Old and New Analgesic Acetaminophen: Pharmacological Mechanisms Compared with Non-Steroidal Anti-Inflammatory Drugs

Abstract

Although it is more than a century since it was first marketed, acetaminophen remains one of the most popular analgesic agents. In addition, acetaminophen has recently been applied to multimodal analgesia in combination with non-steroidal anti-inflammatory drugs, and its consumption significantly increased during the pandemic of coronavirus disease 2019 as well as diclofenac and ibuprofen. However, the detailed mode of analgesic action of acetaminophen is still unclear. In the present study, we comprehensively discuss conventional, recognized, and postulated mechanisms of analgesic acetaminophen and highlight the current mechanistic concepts while comparing with diclofenac and ibuprofen. Acetaminophen inhibits cyclooxygenase with selectivity for cyclooxygenase-2, which is higher than that of ibuprofen but lower than that of diclofenac. In contrast to diclofenac and ibuprofen, however, anti-inflammatory effects of acetaminophen depend on the extracellular conditions of inflamed tissues. Since the discovery of cyclooxygenase-3 in the canine brain, acetaminophen had been hypothesized to inhibit such a cyclooxygenase-1 variant selectively. However, this hypothesis was abandoned because cyclooxygenase-3 was revealed not to be physiologically and clinically relevant to humans. Recent studies suggest that acetaminophen is deacetylated to 4-aminophenol in the liver and after crossing the blood–brain barrier, it is metabolically converted into N-(4-hydroxyphenyl)arachidonoylamide. This metabolite exhibits bioactivities by targeting transient receptor potential vanilloid 1 channel, cannabinoid receptor 1, Cav3.2 calcium channel, anandamide, and cyclooxygenase, mediating acetaminophen analgesia. These targets may be partly associated with diclofenac and ibuprofen. The perspective of acetaminophen as a prodrug will be crucial for a future strategy to develop analgesics with higher tolerability and activity.

1. Introduction

Analgesic and antipyretic activities were first discovered in acetanilide. However, its severe side effects promoted the development of less toxic aniline derivatives, resulting in synthesis of phenacetin (4′-ethoxyacetanilide) and acetaminophen (4′-hydroxyacetanilide, also known as paracetamol) in the 1880s. Although phenacetin was initially more popular than acetaminophen, it was found to cause hemolytic anemia, methemoglobin formation, and kidney damage. Consequently, acetaminophen gained popularity to replace phenacetin and became the most widely used over-the-counter and prescribed medicines to reduce pain and fever [1,2]. Although more than a century has passed since first marketed, acetaminophen still remains one of analgesic agents commonly used to alleviate acute and chronic pain in general population [3].

In addition to the conventional use, acetaminophen is recently used in multimodal analgesia [4], frequently by combining with ibuprofen and diclofena that are in a class of non-steroidal anti-inflammatory drugs (NSAIDs) [5]. Multimodal analgesia involves a combination of two or more different types of analgesic drugs that induce additive and/or synergistic interactions to potentiate analgesia and minimize side effects of each individual drugs. While acetaminophen as a sole drug shows relatively mild to moderate effects, it consistently increases the analgesic efficacy in multimodal analgesia [6]. The multimodal strategy using acetaminophen and NSAIDs has been applied to preoperative, intraoperative, and postoperative pain management [7].

From the emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in December 2019, coronavirus disease 2019 (COVID-19) caused by SARS-CoV-2 spread rapidly all over the world. Because patients infected with SARS-CoV-2 develop headache, myalgia, and fever in most cases, acetaminophen was commonly used for attenuating such COVID-19 symptoms [8] as well as diclofenac and ibuprofen [9]. The early and midterm use of acetaminophen after diagnosis of SARS-CoV-2 infection is effective for COVID-19 patients and does not increase the risk of COVID-19-related hospitalization and death [10]. The analgesic use of acetaminophen became increasingly important, and its consumption significantly increased during the COVID-19 pandemic similarly to that of diclofenac and ibuprofen [11]. Acetaminophen was most frequently used for treating COVID-19 patients, followed by ibuprofen and diclofenac [12]. A retrospective analysis provided an insight that acetaminophen intake is correlated with a lower risk of SARS-CoV-2 infection and a decreased expression of angiotensin-converting enzyme 2, a receptor for the cellular entry of SARS-CoV-2, whereas ibuprofen influences neither COVID-19 occurrence nor the receptor protein level [13].

Despite the fact that acetaminophen has been used in the world for a long time and its application is currently spreading more widely, the detailed mode of action of analgesic acetaminophen remains poorly understood. Different theories and hypotheses have been proposed thus far, including interactions with enzymes, channels, receptors, and lipid membranes [14,15,16]. Excellent reviews on acetaminophen were published by Mallet et al. [17], Ohashi and Kohno [18], and Ayoub [19]. However, given recent advances in acetaminophen pharmacology following an innovative study of Högestätt et al. [20], it is essential to update our knowledge about the mechanisms underlying an analgesic effect of acetaminophen.

Besides the frequent use common to acetaminophen and NSAIDs, acetaminophen structurally shares an aniline moiety with diclofenac that is classified into phenylacetic acid derivatives of NSAIDs together with ibuprofen as shown in Figure 1. However, the pharmacological mechanisms of acetaminophen could be distinct from these NSAIDs as suggested by the combined use of acetaminophen and NSAIDs that induces better pain alleviation compared with either drug alone at the same dose [21]. Acetaminophen may act on different targets or on the same targets but with a different potency compared with NSAIDs, and the inhibition of cyclooxygenase (COX) as the primary target of NSAIDs may not be the main mode of action of acetaminophen.

Pain is divided into three categories: nociceptive, inflammatory, and neuropathic pain based on the initiating stimuli [22]. Nociceptive pain is induced by noxious stimuli generated peripherally, inflammatory pain occurs in inflamed tissues where proinflammatory mediators sensitize nociceptors, and neuropathic pain is caused when the nervous system has some damage. Analgesic agents are considered to attenuate these pains in the peripheral nervous system and in the central nervous system (CNS) by acting on the enzymes, channels, and receptors responsible for nociception and pain transmission. The primary purpose of the present study was to focus on the interactions of acetaminophen and its related compounds with these biomolecules and comprehensively discuss the conventional, recognized, and postulated mechanisms of analgesic acetaminophen, particularly by highlighting the current mechanistic concepts, while making comparisons with NSAIDs such as diclofenac and ibuprofen.

