Kynurenic acid in neurodegenerative disorders—unique neuroprotection or double‐edged sword?

Abstract Aims The family of kynurenine pathway (KP) metabolites includes compounds produced along two arms of the path and acting in clearly opposite ways. The equilibrium between neurotoxic kynurenines, such as 3‐hydroxykynurenine (3‐HK) or quinolinic acid (QUIN), and neuroprotective kynurenic acid (KYNA) profoundly impacts the function and survival of neurons. This comprehensive review summarizes accumulated evidence on the role of KYNA in Alzheimer's, Parkinson's and Huntington's diseases, and discusses future directions of potential pharmacological manipulations aimed to modulate brain KYNA. Discussion The synthesis of specific KP metabolites is tightly regulated and may considerably vary under physiological and pathological conditions. Experimental data consistently imply that shift of the KP to neurotoxic branch producing 3‐HK and QUIN formation, with a relative or absolute deficiency of KYNA, is an important factor contributing to neurodegeneration. Targeting specific brain regions to maintain adequate KYNA levels seems vital; however, it requires the development of precise pharmacological tools, allowing to avoid the potential cognitive adverse effects. Conclusions Boosting KYNA levels, through interference with the KP enzymes or through application of prodrugs/analogs with high bioavailability and potency, is a promising clinical approach. The use of KYNA, alone or in combination with other compounds precisely influencing specific populations of neurons, is awaiting to become a significant therapy for neurodegenerative disorders.

