Acute Administration of Ibuprofen Increases Serum Concentration of the Neuroprotective Kynurenine Pathway Metabolite, Kynurenic Acid: a pilot randomized, placebo-controlled, cross-over study

Abstract

Rationale:

At least six different types of anti-depressant treatments have been shown to either increase the neuroprotective kynurenine pathway (KP) metabolite, kynurenic acid (KynA), or decrease the neurotoxic KP metabolite, quinolinic acid (QA). Non-steroidal anti-inflammatory drugs (NSAIDs) including ibuprofen have shown some efficacy in the treatment of depression but their effects on the KP have not been studied in humans.

Objectives:

To evaluate the effect of ibuprofen on circulating KP metabolites.

Methods:

In a randomized, placebo-controlled, cross-over study, 20 healthy adults (10 women) received a single oral dose of 200mg ibuprofen, 600mg ibuprofen or placebo in a counterbalanced order (NCT02507219). Serum samples were drawn in the mid-afternoon, 5 hours after ibuprofen/placebo administration. KP metabolites were measured blind to visit by tandem mass spectrometry. Data were analyzed with linear mixed effect models. The primary outcome was KynA/QA and the secondary outcome was KynA.

Results:

After Bonferroni correction, there was a significant effect of treatment on KynA/QA. The effect was driven by an increase in KynA concentration after the 600mg dose but not the 200mg dose relative to placebo (Cohen’s d=1.71). In contrast, both the 200mg (d=1.03) and 600mg (d=2.05) doses of ibuprofen decreased tryptophan concentrations relative to placebo.

Conclusions:

Given its KynA-elevating effects, ibuprofen could have neuroprotective effects in the context of depression as well as other neuroinflammatory disorders that are characterized by a reduction in KynA.

INTRODUCTION

The putative role of inflammation in the etiology of mood disorders has led to interest in treating depression with anti-inflammatory medications, including widely-used non-steroidal anti-inflammatory drugs (NSAIDs) which reduce inflammation by inhibiting the cyclooxygenase (COX) enzymes, COX-1 and/or COX-2 (Yagami et al. 2016). A recent meta-analysis reported a therapeutic signal for the COX-2 inhibitor, celecoxib, when used as an add-on for the treatment of depression (Kohler-Forsberg et al. 2019). However, the largest study to date found no efficacy for celecoxib versus placebo in patients with bipolar depression (Husain et al. 2020). In fact, pharmacoepidemiological studies generally suggest that inhibition of COX-2 may be counterproductive but that preferential COX-1 inhibition such as in the form of low-dose aspirin may be protective (Hu et al. 2020; Kessing et al. 2019; Kohler et al. 2015; Stolk et al. 2010; Warner-Schmidt et al. 2011; Zhang et al. 2022). Consistent with these data, we found beneficial effects of adjunctive treatment with low-dose aspirin in the context of bipolar depression (Savitz et al. 2018) although a recent clinical trial of adjunctive low-dose aspirin in youth with major depressive disorder (MDD) was negative (Berk et al. 2020a) and another recent study reported that low-dose aspirin failed to protect against the development of depression in older adults (Berk et al. 2020b).

The anti-depressant effects of ibuprofen a non-selective NSAID that blocks both COX-1 and COX-2 isoforms, have been less well studied. Preclinical work has shown that ibuprofen alleviates depression-like behavior after Bacillus Calmette-Guerin inoculation (Saleh et al. 2014), the forced swim test (Norden et al. 2015), and chronic restraint stress (Seo et al. 2019). An epidemiological study of selective serotonin reuptake inhibitor (SSRI) users showed that concurrent use of ibuprofen reduced the risk of psychiatric contacts over the follow-up period (Kohler et al. 2015) while a pooled analysis of NSAID trials for osteoarthritis showed that 800mg of ibuprofen T.I.D for six weeks was associated with a small but significant reduction in depressive symptoms versus placebo (Iyengar et al. 2013). Data from the FDA’s safety and adverse event reporting program, “MedWatch”, suggested that ibuprofen might have anti-depressant effects in women but not men (Lehrer and Rheinstein 2019). Partially consistent with these data, an observational trial of women at high-risk for depression in the post-partum period reported that utilization of ibuprofen was protective although the effect was marginal (Kapulsky et al. 2021).