2. Selective Cyclooxygenase Inhibition

In response to irritation, injury, or receptor stimulation, arachidonic acid is released from cell membrane phospholipids by phospholipase A₂ and subsequently converted into prostanoids by COX as shown in Figure 2. Among them, specific prostaglandins (PGs) cause pain by activating nociceptors at the peripheral sites and enhancing nociceptive signal transmission in the nervous system, and they mediate inflammatory swelling and pain. The analgesic and anti-inflammatory effects of NSAIDs such as diclofenac and ibuprofen are primarily ascribed to the inhibition of COX that is responsible for the production of proinflammatory PGE₂ and PGI₂ [23]. There are two structurally distinct isoforms of COX, that is, constitutively expressed COX-1 relevant to physiological functions and inducible COX-2 upregulated in pathological conditions. NSAIDs are considered to exert beneficial effects by inhibiting COX-2 but adverse effects by inhibiting COX-1. Because NSAIDs that non-selectively inhibit COX have both effects, COX-2-selective inhibitors (coxibs represented by celecoxib) were developed to lessen or eliminate side effects such as gastrointestinal erosion, ulceration, and bleeding. Although acetaminophen exhibits analgesic and antipyretic activities as well as NSAIDs, it has been claimed that acetaminophen lacks the potent effect to suppress PG production in inflamed tissues. However, different studies suggest that acetaminophen appears to share pharmacological features with COX-2-selective inhibitors.

The inhibitory effects of acetaminophen on COX-1 and COX-2 have been studied thus far together with those of NSAIDs. The COX selectivity of acetaminophen and NSAIDs are shown in Figure 3, which was prepared by using the 50% inhibitory concentration (IC₅₀) for each COX to calculate Log (IC₅₀ for COX-1/IC₅₀ for COX-2). The relative effects to inhibit COX-1 and COX-2 vary depending on the samples and methods used for the assessment of COX activities and from species to species [24]. Therefore, the COX selectivity of acetaminophen and that of NSAIDs were based on the results of human whole blood assays. In such COX assays, drugs were incubated with blood samples collected from healthy donors and then thromboxane B₂ and PGE₂ produced by stimulating with lipopolysaccharide (LPS) was measured to determine the COX-1 activity and the COX-2 activity, respectively [25,26,27,28,29,30,31,32,33]. As shown in Figure 3, acetaminophen inhibits COX-2 selectively as well as coxibs (lumiracoxib, etoricoxib, valdecoxib, rofecoxib, and celecoxib), nimesulide, etodolac, diclofenac, and meloxicam, whereas conventional NSAIDs (flurbiprofen, ketoprofen, aspirin, fenoprofen, ketorolac, indomethacin, naproxen, and ibuprofen) inhibit COX-1 selectively. The COX-2 selectivity of acetaminophen is lower than that of diclofenac but incomparably much higher than that of ibuprofen selective for COX-1.

Hinz et al. [25] conducted in vitro COX activity measurements using human whole blood and demonstrated that the IC₅₀ values of acetaminophen for COX-1 and COX-2 are 113.7 μM and 25.8 μM, respectively. In their following ex vivo experiments using whole blood collected from male subjects who received 1000 mg acetaminophen, the IC₅₀ values for COX-1 and COX-2 were 105.2 μM and 26.3 μM, respectively. These results suggest that acetaminophen is 4.0-fold to 4.4-fold selective for COX-2 in humans. When a 1000 mg single dose of acetaminophen was orally administered to volunteers, the mean plasma concentrations of acetaminophen were greater than or equal to an in vitro IC₅₀ value for COX-2 at least 5 h after administration, whereas they remained below an in vitro IC₅₀ value for COX-1. However, it is questionable whether acetaminophen potently inhibits COX-2 similarly to diclofenac and coxibs as mentioned by Esh et al. [34].

Although acetaminophen is an inhibitor selective for COX-2, its potency to inhibit COX-2 is weaker than that of diclofenac, ibuprofen, and other NSAIDs as shown by the IC₅₀ values for COX-2 in Figure 4, which was prepared by using reported data [25,26,27,29,30,32,33]. Acetaminophen inhibits COX-2 with a mean IC₅₀ value of 28.5 μM, which is comparable to that of naproxen and fenoprofen. Diclofenac is the most potent inhibitor of COX-2 with a mean IC₅₀ value of 0.042 μM, followed by coxibs, oxicams, and conventional NSAIDs that show mean IC₅₀ values ranging from 0.109 μM for ketorolac to 13.9 μM for aspirin.

The analgesic activity of drugs has been frequently evaluated by the writhing responses of rodents to acute nociception. An intraperitoneal injection of diluted acetic acid induces writhing behaviors in mice by releasing PGE₂ and PGI₂ peripherally at the site of noxious stimulation and centrally at the spinal cord dorsal horn and some regions of the brain. PGI₂-mimetic iloprost induces writhing behaviors in mice by directly activating the PGI₂ receptors in peripheral nociceptors without releasing nociceptive PGs, while PGs are still involved in nociceptive transmission at the spinal and supra-spinal levels. When specifying the site of analgesic action of COX inhibitors by comparing the acetic acid-induced and iloprost-induced writhing responses of mice [35,36], it was indicated that acetaminophen preferentially acts on the CNS to inhibit COX, whereas diclofenac acts peripherally. The analgesic effects of acetaminophen in healthy volunteers were investigated by using a non-inflammatory pain model, in which phasic pain was induced by intracutaneously applied brief electrical pulses (20 ms) [37]. Orally administered acetaminophen (1000 mg) reduced pain ratings, cerebral potentials, and electroencephalographic delta power, as did a reference analgesic antipyrine. These results support the central mechanism of acetaminophen analgesia. On the other hand, biochemical and behavioral experiments suggest a different mode of analgesic action [38]. Oral administration of acetaminophen (200 mg/kg, p.o.) did not change PGE₂ contents in the mouse brain, whereas ibuprofen (100 mg/kg, i.p.) caused more than 90% inhibition of PGE₂ production. Both acetaminophen (200 mg/kg, i.p. for mice and 300 mg/kg, i.p. for rats) and ibuprofen (100 mg/kg, i.p. for both) reduced nocifensive behaviors during the second phase of formalin tests, while only acetaminophen inhibited the first phase. Acetaminophen, but not ibuprofen, increased the withdrawal thresholds in mouse tail immersion and von Frey (mechanical nociception) tests.

Given the COX-inhibiting characteristics (selectivity, potency, and site of inhibition), it is reasonable to assume that acetaminophen affects the COX-2 activity, being mechanistically different from NSAIDs as mentioned by Anderson [39].

3. Anti-Inflammatory Effect

Although acetaminophen has the potential to inhibit COX-2, its efficacy against inflammatory symptoms is still a matter of debate. Unlike diclofenac and ibuprofen, acetaminophen has been considered to exert only the limited effects on inflammatory swelling and pain [40]. However, clinically used acetaminophen, especially in dental treatment, shows contradictory results. Bjørnsson et al. [41] conducted a controlled, randomized, double-blind crossover study to evaluate the effects of 1000 mg acetaminophen tablets administered four times daily for 3 days on postoperative inflammatory events. They revealed that acetaminophen is able to reduce the swelling produced after third molar surgery with an efficacy compatible to that of ibuprofen (600 mg tablet four times daily for 3 days). The similar anti-inflammatory effect of acetaminophen was observed in another dental surgery to remove impacted third molars [42]. Brandt et al. [43] determined the effects of acetaminophen and NSAIDs including diclofenac and ibuprofen on the pain scores and effusion and synovial tissue volumes of patients with symptomatic knee osteoarthritis. Treatment with acetaminophen or NSAIDs of ≤4 g/day showed comparable analgesic and anti-inflammatory efficacy.

According to Meirer et al. [44], COX responsible for PG biosynthesis is bifunctional: it comprises the oxygenase function or fatty acid cyclooxygenase activity to catalyze the conversion of arachidonic acid into PGG₂ and the peroxidase function or prostaglandin hydroperoxidase activity to catalyze the conversion of PGG₂ into PGH₂ as shown in Figure 2. NSAIDs compete with arachidonic acid and inhibit the oxygenase reaction, whereas acetaminophen acts as a reducing agent to inhibit the peroxidase reaction [45]. It is presumed that acetaminophen at therapeutic concentrations would more effectively inhibit COX when the levels of arachidonic acid and peroxide are low but less when the levels of arachidonic acid and peroxide are high. In inflamed tissues, not only is arachidonic acid abundant [46] but peroxide is also present at high concentrations [47]. COX-2 inhibition by acetaminophen in interleukin-1α-stimulated human umbilical vein endothelial cells is suppressed or abrogated by adding t-butyl hydroperoxides to the assay system [48]. Relatively low concentrations of peroxide are preferable for acetaminophen to inhibit COX-2, whereas peroxides of relatively high concentrations should interfere with COX-2 inhibition by acetaminophen. The weak anti-inflammatory activity of acetaminophen may be explained by the extracellular conditions of inflamed tissues.

Hinz et al. [25] assessed ex vivo COX inhibition by measuring LPS-induced PGE₂ production in whole blood collected after the oral administration of 1000 mg acetaminophen to volunteers. Acetaminophen inhibited the COX-2 activity by more than 80%, with a degree comparable to inhibition by NSAIDs. In a clinical study of Lee et al. [49], patients received 1000 mg acetaminophen before the surgical removal of two impacted mandibular third molars and then they were subjected to micro-dialysis to collect inflammatory transudate from the surgical site to measure PGE₂ produced by COX-2. Acetaminophen decreased PGE₂ release at the site of injury as well as NSAIDs, suggesting that acetaminophen is a COX-2 inhibitor that acts peripherally. Acetaminophen causes weak COX inhibition in broken cell systems but potently inhibits PG production in intact cells at therapeutic concentrations when a concentration of arachidonic acid is lower than about 5 μM. If arachidonic acid levels are relatively low, PGs are synthesized largely by COX-2 even in cells containing both COX-1 and COX-2. The reported COX-2 selectivity of acetaminophen may be due to the inhibition of a COX-2-dependent pathway to proceed at low rates [50]. It is considered that acetaminophen is definitely able to inhibit COX-2, but the anti-inflammatory efficacy depends on its application conditions.

In addition to high-concentration peroxides in inflamed tissues, the pharmacokinetics of acetaminophen may affect its anti-inflammatory effects. While acidic NSAIDs like diclofenac are highly bound to plasma proteins (mainly albumin) and primarily present as ionized molecules in physiological conditions, free drug fractions and molecules in a non-ionized form are increased at an inflammatory acidic pH, promoting the intracellular diffusion of NSAIDs. Diclofenac accumulates in synovial fluids and its high concentrations persist for a long time after administration, possibly inducing the long-lasting potent inhibition of COX-2 [51]. However, it is not known whether acetaminophen benefits from such inflammation-induced acidosis.

4. Hypothetical Inhibition of Cyclooxygenase-3

After an early study of Flower and Vane [52] suggesting that acetaminophen reduces PG production in the brain more potently than in the spleen of dogs, it had been assumed that acetaminophen inhibits the centrally expressed COX, which is distinct from COX-1 and COX-2 characterized thus far. The third COX isozyme was searched in the CNS for acetaminophen analgesia [53]. Chandrasekharan et al. [54] reported the discovery of a novel COX-1 splicing variant in the canine cerebral cortex and termed it COX-3. Subsequently, COX-3 or COX-1 variant proteins were detected in human tissues [55].

Chandrasekharan et al. [54] compared the inhibitory effects of acetaminophen and NSAIDs on COX-1, COX-2, and COX-3 by measuring PG production in COX-transfected Sf9 insect cells after exposure to arachidonic acid. Selectivity for COX-3 relative to COX-1 or COX-2 of acetaminophen and NSAIDs are shown in Figure 5a, which was prepared by calculating Log (IC₅₀ for COX-1/IC₅₀ for COX-3) or Log (IC₅₀ for COX-2/IC₅₀ for COX-3) in the presence of 5 μM arachidonic acid for acetaminophen and 30 μM arachidonic acid for NSAIDs [54]. When adding 30 μM arachidonic acid to the assay system, acetaminophen inhibited neither COX-1 nor COX-2 even at 1000 μM. Acetaminophen appeared to inhibit COX-3 more selectively than NSAIDs. However, the higher COX-3 selectivity of acetaminophen than that of diclofenac and ibuprofen is ascribable to the false selectivity that was produced by the inappropriate interpretation of the experimental results of COX-1, COX-2, and COX-3 activity measurements. Because COX-3 is a COX-1 splicing variant, acetaminophen is very likely to show selectivity for COX-1 as shown in Figure 5b, which was prepared by calculating Log (IC₅₀ for COX-2/IC₅₀ for COX-1). In the comparative assessment [54], the IC₅₀ values for COX-3 of acetaminophen were 64 μM and 460 μM in the presence of 5 and 30 μM arachidonic acid, respectively, whereas those of NSAIDs ranged from 0.008 μM for diclofenac to 0.24 μM for ibuprofen in the presence of 30 μM arachidonic acid, indicating that the COX-3 inhibitory activity of acetaminophen is not as potent as that of NSAIDs in COX-transfected insect cells.

A study of Chandrasekharan et al. [54] leaves critical issues concerning the methodology and data interpretation of COX-3 as indicated by Esh et al. [34] and Kis et al. [56]. They used insect cells transfected with COX-1, COX-2, or COX-3 to measure the COX activity. However, COX-3 was derived from dogs, whereas COX-1 and COX-2 were derived from mice. Although the COX activity was measured by determining the produced PGs, PGE₂ production from exogenous arachidonic acid in the absence of acetaminophen was so different between COX isozymes that cells containing COX-3 produced 5–25 times less PGE₂ than cells containing COX-1 and COX-2. The concentrations of arachidonic acid significantly influence PGE₂ production, thereby affecting the COX-inhibitory effects of the tested drugs. However, either 5 μM or 30 μM arachidonic acid was pretreated in COX activity assays. It is speculated that the reported greater effects of acetaminophen on COX-3-containing cells do not indicate its selectivity for COX-3 but merely reflect the result of less PGE₂ production [50]. In contrast to its effect on canine COX-3, acetaminophen showed no differences in inhibition potency between human COX-1 and COX-1 splicing variants [55]. Most critically, there have been no studies to follow up the original results on the pharmacological characteristics of COX-3. A splicing variant of COX-1, on which acetaminophen might act selectively, is genetically similar to COX-1 and referred to as putative COX-3; therefore, the term of “COX-3” itself was refuted [57,58]. Since COX-3 is very unlikely to be physiologically and clinically relevant to humans, the “COX-3” mechanistic hypothesis is no longer applicable to the mode of analgesic action of acetaminophen.

5. Conversion of Acetaminophen into Bioactive Compounds

5.1. N-(4-Hydroxyphenyl)arachidonoylamide

Acetaminophen has currently been referred to as a prodrug since Högestätt et al. [20] reported that it is metabolically converted into an active substance in the body. They identified 4-aminophenol and N-(4-hydroxyphenyl)arachidonoylamide (AM404) in the brain 20 min after an intraperitoneal injection of deuterium-labeled acetaminophen (300 mg/kg) into rats, suggesting the possibility that acetaminophen is metabolized to 4-aminophenol, which is further conjugated with arachidonic acid. In their following experiment with rats, acetaminophen (30–300 mg/kg, i.p.) and 4-aminophenol (10–100 mg/kg, i.p.) dose-dependently formed 4-aminophenol and AM404, respectively. When analyzing various tissues obtained from rats after an intravenous injection of acetaminophen (300 mg/kg), the highest concentration of 4-aminophenol and AM404 was found in the liver and the brain, respectively. Furthermore, rat brain homogenates were incubated with 4-aminophenol plus arachidonic acid (100 μM for each), resulting in the formation of a significant amount of AM404, which was prevented by an inhibitor of fatty acid amide hydrolase (FAAH). AM404 was formed by incubating 4-aminophenol in the brain homogenates of wild-type mice, but not in the brain homogenates of FAAH knockout mice. Mallet et al. [38] assessed the effects of orally administered acetaminophen (200 mg/kg) by mouse formalin, tail immersion, and von Frey (mechanical nociception) tests. They demonstrated that acetaminophen shows antinociceptive effects in wild-type mice, but not in FAAH knockout mice. These results suggest that AM404 formation depends on the activity of FAAH. Dalmann et al. [59] proved the involvement of FAAH in acetaminophen analgesia by experiments with mice experiencing inflammatory pain induced by different stimuli. The analgesic effects of acetaminophen were abolished in FAAH knockout mice, but not in wild-type mice. Although FAAH is ubiquitously expressed, global FAAH inhibitor URB597 (0.3 mg/kg) reduced the analgesic action of acetaminophen, but not peripherally restricted FAAH inhibitor URB937 (0.3 mg/kg). However, brain-impermeant FAAH inhibitor URB937 reduced the analgesic action of acetaminophen by intracerebroventricular administration (5 μg/mouse). These results indicate that supra-spinally located FAAH is required for acetaminophen to produce analgesia. Although AM404 is considered to be metabolically formed from 4-aminophenol by FAAH, this enzyme is also responsible for the degradation of AM404. The formation of AM404 in rat and mouse brains was confirmed by using acetaminophen concentrations higher than those clinically used in humans [20,38]. AM404 was also detected in the rat brain after the administration of 1 g acetaminophen corresponding to a human therapeutic dosage [60]. After a 30 min incubation, however, AM404 of 0.1–1 nmol was rapidly degraded by brain membranes from wild-type mice, but not by those from FAAH knockout mice [61]. Whether FAAH mediates AM404 formation or degradation may depend on the substrate concentration.

Moderately lipophilic acetaminophen is absorbed from the small intestine as indicated by a human pharmacokinetic study demonstrating that the plasma maximum concentrations of acetaminophen ranged from 19 to 204 μM after being orally administered at a dose of 1000 mg [62]. Absorbed acetaminophen is deacetylated in the liver to 4-aminophenol, which crosses the blood–brain barrier, and then is converted into AM404 in the CNS by FAAH-mediated conjugation with arachidonic acid as shown in Figure 6. Different experiments demonstrated that AM404 can be formed in the brain of rodents. Muramatsu et al. [60] orally administered acetaminophen (20 mg/kg) to rats and prepared blood and brain samples to measure acetaminophen and AM404, respectively. The determined maximum concentration was 15.8 μg/mL for acetaminophen in the plasma and 150 pg/g for AM404 in the brain. The time to the maximum concentration was 15 min for both, meaning that the blood concentration of acetaminophen reflects the brain content of AM404. They also revealed the possibility that AM404 can be formed from acetaminophen administered at a therapeutic dose (1000 mg) of adult humans. Direct application of AM404, but not acetaminophen, to the spinal cord dorsal horn produced analgesia in rats [63]. Adult patients receiving elective urological surgery under spinal anesthesia were intravenously given a single dose of 1000 mg acetaminophen [64]. Consequently, AM404 was detected at 5.1–57 nM in spinal fluids 10 min after administration, indicating that acetaminophen is centrally converted into AM404 in humans.

As suggested by recent studies, the analgesic effect of acetaminophen is considered to be mediated by its metabolite AM404, which acts in the brain and spinal cord by targeting the transient receptor potential channel, the cannabinoid receptor, the calcium channel, anandamide, FAAH, and COX as shown in Figure 6. Wang et al. [65] investigated the interactions between acetaminophen metabolites and presumable targets by molecular docking and molecular dynamics simulation. Acetaminophen analgesia and AM404 were reviewed by Ohashi and Kohno [18], Ayoub [19], and Mallet et al. [66]. The interactions of AM404 with different targets are discussed below.

In association with acetaminophen analgesia mediated by bioactive metabolites, a question arises as to whether the analgesic effects of NSAIDs are influenced by the metabolic modification of their structures. In pharmacological comparisons, different metabolites of diclofenac were much less effective than their parent structure in inhibiting the phenyl-p-benzoquinone-induced writhing responses of mice and in inhibiting the adjuvant arthritis and carrageenan-induced paw edema of rats [67]. In contrast to acetaminophen, metabolism is not relevant to the analgesic and anti-inflammatory effects of diclofenac. The reactive metabolites of diclofenac are associated with hepatotoxicity rather than with analgesic activity [68].

5.1.1. Transient Receptor Potential Vanilloid 1 Channel

The largest group of receptors responsible for nociception is the transient receptor potential (TRP) channel superfamily, which is classified into TRP vanilloid, TRP ankyrin, and other subfamilies [69]. Among them, the TRP vanilloid 1 (TRPV1) channel and TRP ankyrin 1 (TRPA1) channel behave as polymodal receptors. TRPV1 channels are present both centrally and peripherally as expressed in the spinal cord and brain, the trigeminal and dorsal root ganglia, and various regions relating to pain transmission and modulation. Supra-spinal TRPV1 activation leads to antinociception and analgesia [70]. Dipyrone exerts an analgesic effect in mice by acting on TRPV1 channels in the rostral ventromedial medulla [71], and cannabidiol produces analgesia in rats by activating TRPV1 channels [72].

AM404 structurally resembles TRPV1 agonist anandamide (N-arachidonoylethanolamide) and capsaicin (8-methyl-N-vanillyl-6-nonenamide) as shown in Figure 7. TRPV1 antagonist capsazepine not only inhibits the effect of AM404 activating TRPV1 channels and causing the vasodilation of rat hepatic arteries but also blocks AM404-induced currents in TRPV1-channel-expressed Xenopus oocytes [73], suggesting that AM404 activates TRPV1 channels. Högestätt et al. [20] demonstrated that AM404 activates TRPV1 channels in rat mesenteric arteries at sub-micromolar concentrations almost as potently as capsaicin does and such an effect is inhibited by 3 μM capsazepine, whereas neither acetaminophen nor 4-aminophenol activates TRPV1 channels. An intracerebroventricular injection of AM404 induced antinociception in wild-type mice, which was inhibited by pretreatment with capsazepine, but these effects were not observed in TRPV1 channel knockout mice [38]. Stueber et al. [74] conducted patch-clamp experiments with HEK (human embryonic kidney) 293 cells that expressed different isoforms of recombinant TRPV1 channels. Their results showed that AM404 activates human TRPV1 channels in a concentration-dependent manner at >1 μM. Taken together, AM404 is referred to as a potent activator of TRPV1 channels that mediates acetaminophen analgesia.

In thermal and mechanical rat paw withdrawal tests, NSAIDs including diclofenac attenuated the thermal and mechanical hyperalgesia of rats following TRPV1 channel activation induced by channel agonist capsaicin [75,76]. However, these effects were ascribed to the desensitization of TRPV1 channels. Apart from TRPV1 activation by AM404 and capsaicin, TRPV1 channel blockers were reported to alleviate pain [77]. Ibuprofen not only inhibited TRPV1 channels but also reduced TRPV1 channel expression in the mouse temporomandibular joint after an injection of complete Freund’s adjuvant [78]. Both diclofenac and ibuprofen conjugated with serotonin were found to inhibit TRPV1 channels with IC₅₀ values of 19 μM and 6 μM, respectively [79]. Ibuprofen with a structurally modified carboxylic acid group also exhibited the TRPV1 antagonistic activity to block capsaicin-induced nociception in mice [80].

5.1.2. Cannabinoid Receptor

Cannabinoid receptors are classified into two subtypes, cannabinoid receptor 1 (CB1) and cannabinoid receptor 2 (CB2). CB1 and CB2 are present in the cerebral cortex, spinal cord, and dorsal root ganglia and in microglial and postsynaptic cells, respectively. In the periphery, CB1 is located in nociceptors, whereas CB2 is located in immune tissues. Since CB1 regulates nociception, agents that act on CB1 in the spinal cord, brainstem, or peripheral sensory neurons could exhibit analgesic activity [81].

In the rat hot plate test, both CB1 antagonist AM281 and SR141716A inhibited the analgesic effect of acetaminophen [82]. By thermal, mechanical, and chemical pain tests, Mallet et al. [83] revealed that CB1 antagonist AM251 abolished acetaminophen analgesia in wild-type mice, but not in CB1 knockout mice. Their following experiments showed that an analgesic effect of acetaminophen is suppressed by inhibiting FAAH that mediates the conversion of acetaminophen into AM404. Antinociception induced by acetaminophen in rats was significantly reduced by pretreating with CB1 antagonist AM251, but not with CB2 antagonist AM630 [84]. These results indicate that the activation of CB1 underlies the analgesic effects of acetaminophen and AM404.

In contrast, the antinociceptive effects of ibuprofen in the rat hind paw formalin test were not affected by either AM251 or AM630 [85]. Although diclofenac suppressed hyperalgesia developed by an intraplantar injection of PGE₂ into rats, its effect was not blocked by either AM251 or AM630 [86]. CB1 and CB2 are unlikely to contribute to disclofenac analgesia peripherally. While a microinjection (0.25 μL) of diclofenac (2.5–10 μg) into the medial prefrontal cortex of rats suppressed the neurogenic and inflammatory phases of pain, such antinociceptive effects were inhibited by AM251 [87], suggesting that CB1 is responsible for diclofenac analgesia.

5.1.3. Calcium Channel

Voltage-gated calcium channels are classified into high-voltage activation, comprising Cav1 and Cav2 channels, and low-voltage activation, comprising Cav3 channels that are further sub-classified into Cav3.1, Cav3.2, and Cav3.3 channels. Among these channels, centrally and peripherally expressed Cav3.2 channels are implicated in pain transmission; therefore, the inhibition of Cav3.2 channels leads to pain relief.

Lipoamino acids (anandamide-related endogenous molecules) exhibit analgesic activity by inhibiting Cav3.2 channels [88]. In the mouse tail or paw immersion, von Frey (mechanical nociception), and formalin tests of Kerckhove et al. [89], oral administration of acetaminophen suppressed nociception in wild-type mice but failed in Cav3.2 channel knockout mice. Such an antinociceptive effect was increased by an intrathecal injection of Cav3.2 channel blocker TTA-A2, suggesting that the inhibition of Cav3.2 channels is closely associated with acetaminophen analgesia. In their following experiments, an intracerebroventricular injection of AM404 induced antinociception in wild-type mice, but not in Cav3.2 channel knockout mice. AM404 inhibited Cav3.2 currents in dorsal root ganglion neurons with an IC₅₀ value of 13.7 μM in whole-cell patch-clamp recordings [89]. It is presumed that TRPV1 activation could result in the strong inhibition of Cav3.2 channels, by which AM404 exerts a potent analgesic effect.

No reports on Cav3.2 channel inhibition have been found for analgesic diclofenac and ibuprofen in the literature.

5.1.4. Anandamide Transport and Hydrolysis

Anandamide is an endogenous activator of TRPV1 channels and an endogenous ligand of CB1. Termination of anandamide signaling is regulated by the transport of anandamide into cells, where FAAH hydrolyzes anandamide to arachidonic acid and ethanolamine. Since the analgesic effects of anandamide are terminated by carrier-mediated transport into neurons and astrocytes and the subsequent enzymatic hydrolysis, they are potentiated by inhibiting anandamide cellular transport (or uptake) and FAAH responsible for anandamide hydrolysis.

Beltramo et al. [90] revealed that AM404 inhibits anandamide accumulation in rat cortical neurons and astrocytes with IC₅₀ values of 1 μM and 5 μM, respectively. In their mouse hot plate experiment, AM404 (10 mg/kg, i.v.) significantly enhanced and prolonged analgesia produced by anandamide (20 mg/kg, i.v.). Giuffrida et al. [91] collected blood samples after an injection of AM404 (10 mg/kg, i.p.) into rats. Their high-performance liquid chromatography/mass spectrometric analysis showed that anandamide in plasma is increased after AM404 administration, indicating an elevation of the peripheral anandamide concentration. These results suggest the possibility that AM404 may potentiate the analgesic activity of endogenous anandamide by inhibiting its cellular uptake. However, anandamide transport inhibitors, including AM404, did not inhibit anandamide uptake in neuroblastoma and astrocytoma cells at short time points [92]. AM404 also has the property to inhibit rat brain FAAH with IC₅₀ values of 0.5–6 μM [93]. It is speculated that AM404 increases anandamide levels by inhibiting FAAH rather than inhibiting anandamide transport [92].

Ibuprofen was reported to inhibit FAAH in rat brain homogenates with an IC₅₀ value of 134 μM [94]. While ibuprofen and anandamide synergistically induce antinociception in the rat formalin test, ibuprofen potentiates an analgesic effect of anandamide by blocking its hydrolysis by FAAH [95]. In addition, diclofenac and ibuprofen inhibit anandamide transporters and FAAH to influence the cannabinoid system [96].

5.1.5. Cyclooxygenase

Högestätt et al. [20] determinined PGE₂ production by RAW264.7 macrophages exposed to LPS and demonstrated that AM404 inhibits COX-2 at 0.1–1.0 μM with a potency almost similar to that of selective COX-2 inhibitor NS-398. N-(3-Hydroxyphenyl)arachidonoylamide), a metaisomer of AM404 (N-(4-hydroxyphenyl)arachidonoylamide), was also reported to inhibit purified COX-2 at 25 μM [97]. Saliba et al. [98] treated primary mouse and rat microglial cell cultures with AM404 and found that AM404 prevents PGE₂ production induced by LPS in both cell cultures at concentrations of 1–10 μM. Neither TRPV1 antagonist capsazepine (10 μM) nor CB1 antagonist AM251 (10 μM) affected such preventive effects of AM404 (5 μM) in rat microglial cells. AM404 (1–10 μM) also inhibited PGE₂ production induced by LPS in primary microglial cells from TRPV1 channel knockout mice. Furthermore, they measured the COX activity of primary rat microglial cells to compare the effects of AM404 on COX-1 and COX-2. AM404 inhibited COX-2 potently at 0.1–10 μM as well as 10 μM diclofenac but inhibited COX-1 mildly.

5.1.6. Other Bioactivities of AM404

As applied to the treatment of COVID-19 symptoms [99], repurposing approved drugs outside the scope of their original clinical indication has been attracting attention as a strategy for developing novel medications. Repurposing studies suggest that AM404 possesses bioactivities other than analgesic activity, which are mechanistically independent of the interactions with channels, receptors, and enzymes described above.

• Antibacterial Activity

In order to repurpose existing drugs into the treatment of oral infections, Gerits et al. [100] screened an NIH clinical library and selected AM404 as an antibacterial candidate against periodontopathic Porphyromonas gingivalis [101]. They assessed the effects of AM404 on oral and nonoral pathogenic bacteria [102]. Consequently, AM404 was found to inhibit the growth of Porphyromonas gingivalis with a minimum inhibitory concentration and minimum bactericidal concentration of 12.5 μM for both, but not other bacterial species. AM404 was also effective in inhibiting Porphyromonas gingivalis biofilm formation on the surfaces of titanium disks used as a model for dental implants. These results suggest that AM404 reduces the risk of oral infection with periodontopathic bacteria and the inflammatory pain associated with periodontal diseases. AM404 specifically interacted with bacterial membranes to increase outer and inner membrane permeability, disrupting the membrane integrity of Porphyromonas gingivalis.

Al-Janabi [103] reported that ibuprofen inhibits the growth of Staphylococcus aureus and Paracoccus yeei with a minimum inhibitory concentration of 1.25 mg/mL for both. Diclofenac also showed an antibacterial effect on Staphylococcus aureus [104]. When a drug repurposing study was conducted for NSAIDs, diclofenac and ibuprofen exhibited the antibiofilm activity, suggesting their potential as the adjunctive therapy of biofilm-related infections [105]. Diclofenac also resensitized methicillin-resistant Staphylococcus aureus to β-lactam antibiotics [106]. Diclofenac and ibuprofen have membrane interactivity [107] as AM404 interacts with bacterial membranes to cause membrane permeabilization [102]. Ibuprofen also inhibits the growth of two Candida albicans strains and 10 Candida non-albicans strains with minimum inhibitory concentrations of 1–3 mg/mL [108].

• Antiviral Activity

Dengue is an infectious disease caused by dengue virus (DENV) comprising antigenically distinct DENV-1, DENV-2, DENV-3, and DENV-4. Among DENV RNA genome-encoding proteins, non-structural protein NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5 are essential for viral replication, assembly, and maturation; therefore, these proteins are the targets for treating dengue [109]. Van Cleef et al. [110] suggested the ability of AM404 to block DENV replication by interacting with the viral NS4B protein. By using DENV-infected HeLa cells, they revealed that AM404 reduces the viral RNA accumulation of DENV-2 (3-fold and 25-fold reduction at 48 and 72 h postinfection, respectively) and DENV-1 (16-fold and 19-fold reduction at 48 and 72 h postinfection, respectively).

NSAIDs exhibit antiviral activity by inhibiting viral replication and cellular entry, competing with RNA binding to viral nucleoprotein, and increasing nitric oxide production [111]. Paul et al. [112] assessed the effects of NSAIDs on DENV-2 by in silico and in vitro screening assays. Their molecular docking study suggested that ibuprofen and diclofenac have binding affinity for DENV-2 NS2B and NS3 proteins. Ibuprofen and diclofenac inhibited tDENV cellular entry at 6.25 and 25 μg/mL, respectively.

• Anticancer Activity

Ahmed et al. [113] screened an NIH clinical collection of a small-molecule compound library for repurposing antibacterial/anti-inflammatory agents as an anticancer drug to target colorectal cancer stem-like cells. Among the selected compounds, AM404 effectively prevented the stemness/de-differentiation, migration, and drug-resistance of colorectal cancer cells. Caballero et al. [114] revealed that AM404 has the property to inhibit NFAT (nuclear factor of activated T cells) and NF-κB (nuclear factor κ-B) signaling pathways and impair the migration and invasiveness of human neuroblastoma cell line SK-N-SH.

Since elevated PGE₂ production stimulates tumor cell proliferation, COX-2 plays a critical role in cancer progression. NSAIDs that inhibit COX-2 reduce the growth of cancer cells and the risk of various types of cancer [115], while COX-independent mechanisms have been also suggested for the chemopreventive effects of NSAIDs [116]. Ibuprofen, diclofenac, celecoxib, aspirin, etc., are all known to possess promising anticancer activity, and some of them have already been used in clinical trials for cancer treatment [117]. Leidgens et al. [118] reported that diclofenac and ibuprofen inhibit the proliferation and migration of human glioma cells. There is preclinical evidence that diclofenac is effective against colorectal, neuroblastoma, pancreatic, ovarian, glioma, melanoma, prostate, and breast cancer [119]. Marinov et al. [120] investigated the effects of diclofenac on tumor cell lines originating from breast (MCF-7), cervical (HeLa), and colorectal cancer (HT-29). After treatment for 48 h, diclofenac decreased the cell viability of MCF-7 (IC₅₀ value of 46.5 μg/mL) most potently, followed by HT-29 (79.0 μg/mL) and HeLa tumor cells (174 μg/mL).

5.2. N-Acetyl-p-benzoquinone Imine

Acetaminophen is oxidatively metabolized to N-acetyl-p-benzoquinone imine (NAPQI) in the liver [121], which is catalyzed by cytochrome P450 (CYP450) monooxygenases, mainly CYP2E1 and CYP3A4. NAPQI has the property to readily react with thiol compounds, particularly glutathione, and with hepatic proteins. If acetaminophen is administered at high doses or overdoses, NAPQI formed in a large amount depletes glutathione and reacts with cellular macromolecules, thereby causing severe liver damage.

CYP450 monooxygenase activity and immunoreactivity found in the rat and human spinal cord suggest the possibility that NAPQI is formed in regions of the spinal cord. NAPQI was detected in the mouse spinal cord after the systemic administration of acetaminophen [122], although it was recently suggested that NAPQI is absent in the mouse brain after acetaminophen overdosing [123]. In addition to hepatotoxicity, NAPQI may partly contribute to analgesia [124], as shown in Figure 6. NAPQI acts on different channels as described below.

5.2.1. Transient Receptor Potential Vanilloid 1 Channel

Eberhardt et al. [125] explored the effects of acetaminophen metabolites by patch-clamp experiments. They found that NAPQI, but not acetaminophen, irreversibly activates and sensitizes the human, rat, and rabbit TRPV1 channels expressed in HEK (human embryonic kidney) 293 cells. NAPQI was also reported to enhance the TRPV1-activating potency of AM404 [74]. NAPQI may potentiate the analgesic effect of acetaminophen cooperatively with AM4014 though TRPV1 activation.

5.2.2. Transient Receptor Potential Ankyrin 1 Channel

TRPA1 channels are co-expressed with TRPV1 channels in the dorsal root, vagal, and trigeminal ganglia and the brain and share pharmacological features with TRPV1 channels. TRPA1 channels implicated in nociception and pain transmission can be a drug target to alleviate pain [126]. When injected intrathecally, NAPQI dose-dependently induced spinal antinociception in wild-type mice as well as TRPA1 agonist cinnamaldehyde, but not in TRPA1 channel knockout mice, which were assessed by hot plate, paw pressure, cold plate tests of Andersson et al. [122]. TRPA1 channel activation by locally formed NAPQI could contribute to acetaminophen analgesia in addition to TRPV1 activation by centrally formed AM404 [127].

Diclofenac was also suggested to activate TRPA1 channels as well as NAPQI [128].

5.2.3. Potassium Channel

The activation of neuronal voltage-gated potassium Kv7 channels results in pain alleviation [129]. Ray et al. [130] conducted electrophysiological experiments with dissociated rat dorsal root ganglion and spiral dorsal horn neurons in a perforated current-clamp mode. NAPQI, but neither acetaminophen nor AM404, acted on Kv7 channels to reduce membrane excitability and dampened the excitability of first- and second-order neurons in the pain pathway. Kv7 channel blocker XE991 inhibits the effect of Kv7 agonist retigabine to attenuate complete Freund’s adjuvant-caused inflammatory pain in rats [131]. Systemically, but not intrathecally, administered XE991 antagonized the effect of acetaminophen to reduce rat hind paw nociception induced by a carrageenan injection [132]. It is presumed that Kv7 channel activation by NAPQI at the peripheral site is responsible for acetaminophen analgesia.

There have been no studies relevant to Kv7 channel activation by diclofenac and ibuprofen in the literature.

6. Analgesic Strategy from the Perspective of Acetaminophen

When acetaminophen is administered repeatedly, its metabolite NAPQI can potentially damage the liver. Hepatotoxicity associated with acetaminophen is one of the major causes of drug-related liver injury, and the incidence of acetaminophen-induced liver intoxication has been increasing over the past few decades around the world [133].

Besides the intrinsic side effects, acetaminophen and NSAIDs have ecotoxicological challenges as they are currently regarded as one of the most emerging xenobiotics [134]. Acetaminophen, diclofenac, and ibuprofen are included in globally prevalent contaminants in natural matrices. It is assumed that acetaminophen exhibits acute and chronic toxicity against freshwater invertebrates together with diclofenac and ibuprofen [135]. The universal accumulation of acetaminophen in such aquatic environments as surface water, wastewater, and drinking water became a critical issue due to increases in its global production and consumption, particularly during the COVID-19 pandemic, because acetaminophen and its toxic metabolites are ubiquitously detected [136], requiring a reduction in the use of acetaminophen and NSAIDs.

From the perspective of acetaminophen as a prodrug, analgesic molecules have been recently investigated to suppress or eliminate the adverse toxicity but keep or improve the beneficial activity by two strategies: biotransformation into AM404 or its analogs and the structural modification of acetaminophen.

6.1. Biotransformation into AM404 or Its Analogs

Barrière et al. [137] studied whether 4-aminophenol (deacetylated metabolite of acetaminophen) and 4-(aminomethyl)-2-methoxyphenol (AMMP, structural analog of 4-aminophenol) are biotransformed into analgesic compounds in the brain because the manner in which none of these phenol derivatives are directly oxidized to hepatotoxic NAPQI was considered. When intraperitoneally injected into mice, 4-aminophenol (30 and 100 mg/kg) and AMMP (100 and 300 mg/kg) dose-dependently formed AM404 plus N-(4-hydroxyphenyl)-9Z-octadecenamide (HPOA) and N-vanillylarachidonamide (Arvanil) plus N-vanillyloleamide (Olvanil), respectively, in the brain 20 min after injection as shown in Figure 8. AM404 and its structural analogs showed analgesic effects, which were inhibited by TRPV1 antagonist capsazepine. The order of potency to activate TRPV1 channels was Arvanil = Olvanil >> AM404 > HPOA. The brain contents of AM404, Arvanil, and Olvanil but not HPOA were reduced in FAAH knockout mice, indicating that the formation of Arvanil, Olvanil, and AM404 is mediated by FAAH.

Å Nilsson et al. [138] verified whether the structural analogs of acetaminophen metabolite 4-aminophenol undergo FAAH-dependent N-arachidonoyl conjugation to form TRPV1-activating compounds with antinociceptive activity. They investigated different primary amines by in vitro and in vivo experiments using rodents. Consequently, it was revealed that N-arachidonoyl conjugates are formed from 5-amino-2-methoxyphenol (AMP) and 5-aminoindazole (AI) in a mouse brain 20 min after intraperitoneal administration of them (100 mg/kg for each) as shown in Figure 8. The FAAH-dependent formation of both conjugates was confirmed by using FAAH knockout mice. In the mouse formalin test, AMP and AI (100 mg/kg, i.p. for each) showed antinociceptive effects, which were inhibited by an intracerebroventricular injection of FAAH inhibitor URB937 and TRPV1 antagonist capsazepine. Acetaminophen (300 mg/kg, i.p.) caused liver necrosis in mice, whereas 4-aminophenol, AMMP, AMP, and AI (300 mg/kg, i.p. for each) were not hepatotoxic.

