this is the full text: Its pretty short actually Schwiebert C, Irving C, Gillman PK. Small doses of methylene blue, previously considered safe, can precipitate serotonin toxicity. Anaesthesia. Aug 2009;64(8):924.
Ramsay RR, Dunford C, Gillman PK. Methylene blue and serotonin toxicity:inhibition of monoamine oxidase A (MAO A) confirms a theoretical prediction. BR J Pharmocol. 2007/11;152:946-951.
RESEARCH PAPER
Methylene blue and serotonin toxicity: inhibition of monoamine oxidase A (MAO A) confirms a theoretical prediction
RR Ramsay1, C Dunford1 and PK Gillman2 1Centre for Biomolecular Sciences, University of St Andrews, North Haugh, St Andrews, Fife, Scotland, UK and 2PsychoTropical
Research, Bucasia, Queensland, Australia
Background and purpose: Monoamine oxidase inhibitors (MAOI) are known to cause serotonin toxicity (ST) when administered with selective serotonin reuptake inhibitors (SSRI). Methylene blue (methylthionium chloride, MB), a redox dye in clinical use, has been reported to precipitate ST in patients using SSRI. MB was assessed for MAO inhibition and so for its potential to precipitate ST.
Experimental approach: Inhibition of purified human MAO was quantified using kinetic assays and visible spectral changes to study the interactions of MB with MAO A. Key results: MB was a potent (tight binding) inhibitor for MAO A. It also inhibited MAO B but at much higher concentration. Interactions of MB with the active site of MAO A were confirmed by its action both as an oxidising substrate and as a one-electron reductant.
Conclusions and implications: MB is a potent reversible inhibitor of MAO A with implications for gut uptake of amines when administered orally. At concentrations reported in the literature after intravenous administration, MAO B would be partially inhibited but MAO A would be completely inhibited. This inhibition of MAO A would be expected to lead to perturbations of 5-hydroxytryptamine metabolism and hence account for ST occurring when administered to patients on SSRI treatment. British Journal of Pharmacology (2007) 152, 946–951; doi:10.1038/sj.bjp.0707430; published online 27 August 2007
Keywords: methylene blue; methylthionium chloride; serotonin toxicity; monoamine oxidase A; MAOI; tight binding
Abbreviations: MAO, monoamine oxidase; MAOI, monoamine oxidase inhibitor; MB, methylene blue (methylthionium chloride); SSRI, selective serotonin reuptake inhibitor; ST, serotonin toxicity
Introduction
Serotonin toxicity (ST), formerly referred to as serotonin syndrome, is one of very few drug–drug interactions involving usual therapeutic doses of commonly used drugs that are capable of developing rapidly and causing death within 24h. A mixture of therapeutic doses of any drug with significant monoamine oxidase (MAO) inhibitor (MAOI) properties with another drug that has potency as a selective serotonin reuptake inhibitor (SSRI) produces a high risk of precipitating this rapidly worsening interaction (reviewed in Gillman, 2006a). The typical clinical features of ST are (i) neuromuscular hyperactivity: tremor, clonus, myoclonus and hyperreflexia, and, in the advanced stage, pyramidal rigidity; (ii) autonomic hyperactivity: diaphoresis, fever, tachycardia, tachypnoea and mydriasis; and (iii) altered mental status: agitation, excitement,
Correspondence: Dr RR Ramsay, Centre for Biomolecular Sciences, University of St Andrews, North Haugh, St Andrews, Fife, Scotland KY16 9ST, UK. E-mail: rrr@st-and.ac.uk Received 29 May 2007; revised 11 July 2007; accepted 17 July 2007; published online 27 August 2007
with confusion in the advanced stage. There have been several deaths or near deaths, reported recently where a single dose of a SSRI has been inadvertently added to a MAOI (Otte et al., 2003; Cassens et al., 2006; Zonneveld et al., 2006).
