Inhibition of the Kynurenine-NAD+ Pathway Leads to Energy Failure and Exacerbates Apoptosis in Pneumococcal Meningitis

Home , Kynurenine, Kynurenine pathway

J Neuropathol Exp Neurol Vol. 69, No. 11 Copyright Ó 2010 by the American Association of Neuropathologists, Inc. November 2010 pp. 1096Y1104

ORIGINAL ARTICLE

Caroline L. Bellac, PhD, Roney S. Coimbra, PhD, Stephan Christen, PhD, and Stephen L. Leib, MD Downloaded from https://academic.oup.com/jnen/article/69/11/1096/2917121 by guest on 28 September 2021 We previously showed that in experimental pneumo- Abstract coccal meningitis, the kynurenine (KYN) pathway is acti- Pneumococcal meningitis causes neurological sequelae, including vated, and that this leads to accumulation of KYN metabolites learning and memory deficits in up to half of the survivors. In both in the hippocampus (11). Activation of the KYN pathway humans and in animal models of the disease, there is apoptotic cell and the resulting increase in KYN intermediates have been death in the hippocampus, a brain region involved in learning and documented in other neuroinflammatory diseases and in neuro- memory function. We previously demonstrated that in an infant rat degenerative diseases (12Y14). The pathogenesis of brain in- model of pneumococcal meningitis, there is activation of the kynuren- jury in these disorders may involve the accumulation of ine (KYN) pathway in the hippocampus, and that there was a positive KYN metabolites such as 3-hydroxykynurenine (3-HKYN), correlation between the concentration of 3-hydroxykynurenine and the 3-hydroxyanthranilic acid (3-HAA), or quinolinic acid extent of hippocampal apoptosis. To clarify the role of the KYN path- (QUINA) (15Y17). We also found that 3-HKYN concentra- way in the pathogenesis of hippocampal apoptosis in pneumococcal tions in pneumococcal meningitis correlated to the extent of meningitis, we specifically inhibited 2 key enzymes of the KYN path- hippocampal apoptotic damage (11). Moreover, 2 recent stud- way and assessed hippocampal apoptosis, KYN pathway metabolites, ies of parasitic brain infections (African trypanosomiasis and and nicotinamide adenine dinucleotide (NAD) concentrations by high- cerebral malaria) in mice showed that inhibition of the KYN performance liquid chromatography. Pharmacological inhibition of kyn- pathway by Ro-61-8048 reduced brain inflammation and pro- urenine 3-hydroxylase and kynureninase led to decreased cellular NAD+ longed survival, suggesting that KYN pathway inhibition may levels and increased apoptosis in the hippocampus. The cerebrospi- represent a therapeutic option in neuroinflammatory diseases nal fluid levels of tumor necrosis factor and interleukin-1> and -A (18, 19). were not affected. Our data suggest that activation of the KYN path- The KYN pathway is the major route of tryptophan way in pneumococcal meningitis is neuroprotective by compensating catabolism and the only de novo synthesis pathway of the for an increased NAD+ demand caused by infection and inflamma- essential coenzyme nicotinamide adenine dinucleotide tion; this mechanism may prevent energy failure and apoptosis in the (NAD) (Fig. 1). Activation of the KYN pathway has been hippocampus. reported to counteract NAD+ depletion in activated macrophages and astrocytes (12, 20). The NAD+ is an essential Key Words: Apoptosis, Energy failure, Hippocampus, Kynurenine, cofactor in various cellular reactions, including adenosine Meningitis, Pharmacological inhibition, Streptococcus pneumoniae. triphosphate synthesis and DNA repair. Intracellular NAD+ and adenosine triphosphate levels are known to be critical for cell survival (21, 22). Although NAD+ can be recycled from INTRODUCTION nicotinic acid or nicotinamide, de novo synthesis of NAD+ Pneumococcal meningitis is associated with a mortality occurs by the KYN pathway. rate of up to 30% and persistent neurological sequelae in up Here, we assessed the role of KYN pathway activation to 50% of survivors because of different degrees and extents in hippocampal apoptosis by pharmacological inhibition of of brain damage (1, 2). Apoptosis of hippocampal neurons the pathway at the level of kynurenine 3-hydroxylase (HK) occurs in pneumococcal meningitis in humans and is asso- and kynureninase (KZ) in an infant rat model of pneumo- ciated with learning and memory deficits in corresponding coccal meningitis. animal models of the disease (3Y7). Both the pathogen and the host inflammatory reaction may contribute to the development MATERIALS AND METHODS of hippocampal neuronal injury (1, 8Y10). Materials Tryptophan, KYN, 3-HKYN, kynurenic acid (KYNA), From the Institute for Infectious Diseases, University of Bern, Bern, Switzerland. magnesium chloride hexahydrate, bathophenanthroline disul- Send correspondence and reprint requests to: Stephen L. Leib, MD, Institute fonic acid, and the oxidized and reduced forms of NAD (NAD+/ for Infectious Diseases, University of Bern, Friedbuehlstrasse 51, P.O. NADH) were obtained from Sigma (St Louis, MO). An- Box 61, CH-3010 Bern, Switzerland; E-mail: [email protected] This work was supported by grants 31-120725 and 31-116257 from the Swiss thranilic acid (AA), zinc acetate dihydrate, TRIS hydrochlo- National Science Foundation and by Grant mkl/stm 55-2005 from the ride, ammonium acetate, tetrabutylammonium hydrogen sulfate Roche Research Foundation. and perchloric acid were from Fluka (Buchs, Switzerland).