To overcome the hepatotoxicity of NAPQI, Bazan et al. [139] explored structural analogs or derivatives of acetaminophen. In a series of studies, they synthesized N,N-diethyl-2-[[2-(4-hydroxyanilino)-2-oxo-ethyl]sulfamoyl]benzamide (SRP-001), as shown in Figure 8, and they comprehensively studied its antinociceptive activity, hepatotoxicity, and the mode of its action while comparing it with acetaminophen [140]. Consequently, SRP-001 was found to exert antinociceptive effects comparable to those of acetaminophen in mouse and rat hyperalgesia experiments and exhibit antipyretic activity in the mouse LPS-induced fever model similarly to acetaminophen. It was also indicated that SRP-001 produces analgesia via AM404 formation in the midbrain periaqueductal gray. Interestingly, SRP-001 formed larger amounts of AM404 than acetaminophen following an intraperitoneal injection. Although acetaminophen is metabolized to hepatotoxic NAPQI in the liver by CYP450, SRP-001 did not follow such a metabolic pathway; therefore, it lacked hepatotoxicity due to NAPQI. In addition, SRP-001 preserved hepatic tight junction integrity at high doses. When equimolar doses of 992–5954 mmol/kg, corresponding to 150–900 mg/kg for acetaminophen and 402–2414 mg/kg for SRP-001, were orally administered to CD-1 male mice, the Kaplan–Meier survival curves that indicate the probability of survival over time showed a dose-dependent increase in mortality by 72 h after administration in acetaminophen-treated groups (1/10 to 7/10), but no mortality in SRP-001-treated groups. The phase 1 trial revealed that SRP-001 has safety, tolerability, and favorable pharmacokinetics (a half-life of 4.9–9.8 h). SRP-001 could be a promising analgesic alternative to acetaminophen.

Given these results, biotransformation into AM404 or its analogs will be a future strategy to develop novel analgesic agents with higher activity and tolerability.

6.2. Structural Modification of Acetaminophen

Since adamantane derivatives have been used for drug design, Fresno et al. [141] replaced a phenyl ring of acetaminophen with an adamantane ring to synthesize 1,3-disubstituted, (E)-1,4-disubstituted, and (Z)-1,4-disubstituted derivatives, as shown in Figure 9. They assessed the pharmacological effects and mechanisms of 1-acetylamino-3-hydroxyadamantane (AHA) and 1-hydroxy-4-acetylaminoadamantane (HAA). In the mouse acetic acid writhing test, AHA, HAA (a stereoisomeric mixture), and acetaminophen were intraperitoneally administered at 100–500 mg/kg, 250–750 mg/kg, and 100–200 mg/kg, respectively. Both AHA and HAA showed antinociceptive effects with greater potency in HAA. HAA acted on TRPA1 channels as a selective antagonist, while it neither activated cannabinoid receptors nor inhibited COX. Adamantane derivatives not only exhibited more improved analgesic activity compared with acetaminophen but also displayed biocompatibility, without producing toxic effects in chronic treatments.

Sinning et al. [142] shortened the acyl chain of AM404 from C20 to C2 to synthesize three acetaminophen analogs, N-(1H-indazol-5-yl)acetamide (IA), N-(4-hydroxybenzyl)acetamide (HBA), and N-(4-hydroxy-3-methoxyphenyl)acetamide (HMPA), as shown in Figure 9. None of these analogs displayed the pronounced affinity for either CB1 or CB2. IA selectively inhibited COX-2, and HMPA inhibited both COX-1 and COX-2 in human whole blood assays. When assessed by the mouse formalin test, IA and HBA exerted analgesic effects when intraperitoneally injecting 50 mg/kg and 275 mg/kg, respectively. HMPA was excluded from the assessment because of its significant toxicity to mice. The analgesic potencies of IA and HBA were greater than or comparable to that of acetaminophen (200 mg/kg, i.p.).

A structural analog of acetaminophen, N-(4-hydroxyphenyl)-5-methyl-1H-pyrazole-3-carboxamide (JNJ-10450232/NTM-006), as shown in Figure 9, was designed to reduce the hepatotoxicity of acetaminophen. JNJ-10450232/NTM-006 exhibits analgesic efficacy and potency comparable to that of acetaminophen in rat inflammatory pain experiments but not hepatotoxicity in mouse liver injury experiments, and the systemic administration of JNJ-10450232/NTM-006 and acetaminophen results in comparable peripheral levels of 4-aminophenol and brain levels of AM404 [143]. By a double-blind, placebo-controlled, first-in-human study, Gelotte et al. [144] revealed that JNJ-10450232/NTM-006 is safe and well tolerated in healthy male volunteers following single (50–6000 mg) and multiple (250–2500 mg twice daily for 8 days) doses. In a phase 2 dental pain study of Gelotte et al. [145], JNJ-10450232/NTM-006 produced relatively long-lasting analgesia in patients undergoing a third molar extraction when evaluating efficacy during 24 h following administration of 250 mg and 1000 mg. A preclinical safety assessment indicated that JNJ-10450232/NTM-006 does not cause hepatotoxicity at supratherapeutic doses in a mouse model sensitive to acetaminophen hepatotoxicity and in rat, dog, and non-human primate 28-day repeat-dose toxicity studies [146]. Regarding metabolism in rats, dogs, monkeys, and humans, JNJ-10450232/NTM-006 was found to differ from acetaminophen in biotransformation with the reduced potential to form reactive quinone imine [147], which may explain its lower hepatotoxicity.

These results support the possibility that the structural modification of acetaminophen is a promising strategy to develop novel analgesic agents.

7. Conclusions

We have summarized the conventional, recognized, and hypothetical mechanisms of analgesic acetaminophen and highlighted the current mechanistic concepts.

Acetaminophen inhibits COX-2 more potently than COX-1, although its COX-2 inhibitory effect is weaker than that of NSAIDs. The COX-2 selectivity of acetaminophen is higher than that of ibuprofen but lower than that of diclofenac. Both NSAIDs more potently inhibit COX-2 than acetaminophen, so they exhibit high efficacy as an anti-inflammatory drug, whereas acetaminophen has limited effects on inflammatory swelling and pain depending on its application conditions. After the discovery of COX-3 in the brain of canines, it had been hypothesized that acetaminophen selectively inhibits this COX-1 splicing variant. However, COX-3 inhibition by acetaminophen was not as potent as that by diclofenac and ibuprofen. COX-3 itself was revealed to not be physiologically and clinically relevant to humans; consequently, the mechanistic “COX-3” hypothesis has no longer applied to acetaminophen analgesia. Recent studies suggest that acetaminophen is deacetylated to 4-aminophenol in the liver, and after crossing the blood–brain barrier, this metabolite is subjected to FAAH-mediated conjugation with arachidonic acid and converted into AM404 in the CNS. AM404 exhibits diverse bioactivities, including analgesic activity, by activating TRPV1 channels, activating CB1, inhibiting Cav3.2 calcium channels, inhibiting anandamide transport and hydrolysis, and inhibiting COX. AM404-interacting channels and receptors may partly be the targets of diclofenac and ibuprofen. Metabolically formed AM404 is considered to mediate an analgesic effect of its parent structure acetaminophen. Acetaminophen administered at overdoses or even at the therapeutic dose is metabolized to hepatotoxic NAPQI, which is a critical issue in the practical use of acetaminophen. Biotransformation into AM404 or its analogs and the structural modification of acetaminophen have been investigated for the purpose of obtaining novel molecules that do not exert adverse effects like hepatotoxicity by producing NAPQI but possess an improved property of alleviating pain. The perspective of acetaminophen as a prodrug will be crucial for developing analgesic agents with higher tolerability and activity, which would be an analgesic alternative to acetaminophen and NSAIDs. Such a perspective may be applied to the discovery of drug leads beyond analgesia because AM404 exhibits bioactivities other than analgesic activity.