In the past the only drugs recognised as possessing these properties were antidepressants used by specialists so that the risk of them being combined inadvertently was small. However, there are now many drugs on the market that act as 5-hydroxytryptaminergic agents (mostly SSRIs), or MAOIs, which are neither advertised, nor generally recognised, as 5-hydroxytryptaminergic drugs (for example, sibutramine and linezolid). Despite the clear understanding of ST interactions that has developed over the past decade, there is still misinformation in standard texts, such as the British National Formulary, and in pharmaceutical company product information (Gillman, 2005). Recent reviews deal with this complex topic in detail (Gillman, 1998, 2006a, b, c; Dunkley et al., 2003; Whyte et al., 2003; Whyte, 2004; Isbister and Buckley, 2005).
+
Figure 1
N H
N S+ N
Methylene blue inhibits monoamine oxidase A
RR Ramsay et al 947
tyrosines so that the amine is close to the N-5 of the flavin. The active site of MAO A is wider than that of MAO B (De Colibus et al., 2005) and, in functional studies, MAO A accommodates larger non-flexible ligands (Efange et al., 1993; Medvedev et al., 1998). Structure–function studies have shown that MAO substrates and inhibitors are some- what lipophilic with an aromatic ring and a nitrogen (or oxygen or sulphur) close to that ring (reviewed in Wouters, 1998; Kalgutkar et al., 2001; Ramsay and Gravestock, 2003). MB fits this description and indeed was an inhibitor and electron acceptor for MAO B (Shumakovich et al., 2004). Other tricyclic heterocycles, such as the b-carbolines (Kim et al., 1997), are also known as effective inhibitors, particularly of MAO A.
The experiments reported below show that MB inhibits both MAO A and MAO B. In particular, it is a tight-binding inhibitor of MAO A that interacts with the active site as shown by its action as an oxidative substrate and by its redox effects on the flavin of MAO A.
Methods
MAO assays
Recombinant human liver MAO A and MAO B were purified as reported previously (Tan and Ramsay, 1993).
MAO A (usually about 50nM) was assayed spectrophoto- metrically at 314 nm in 50 mM potassium phosphate, pH 7.2–0.2%, Triton-X100 at 301C using kynuramine (0.1– 0.9 mM) as the substrate (Tan and Ramsay, 1993). The kinetic constants were calculated by the Shimazu kinetic software. Apparent Km values in the presence of MB (30–200 nM) were also determined. This experiment revealed an apparent Ki of 69 nM, only slightly higher than the enzyme concentration, indicating that Michaelis–Menten analysis could not be used because the concentration of free inhibitor would be depleted by binding to the enzyme. To obtain the Ki, the initial rates in the absence and presence of MB were analysed by the equation developed by Morrison (1968) for tight- binding inhibitors.
MAO B was assayed polarographically using the same buffer as for MAO A (Tan and Ramsay, 1993). Inhibition was measured for 0–20 mM MB in the presence of 0.6 mM benzylamine (2 Km ) as shown in Figure 2. Further analysis of the kinetics of interaction with MAO B would require varying both the amine and oxygen substrates, because the Km for oxygen is 0.28mM (Husain et al., 1982; Ramsay et al., 1987).
Spectrophotometric studies of MAO
Spectral data were obtained using MAO A in 50 mM potassium phosphate, pH 7.2, 0.05% Brij, as in previous work (Ramsay, 1991; Jones et al., 2005). The enzyme was prepared for the experiments by centrifugal gel filtration, diluted to about 15mM, and its concentration determined from the absor- bance at 456 nm (extinction coefficient, 12 300 M1 cm1). For anaerobic experiments, oxygen was replaced with argon by cycling the custom-made anaerobic cuvette between vacuum and argon.
2H+ + 2e
2H+ + 2e
N
Blue
Colourless
NH S N
Oxidised and reduced forms of methylene blue.
The spectrum concept of ST predicts clearly that severe, life-threatening, degrees of toxicity are likely to develop only after the co-administration of SSRIs and MAOIs. A case report of neurological toxicity associated with the administration of methylene blue (methylthionium chloride, MB), which precipitated the work in this article (Rosenbaum, 2006) seemed an exception. However, the patients in that and subsequently uncovered reports (Gillman, 2006c) had all been taking an SSRI antidepressant before surgery involving the intravenous infusion of MB. All these patients experienced severe toxicity of the degree expected to result only from a combination of MAOI and SSRI, and too severe to be the result of an SSRI alone. A search of the literature revealed no report of such toxicity from MB alone. This led to the prediction that MB (Figure 1) must possess significant MAOI potency.