1096 J Neuropathol Exp Neurol Volume 69, Number 11, November 2010

Copyright © 2010 by the American Association of Neuropathologists, Inc. Unauthorized reproduction of this article is prohibited. J Neuropathol Exp Neurol Volume 69, Number 11, November 2010 Kynurenine-NAD+ Pathway in Pneumococcal Meningitis

animals (n = 24) were injected with 10 KL of sterile saline solution. At 18 hours after injection of bacteria or saline, the animals were clinically assessed by scoring disease se- verity (1 = coma; 2 = does not turn upright; 3 = turns upright within 30 seconds; 4 = minimal ambulatory activity, turns upright in less than 5 seconds; and 5 = normal). Cerebrospinal fluid (CSF) was obtained by puncture of the cisterna magna, and 5 KL was serially diluted and cultured to determine bacterial CSF titers (24). All animals then received antibi- otic therapy (ceftriaxone, 100 mg/kg subcutaneously; Roche Pharma, Reinach, Switzerland) and were killed at prede- termined time points (i.e. at 24 hours or 36 hours after infection) by an overdose of intraperitoneally injected pento- Downloaded from https://academic.oup.com/jnen/article/69/11/1096/2917121 by guest on 28 September 2021 barbital (150 mg/kg). Immediately after death, the animals were perfused with 30 mL of ice-cold PBS via the left cardiac ventricle. Animal studies were approved by the Animal Care and Experimentation Committee of the Canton Bern, Switzerland and followed the National Institutes of Health’s guidelines for the performance of animal experiments. Pharmacological Inhibition of the KYN Pathway by Ro-61-8048 and oMBA To test the contribution of 3-HKYN on neuronal apoptosis, we administered the specific HK inhibitor Ro-61-8048. Preliminary dose-finding tests directed us to use 75 mg/kg, intraperitoneally injected at 8 hours, 16 hours, 24 hours, and 32 hours after infection. In contrast to humans, who synthesize 3-HAA exclusively from 3-HKYN, rats synthesize 3-HAA also from an FIGURE 1. Schematic of the kynurenine pathway, the de novo intermediate metabolite AA, albeit to a lesser degree (Fig. 1). synthesis pathway of nicotinamide adenine dinucleotide To assess whether potentially neurotoxic metabolites further (NAD+). Tryptophan (TRP) is metabolized over different meta- downstream of 3-HKYN (e.g. 3-HAA and QUINA) contrib- bolic steps into quinolinic acid (QUINA) and culminates with ute to apoptosis, we inhibited the enzyme KZ by administer- + the formation of NAD . Different points for pharmacological ing oMBA in parallel. This was done by administration of intervention are indicated (gray shaded ovals). The compound 75 mg/kg Ro-61-8048 and 50 mg/kg oMBA intraperitoneally Ro-61-8048 inhibits the transformation of kynurenine into 3- at 8 hours, 16 hours, 24 hours, and 32 hours after infection. hydroxykynurenine (3-HKYN) by blocking the enzyme kynurenine 3-hydroxylase (HK). The activity of kynureninase (KZ) Vehicle solution consisted of aqua ad injectabilia with 0.1% is decreased by the inhibitor of kynureninase (oMBA), resulting Tween 80. in lower levels of anthranilic acid (AA) and 3-hydroxyanthranilic For quantification of CSF cytokine levels and determi- + acid (3-HAA). HAD, anthranilate 3-monooxygenase; IDO, in- nation of hippocampal NAD levels, animals were killed at doleamine 2,3-dioxygenase; KAT, kynurenine aminotransfer- 24 hours after infection. For evaluation of hippocampal ase; KYNA, kynurenic acid. apoptosis in 1 hemisphere and determination of hippocampal KYN metabolites in the other hemisphere, animals were killed Acetonitrile, methanol, potassium cyanide, and all other re- at 36 hours after infection. These time points were based on agents were of high-performance liquid chromatography our previous demonstration that apoptotic damage peaks at (HPLC) grade and purchased from Merck (Darmstadt, Ger- approximately 36 hours after infection (25). Because the pathophysiological events leading to apoptosis precede the many). Analytical-grade water was prepared using a Milli-Q + synthesis purification system (Millipore, Zug, Switzerland). All histological appearance of apoptosis, we measured NAD mobile phases were filtered through 0.22-KmDuraporemem- levels at the earlier time point, that is, between 18 hours branes (Millipore) and degassed by sonication before use. The and 24 hours after infection. Although apoptosis almost exclusively occurs in the dentate gyrus, KYN metabolites inhibitors Ro-61-8048 (inhibitor of kynurenine-3-hydroxylase) + and oMBA (inhibitor of KZ) were kindly provided by Dr S. and NAD levels were measured in homogenates of whole Roever (Roche Pharma AG, Basel, Switzerland). hippocampi. For analysis at 24 hours after infection, animals were Infant Rat Model of Pneumococcal Meningitis treated at 8 hours and 16 hours after infection with Ro- Eleven-day-old Wistar rats (n = 160) were infected 61-8048 (n = 5), with the combination of inhibitors (n = 7), or intracisternally with 10 KL of saline containing Streptococcus with vehicle solution alone (n = 14). Mock-infected animals pneumoniae type 3 (log10 6.5 T 0.5 colony forming units were treated with the combination (n = 4) or vehicle solution [cfu]/mL), with a 32-gauge needle (23, 24). Mock-infected (n = 8). At 24 hours after infection, CSF was taken and the

Ó 2010 American Association of Neuropathologists, Inc. 1097

Copyright © 2010 by the American Association of Neuropathologists, Inc. Unauthorized reproduction of this article is prohibited. Bellac et al J Neuropathol Exp Neurol Volume 69, Number 11, November 2010 brains were dissected. Hippocampal samples were frozen on droxide, 1 mmol/L bathophenanthroline disulfonic acid). After dry ice, weighed, and stored at j80-C for further analysis by 5 minutes at room temperature, the homogenate was extracted HPLC and multiplex suspension array. twice with chloroform. The aqueous phase was diluted 1:5 with For analysis at 36 hours after infection, infected animals mobile phase, filtered through 0.45-Km positively charged were treated with the HK inhibitor Ro-61-8048 (n = 19) or the nylon filters (Mettler Toledo, Greifensee, Switzerland); 80 KL combination of both inhibitors (n = 8) at 8 hours, 16 hours, of the filtrate were then injected onto the system described in 24 hours, and 32 hours after infection. One group of infected Ghielmetti et al (27). animals was treated with vehicle solution alone (n = 27). Mock-infected animals were treated with the combination Multiplex Analysis of CSF (n = 9) or with vehicle solution alone (n = 8). The levels of the proinflammatory cytokines tumor necrosis factor (TNF), macrophage-specific interleukin-1> Determination of KYN Metabolites (IL-1>), and IL-1A in CSF were measured by multiplex sus- After perfusion with PBS, the brains were dissected, pension array using the Luminex 200 system (Luminex Corp, Downloaded from https://academic.oup.com/jnen/article/69/11/1096/2917121 by guest on 28 September 2021 rolled on a filter paper to remove the meninges, and the brains Austin, TX) with rat Fluorokine MAP Multiplex reagents divided in the midline. Hippocampi were isolated in ice-cold (R&D Systems Inc, Minneapolis, MN). The CSF samples PBS from the left half of each brain. Tissue samples were were diluted 1:5 in sample assay buffer and were then pro- immediately frozen on dry ice, weighed, and stored separately cessed and measured according to the manufacturer’s at j80-C. instructions. Data were analyzed using a 5-parametric logistic Frozen tissue samples were homogenized 1:6 (wt/vol) curve fitting using Bioplex manager software 4.01 (BioRad in Milli-Q water, and proteins in the homogenates were pre- Laboratories Inc, Hercules, CA). cipitated with 6% perchloric acid (to every 100 KLof homogenate, 25 KL of perchloric acid 6% was added). After centrifugation at 10,000 Â g for 10 minutes at 4-C, the RESULTS supernatant was filtered (HPLC filter, 0.2 Km; nylon mem- Clinical Outcome of Experimental brane, Infochroma AG, Zug, Switzerland; 3 mL syringe, K Pneumococcal Meningitis LUER, Rodby, Denmark) and 80 L of the filtrate was All infected animals developed meningitis as indicated applied onto the HPLC system, as previously described (11). by positive bacterial CSF titers at 18 hours after infection Histopathologic Examination (vehicle, log 10 7.5 cfu/mL). Administration of KYN pathway The right hemisphere of each animal killed at 36 hours inhibitors did not affect CSF bacterial titers (HK inhibited, after infection was used to assess apoptotic damage in the log 10 7.5 cfu/mL; combined inhibited, log 10 7.6 cfu/mL). At hippocampus. After perfusion with PBS, the hemisphere was 18 hours after infection, clinical signs of meningitis were fixed in 4% formaldehyde in PBS for 12 hours, followed present in all infected animals. At this time point, clinical by submersion in 18% sucrose (in PBS) for 24 hours. Four scores did not significantly differ between the treatment coronal cryosections (45 Km) from the dorsal brain region of groups (vehicle, 3.8; HK inhibited, 3.5; and combined in- each animal were mounted onto polylysine-coated slides and hibited, 3.4). However, during a 36-hour period, infected stained with Nissl (cresyl violet) stain. Apoptotic damage in animals treated with the HK inhibitor alone or the combi- the hippocampus was evaluated by histomorphometry, as nation of both inhibitors showed significantly higher mor- previously described (23, 26). Apoptosis in the granule cell tality versus vehicle-treated animals (Kaplan-Meier survival layer of the hippocampus, defined as cells showing markedly curves with Mantel-Cox test; vehicle, n = 59, vs HK in- shrunken, condensed, or fragmented nuclei, was counted at hibition, n = 55, p G 0.01; vehicle vs combined inhibi- 400Â in 3 visual fields for each of the 4 blades of the dentate tion, n = 34, p G 0.01; HK inhibition vs combined inhibition, gyrus. Averages per dentate gyrus (6 visual fields) per animal p G 0.05; Fig. 2). were calculated from all slides evaluated. Apoptotic damage Effect of KYN Pathway Inhibitors on was scored as follows: 0 = 0 to 5 apoptotic cells; 1 = 6 to 20 apoptotic cells; 2 = more than 20 apoptotic cells. Apoptosis Hippocampal Concentrations of scores generated by this method closely correlate with data KYN Metabolites obtained by other methods for the detection of apoptosis, for Quantification of KYN metabolites in the hippo- example, detection of DNA fragmentation by terminal deox- campus was used to document the efficacy of the inhibitors ynucleotidyl transferase dUTP nick end labeling stain and (Table 1). As in previous reports in which Ro-61-8048 was immunostaining for active caspase-3 (25). Histopathologic defined to be highly specific with an IC50 value in the nano- evaluation was done by an investigator blinded to the clinical, molar range (37 nmol/L) (29), treatment with Ro-61-8048 microbiological, and treatment data of the animals. resulted in a 50-fold increase in accumulation of the upstream substrate KYN compared with vehicle-treated infected animals + Quantification of NAD (104 Kmol/L, n = 6, vs 2 Kmol/L, n = 7; p G 0.01). The The reduced and oxidized forms of NAD were deter- potency of Ro-61-8048 was further demonstrated by a 10-fold mined by ion-pairing HPLC with fluorescence detection as reduction in 3-HKYN concentrations in infected inhibited described (27, 28). The whole frozen hippocampi were compared with infected vehicle-treated animals (21 nmol/L, homogenized 1:20 (wt/vol) in ice-cold homogenization buffer n = 6, vs 348 nmol/L, n = 7; p G 0.01). At the same time, AA (200 mmol/L potassium cyanide, 60 mmol/L potassium hy- concentrations in infected animals treated with Ro-61-8048

1098 Ó 2010 American Association of Neuropathologists, Inc.

Combined inhibition shifted the metabolic ﬂux of the KYN pathway even further away from 3-HKYN to KYNA production (3,219 nmol/L, n = 7, vs 20 nmol/L, n = 6) (Table 1).

Effect of KYN Pathway Inhibition on Apoptotic Damage Inhibition of HK in infected animals caused a significant increase in the extent of apoptotic damage, resulting in a 2.5-fold increase in apoptosis scores (0.5 in vehicle group, n = 27, vs 1.3 in HK-inhibited group, n = 19; p G 0.01). Combined inhibition of both HK and KZ also lead to a significantly increased occurrence of apoptosis (1.5, n = 8, vs Downloaded from https://academic.oup.com/jnen/article/69/11/1096/2917121 by guest on 28 September 2021 0.5 [vehicle], n = 31; p G 0.01); the latter was higher than that in singly inhibited animals, but the difference did not reach statistical significance (Figs. 3 and 4A). In mock-infected animals, combined treatment had no significant effect on apoptosis versus vehicle-treated animals (score = 0.12, n = 9, vs score = 0 [vehicle]; n = 6; p = not significant). Effect of KYN Pathway Inhibition on NAD+ Levels There was no difference in NAD+ levels between mock- infected and infected vehicle-treated animals (Fig. 4B). In FIGURE 2. The impact of kynurenine pathway inhibition on infected animals treated with the HK inhibitor alone, there mortality in infected animals treated with the combination + of the inhibitors Ro-61-8048 and oMBA showed a statistically was a reduction of NAD levels, although the effect was not G statistically significant (Fig. 4B). In infected animals treated significant increased mortality (p 0.05; Kaplan-Meier sur- + vival curves with Mantel-Cox test) compared with infected with both inhibitors, however, NAD levels were significantly vehicle-treated animals and Ro-06108048/treated (kynurenine lower versus infected vehicle-treated animals (90 Kmol/L, 3-hydroxylaseYinhibited) animals. n = 7, vs 148 Kmol/L, n = 14; p G 0.05; Fig. 4B). Inhibition of the enzymes HK and KZ did not affect the ratio of NADH/ NAD+ compared with untreated animals. The NAD+ levels increased 111-fold (1672 nmol/L, n = 6, vs 15 nmol/L, n = 7; in mock-infected animals treated with both inhibitors were p G 0.01) and KYNA concentrations 129-fold (2,577 nmol/L, G not significantly different from mock-infected vehicle-treated n = 6, vs 20 nmol/L, n = 7; p 0.01) when compared animals (data not shown). These findings suggest that acti- with infected vehicle-treated animals. This demonstrates vation of the KYN pathway during the acute disease con- that 3-HKYN production is the main route of KYN deg- tributes to the maintenance of cellular NAD+ homeostasis in radation in this infant rat model of pneumococcal meningi- the hippocampus during pneumococcal meningitis. tis, similarly to what has been observed in other studies in rats (30). Although the KZ inhibitor oMBA is less potent than Effect of KYN Pathway Inhibition on Cytokine Ro-61-8048 (IC50 of oMBA = 3 Kmol/L) (30), the combi- Levels in CSF nation of oMBA with Ro-61-8048 substantially inhibited the Inhibition of HK and KZ had no significant effect on the increased accumulation of AA resulting from single inhibition CSF concentrations of TNF, IL-1>,orIL-1A (Table 2). of HK (199 nmol/L, n = 6, vs 1,672 nmol/L, n = 7; p G 0.01). The effect of simultaneous inhibition was most clearly re- flected by a further increase in KYN, the substrate of both DISCUSSION enzymes targeted by the treatment (140 Kmol/L, n = 7, vs The KYN pathway has primarily been investigated with 2 Kmol/L in infected/vehicle-treated animals, n = 7; p G 0.01). a focus on the potential neurotoxic effect of its metabolites

TABLE 1. Concentrations of Kynurenine Metabolites in Hippocampi at 36 Hours After Infection Treatments KYN Metabolites Mock Infected + Vehicle Infected + Vehicle Infected + Ro-61-8048 Infected + Ro-61-8048 + oMBA KYN, Kmol/L 1.0 T 0.5 2.2 T 1.4 103.5 T 3.0 140.7 T 29.6 3-HKYN, nmol/L 34.1 T 10.10 347.9 T 159.1 21.2 T 3.4 172.3 T 53.7 AA, nmol/L 11.7 T 4.6 14.9 T 2.9 1,672 T 470.9 198.8 T 32.0 KYNA, nmol/L 5.0 T 4.6 20.0 T 5.4 2,577 T 336.4 3,219 T 452.1

3-HKYN, 3-hydroxykynurenine; AA, anthranilic acid; KYN, kynurenine; KYNA, kynurenic acid; oMBA, inhibitor of kynureninase, Ro-61-8048, inhibitor of the enzyme kynurenine 3-hydroxylase.

Ó 2010 American Association of Neuropathologists, Inc. 1099

FIGURE 3. Histological evidence of apoptotic damage in the dentate gyrus of infected infant rats. (AYD) At 36 hours after mock infection (A) or infection (BYD). Infected animals were treated with vehicle solution (B), kynurenine 3-hydroxylase (HK) inhibitor Ro-61-8048 (C), or the combination of Ro-61-8048 and oMBA (D). Mock-infected animals were treated with vehicle solution (A). There was greater apoptotic damage in (C) than in (B), and there was a greater difference between (B) and (D). Scale bars = 100 Km; insert, 30 Km. Cryosections were stained with cresyl violet (Nissl).

