¹Knoll Limited Research and Development, Nottingham, UK

²Knoll GmbH, 50 Knollstrasse, D-67061, Ludwigshafen, Germany

Correspondence to: I C Kilpatrick, Knoll Limited, St Nicholas Court, 25-27 Castle Gate, Nottingham NG1 7AR, UK. E-mail: Ian.Kilpatrick@knoll.co.uk

^†This article is dedicated to the late Professor Michael J Stock, colleague and friend.

OBJECTIVE AND DESIGN: It has been proposed that the anti-obesity agent, phentermine, may act in part via inhibition of monoamine oxidase (MAO). The ability of phentermine to inhibit both MAO_A and MAO_B in vitro has been examined along with that of the fenfluramine isomers, a range of selective serotonin reuptake inhibitors and sibutramine and its active metabolites.

RESULTS: In rat brain, harmaline and lazabemide showed potent and selective inhibition of MAO_A and MAO_B, their respective target enzymes, with IC₅₀ values of 2.3 and 18 nM. In contrast, all other drugs examined were only weak inhibitors of MAO_A and MAO_B with IC₅₀ values for each enzyme in the moderate to high micromolar range. For MAO_A, the IC₅₀ for phentermine was estimated to be 143 µM, that for S(+)-fenfluramine, 265 µM and that for sertraline, 31 µM. For MAO_B, example IC₅₀s were as follows: phentermine (285 µM), S(+)-fenfluramine (800 µM) and paroxetine (16 µM). Sibutramine was unable to inhibit either enzyme, even at its limit of solubility.

CONCLUSION: We therefore suggest that MAO inhibition is unlikely to play a role in the pharmacodynamic properties of any of the tested drugs, including phentermine. Instead, the lack of potency of these drugs as MAO inhibitors is contrasted with their powerful ability either to inhibit the uptake of one or more monoamines (fluoxetine, paroxetine, sertraline, sibutramine's active metabolites) or to evoke the release of one or more monoamines (S(+)-fenfluramine, S(+)-norfenfluramine, phentermine). These differences in mode of action may be linked to the adverse cardiovascular events experienced with some of the releasing agents.

International Journal of Obesity (2001) 25, 1454-1458

phentermine; fenfluramine; monoamine oxidase; serotonin reuptake; serotonin release; cardiovascular

Introduction

Phentermine is an amphetamine derivative that has been used to treat obesity for over three decades and from 1992, it was often used for this purpose in combination with either fenfluramine or D-fenfluramine (S(+)-fenfluramine) until an association of fenfluramines with both heart-valve disease^1,2 and pulmonary hypertension³ led to the withdrawal of fenfluramines in 1997. Whilst the pathobiology of both events remains uncertain, a link with elevated plasma serotonin (5HT) has been proposed on more than one occasion^3,4,5,6 and for valvulopathy, 5HT receptor mechanisms on valve tissue are also postulated to be relevant.⁷ Aside from the known ability of phentermine to promote dopamine release in the brain,^8,9 recent discussion on its mode of action has centred on a proposal that phentermine may inhibit monoamine oxidase (MAO).¹⁰ Within this proposal, the authors not only inferred that such an action could boost free 5HT concentrations in plasma beyond those evoked by 5HT-releasing agents such as fenfluramine and S(+)-fenfluramine, but for the same reason they also warned against the possible combined use of phentermine with selective serotonin reuptake inhibitors (SSRIs) such as fluoxetine.¹⁰ However, such a role of phentermine as an MAO inhibitor (MAOI) has been discounted based on a review of some basic and clinical biochemical findings.¹¹

Whilst such warnings undoubtedly generate concern among physicians and patients alike, neither report provides an unequivocal answer to the fundamental question of whether MAO inhibition is relevant to the actions of phentermine. The problem is compounded because until the recent study,¹⁰ no published work from within the last two decades has examined the interaction between phentermine and MAO. This prompted us to compare the abilities of some anti-obesity agents including phentermine, S(+)-fenfluramine and the 5HT and noradrenaline reuptake inhibitor (SNRI), sibutramine, along with a series of SSRIs, to influence MAO activity directly. In an attempt to shed light on the possible link between elevated 5HT levels and cardiovascular disease, comparisons of the potency of these drugs derived here against the isoenzymes of MAO (MAO_A, which is chiefly responsible for 5HT degradation and MAO_B) are then made with their known abilities to inhibit 5HT reuptake and to evoke 5HT release. Some of these data have been published in preliminary form.¹²

Methods

Monoamine oxidase A and B activity and its inhibition

We used modifications to a published method¹³ for MAO estimation that allowed us to selectively assay MAO_A and MAO_B in samples of rat brain. Briefly, male Sprague-Dawley rats (190-200 g) were killed by cervical dislocation, their brains dissected free and the cerebella removed from them. Each brain was weighed, suspended in 50 mM potassium phosphate buffer (9 mlg^-1 wet weight; pH 7.4) at 2°C and homogenised in a PolytronÒ PT. The homogenate was filtered through four layers of gauze and diluted 1 in 40 vol/vol by the same ice-cold buffer. This preparation was apportioned into aliquots and stored at -80°C until use. One hundred microlitres of homogenate were pre-incubated with 100 µl of phosphate buffer (containing test compounds as required) for 5 min at 37°C. A 100 µl aliquot of substrate solution (see below) was then added and the mixture incubated at 37°C for 20 min. The reaction was stopped by adding 1.5 ml of 0.1 M HCl. All assays were performed in triplicate and controls were obtained by adding HCI immediately after adding substrate solution without further incubation.

The substrate for MAO_A was [side chain-2-¹⁴C]-tryptamine (NEN Life Science Products; specific activity 1900 MBq/ mmol). The compound was diluted to obtain 10 000-30 000 dpm/100 µl corresponding to 0.3-1.0 µM final concentration in the assay mixture. The substrate for MAO_B was 2-phenyl-[1-¹⁴C]-ethylamine (NEN Life Science Products; specific activity 1550 MBq/mmol). The compound was diluted to obtain 10 000-30 000 dpm/100 µl corresponding to 0.35-1.05 µM final concentration in the assay mixture. Under the conditions employed, substrate turnover in control assays was in the range of 20-30% for MAO_A and 30-40% for MAO_B, respectively.

After stopping the reaction, 3.5 ml toluene-based scintillant were added and the samples were shaken for 30 min in order for the MAO-generated non-basic metabolites to be extracted into the organic phase, leaving the unreacted substrates in the non-scintillating aqueous phase. Once the aqueous and organic phases had separated, the vials were transferred to a Packard TriCarb 2000Ò scintillation counter.

Data presentation and statistical analysis

Inhibition of enzyme activity was expressed as IC₅₀ values calculated by linear regression of the logit-transformed inhibition values vs log concentration values. Since MAO_A is largely responsible for the metabolism of 5HT in rat brain, an estimate of the relative potency of each drug to inhibit this isoenzyme was made in comparison to the inhibitory potency of phentermine. For this analysis, data were log₁₀-transformed and subjected to one-way ANOVA followed by Dunnett's test. A P-value lower than 0.05 was taken to be significant.

Drugs

Fluoxetine hydrochloride, paroxetine hydrochloride, sertraline hydrochloride, S(+)-fenfluramine, R(-)-fenfluramine, S(+)-norfenfluramine, sibutramine (N-(1-[1-(4-chlorophenyl)cyclobutyl]-3-methylbutyl)-N,N-dimethylamine hydrochloride monohydrate) and its active metabolites, metabolite 1 (BTS 54 354; N-(1-[1-(4-chlorophenyl)cyclobutyl]-3-methylbutyl)-N-methylamine hydrochloride) and metabolite 2 (BTS 54 505; 1-[1-(4-chlorophenyl)cyclobutyl]-3-methylbutylamine hydrochloride, were synthesised by the Medicinal Chemistry Department of Knoll Limited Research & Development, Nottingham, UK. Lazabemide was synthesised by the Medicinal Chemistry Department of Knoll GmbH Research & Development, Ludwigshafen, Germany. Phentermine hydrochloride and harmaline hydrochloride were purchased from Sigma. All general reagents were purchased from Merck, Darmstadt.

Results

Reference agents

Under the present conditions of low (<1.0 µM) substrate concentration (¹⁴C-tryptamine for MAO_A and ¹⁴C--phenylethylamine for MAO_B), the high potencies and selectivities of reference inhibitors of MAO_A and MAO_B were confirmed. Thus, the data in Table 1 show that the reversible and selective MAO_A inhibitor, harmaline,¹⁴ yielded an IC₅₀ value of 2.3 nM against MAO_A compared with 59 000 nM against MAO_B. On the other hand, the reversible and selective MAO_B inhibitor, lazabemide,¹⁵ produced an IC₅₀ value of 18 nM against MAO_B compared with 125 000 nM against MAO_A.

Anti-obesity agents and SSRIs

Table 1 shows that phentermine is a very weak inhibitor of MAO_A and MAO_B with IC₅₀ values in the high micromolar range. The two enantiomers of fenfluramine are similarly weak inhibitors of each enzyme, although S(+)-fenfluramine is slightly but significantly less active than phentermine as a MAO_AI whilst R(-)-fenfluramine is slightly but significantly more potent than phentermine in this regard. The active metabolite of S(+)-fenfluramine, S(+)-norfenfluramine, is significantly more active than phentermine as a MAO_AI, being four times as potent here.

Various SSRIs along with metabolites 1 and 2, the active metabolites of the SNRI, sibutramine, are also weak inhibitors of MAO_A and MAO_B with IC₅₀ values in the micromolar range. Of the SSRIs tested, sertraline was the most potent MAO_AI although each SSRI was significantly more potent than phentermine as a MAO_AI. Whilst neither metabolite of sibutramine differed significantly from phentermine in its potency as an inhibitor of MAO_A, sibutramine itself was ineffective as an inhibitor of either isoenzyme, even at its limit of solubility (Table 1).

Discussion

Methodological considerations

The assay was validated by the use of the reversible MAO inhibitors, harmaline and lazabemide, whose selectivity for MAO_A and MAO_B and high inhibitory potency within the low nanomolar range was confirmed.^14,15 Given the reported differences in assay conditions (eg substrates, substrate concentrations, incubation time), the various literature values^16,17,18 for the potency of fluoxetine as an inhibitor of MAO_A and MAO_B and that reported here concur well. The early report¹⁹ that sibutramine did not inhibit rat brain MAO either at its highest tested in vitro concentration of 100 µM or in ex vivo studies is also borne out by the current findings.

Phentermine and fenfluramines as putative MAO inhibitors

The present estimates of the inhibitory potency of phentermine are in accord with IC₅₀ data from nearly three decades ago which ranged from 180 to 1000 µM^20,21 obtained from studies using rat brain or liver that were unable to differentiate between the isozymes of MAO. The current data also concur with those from the very recent study of Ulus and colleagues,¹⁰ in which estimates of the IC₅₀ of phentermine as a MAO_AI in brain tissue are 100-200 µM although their IC₅₀ estimate of 600-900 µM for phentermine as a MAO_BI in brain is slightly less potent than that observed in the current study. In comparison with the potency of the reference inhibitors used in the present study, phentermine is approximately 62 000 times less active as an inhibitor of MAO_A and some 16 000 times less active as an inhibitor of MAO_B.

In fact, the inhibitory potencies of harmaline and lazabemide for their respective non-target enzyme are slightly greater than those of phentermine for these same enzymes and are similar to those of the SSRIs. However, moclobemide, an established antidepressant and reversible inhibitor of MAO_A, has a relatively weak potency in vitro (IC₅₀»6 µM²²). This potency, which is at least partly reflected in the large daily dosing requirements of patients, is barely an order of magnitude greater than that displayed by some SSRIs, and yet none of these is endowed with properties of MAO inhibition at therapeutic doses (see below). As phentermine is less potent still, we would argue that phentermine should not be classed as an MAOI, especially in view of its estimated plasma level in man of 0.7 µM following recommended doses.¹¹

MAO inhibition as a link with drug-associated cardiovascular disease

One clear omission from the work described by Ulus et al¹⁰ was a failure to consider the potency of the fenfluramines as MAOIs. This was surprising because fenfluramine and its isomers were attributed with weak MAO inhibitory properties (IC₅₀»1000 µM)²³ 3 y prior to the earliest studies on phentermine and the results of these studies were later reinforced.²⁴ Using the more sophisticated assay conditions here, the present data show that phentermine is barely twice as potent as S(+)-fenfluramine as a MAO_AI, whilst a more striking finding is that the major active metabolite of S(+)-fenfluramine, ie S(+)-norfenfluramine, is four times as potent as phentermine as a MAO_AI. Thus, coupled with earlier data that placed S(+)-fenfluramine 10-fold more active than phentermine in this regard,^{20,21,22,23,24} there is no reason why, if phentermine were to be recognised as a MAOI, S(+)-fenfluramine should not be similarly treated. Indeed, MAO inhibition has already been proposed²¹ as a causal link between the incidence of pulmonary hypertension and treatment with aminorex or chlorphentermine, two early anorectic agents that were withdrawn over 25 y ago. However, neither phentermine nor compounds with significantly greater MAO inhibition than phentermine (Table 1) have been associated with such cardiovascular disease. As discussed below, it is therefore considered unlikely that MAO inhibition contributes to fenfluramine's pharmacological actions.

Alternative sources of 5HT elevation by phentermine and fenfluramine?

Whether valvulopathy and/or pulmonary hypertension are linked to prolonged or even brief rises in circulating 5HT is unresolved. The acute ability of S(+)-fenfluramine and its active metabolite to release 5HT from blood platelets^25,26 could provide a peripheral source of 5HT that is relevant to the pathobiology. However, plasma 5HT changes after drug treatment are difficult to interpret since it is very likely that repeated fenfluramine treatment will deplete platelets of their 5HT stores²⁷ and may even cause plasma 5HT levels to fall, as demonstrated both in rats given S(+)-fenfluramine for 14 days²⁷ and in patients treated with fenfluramine and phentermine for at least 2 months.^28,29 Alternatively, direct 5HT receptor stimulation by S(+)-norfenfluramine⁷ could trigger mechanisms that lead to valvulopathy. In any event, it is clear from Tables 1 and 2 that S(+)-fenfluramine and its active metabolite, S(+)-norfenfluramine, are more potent releasers of 5HT from brain slices³⁰ than they are inhibitors of MAO and that this 5HT-releasing action is not shared by phentermine, the SSRIs, the SNRI, sibutramine, or its two active metabolites (Table 2).

However, these comparisons of potency ratios pale into insignificance when the MAO inhibitory activities of each of the tested SSRIs are placed alongside their ability to inhibit reuptake of radiolabelled 5HT by rat frontal cortex synaptosomes.³¹ As shown in Table 2, this analysis reveals that the SSRIs inhibit 5HT reuptake by 50% at concentrations which are 2000-162 000-fold less than their IC₅₀ values for MAO inhibition. Unlike the SSRIs, both S(+)-fenfluramine and its metabolite are extremely weak inhibitors of 5HT reuptake, and phentermine is essentially inactive with a K_i of 11 µM^30,31,32(see Table 2). Importantly, the SSRIs, none of which releases 5HT, are more potent MAO inhibitors than either S(+)-fenfluramine or phentermine and each is of similar potency to S(+)-norfenfluramine.

These data illustrate two vital points. First, the compounds that are the most potent 5HT reuptake inhibitors, the SSRIs, also happen to be the most potent MAO inhibitors. Yet, if MAO inhibition was of major pharmacological relevance to the action of the SSRIs and SNRIs, they would never have been developed because of the risks of inducing, for example, the serotonin syndrome. That this does not occur is simple testament to the many orders of magnitude between their potency as SSRIs/SNRIs and MAO inhibitors. Second, the data stress that 5HT reuptake inhibition, as borne out by enormous clinical experience, is not associated with these cardiovascular disorders because they have not been reported even with prolonged use of the SSRIs.

Conclusions

For any drug that has multiple possible targets, relative potency is the most important parameter to consider. On these grounds alone, the hypothesis¹⁰ that MAO inhibition by phentermine may contribute to cardiovascular disorders has little to commend it given the many orders of magnitude that separate the potency of SSRIs and SNRIs as 5HT reuptake inhibitors vs MAOIs. Whilst none of the SSRIs or SNRIs releases 5HT, all are very weak MAOIs and some are marginally more potent than phentermine, the fenfluramine enantiomers or S(+)-norfenfluramine.

In summary, the data presented indicate that MAO inhibition by phentermine or fenfluramines is unlikely to play a role in the development of fenfluramine-linked cardiovascular disorders.

Acknowledgements

We are grateful for the skilful assistance of Willi Hoffman and Ingrid Volk.

1 Connolly HM, McGoon MD. Obesity drugs and the heart. Curr Probl Cardiol 1999; 24: 751-792,

2 Gardin JM, Schumacher D, Constantine G, Davis KD, Leung C, Reid CL. Valvular abnormalities and cardiovascular status following exposure to dexfenfluramine or phentermine/fenfluramine. JAMA 2000; 283: 1703-1709, MEDLINE

3 Abenhaim L, Moride Y, Brenot F, Rich S, Benichou J, Kurz X, Higenbottam T, Oakley C, Wouters E, Aubier M, Simmoneau G, Bégaud B. Appetite-suppressant drugs and the risk of primary pulmonary hypertension. New Engl J Med 1996; 335: 609-616, MEDLINE

4 Fishman AP. Aminorex to Fen/Phen. An epidemic foretold. Circulation 1999; 99: 156-161, MEDLINE

5 Hervé P, Launay J-M, Scrobohaci M-L, Brenot F, Simonneau G, Petitpretz P, Poubeau P, Cerrina J, Duroux P, Drouet L. Increased plasma serotonin in primary pulmonary hypertension. Am J Med 1995; 99: 249-254, MEDLINE

6 Robiolio PA, Rigolin VH, Wilson JS, Harrison JK, Sanders LL, Bashore TM, Feldman JM. Carcinoid heart disease. Correlation of high serotonin levels with valvular abnormalities detected by cardiac catheterization and echocardiography. Circulation 1995; 92: 790-795, MEDLINE

7 Fitzgerald LW, Burn TC, Brown BS, Patterson JP, Corjay MH, Valentine PA, Sun J-H, Link JR, Abbaszade I, Hollis JM, Largent BL, Hartig PR, Hollis GF, Meunier PC, Robichaud AJ, Robertson DW. Possible role of valvular serotonin 5-HT_2B receptors in the cardiopathy associated with fenfluramine. Mol Pharmac 2000; 57: 75-81,

8 Baumann MH, Ayestas MA, Dersch CM, Brockington A, Rice KC, Rothman RB. Effects of phentermine and fenfluramine on extracellular dopamine and serotonin in rat nucleus accumbens: therapeutic implications. Synapse 2000; 36: 102-113, MEDLINE

9 Rowley HL, Butler SA, Prow MR, Dykes SG, Aspley S, Kilpatrick IC, Heal DJ. A comparison of the effects of sibutramine and other weight-modifying drugs on extracellular dopamine in the nucleus accumbens of freely-moving rats. Synapse 2000; 38: 167-176, MEDLINE

10 Ulus IH, Maher TJ and Wurtman RJ. Characterization of phentermine and related compounds as monoamine oxidase (MAO) inhibitors. Biochem Pharmac 2000; 59: 1611-1621,

11 Rothman RB. Is phentermine an inhibitor of monoamine oxidase? A critical appraisal. Synapse 1999; 32: 141-145, MEDLINE

12 Kilpatrick IC, Traut M, Heal DJ. Monoamine oxidase inhibition determined in rat brain is unlikely to account for cardiovascular risks associated with fenfluramine and phentermine. Br J Pharmac 1999; 128: 66P,

13 Wurtman RJ, Axelrod J. A sensitive and specific assay for the estimation of monoamine oxidase. Biochem Pharmac 1963; 12: 1439-1441,

14 Fuentes JA, Neff NH. Selective monoamine oxidase inhibitor drugs as aids in evaluating the role of type A and B enzymes. Neuropharmacology 1975; 14: 819-825, MEDLINE

15 Henriot S, Kuhn C, Kettler R, Da Prada M. Lazabemide (Ro 19-6327), a reversible and highly sensitive MAO-B inhibitor: preclinical and clinical findings. J Neural Transm 1994; 41(Suppl): 321-325,

16 Leonardi ETK, Azmitia EC. MDMA (Ecstasy) inhibition of MAO type A and type B: comparisons with fenfluramine and fluoxetine (Prozac). