the SK lesion because viral Ag does not persist to drive lesions. However, in the autoimmune disease and graft-versus-host disease system, the inflammatory cells in the lesions are under consistent activation.

One issue yet to be explained was the discrepancy with the dose of 2DG used to achieve therapeutic effects in vivo with that found to selectively inhibit inflammatory T cell responses in vitro. However, we could not measure the absolute concentrations of 2DG in the blood and DLNs after in vivo administration, and it is well known that 2DG is rapidly metabolized in vivo (40). Another perplexing issue was the observation that 2DG could inhibit the effects of glucose in vitro when the latter was far in excess of 2DG concentrations. Data from studies with yeast at least partially explain such observations. Thus 2DG may inhibit the function of glucose transporters (41) and it is conceivable that 2DG has additional off-target effects not mediated by glucose itself (42). These issues needs to be further clarified.

Another unexpected observation made in this report was the dramatic consequence of 2DG therapy administered during the acute phase of ocular HSV infection. In such experiments, many of the animals succumbed to lethal consequences with the virus spreading to and replicating in the CNS. Herpes simplex encephalitis (HSE) is a very rare outcome of HSV infections in adult humans unless they are genetically compromised in some immune component (43–46). In animal models for HSV infection, HSE is more common and can occur at higher doses of infection, with some virulent strains of virus, or if animals have one of several defects of either innate or adaptive immunity (22, 47–50). We strongly suspect that the lethal consequences of HSV infection in 2DG-treated mice could be the result of the impaired function of a protective component of innate immunity. So far we have minimal support for this idea, but could show that the number of IL-1β–expressing macrophages but not neutrophils was reduced at the site of infection as a consequence of 2DG therapy. In addition, animals can suffer lethal consequences when treated with 2DG as was noted with another viral model (16). Accordingly, with influenza, the lethal effects were attributed to the increased expression of the endoplasmic reticulum stress-induced transcription factor CHOP protein and its target gene Gadd34 in the hindbrains of mice treated with 2DG. The increased expression of CHOP protein caused apoptosis of neurons, leading to neuronal dysfunction and death. This phenomenon has yet to be evaluated in our HSE model, but we suspect to find differences because HSV replication occurs in the CNS and can be very destructive resulting from both direct viral and immune-pathological destructive events (22, 48, 49).

To our knowledge, this study is one of the first to evaluate the use of a drug that influences metabolic processes used by different immune components responding to an infection. We show that modulating the main energy generating component of inflammatory T cells has value to modulate the extent of viral immune inflammatory lesions. In our system, the result was particularly effective because 2DG therapy inhibited proinflammatory T effectors but had no direct effect on cells such as Treg that play a protective function in SK lesions. Our results also indicate that metabolic-modifying drugs should be used with caution especially during virus infections. Thus, when 2DG therapy was used when the virus was still replicating, the viral replication was enhanced, which could have had lethal consequences as a result of virus spreading to the brain.

The authors have no financial conflicts of interest.