3-HKYN, 3-HAA, and QUINA (15Y17). Previous studies perspective on the physiological relevance of the KYN-NAD+ have also shown that HK inhibition by Ro-61-8048 reduces pathway, namely, that its activation contributes to cellular brain inﬂammation in mouse models of parasitic brain infec- NAD+ homeostasis and thereby attenuates hippocampal neu- tion (18, 19). The effect of KYN-NAD+ pathway inhibition on ronal apoptosis. neuronal apoptosis (speciﬁcally in bacterial meningitis) has In a previous study of experimental pneumococcal not been examined to date. Our present results provide a new meningitis, we found accumulation of 3-HKYN and 3-HAA

1100 Ó 2010 American Association of Neuropathologists, Inc.

FIGURE 4. Quantitation of apoptosis and oxidized and reduced forms of nicotinamide adenine dinucleotide (NAD+)/(NADH) in pneumococcal meningitis. (A) The apoptotic scores in the dentate gyrus at 36 hours after infection after different pharmacological treatments: mock/veh, mock-infected vehicle-treated animals (n = 8); inf/veh, infected vehicle-treated animals (n = 27); inf/HK-inh, infected animals treated with the kynurenine 3-hydroxylase (HK) inhibitor Ro-61-8048 (n = 19); inf/co-inh, infected animals treated with the combination of HK and kynureninase (KZ) inhibitors (n = 8). Kynurenine 3-hydroxylase inhibition by Ro-61-8048 resulted in increased apoptosis versus inf/veh. Simultaneous inhibition of both HK and KZ led to a significantly increased apoptosis score. Values = mean T SEM. Statistic differences between inf/veh and inf/HK-inh or inf/co-inh are indicated (**), p G 0.01. (B) The NAD+ levels were determined in hippocampi at 24 hours after infection and compared among the 4 treatment groups. There was no difference between mock/veh (n = 8) and inf/veh (n = 14), whereas inhibition by the combination of the 2 inhibitors (n = 7) led to a significant decrease in NAD+ levels. Values are mean T SEM. (*) p G 0.05. Bars represent the total concentration of hippocampal NAD (sum of NADH plus NAD+). in brain regions prone to injury as a result of an increased KZ led to even more pronounced apoptosis in the hippocam- activity of HK (11). Although hippocampal 3-HKYN con- pal dentate gyrus than inhibition of HK alone (Figs. 3 and 4). centrations correlated with the extent of apoptotic damage in Thus, contrary to findings in other paradigms of brain damage the hippocampus (11), a direct neurotoxic effect of 3-HKYN (31, 32), the present results in pneumococcal meningitis do not was not demonstrated (11). Here, we show that when HK is support a direct neurotoxic role for 3-HKYN or further down- inhibited by Ro-61-808 in infected animals, the metabolic flux stream metabolites in the pathophysiology of apoptotic hippo- through the KYN pathway is shifted toward AA production campal brain injury. (Table 1). Because in rats 3HAA and further downstream The detrimental effect of combined inhibition in infec- metabolites can also be produced via AA, we inhibited KZ tion was also evident by the increased mortality in com- with oMBA. Although this inhibitor alone was not as potent parison with vehicle-treated animals (Fig. 2). Inasmuch as as Ro-16-8048, cotreatment with oMBA resulted in a nearly none of the mock-infected animals died when treated with 10-fold reduction of the elevated AA accumulation mediated the same doses of inhibitors and did not show increased ap- by inhibition of HK and caused a further increase in KYN optosis, a toxic effect of the compounds (at least under accumulation. Combined inhibition of both enzymes HK and physiological conditions) can be excluded. However, in brain

TABLE 2. Cytokine Levels in CSF at 24 Hours After Infection Treatments Mock-Infected Vehicle Infected Vehicle Infected + Ro-61-8048 Infected + Ro-61-8048 + oMBA Cytokine, pg/mL (n = 5) (n = 10) (n = 8) (n = 8) IL-1> 62 T 16 788 T 372 1,423 T 1,550 917 T 598 IL-1b 22 T 9 6,338 T 3,696 9,145 T 6,910 9,205 T 5,488 TNF 51 T 8 1,446 T 837 1,368 T 432 1,780 T 820

IL-1, interleukin-1; HK inhibition, treatment with Ro-61-8048, inhibitor of the enzyme kynurenine 3-hydroxylase; TNF, tumor necrosis factor.

Ó 2010 American Association of Neuropathologists, Inc. 1101

Copyright © 2010 by the American Association of Neuropathologists, Inc. Unauthorized reproduction of this article is prohibited. Bellac et al J Neuropathol Exp Neurol Volume 69, Number 11, November 2010 malaria infection, unexpected death occurred in up to 30% of exogenous KYN was previously used to quantify the meta- animals treated with Ro-61-8048 (18). In these studies, death bolic flux through the KYN pathway (42, 43). Tracking the might be attributed to chronic toxicity of Ro-61-8048 (33). In incorporation of radiotraced metabolites into newly synthe- the model of meningitis used here, the reasons for sponta- sized NAD+ molecules as a reflection of NAD+ supplemen- neous mortality are likely multifactorial, involving systemic tation over the KYN pathway may potentially localize major (e.g. reduced systemic blood pressure and sepsis) and brain- NAD+-consuming enzymes. Consistent with the concept that specific factors (e.g. edema and ischemia). Animals with acute NAD+ homeostasis is critical for neuronal survival, supple- pneumococcal meningitis are clinically unstable and highly mentation of NAD+ with exogenous NAD+ precursors has susceptible to any additional perturbation. shown varying degrees of neuroprotection in different para- Because 3-HKYN production from KYN is the major digms of brain injury (44, 45). route of KYN degradation, this metabolic path is a key con- Although NAD+ replenishment may be one way how tributor to NAD+ production, particularly in infection when HK activation of the KYN pathway exhibits protective effects on enzymatic activity is increased. In the present study, single neuronal survival, inhibition of the pathway has additional Downloaded from https://academic.oup.com/jnen/article/69/11/1096/2917121 by guest on 28 September 2021 inhibition of HK had a major detrimental impact on the clini- consequences unrelated to NAD+ homeostasis that might in- cal status and NAD+ levels of infected animals, and additional fluence the extent of hippocampal apoptosis. Kynurenic acid KZ inhibition by oMBA led to a further decrease in cellular in infected animals treated with combined inhibitors accu- NAD+ levels and an increased mortality. A decrease in cellular mulated more than 100-fold versus vehicle-treated animals, energy storage levels by inhibition of the KYN-NAD+ pathway reaching a low micromolar concentration (Table 1). Kynur- in a clinically unstable disease state likely contributed to the enic acid acts as an antagonist on 2 different receptors, that is, increase in mortality observed (Figs. 2 and 4). on the glycine coagonist site of the N-methyl-D-aspartic acid Cellular NAD+ levels are critical factors for cell sur- (NMDA) receptor and on the allosteric-potentiating ligand vival. Energy failure upon depletion of NAD+ stores triggers site of the nicotinic acetylcholine receptor (46). Both recep- cell death because cellular adenosine triphosphate synthesis tors are involved in brain excitation. Previous studies showed cannot be maintained. The de novo synthesis of NAD+ a neuroprotective effect for KYNA in neuroinflammation at depends exclusively on the KYN-NAD+ pathway, and thus micromolar concentrations (47, 48). Kynurenic acid might + + the reduced hippocampal NAD levels upon KYN-NAD antagonize the excitotoxic effects of glutamate and D-serine pathway inhibition in infected animals is a likely explanation that are both increased in neuroinflammatory conditions for the observed increase in hippocampal apoptosis. Con- (49). In an infant rat model of meningitis caused by group B versely, activation of the KYN pathway has been reported streptococci, we previously found that KYNA treatment to counteract NAD+ depletion in activated macrophages and attenuated neuronal injury (50). Recent reports propose, astrocytes (20, 34). The NAD+ is consumed by 4 major however, that NMDA antagonists may enhance apoptotic groups of enzymes that may influence NAD+ levels in in- neurodegeneration of hippocampal neurons through a reduc- fection, that is, poly(ADP-ribose) polymerases (PARPs), tion of glutamate stimulation of excitatory receptors (51Y54). mono(ADP-ribosyl) transferases, bifunctional ADP-ribosyl Indeed, we previously found an aggravation of apoptotic cyclases/cyclic ADP-ribose hydrolases, and NAD+-dependent neuronal death in the hippocampal dentate gyrus by adju- histone deacetylases (the Sir2 family proteins [sirtuins]). Little vant treatment with dextromethorphan, a noncompetitive is known about the involvement of these enzymes in the NMDA receptor antagonist in the infant rat model of pneu- pathophysiology of bacterial meningitis, but pharmacological mococcal meningitis (55). The inflammatory cytokines TNF, inhibition of PARP-1 attenuated CNS complications and IL-1>, and IL-1A in the CSF remained largely unaffected improved the clinical status in 2 studies (35, 36). Inhibition by KYN-NAD+ pathway modulation with both inhibitors of PARP is generally considered to be a strategy for the used (Table 2), indicating that the increase in hippocampal attenuation of neuronal damage by preventing cellular NAD+ apoptosis under KYN-NAD+ pathway inhibition in pneu- depletion (37Y39). Interdependence between depleted cellular mococcal meningitis is not caused by modulation of CSF NAD+ levels and increased brain damage has also been inflammation. demonstrated in a model of focal brain ischemia (40, 41). In summary, the present data indicate that inhibition of Comparing NAD+ levels in animals treated with the KYN- the KYN-NAD+ pathway in pneumococcal meningitis, a NAD+ pathway inhibitors with the levels found in untreated condition with inflammation-induced increase in NAD+ con- ones revealed the importance of the KYN-NAD+ pathway for sumption, causes depletion of cellular NAD+ and increases the maintenance of NAD+ homeostasis. A significant deple- hippocampal apoptosis. This suggests that maintaining cel- tion of cellular NAD+ stores was found in the hippocampus of lular energy homeostasis by activation of the KYN-NAD+ infected animals treated with both inhibitors versus vehicle pathway intrinsically attenuates hippocampal apoptosis in (Fig. 4). Consequently, the observation that hippocampal bacterial meningitis. The finding of a detrimental effect of NAD+ levels did not differ between infected and uninfected KYN-NAD+ pathway inhibition is also likely relevant to other animals (Fig. 4) indicates that enhanced tryptophan degrada- neuroinfectious and neuroinflammatory diseases. tion by KYN pathway activation leads to increased de novo synthesis of NAD+ and prevents cellular energy failure and neuronal apoptosis in the hippocampus. In studies inves- ACKNOWLEDGMENTS tigating the activation of the KYN pathway under neuro- The authors thank Franziska Simon, Angela Bu¨hlmann, inflammatory conditions, intrastriatal injection of radiolabeled and Kevin Oberson for excellent technical support.

1102 Ó 2010 American Association of Neuropathologists, Inc.

REFERENCES 26. Pfister LA, Tureen JH, Shaw S, et al. Endothelin inhibition improves 1. Grandgirard D, Leib SL. Strategies to prevent neuronal damage in cerebral blood flow and is neuroprotective in pneumococcal meningitis. Y paediatric bacterial meningitis. Curr Opin Pediatr 2006;18:112Y18 Ann Neurol 2000;47:329 35 2. Weisfelt M, van de Beek D, Spanjaard L, et al. Clinical features, com- 27. Ghielmetti M, Ren H, Leib SL, et al. Impaired cortical energy metabolism plications, and outcome in adults with pneumococcal meningitis: A but not major antioxidant defenses in experimental bacterial meningitis. Y prospective case series. Lancet Neurol 2006;5:123Y49 Brain Res 2003;976:139 48 3. Free SL, Li LM, Fish DR, et al. Bilateral hippocampal volume loss in 28. Klaidman LK, Leung AC, Adams JD Jr. High-performance liquid chro- patients with a history of encephalitis or meningitis. Epilepsia 1996;37: matography analysis of oxidized and reduced pyridine dinucleotides in Y 400Y5 specific brain regions. Anal Biochem 1995;228:312 17 4. Hoogman M, van de Beek D, Weisfelt M, et al. Cognitive outcome in 29. Rover S, Cesura AM, Huguenin P, et al. Synthesis and biochemical adults after bacterial meningitis. J Neurol Neurosurg Psychiatry 2007;78: evaluation of N-(4-phenylthiazol-2-yl)benzenesulfonamides as high- 1092Y96 affinity inhibitors of kynurenine 3-hydroxylase. J Med Chem 1997;40: Y 5. Leib SL, Heimgartner C, Bifrare YD, et al. Dexamethasone aggravates 4378 85 hippocampal apoptosis and learning deficiency in pneumococcal menin- 30. Chiarugi A, Carpenedo R, Molina MT, et al. Comparison of the neu-

gitis in infant rats. Pediatr Res 2003;54:353Y57 rochemical and behavioral effects resulting from the inhibition of Downloaded from https://academic.oup.com/jnen/article/69/11/1096/2917121 by guest on 28 September 2021 6. Loeffler JM, Ringer R, Hablutzel M, et al. The free radical scavenger kynurenine hydroxylase and/or kynureninase. J Neurochem 1995;65: Y alpha-phenyl-tert-butyl nitrone aggravates hippocampal apoptosis and 1176 83 learning deficits in experimental pneumococcal meningitis. J Infect Dis 31. Guidetti P, Schwarcz R. 3-Hydroxykynurenine potentiates quinolinate 2001;183:247Y52 but not NMDA toxicity in the rat striatum. Eur J Neurosci 1999;11: Y 7. Nau R, Soto A, Bru¨ck W. Apoptosis of neurons in the dentate gyrus in 3857 63 humans suffering from bacterial meningitis. J Neuropathol Exp Neurol 32. Okuda S, Nishiyama N, Saito H, et al. 3-Hydroxykynurenine, an 1999;58:265Y74 endogenous oxidative stress generator, causes neuronal cell death with 8. Koedel U, Scheld WM, Pfister HW. Pathogenesis and pathophysiology apoptotic features and region selectivity. J Neurochem 1998;70: Y of pneumococcal meningitis. Lancet Infect Dis 2002;2:721Y36 299 307 9. Mitchell L, Smith SH, Braun JS, et al. Dual phases of apoptosis in 33. Cozzi A, Carpenedo R, Moroni F. Kynurenine hydroxylase inhibitors re- pneumococcal meningitis. J Infect Dis 2004;190:2039Y46 duce ischemic brain damage: Studies with (m-nitrobenzoyl)-alanine (mNBA) 10. Zweigner J, Jackowski S, Smith SH, et al. Bacterial inhibition of phos- and 3,4-dimethoxy-[-N-4-(nitrophenyl)thiazol-2yl]-benzenesulfonamide (Ro phatidylcholine synthesis triggers apoptosis in the brain. J Exp Med 61-8048) in models of focal or global brain ischemia. J Cereb Blood Flow Y 2004;200:99Y106 Metab 1999;19:771 77 11. Bellac CL, Coimbra RS, Christen S, et al. Pneumococcal meningitis 34. Grant RS, Naif H, Espinosa M, et al. IDO induction in IFN-gamma causes accumulation of neurotoxic kynurenine metabolites in brain activated astroglia: A role in improving cell viability during oxidative Y regions prone to injury. Neurobiol Dis 2006;24:395Y402 stress. Redox Rep 2000;5:101 4 12. Guidetti P, Bates GP, Graham RK, et al. Elevated brain 3-hydroxy- 35. Clark RS, Vagni VA, Nathaniel PD, et al. Local administration of the kynurenine and quinolinate levels in Huntington disease mice. Neurobiol poly(ADP-ribose) polymerase inhibitor INO-1001 prevents NAD+ Dis 2006;23:190Y97 depletion and improves water maze performance after traumatic brain 13. Guillemin GJ, Brew BJ, Noonan CE, et al. Indoleamine 2,3 dioxygenase injury in mice. J Neurotrauma 2007;24:1399Y405 and quinolinic acid immunoreactivity in Alzheimer’s disease hippo- 36. Koedel U, Winkler F, Angele B, et al. Meningitis-associated central ner- campus. Neuropathol Appl Neurobiol 2005;31:395Y404 vous system complications are mediated by the activation of poly(ADP- 14. Sanni LA, Thomas SR, Tattam BN, et al. Dramatic changes in oxida- ribose) polymerase. J Cereb Blood Flow Metab 2002;22:39Y49 tive tryptophan metabolism along the kynurenine pathway in exper- 37. Barr TL, Conley YP. Poly(ADP-ribose) polymerase-1 and its clinical imental cerebral and noncerebral malaria. Am J Pathol 1998;152: applications in brain injury. J Neurosci Nurs 2007;39:278Y84 611Y19 38. Lenzser G, Kis B, Snipes JA, et al. Contribution of poly(ADP-ribose) 15. Schwarcz R. The kynurenine pathway of tryptophan degradation as a polymerase to postischemic blood-brain barrier damage in rats. J Cereb drug target. Curr Opin Pharmacol 2004;4:12Y17 Blood Flow Metab 2007;27:1318Y26 16. Stone TW. Kynurenines in the CNS: From endogenous obscurity to 39. Moroni F. Poly(ADP-ribose)polymerase 1 (PARP-1) and postischemic therapeutic importance. Prog Neurobiol 2001;64:185Y218 brain damage. Curr Opin Pharmacol 2008;8:96Y103 17. Stone TW, Mackay GM, Forrest CM, et al. Tryptophan metabolites and 40. Ying W. NAD+ and NADH in neuronal death. J Neuroimmune Phar- brain disorders. Clin Chem Lab Med 2003;41:852Y59 macol 2007;2:270Y75 18. Miu J, Ball HJ, Mellor AL, et al. Effect of indoleamine dioxygenase-1 41. Ying W. NAD+ and NADH in brain functions, brain diseases and brain deficiency and kynurenine pathway inhibition on murine cerebral aging. Front Biosci 2007;12:1863Y88 malaria. Int J Parasitol 2009;39:363Y70 42. Guidetti P, Eastman CL, Schwarcz R. Metabolism of [5-3H]kynurenine 19. Rodgers J, Stone TW, Barrett MP, et al. Kynurenine pathway inhibition in the rat brain in vivo: Evidence for the existence of a functional reduces central nervous system inflammation in a model of human kynurenine pathway. J Neurochem 1995;65:2621Y32 African trypanosomiasis. Brain 2009;132:1259Y67 43. Amori L, Guidetti P, Pellicciari R, et al. On the relationship between the 20. Grant RS, Passey R, Matanovic G, et al. Evidence for increased de novo two branches of the kynurenine pathway in the rat brain in vivo. J Neu- synthesis of NAD in immune-activated RAW264.7 macrophages: A rochem 2009;109:316Y25 self-protective mechanism? Arch Biochem Biophys 1999;372:1Y7 44. Tsuchiyama R, Sozen T, Manaenko A, et al. The effects of nicotinamide 21. Eguchi Y, Shimizu S, Tsujimoto Y. Intracellular ATP levels deter- adenine dinucleotide on intracerebral hemorrhage-induced brain injury in mine cell death fate by apoptosis or necrosis. Cancer Res 1997;57: mice. Neurol Res 2009;31:179Y82 1835Y40 45. Holland MA, Tan AA, Smith DC, et al. Nicotinamide treatment provides 22. Leist M, Single B, Castoldi AF, et al. Intracellular adenosine triphosphate acute neuroprotection and GFAP regulation following fluid percussion (ATP) concentration: A switch in the decision between apoptosis and injury. J Neurotrauma 2008;25:140Y52 necrosis. J Exp Med 1997;185:1481Y86 46. Lopes C, Pereira EF, Wu HQ, et al. Competitive antagonism between 23. Bifrare YD, Gianinazzi C, Imboden H, et al. Bacterial meningitis causes the nicotinic allosteric potentiating ligand galantamine and kynurenic two distinct forms of cellular damage in the hippocampal dentate gyrus in acid at alpha7* nicotinic receptors. J Pharmacol Exp Ther 2007;322: infant rats. Hippocampus 2003;13:481Y88 48Y58 24. Leib SL, Kim YS, Black SM, et al. Inducible nitric oxide synthase and 47. Obrenovitch TP, Urenjak J. In vivo assessment of kynurenate neuro- the effect of aminoguanidine in experimental neonatal meningitis. J Infect protective potency and quinolinate excitotoxicity. Amino Acids 2000;19: Dis 1998;177:692Y700 299Y309 25. Gianinazzi C, Grandgirard D, Imboden H, et al. Caspase-3 mediates 48. Urenjak J, Obrenovitch TP. Neuroprotective potency of kynurenic acid hippocampal apoptosis in pneumococcal meningitis. Acta Neuropathol against excitotoxicity. Neuroreport 2000;11:1341Y44 2003;105:499Y507 49. Vesce S, Rossi D, Brambilla L, et al. Glutamate release from astrocytes

Ó 2010 American Association of Neuropathologists, Inc. 1103

in physiological conditions and in neurodegenerative disorders char- plasticity, and cognitive behavior. Neuropsychopharmacology 2010;35: acterized by neuroinflammation. Int Rev Neurobiol 2007;82:57Y71 1734Y42 50. Leib SL, Kim YS, Ferriero DM, et al. Neuroprotective effect of excitatory 53. Ikonomidou C, Stefovska V, Turski L. Neuronal death enhanced by amino acid antagonist kynurenic acid in experimental bacterial menin- N-methyl-D-aspartate antagonists. Proc Natl Acad Sci U S A 2000;97: gitis. J Infect Dis 1996;173:166Y71 12885Y90 51. Biegon A, Fry PA, Paden CM, et al. Dynamic changes in N-methyl- 54. Pohl D, Bittigau P, Ishimaru MJ, et al. N-methyl-D-aspartate antagonists D-aspartate receptors after closed head injury in mice: Implications for and apoptotic cell death triggered by head trauma in developing rat brain. treatment of neurological and cognitive deficits. Proc Natl Acad Sci U S A Proc Natl Acad Sci U S A 1999;96:2508Y13 2004;101:5117Y22 55. Sellner J, Ringer R, Baumann P, et al. Effect of the NMDA-receptor 52. Potter MC, Elmer GI, Bergeron R, et al. Reduction of endogenous kyn- antagonist dextromethorphan in infant rat pneumococcal meningitis. Curr urenic acid formation enhances extracellular glutamate, hippocampal Drug Metab 2008;9:83Y88 Downloaded from https://academic.oup.com/jnen/article/69/11/1096/2917121 by guest on 28 September 2021

1104 Ó 2010 American Association of Neuropathologists, Inc.