Existing clinical literature (reviewed in Gillman, 2006c), including a report of antidepressant effect of MB (Naylor et al., 1987), appeared to support the possibility that MB would inhibit MAO, as did the well-known properties of MB as an electron acceptor. Effective inhibition of amine oxidases, including MAO, was reported to play a role in the prevention by MB of ifosfamide encephalopathy (Aeschlimann et al., 1996; Kupfer et al., 1996). However, the only reported Ki for inhibition of MAO B was rather high at 5.6mM (Bachurin et al., 2001; Shumakovich et al., 2004). This is within the range for whole-blood concentration immediately after a 100 mg oral dose in humans, and intra- venous administration gave more than 10-fold higher sustained concentrations (Peter et al., 2000). In another study, the concentration of MB in rat brain was estimated at 0.5 mM from 1 mg kg1 intraperitoneally (Callaway et al., 2004).
The role of MAO in the brain, its involvement in disease and the therapeutic potential of MAOIs has recently been reviewed in the light of new structural information (Youdim et al., 2006). Crystal structures of MAO A (De Colibus et al., 2005) and MAO B (Binda et al., 2002) have shown that the active site is a long, narrow hydrophobic cavity penetrating deep into the molecule where the substrate is aligned by two
British Journal of Pharmacology (2007) 152 946–951
Methylene blue inhibits monoamine oxidase A
RR Ramsay et al
Dose–response curves for the inhibition of MAO A (left) and MAO B (right) by methylene blue (MB). MAO A (122 nM) was assayed spectrophotometrically with 0.3 mM (2 Km) kynuramine, and 25 concentrations of MB (some points omitted for clarity). MAO B was assayed polarographically with 0.6 mM (2 Km) benzylamine.
Data analysis
Parameters are shown as the values calculated7s.e. from the non-linear regression fit. The calculation of Ki used the equation for tight binding inhibitors (Morrison, 1968) in PRISM (from Graphpad Software, Inc., http://www.graphpad. com/curvefit/). Ki values are the mean7s.d. mean of values from four experiments.
Materials
All chemicals were purchased from Sigma-Aldrich (Poole, Dorset, UK). The purity of the MB was checked by determining the ratio of the absorbance at 664–610nm using 10 mM MB in water (Fujimoto et al., 1994). The ratio was 2.02 so the MB was used as supplied. It should be noted that the molar absorbance of MB varies with concentration (Lewis et al., 1943). The oxidised and reduced forms of MB are shown in Figure 1.
Results
MAO A and MAO B have very different sensitivity to MB
Figure 2 shows the inhibition of the two forms of MAO by MB, each assayed at subsaturating substrate concentration (2 Km). The IC50 value for MB inhibition of MAO A was 16478 nM under this specific condition with 122 nM MAO A. In contrast, no inhibition of MAO B was seen in the spectrophotometric assay at 100nM. Data from the polaro- graphic assays of MAO B are shown in Figure 2 (the right- hand curve) and the IC50 value for MAO B inhibition by MB was 5.571.7 mM.
MB is a tight-binding inhibitor of MAO A
The IC50 value of 16478nM (Figure 2) was obtained using 122 nM MAO A in the assays. Clearly, the normal Michaelis– Menten assumption that the inhibitor concentration would not be depleted by the enzyme did not hold. When
948
Figure 2
both kynuramine and MB were varied, the IC50 value increased as the substrate concentration increased (data not shown), consistent with competitive inhibition (Henderson, 1972). Tight-binding analysis of data for MAO A according to Morrison (1968) gave a Ki value of 2773 nM (mean7s.d.; n 1⁄4 4).
MB binds to the active site of MAO A
Inhibitors binding in the active site of MAO A influence the visible spectrum of the covalently bound cofactor flavin adenine dinucleotide (FAD) (Hynson et al., 2003). When MAO A was titrated with MB (Figure 3), only a very small and featureless change in the difference spectrum of the MAO A FAD (MAO A with MB—MAO A alone) at around 450 nm was detected (Figure 3b). In previous work, the KD for inhibitor binding was calculated from inhibitor-induced changes in the flavin spectrum (Hynson et al., 2003). However, the influence of MB on the flavin spectrum was too small to analyse accurately, with a maximum decrease of 0.01 that was compromised by the large adjacent absorbance peaks of the MB.
In contrast, the absorbance of MB was greatly (about 50%) decreased in the presence of enzyme (Figure 3b) compared with the value of 61 500 M1 cm1 determined for the conditions used here, an aqueous solution of 50mM potassium phosphate, pH7.4, containing 0.05% Brij. The molar extinction coefficient for MB is influenced by the concentration at which it is determined (Lewis et al., 1943; Bergmann and O’Konski, 1963), by the salt concentration in aqueous solution (Fujimoto et al., 1994), and by the presence of detergent (Yamamoto et al., 2007). Binding to or association with surfaces or other compounds, such as serum albumin (Hu et al., 2005), also alters the absorption spectrum. Thus, the decrease in absorbance of MB induced by enzyme was likely to be due to nonspecific absorption rather than specific insertion into the hydrophobic active site and was thus not examined further.
MB acts as an oxidising substrate for MAO A
Proximity of MB to the flavin was confirmed by the reduction of MB by kynuramine in the presence of MAO A under anaerobic conditions (Figure 4). MAO A with 10mM MB was made anaerobic and excess kynuramine (0.3mM final concentration) was added from the sidearm. Spectra were collected every 30 min overnight (Figure 4a). The time courses for the absorbance of MB (664 nm) and for the flavin (456nm) show that the flavin was rapidly reduced by the excess of substrate, then reoxidised until MB reduction reached steady state (Figure 4b). The steady-state rate of reduction of MB over 12 h (Figure 4b) was 0.35 nmol h1 and the rate of oxidation of kynuramine was 0.39nmolh1. To compare this rate with that where oxygen is the electron acceptor, as in Figure 2, the observed velocities were converted to rate constants to normalise for the different enzyme concentrations. The rate constant for the oxidation of kynuramine by MB was 0.045h1, extremely slow compared to the rate constant with oxygen at
British Journal of Pharmacology (2007) 152 946–951
Methylene blue inhibits monoamine oxidase A
RR Ramsay et al 949
Figure 3 Titration of MAO A with methylene blue (MB). Increments of MB (1 mM) were added to MAO A (7.6 mM) and the spectra recorded for MAO A with final concentrations of 0–10 mM MB as indicated. Arrows indicate the increase at 665 nm due to MB and the decrease in the flavin region. (a) The difference between the spectrum for MAO A with 6 mM or 10 mM MB and the spectrum for MAO A alone is shown in (b). © The difference spectra are the same as (b) but on an expanded scale to show the change in the absorbance of the MAO A flavin.
Figure 4 Methylene blue (MB) acts as the oxidising substrate for MAO A under anaerobic conditions. (a) MAO A (8 mM) with 10 mM MB was made anaerobic and 0.3 mM kynuramine added at time 0. The absorbance of MB at 665 nm decreased with time as indicated by the arrow. The flavin absorbance at 456 nm decreased during the mixing time but returned to its starting value over 3 h. (b) The steady-state rate of oxidation of kynuramine (314 nm) and of reduction of MB (665 nm) were compared between 4 and 12 h when the flavin absorbance at 456 nm was constant.
4950 h1. Others have previously reported that MB gave a 500-fold slower rate than oxygen with MAO B (Shumakovich et al., 2004).
MB in the presence of reducing agents reduces MAO A
MB is commonly used as a photosensitiser. Light excites MB into a triplet state that can extract electrons from amines
British Journal of Pharmacology (2007) 152 946–951
Methylene blue inhibits monoamine oxidase A
RR Ramsay et al
Figure 5 Reduction of MAO A by methylene blue (MB) in the presence of D-amphetamine and dithiothreitol (DTT). MAO A was diluted eightfold from stock to 27 mM so that the concentration of D-amphetamine was 60mM and of DTT about 0.35mM, was made anaerobic, and 30 mM MB added both to the MAO A cuvette and to the blank cuvette. The spectral change after the addition showed some reduction of MB (decrease at 665 nm) and appearance of the features of the MAO A anionic semiquinone (increase at 380 and 412nm, decrease at 456nm). After 2h at 201C, the spectrum indicated almost complete conversion of MAO A to the semiquinone form. No further change was observed even 24 h later.
generating reduced MB (Kim and Scranton, 2004). Thiol- reducing agents are often used to produce reduced MB (for example, Jonnalagadda and Tshabalala, 1992). When MB was added anaerobically to MAO A-containing dithiothreitol (normally present for storage) at about 0.3 mM and the inhibitor amine, D-amphetamine, at 50mM (Ki for MAO A is 20 mM (Ramsay and Hunter, 2002)), reduction of MAO A was observed. Figure 5 shows the decrease at 456nm and the appearance of the 412 nm peak typical of the anionic flavosemiquinone in MAO A in the presence of 30mM MB. After 2 h, the one-electron reduction of MAO A was complete. No further reduction to the dihydroflavin was observed because of the presence of D-amphetamine that stabilises the MAO semiquinone form (Ramsay and Hunter, 2002). Only about 1% radical was generated in the absence of dithiothrei- tol, so the radical generation requires the thiol reagent.
Discussion
MB has been in medical use for many years with few reports of toxic effects. In early experiments, oral MB was found to be innocuous in 15 kg dogs at 2 g per day, but an intravenous dose of 0.5 g was lethal (Agell and Agell, 1929). The recent case reports of neurological sequelae following MB infusions summarised by Gillman (2006c) confirm that MB is not an innocuous substance. Indeed, MB is cytotoxic to bacteria and has been reported to increase oxidative stress (May et al., 2003). Furthermore, its reactive redox chemistry led to its use as a photosensitiser. However, the particular clinical observa- tions in ST and their occurrence only in patients on SSRIs indicated that MB must also influence the 5-hydroxytrypta- mine system.
The high Ki value for MAO B of 5.6mM (Bachurin et al., 2001), confirmed by the IC50 reported here, was not
950
British Journal of Pharmacology (2007) 152 946–951
consistent with MAO B as a target for the action of MB. Despite this, in rats, where the brain concentration of MB was estimated to be about 0.5 mM (Callaway et al., 2004), both MAO inhibition and increase of 5-hydroxytryptamine as a result of MB treatment have been reported (Lerch et al., 2006). If MAO B is not inhibited, then MAO A could be the target. At the molecular level, both the crystal structure (Binda et al., 2002) and the inhibitor profile for MAO B (Efange et al., 1993) indicate that the active site is relatively narrow. In contrast, MAO A accommodates (De Colibus et al., 2005) and is strongly inhibited by larger molecules such as b-carbolines (Kim et al., 1997) and oxazolidinones such as linezolid (Jones et al., 2005). The data presented here establish that MB is a very potent inhibitor of MAO A with a Ki of 27 nM and that it binds in the active site of MAO A.
The prediction based on the clinical observations, that MAO A would be inhibited by MB, has been proven. It is consistent with structural information about the two MAO isoforms and is supported by information from studies on rat brain. MB is an effective MAO A inhibitor and thus will influence 5-hydroxytryptamine levels at doses expected during MB infusion. For comparison, linezolid, for which the risk of ST in patients receiving SSRIs is established (Morales-Molina et al., 2005), has a Ki for MAO A of 20 mM (Jones et al., 2005). We conclude that the use of MB by infusion in combination with other 5-hydroxytryptaminergic agents could lead to the ST syndrome as a result of MAO A inhibition.
Acknowledgements
Some of the experiments formed part of CD’s undergraduate thesis (University of St Andrews). Some initial experiments were performed by Mr Aleksey Zholobenko, University of St Andrews, during summer work experience and were reported in a poster presented at the 12th Amine Oxidase Workshop, August 2006 and at the Biochemical Society focus meeting on Health Implications of Dietary Amines, October 2006 (http://www.biochemis...ings/postertoc/ SA062/default.htm).