Like other NSAIDs, ibuprofen inhibits the COX enzymes that promote inflammation by catalyzing the formation of prostaglandins and thromboxanes from arachidonic acid (although it likely has many other “off-target effects”). One consequence of the inflammatory process is activation of the kynurenine pathway (KP) leading to the production of several neuroactive metabolites that modulate a range of physiological processes, including glutamatergic and dopaminergic neurotransmission (Dantzer et al. 2011; Schwarcz et al. 2012) (Figure 1). We and others have shown that depression is associated with a decrease in the circulating concentration of the NMDA receptor antagonist, kynurenic acid (KynA) relative to the NMDA receptor agonist, quinolinic acid (QA) (Bartoli et al. 2021; Bay-Richter et al. 2015; Myint et al. 2007; Paul et al. 2022; Savitz 2020; Wurfel et al. 2017). Some groups have also reported an increase in QA in post-mortem brain samples from patients with severe depression (Steiner et al. 2011) and in the cerebrospinal fluid (CSF) of suicide attempters (Erhardt et al. 2013). Further, several forms of anti-depressant therapy including escitalopram (Halaris et al. 2015), ketamine (Kadriu et al. 2019; Verdonk et al. 2019; Zhou et al. 2018), electroconvulsive therapy (Guloksuz et al. 2015; Schwieler et al. 2016), cognitive behavior therapy (Savitz et al. 2020), aerobic exercise (Agudelo et al. 2014; Javelle et al. 2021) and real-time fMRI neurofeedback (Tsuchiyagaito et al. 2021) have been shown to either increase KynA or decrease QA, raising the possibility of a common underlying mechanism of action. Given these data and the anti-depressant potential of NSAIDs (albeit equivocal) we were interested in the effect of acute ibuprofen administration on the KP. Here, we tested the effect of a single dose of ibuprofen on serum KP metabolites in healthy adults using a cross-over design with two different doses of ibuprofen (600 mg and 200 mg) and one dose of placebo. Consistent with our recent work (Savitz et al. 2020; Tsuchiyagaito et al. 2021; Zheng et al. 2022) we selected KynA/QA as the primary outcome in order to evaluate the balance between the putative neuroprotective and neurotoxic arms of the kynurenine pathway. We hypothesized that ibuprofen, especially at the 600mg dose, would increase serum concentrations of KynA and/or decrease concentrations of QA relative to placebo.

Figure 1,

reproduced from (Savitz 2020), illustrates the main branches of the kynurenine pathway (KP). Approximately 95% of tryptophan is metabolized into kynurenine by the enzymes, indoleamine dioxygenase (IDO) and tryptophan dioxygenase (TDO). During inflammation IDO is upregulated and kynurenine is in turn preferentially converted down the quinolinic acid (QA) pathway at the expense of the neuroprotective metabolite, kynurenic acid (KynA). Theoretically, this rerouting of metabolism is in order to generate NAD for activated immune cells but may have adverse consequences in the form of the generation of neurotoxic metabolites, including the NMDA receptor agonist QA.

METHODS

Methodological details have been published elsewhere (Burrows et al. 2022; Cosgrove et al. 2021) and are summarized here. The research was conducted in accordance with the Helsinki Declaration of 1975. Participants provided written informed consent after receiving a full explanation of the study procedures and risks, as approved by the Western IRB. Participants were recruited between May and October of 2015. Twenty-two healthy participants (11 women) aged 18–50 years participated in a double-blind, randomized, crossover study (NCT02507219; Table 1). Two participants withdrew before completion, so that data from 20 participants were analyzed (n=10 women). The CONSORT diagram is shown in Figure 2. Exclusion criteria were as follows: (1) history of any psychiatric disorder (2) current or past 6-month alcohol or drug abuse; (3) regular use (> 15 days for past 30 days) of NSAIDS; (4) history of clinically significant hepatic, cardiac, renal, neurologic, cerebrovascular, metabolic, gastric, or pulmonary disease; (5) history of seizure disorders; (6) pregnancy or plan to become pregnant within the next 18 weeks; (7) women who are currently menstruating; (8) fMRI-related exclusion criteria (e.g. claustrophobia, metal implants).

Table 1.

Characteristics of the Study Participants

Variable
N	20
Age (years)	32.4 ± 6.7
Sex (% male)	10 (50.0)
Body Mass Index (kg/m2)	26.6 ± 5.8
Ethnicity
White	12 (100%)
Hispanic or Latino	0 (0%)
Educational status
Some college, no degree	3
Associate degree	11
Bachelor’s degree	2
Master’s degree	2
Missing data	2
Concomitant medication
Birth Control	3
OTC allergy medications	3
Phentermine	1
Clonazepam	1
Levothyroxine	1

Figure 2. CONSORT Flow Diagram.

illustrates the number of participants recruited, randomized, and analyzed in the study.

Each participant completed a baseline screening assessment and was scheduled for three subsequent visits. During each visit, scheduled 2 weeks apart to preclude any potential carryover effects, participants received one oral pill containing placebo, 200 mg of ibuprofen, or 600 mg of ibuprofen, so that all participants received a total of 3 different doses. Ibuprofen was capsuled, and identical placebo capsules were produced by a local compounding pharmacy. The medication was administered together with a snack after an overnight fast at either 8am or 10am. Subsequently, participants completed an MRI scan, rating scales, and behavioral tasks. Blood was drawn approximately 5 hours post administration of ibuprofen/placebo in the mid-afternoon. Administration order was counterbalanced across participants, and experimenters were blinded to the medication codes. Although participants reported occasional over-the-counter NSAID use during the study period, they never took an NSAID within 48 hours of a study visit.

Serum samples were collected with BD Vacutainer serum tubes, processed according to the standard BD Vacutainer protocol, and stored at −80 C. Serum concentrations of tryptophan (TRP), kynurenine, kynurenic acid (KynA), 3-hydroxykynurenine (3HK), and quinolinic acid (QA) were measured blind to visit by Charles River Laboratories in January 2018. The KP metabolite concentrations were determined by high performance liquid chromatography (HPLC) with tandem mass spectrometry (MS/MS) detection using their standard protocols. The lowest level of quantification and intra-assay percentage of coefficient of variation for each of the KP metabolites were as follows: TRP: 3 μM, 1.4%; kynurenine: 0.225 μM, 1.6%; KynA: 7.5 nM, 6.6%; 3HK: 5 nM, 3.9%, and QA: 50 nM and 8.6%.

Data were analyzed with linear mixed effect models with treatment condition (i.e. 600mg, 200 mg, or placebo) as the independent variable and participant as a random factor. Covariate inclusion was determined by Bayesian Information Criterion (BIC) model comparison for all combinations of age, sex, and body mass index (BMI). Carry over effects were ruled out by testing the interaction between visit number and treatment condition for each outcome variable. Given the above-mentioned effect of antidepressant treatments on the KP, we selected the ratio of KynA to QA as the primary outcome. The secondary outcome was the change in KynA and the other KP metabolites were included as exploratory outcomes. We performed a Bonferroni correction to control for multiple comparisons (P<0.006, 2-tailed test).

RESULTS

There was a significant effect of treatment on the primary outcome, KynA/QA (F_2,38 =5.79, p=0.006) such that KynA/QA concentration was increased after the 600mg dose of ibuprofen (Cohen’s d=0.99) relative to placebo (Table 2; Figure 3). The effect was driven by a significant increase in the secondary outcome variable, KynA (F_2,38 =8.03, p=0.001) after the 600mg dose of ibuprofen (Cohen’s d=1.71) relative to the 200mg dose and placebo (Figure 3). Consistent with these results, there was a significant increase in KynA/Kyn (F_2,38 =9.87, p<0.001) after the 600mg dose of ibuprofen (Cohen’s d=1.42) relative to the 200mg dose and placebo. There was no significant effect of treatment condition on QA, 3-HK, kynurenine, and Kyn/TRP concentrations (all p’s >0.01). There was, however, a significant effect of treatment on TRP concentrations (F_2,38 = 10.1, p<0.001) such that TRP concentrations were decreased after both the 200mg dose (Cohen’s d=1.03) and 600mg dose (Cohen’s d=2.05) of ibuprofen relative to placebo (Figure 3).

Table 2:

Kynurenine Metabolites by IBU Dosage. Mean (SEM) and Cohen’s D of each metabolite and ratio grouped by IBU dose condition. Unadjusted Cohen’s D statistics are given for the difference between the IBU dose condition and placebo (reference). Units in ratios converted to nM where necessary.

IBU Dose	Placebo	200mg		600mg
	Mean(SEM)	Mean(SEM)	Cohen’s D	Mean(SEM)	Cohen’s D
TRP (μM)	51.5 (2.1)	46.0 (1.8)	0.64	44.0 (1.9)	0.86
KYN (μM)	1.64 (0.07)	1.58 (0.06)	0.22	1.60 (0.07)	0.13
KynA (nM)	30.4 (2.5)	31.3 (2.0)	0.09	37.0 (2.7)	0.57
3HK (nM)	22.8 (1.9)	22.3 (0.9)	0.06	24.2 (1.3)	0.19
QA (nM)	262 (18)	261 (14)	<0.01	264 (15)	0.03
KYN/TRP	0.032 (0.001)	0.035 (0.02)	0.39	0.07 (0.002)	0.63
KynA/KYN	0.019 (0.001)	0.020 (0.002)	0.28	0.024 (0.002)	0.71
KynA/3HK	1.43 (0.11)	1.45 (0.12)	0.04	1.57 (0.12)	0.29
KynA/QA	0.12 (0.01)	0.13 (0.01)	0.04	0.15 (0.02)	0.45

IBU = ibuprofen, TRP = tryptophan, KYN = kynurenine, KynA = kynurenic acid, 3HK = 3-hydroxykynurenine, QA = quinolinic acid, SEM = standard error of the mean

Figure 3.

Boxplots illustrating the mean concentrations of KynA/QA, KynA, TRP, Kyn, 3-HK, and QA in serum approximately five hours after a single oral dose of placebo, 200mg of ibuprofen, or 600mg of ibuprofen. The solid line in BOLD represents the median, the boxes represent the 25th-75th percentiles, and the tails, the 95% confidence interval.

DISCUSSION

This study revealed that oral ingestion of a 600mg dose of ibuprofen induced a significant (~25%) increase in serum KynA concentrations relative to placebo. This ibuprofen-associated increase in circulating KynA is consistent with several preclinical studies. Edwards et al. initially showed that a single dose of the non-selective COX inhibitor, diclofenac (40mg/kg, s.c.), increased KynA levels in the rat brain one hour after administration (Edwards et al. 2000). Subsequent work confirmed this effect at 50mg/kg, i.p. 3.5 hours after diclofenac administration and extended this finding to indomethacin (50 mg/kg, i.p.), a preferential COX-1 inhibitor (Schwieler et al. 2005). A follow-up study showed that the same dose of indomethacin increased KynA by 150–300% and increased the tonic firing of dopamine neurons in the ventral tegmental area (VTA) (Schwieler et al. 2006). More recently, KynA concentrations in both plasma and hippocampus were shown to be significantly increased after a single ibuprofen dose in rats (Maciejak et al. 2013). The increase in KynA was significant for both 75 and 150mg/kg i.p. doses but the effect was strongest at the higher dose. In contrast, the selective COX-2 inhibitors, parecoxib and meloxicam decreased brain concentrations of KynA (Schwieler et al. 2005) and decreased firing rate and burst firing activity in the VTA (Schwieler et al. 2006). The decrease in KynA observed in the case of selective COX-2 inhibitors versus the increase in KynA associated with non-selective NSAIDs could conceivably explain why selective COX-2 inhibitors have been reported to worsen depression while COX-1 inhibitors or non-selective NSAIDs may have anti-depressant properties (Hu et al. 2020; Kessing et al. 2019; Kohler et al. 2015; Savitz et al. 2018; Stolk et al. 2010; Warner-Schmidt et al. 2011; Zhang et al. 2022). Further research is needed to test this hypothesis.

The mechanism through which ibuprofen increases KynA is unclear. Schwieler and colleagues found that they could block the indomethacin-induced increase in brain KynA with the prostaglandin E1/E2 agonist, misoprostol, suggesting that the effect may be mediated by the inhibitory action of prostaglandins on the synthesis of KynA or its precursor, kynurenine (Schwieler et al. 2005). Thus, it is possible that KynA concentrations are altered because of upstream changes in the KP. However, we did not observe a significant effect of treatment on kynurenine concentration in this study. Third, there is some evidence that the non-selective NSAID, diclofenac inhibits kynurenine monooxygenase (KMO), the enzyme that converts kynurenine to 3-HK (Modoux et al. 2021). Auer and colleagues compared the molecular structure of known KMO inhibitors to a database of existing medications and found a match with diclofenac. The in silico data were then confirmed with an in vitro KMO cell lysate enzymatic inhibition assay which showed that diclofenac had inhibitory effects on KMO (Shave et al. 2018). It is conceivable that ibuprofen and perhaps other non-selective NSAIDs also possess inhibitory activity against KMO which would then bias metabolism of kynurenine towards KynA. However, in a post-hoc analysis, we observed that ibuprofen was associated with a non-significant increase in 3HK/Kyn, which we would not expect if ibuprofen was inhibiting KMO.

The other result of interest was the decrease in serum TRP observed with both the 200mg and 600mg doses of ibuprofen relative to placebo. The decrease in TRP concentration is consistent with a preclinical study that showed that single doses of both 75mg/kg and 150mg/kg i.p. of ibuprofen resulted in significant decreases in plasma TRP in rats (Maciejak et al. 2013). The decrease in serum TRP concentration is likely a consequence of TRP being displaced from its binding site on albumin by ibuprofen, resulting in greater renal clearance. Albumin is a serum protein that serves as a transport vehicle for many ligands and drugs, including TRP and ibuprofen, respectively. A recent study showed that infusion of 800mg of ibuprofen during hemodialysis treatment could displace several toxins as well as TRP from albumin augmenting their removal and decreasing their concentrations in serum (Madero et al. 2019).

According to the monoamine theory of depression, a decrease in TRP, the precursor of serotonin, would likely worsen depressive symptoms (Schildkraut and Kety 1967). However, this hypothesis has fallen into disfavor in recent years, in part because no convincing mechanism of monoamine loss has been discovered. Interest in the KP was originally viewed through the lens of the monoamine theory of depression. That is, a key pathophysiological mechanism was thought to be the preferential breakdown of TRP into kynurenine at the expense of serotonin (Capuron et al. 2002; Lapin and Oxenkrug 1969; Widner et al. 2002). However, several preclinical studies have demonstrated that inflammatory stimuli actually increase rather than decrease brain TRP and serotonin (Dunn and Welch 1991; O’Connor et al. 2009). Moreover, Dantzer and colleagues demonstrated that the depressive effects of lipopolysaccharide (LPS) could be blocked with an inhibitor of indoleamine 2’3 dioxygenase (IDO) without affecting brain TRP and serotonin turnover (O’Connor et al. 2009). Thus, the current model postulates that an imbalance between downstream neuroprotective and neurotoxic KP metabolites such as KynA and QA is the critical pathophysiological mechanism underlying depression (Myint and Kim 2003; Savitz 2020; Savitz et al. 2015; Wichers et al. 2005). If this hypothesis is correct and ibuprofen does increase KynA concentrations in the brain, then this raises the possibility that ibuprofen may have neuroprotective effects not just in the case of depression, but in several other neuroinflammatory disorders characterized by a decrease in KynA. For instance, epidemiological data suggest that ibuprofen reduces the risk for Parkinson’s Disease (PD) (Ascherio and Schwarzschild 2016; Gao et al. 2011) and decreased concentrations of KynA have been reported in the cerebrospinal fluid and plasma of patients with PD (Heilman et al. 2020; Lim et al. 2017; Sorgdrager et al. 2019). Conceivably, ibuprofen could be protecting against PD by increasing KynA levels and thereby limiting the excitotoxic effects of QA on the NMDA receptor.

The main limitation of this work is that the data were collected from medically healthy individuals with no history of psychiatric disorders. Thus, we were unable to test whether ibuprofen had any anti-depressant effects and whether these effects were mediated by changes in inflammatory mediators or KP metabolites. Similarly, we were unable to evaluate the effect of ibuprofen on pain. For instance, Schwieler and colleagues have previously proposed that increased concentrations of brain KynA may contribute to the analgesic effects of non-selective NSAIDs (Schwieler et al. 2005). Second, we did not obtain measurements of COX activity or prostaglandin and thromboxane metabolites although given the healthy nature of the sample we may have had limited sensitivity. Third, since this was a pilot study, the sample size was small, and we did not perform an a priori power analysis. Fourth, KP concentrations were measured in the serum and do not necessarily extrapolate to the brain parenchyma - for a recent discussion of this issue see (Savitz 2022). Nevertheless, it is worth noting that ibuprofen was reported to increase KynA concentrations in both the hippocampus and plasma of rats (Maciejak et al. 2013). Finally, we used a single-dose design which eliminated the problem of compliance but raises the issue of whether similar effects on the KP would also occur after chronic treatment with ibuprofen.

In sum, this randomized, placebo controlled, cross-over design provides preliminary evidence that ibuprofen acutely increases the serum concentration of the neuroprotective KP metabolite, KynA. This result raises the possibility that ibuprofen could have therapeutic benefits in the context of diseases that are characterized by reductions in KynA such as mood disorders. Further, the data raise the possibility that ibuprofen could be used as an experimental tool to manipulate the KP in order to better understand its role in the pathophysiology of neurological and psychiatric disorders.