route. 1,2 The discoveries of last decades strongly support the concept of viewing the disturbed KP as an important link in the cycle of events leading to the development of brain pathologies. Various kynurenines are of substantial biological importance due to their ability to modify neurotransmission and to alter the immune response. 3,4 The family of KP metabolites comprises compounds acting in a divergent way and considered to be either neuroprotective or neurotoxic. 5 The synthesis of specific compounds is tightly regulated and may considerably vary under physiological and pathological conditions. 5 Kynurenic acid (KYNA), the main neuroprotective compound of the path, was discovered in the 19 th century as a constituent of canine urine and initially regarded merely as a by-product of tryptophan degradation. 6 The molecular structure of KYNA was unraveled at the beginning of the 20 th century, 7 yet the particular steps of the KP leading to KYNA formation were determined much later. Discoveries of the 1980s revealed the ability of KYNA to block the excitatory amino acid receptors under in vitro and in vivo conditions. 8,9 Soon, abnormalities in cerebral KYNA synthesis have been implicated in the pathogenesis of neurodegeneration. 10 Intensified research during last four decades revolutionized our knowledge about KYNA and brought valuable data supporting the significant role of KYNA as an exceptional tryptophan metabolite in the mammalian brain. 11 This review aimed to discuss the involvement of altered KYNA metabolism in the development of neurodegenerative diseases, as well as the future of pharmacological manipulations aimed to boost brain KYNA as potential therapeutic agents.
The KP is functional in the brain and in the periphery. 12 The first step of tryptophan metabolism is catalyzed by the step-limiting enzymes, indoleamine 2,3-dioxygenases (IDO-1 and 2) and tryptophan 2,3-dioxygenase (TDO), yielding N-formyl-kynurenine ( Figure 1). Nformyl-kynurenine is further converted to a direct precursor of KYNA, L-kynurenine, by formamidase. In the periphery, the constitutive expression of IDO-1 is restricted and was described mostly within endothelial, pancreatic, placental, or antigen-presenting cells. Interestingly, F I G U R E 1 Scheme of kynurenine pathway IDO-1 manifests pronounced susceptibility to the induction by proinflammatory molecules, such as interferonγ, tumor necrosis factor α (TNFα), interleukin-6 (IL-6), or IL-10, in a variety of cells. 13,14 IDO-2 and TDO show higher tissue specificity, mainly restricted to liver, and much lower expression level. 15 The enzymatic activity of TDO can be induced by estrogens, glucocorticoids, and tryptophan itself. 2 In the brain, striatal neurons and astrocytes express high levels of IDO-1 mRNA. 16 TDO protein and its mRNA are also detectable in neurons and astrocytes. [17][18][19] The major central pool of KYNA is formed locally, from its precursor, L-kynurenine. 20,21 L-kynurenine, on the contrary, originates mostly from peripheral sources (60-70%), whereas the remaining 30-40% is produced in situ. 21 L-kynurenine can be also converted along another arm of the KP to neurotoxic 3-hydroxykynurenine (3-HK), QUIN, and further down to NAD. 22 The fate of L-kynurenine degradation and its availability for the synthesis of KYNA is determined by a number of factors, including tissue and cell type. Central KYNA production occurs mostly in astrocytes and endothelial cells and to a much lesser extent within neurons. [23][24][25][26] In contrast, neurotoxic QUIN is generated in the human brain mainly by the microglial cells and macrophages. 27 The principal route of KYNA synthesis is based on an irreversible transamination of L-kynurenine catalyzed by kynurenine transaminases (KATs). 28 KYNA is produced by various tissues and organs, including liver, kidneys, intestines, or endothelium. 29,30 In the brain, KATs are expressed mainly in astrocytes and to a lesser degree in neuronal cells, for example, in hippocampus, substantia nigra, or striatum. 24,25,[31][32][33] KATs are characterized by a different level of specific activity in various brain regions. 34,35 In humans and rodents, four isoforms of KATs, using L-kynurenine as a donor for amino group, were characterized and include KAT I (glutamine transaminase K/cysteine conjugate beta-lyase 1), KAT II (α-aminoadipate aminotransferase), KAT III (glutamine transaminase L/cysteine conjugate beta-lyase 2), and KAT IV (the mitochondrial aspartate aminotransferase/glutamic-oxaloacetic transaminase 2). 36 Each KAT enzyme has an optimal pH range and a distinct substrate profile, despite sharing a number of amino acid and α-keto acid substrates. [37][38][39] KATs manifest relatively low affinity for L-kynurenine (K m approx. 1 mM). Under physiological conditions, KAT II is considered a major biosynthetic enzyme responsible for KYNA formation. 40 A targeted deletion of KAT II in mice leads to an early and transitory decrease in brain KAT activity and KYNA levels with commensurate behavioral and neuropathological changes. 41 In KAT II-deficient mice, striatal KYNA level was transiently reduced around the 2 nd week of age and the degree of neuronal loss following the local administration of QUIN was strongly enhanced. 42 Later on, however, KYNA levels were normalized, possibly as a result of compensatory changes. 41,42 Indirectly, the activity of kynurenine monooxygenase (KMO), synthesizing 3-HK and displaying a much lower K m value for Lkynurenine, also impacts the synthesis of KYNA. Inhibition of KMO activity increases the pool of L-kynurenine available for KATs. This, in turn, may easily shift the KP and direct it to the neuroprotective branch; conversely, an enhanced activity of KMO stimulates metabolism of tryptophan along the neurotoxic arm of the pathway. 43 As shown under in vitro and in vivo conditions, the composition of the extracellular milieu, the availability of oxygen and glucose, or level of ammonia and amino acids may influence the synthesis of KYNA. [44][45][46][47] Notably, neurotoxic compounds such as mitochondrial toxins or pyrethroid pesticides inhibit, whereas a number of therapeutic agents, including beta-adrenoceptor agonists, nitric oxide donors, memantine, antidepressants, or some antiepileptics, stimulate KYNA production in the brain 48-53 .

| Other sources of brain KYNA
Apart from the canonical KAT-related synthesis, alternative mechanisms were implicated in the synthesis of KYNA. 54,55 Indole-3-pyruvic acid, the keto-analog of tryptophan, increases KYNA content in various organs, including brain. 56 Indole-3-pyruvic acid is effectively converted to KYNA in a non-enzymatic reaction requiring ample presence of oxygen. Reactive oxygen species (ROS) target the enol form of indole-3-pyruvic acid, which undergoes pyrrole ring cleavage and subsequently forms KYNA. 57,58 L-L-kynurenine may also yield KYNA when incubated in the presence of H 2 O 2 , with or without peroxidases. 55 It is a pH-dependent process, with the highest conversion of L-kynurenine occurring at the pH between 7.4 and 8.0. 55 The contribution of alternative routes to the overall KYNA production still remains unclear. However, in the altered redox environment and when the antioxidant system is defective, as often is the case in neurodegenerative disorders, their significance may increase. Indeed, the lack of correlation between KATs activities and KYNA levels was reported in lead intoxication, Down syndrome, and disturbances of thyroid hormone levels. [59][60][61][62] Furthermore, although peripherally synthesized KYNA poorly passes through the blood-brain barrier, certain conditions may facilitate its penetration into the brain. Systemic administration of KYNA prior to the cerebral ischemia potently increased its brain concentrations, 63 possibly as a result of passive diffusion. 21 In addition, KYNA was identified as a high-affinity substrate for organic anion transporters, OAT1 and OAT3. 64,65 Experimental use of probenecid, a non-selective inhibitor of Oat1, was shown to increase the brain level of KYNA. 66,67 Interestingly, thyroid hormones may enhance the removal of KYNA and modulate its brain level via diverse mechanisms, including the action of Oat. 59 Finally, apart from the de novo synthesis by the mammalian tissues, KYNA can be generated in the digestive system by microbiota and exogenously delivered with food products. 68,69

| Neurotoxic branch of kynurenine pathway
Three major neurotoxic metabolites of the KP include 3-HK, QUIN, and 3-hydroxyanthranilic acid (3-HANA). 3-HK is an immediate product of L-kynurenine conversion carried out by KMO. Metabolism of 3-HK by kynureninase yields 3-HANA, which, in two enzymatic steps, can be further converted to QUIN. The toxicity of 3-HK has been attributed mainly to the formation of free radicals and an induction of apoptotic neuronal death. 70,71 The results of numerous research clearly indicate that QUIN is capable of acting as an endogenous excitotoxin. QUIN-evoked neuronal loss is mostly associated with an excessive stimulation of NR2A and NR2B subunits of N-methyl-D-aspartate (NMDA) receptor at agonist-binding site. In the brain, physiological concentrations of QUIN are in nM range (~50-100 nM) and are approx. 20 times lower than in the periphery. 72 At low concentrations, QUIN induces proliferation of the stem cells and is an intermediate metabolite along the pathway yielding NAD+in human brain cells. 73,74 At high, close to millimolar concentrations, QUIN induces selective, axon-sparing neuronal loss under various experimental conditions. The susceptibility of neurons to the QUIN-induced damage depends on the brain area, with cortical, striatal, and hippocampal neurons being the most sensitive. 3 It was debated whether endogenous levels of QUIN are sufficient to cause neurotoxicity, yet, in the view of accumulated data, the compound undoubtedly may evoke neuronal death. In human brain, QUIN levels increase during inflammation or cerebral insults up to the micromolar values. 75 Locally, QUIN concentration may be much higher. 3 Notably, even low concentrations of QUIN may induce neuronal loss, providing that the exposure is prolonged. In organotypic corticostriatal, but not in caudate nucleus, cultures, exposed to low (100 nM) concentration of QUIN for up to 7 weeks, a clear focal neurodegeneration was developed. 76 QUIN was also demonstrated to enhance the glutamate release, to inhibit glutamate reuptake, and to stimulate lipid peroxidation. 77 Such local rise evokes depolarization of the postsynaptic membrane sufficient to remove the Mg 2+ block of NMDA receptor-linked ion channel. Moreover, QUIN may impair the function of blood-brain barrier, induce nitric oxide production, and cause hyperphosphor-

| B I OLOG I C AL K YNA TARG E TS IN THE B R AIN
The level of KYNA in the central nervous system (CNS) depends on the species, studied region, and the ontogenetic stage of development. 81 In human brain, KYNA occurs in low micromolar range (approx. 0.1-1.5 μM), which is 20-50 times higher than in rodent CNS (0.001-0.05 μM). 34,[82][83][84][85] The content of KYNA was reported to be the lowest in cerebellum and medulla (0.1-0.3 pmol/mg), intermediate in cortical areas and substantia nigra (0.2-0.6 pmol/mg), and the highest in putamen and globus pallidus (0.7-1.4 pmol/mg). 34,82 In human CSF, KYNA concentration is low (0.001-0.01 μM), yet it steadily increases with age. 81,86,87 In other species, the brain content of KYNA also rises with age. 88 Over 50-fold increase in brain KYNA between 1 st week and 18 th month of age was reported in rats. 89 Others demonstrated a threefold increase between the 3 rd and 24 th months of age. 90 KYNA is quickly liberated from the cell and is not a subject of enzymatic degradation or reuptake processes. 11 [94][95][96][97] Notably, KYNA attenuates the morphological and behavioral consequences of experimental administration of its kynureninergic alter ego-excitotoxic QUIN.
As KYNA may impact the extracellular levels of glutamate, acetylcholine, GABA, and dopamine, neuromodulation is an important aspect of its role. [98][99][100][101] In striatal preparations, low nanomolar concentrations of KYNA reduced glutamate release in caudate nucleus and impaired the neurotransmitter release. 102 Experimental studies in vivo confirmed that fluctuations of KYNA level may alter glutamine, acetylcholine, and dopamine release. 98,99,101 KYNA has also been identified as a ligand of formerly orphan G protein-coupled receptor, GPR35, 103 broadly expressed in various immune cells. Apart from the regulation of immune response, KYNA-GPR35 interaction may inhibit Ca 2+ channels in sympathetic neurons and reduce synaptic activity in hippocampal neurons. 104,105 Therefore, KYNA capability to activate GPR35 might represent another way to reduce the excitatory transmission. 105,106 KYNA is also targeting xenobiotic receptor, the aryl hydrocarbon receptor (AHR). 107 KYNA-related AHR stimulation increases the interleukin-6 expression, which is often associated with promoting carcinogenesis and tumor outgrowth. 107,108 Moreover, KYNA displays the scavenging ability toward ROS. In the homogenates of rat brains, KYNA decreased the production of free radicals and lipid peroxidation. 109 It has been postulated that KYNA targets also α7-nicotinic acetylcholine receptors (α7nAChR); however, this mechanism is still being controversial. [110][111][112] The inflammation emerged as one of the key factors contributing to the neuronal loss and compromised regeneration and thus was implicated in the pathogenesis of neurodegenerative disorders. An important link between proinflammatory status and the activation of KP is well substantiated. 11,86 Importantly, the metabolites of KP may act as pro-and antiinflammatory compounds. Genomic interventions aimed to eliminate IDO, TDO, or KMO were shown to alleviate the course of chronic inflammation, reduce viral replication, or change the expression of proinflammatory molecules. 113 On the contrary, a number of kynurenines, including KYNA, emerged as antiinflammatory compounds. 11,113 KYNA was demonstrated to attenuate inflammation by several ways including the reduction in TNF expression, diminished interleukin-4 and α-defensin secretion, or inhibition of Th17 cell differentiation, at least in part through activation of GPR35. 11 The interplay between immune activation and the KP activity results in a delicate balance, which may easily be shifted either to or away from neuroprotective KYNA ( Figure 3).
However, an excessive blockade of glutamate-mediated neurotransmission may impair cognition and memory processes. [114][115][116][117][118] Thus, manipulations of the endogenous KYNA level may exert dual, conflicting effects-beneficial neuroprotection and unfavorable cognitive dysfunction. Considering the chronic nature of neurodegenerative disorders, neuroprotection seems to be essential, as it may slow the progress of disease. Maintaining adequate levels of brain KYNA seems vital to obtain neuroprotection without cognitive adverse effects. The optimal therapeutic intervention would include a regionselective increase in KYNA; however, such pharmacological tools are not available yet.  132,137 In the advanced stages of HD, a reduction (−1.6-fold) in KYNA CSF levels was observed. 136 In the periphery, the baseline L-kynurenine levels were higher in HD and the difference remained obvious despite tryptophan depletion or loading. 138 Serum KYNA level in HD was not altered in comparison with control; however, the KYNA/L-kynurenine ratio was lower. 138 In a co-

| Parkinson's disease
Parkinson's disease (PD) is a common, progressive neurodegenera-  145,146 This highly selective neurotoxin causing nigral degeneration, followed by a classical PD-like behavioral pattern in various species, including rodents and primates, quickly became a valuable research tool. 146 The mechanisms underlying selective toxicity depend primarily on the glial conversion of MPTP to pyridinium metabolite (MPP+), which, upon release from astrocytes, inhibits neuronal mitochondrial respiratory chain and constitutes a source of free radicals. 147 Interestingly, MPP + was discovered to inhibit the cortical KAT activity and to reduce KYNA formation in vitro in rat cortical slices. 126 The effect was confirmed in vivo, as MPTP decreased KYNA synthesis and the density of KAT I immunoreactive nigral neurons in mice. 148,149 Consistently, KYNA pretreatment was shown to reduce the apoptosis of neurons by downregulating Bax expression and maintaining mitochondrial function, in human neuroblastoma cell line exposed to MPP + . 150 Human studies mostly demonstrate that in the brain of PD victims, the metabolism of tryptophan is shifted toward neurotoxic kynurenines with ensuing deficiency of KYNA. Postmortem studies reported diminished KYNA and L-kynurenine levels in frontal cortex, putamen, and substantia nigra, without change in tryptophan/ L-kynurenine and L-kynurenine/KYNA ratios, in the brains of PD victims. 151,152 In caudate and precentral cortical gyrus, KYNA content did not differ from control values. 80 In the periphery, the results are not consistent. In erythrocytes obtained from PD patients, higher levels of KYNA and enhanced activity of KAT II, but not of KAT I, were detected. In serum, KYNA level remained unchanged, while KAT I and KAT II activities were lower. 153 Similarly, an increase in L-kynurenine/tryptophan ratio, depletion of plasma tryptophan level, and increase in L-kynurenine and KYNA were reported. 154 Increase in serum KYNA was also observed among patients without dyskinesia, but not in dyskinetic PD patients. 155 In contrast, the deficiency of KYNA was revealed in a metabolomic study performed on a larger cohort of PD patients. Findings included lower plasma KYNA/L-kynurenine ratio, higher QUIN level, and increased QUIN/KYNA ratio. 156 Similarly, lower KYNA, higher QUIN, and an elevated QUIN/picolinic acid ratio in CSF, as well as high 3-HK in plasma, were detected. 157 The above data suggest that deficient KYNA synthesis seems to be limited to the brain in the course of PD, whereas in the periphery,

| Alzheimer's disease
Alzheimer's disease (AD) is the major cause of age-related dementia among elderly population. This progressive neurodegenerative disorder leads inevitably to a severe deterioration of cognitive functions and exerts dramatic negative impact on patients' quality of life. The characteristic neuropathological features of AD include senile plaques composed of beta-amyloid aggregates and neurofibrillary tangles built from hyperphosphorylated tau proteins. 158,159 Cholinergic neurons of the forebrain and hippocampal and cortical glutamatergic neurons are among the most affected areas. 160 In a transgenic mouse model of AD, a decrease in brain KYNA was confirmed. 161 However, the data on KYNA levels in AD patients are not consistent. 160,[162][163][164] Up to our knowledge, the data from brains of AD patients are very limited. In a small study involving postmortem analyses of specimens obtained from 11 patients with an advanced stage of AD, KYNA concentration was not altered in cortical areas, and increased in putamen (1.92-fold) and caudate nucleus (1.77-fold). 160 In latter structures, elevated KYNA correlated with the KAT I activity. 160 Lower KYNA levels was also detected in 5 brain structures obtained from AD victims. 83 Analyses of KYNA content in CSF of AD patients is not conclu- The potential role of disturbed KYNA formation in the pathogenesis of dementias is not fully understood. It is important to note that disproportionately high KYNA production, in our opinion secondary to the neuronal loss, may be aimed to further prevent the death of neurons in AD. Unfortunately, at high concentrations KYNA may act as a double-sword and actually impair working memory and contextual learning. [174][175][176][177] An increase in error frequency has been reported in rats treated intraperitoneally with L-kynurenine and manifesting high levels of brain KYNA, produced de novo within the brain from its precursor. 116 Similarly, adult rats treated throughout their adolescence with L-kynurenine exhibited deficits in contextual fear memory, a novel object recognition memory, but not cue-specific fear memory. 175 Adult rats exposed pre-and postnatally (gestation day 15-postnatal day 21) to L-kynurenine manifested a threefold increase in forebrain KYNA levels, a 2.5-fold increase in prefrontal cortex KYNA, and deficits in initial reversal learning and extradimensional shift. 176 Hence, KYNA, a metabolite of KP with neuroprotective effects at physiological concentrations, may exacerbate cognitive dysfunction and memory impairment in AD. However, as discussed above, an increase of brain KYNA levels most probably results from and is not a cause of neurodegeneration. In order to clarify this issue, longitudinal studies assessing the level of KYNA prior to and during the occurrence of overt symptoms of AD and dementia should be performed.

| THER APEUTI C PER S PEC TIVE S OF IN CRE A S ING K YNA LE VEL S IN NEURODEG ENER ATIVE DISORDER S
In the past, given the very limited penetration of KYNA through the BBB and its rapid removal from the brain and body, the use of KYNA in the treatment of neurodegenerative diseases seemed virtually impossible. 178 Various attempts aimed to refine the bioavailability of KYNA have brought promising results. The most successful approaches are based on the use of KYNA analogs penetrating through the blood-brain barrier, or modulation of the KP aimed to increase the concentration of KYNA substrate, L-kynurenine, in the periphery through an inhibition of selected key enzymes. The latter approach results in an enhanced availability of L-kynurenine for brain KYNA synthesis, as this KP metabolite easily enters central compartment.
As the current reports on the KYNA therapeutic abilities in vivo seem rather optimistic, bypassing the main obstacle by improving its bioavailability may be a milestone in introducing KYNA to the treatment of neurodegenerative diseases.

| Inhibition of KMO
Modulation of the KP-controlling enzymes is a crucial step toward the increased production of neuroprotective metabolites with a simultaneous reduction in neurotoxic QUIN and 3-HK in the brain.
Due to the fact that astrocytes do not express KMO, the major astrocytic product of tryptophan catabolism is KYNA. 179 In contrast, microglia and macrophages convert tryptophan along both arms of the KP-neuroprotective and neurodegenerative. 70 In such scenario, proinflammatory environment, consistently implied as one of the factors contributing to the development of neurodegeneration, leads to an ample production of neurotoxic kynurenines, such as QUIN or 3-HK. 180 Thus, inhibition of KMO allows astrocytes to retain more of L-kynurenine and to produce larger amounts of KYNA, sufficient to antagonize the glutamate and QUIN excitotoxicity. Indeed, it is broadly documented that KMO activity is crucial for directing the metabolic fate of L-kynurenine, and thus influences the QUIN/KYNA ratio the most. 181 A number of KMO inhibitors were synthesized and tested in various experimental models. 182,183 The initial studies were carried out prior to the identification of crystal structure of KMO; thus, the design of earliest inhibitors was based on the structure of L-kynurenine. 182 One of the first KMO inhibitors used experimentally was nicotinylalanine. 184,185 The development of selective KMO inhibitors started in the 1990s, with introduction of m-nitrobenzoylalanine (mNBA) showing the IC 50 = 0.9 μM. 186 Experimental administration of 400 mg/kg of mNBA to rats resulted in a substantial increase of L-kynurenine and KYNA levels-13-and fivefold in the brain, fivefold and 2.4-fold in the blood, and sixfold and 3.5-fold in the liver, respectively. 183  provided neuroprotection in the rat and gerbil ischemia models, displayed antiepileptic activity against electroshock-induced seizures in mice and rats, and reduced the cerebral QUIN accumulation in mice subjected to immune activation. 66,[190][191][192] Due to the relative instability of Ro 61-8048, its slow-release prodrug form,

| Parkinson's disease
Neuroprotective and antiparkinsonian effects of glutamate receptor antagonists are well documented. 143,198 In line with these data, either administration of KYNA itself or use of pharmacological tools increasing the availability of L-kynurenine and its conversion to KYNA may reduce neuronal loss and behavioral symptoms in experimental PD models. 185,[199][200][201][202] Modifications of tryptophan metabolism seem to exert dual therapeutic benefit in PD, neuroprotection, and prevention of L-DOPA-induced motor side effects. 203 Direct application of KYNA into the medial segment of the globus pallidus reduced the behavioral symptoms in MPTP-induced PD model. 199 Similarly, in monkeys with hemiparkinsonism induced by unilateral, intraarterial administration of MPTP, KYNA infusion into the contralateral globus pallidus internus alleviated the disease symptoms. 200 The intracerebroventricular infusion of nicotinylalanine, inhibiting kynureninase and L-kynurenine hydroxylase activity, combined with L-kynurenine and probenecid, an inhibitor of organic acid transport, substantially increases KYNA level in rodent brain. This approach was used to raise KYNA content in rat substantia nigra and appears to be sufficient to protect neurons from QUIN-induced toxicity. 185  shown to exert neuroprotective effects in experimental PD models is also able to increase KYNA formation and to prevent the MPP +induced decline in KYNA synthesis. 207,208 Experimental data consistently imply that shift of the KP to neurotoxic branch producing 3-HK and QUIN formation, with relative or absolute deficiency of KYNA, is an important factor contributing to the development and progress of PD. We are still awaiting the synthesis of more precise pharmacological tools, able to modulate the KP within basal ganglia in a selective manner, which may become a promising therapeutic option for PD.

| Alzheimer's disease
Only limited studies exploited KMO inhibitors as possible therapy for AD. Chronic oral therapy with JM6, inhibitor of KMO, was demonstrated to rise brain KYNA, due to de novo synthesis of the compound from L-kynurenine, and to reduce extracellular glutamate in a transgenic mouse model of AD. 161 JM6 did not exert significant effects on Aβ plaque formation; however, it prevented spatial memory deficits. The compound also extended life span, prevented synaptic loss, and decreased microglial activation. 161

| KMO inhibitors-limitations
Despite a large therapeutic potential, there are important drawbacks of some currently available KMO inhibitors. Certain compounds, such as mNBA and UPF-648, were found to act as uncouplers of NADPH oxidation, which may actually potentiate neuronal loss via generation of cytotoxic hydrogen peroxide. 209 Therefore, precise design of novel compounds, effectively increasing brain KYNA levels, yet devoid of harmful production of free radicals, remains an important goal in the development of drugs against neurodegenerative disorders.

| Analogs and prodrugs of KYNA
The goal of creating new KYNA analogs and prodrugs was to overcome the obstacle of poor BBB penetration by KYNA itself and to synthesize precursors that preferentially would not be metabolized to neurotoxic kynurenines. As a result, chlorokynurenines, including 4-chlorokynurenine and 4,6-dichlorokynurenine, emerged, meeting the above criteria, including fast delivery into the CNS and, once in the brain parenchyma, an easy conversion to potent NMDA antagonists acting at the glycine site, 7-chlorokynurenic acid, or 5,7-dichlorokynurenic acid. 210 Similar parameters characterize also esterified analogs and esterified 4amino analogs. 94 Another approach to improve the BBB penetration was based on utilizing D-glucose ester of 7-chlorokynurenic acid. The conjugate manifests improved BBB penetration as a result of an active transport by the glucose transporter GLUT1. 211 Indeed, systemic administration of the conjugates resulted in an anticonvulsant effect in mice affected by NMDA-associated seizures.

| Parkinson's disease
Up to our knowledge, KYNA analogs were not studied in experimental models of Parkinson's disease. However, a number of compounds able to increase KYNA levels, through mechanisms distinct from interference with KP, successfully ameliorated L-DOPA-induced dyskinesias. One of the interesting therapeutic options for PD seems to be the antiepileptic drug zonisamide, shown to reduce motor symptoms in patients with L-DOPA-induced dyskinesias. 215 Zonisamide, apart from his broad pharmacological effects including inhibition of voltagegated sodium channels, T-type calcium channels, and monoamineoxidase, has been shown to increase KYNA production. 216

ACK N OWLED G EM ENTS
This study was supported by the Medical University in Lublin grants DS 450/20, 450/21.

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflicts of interest.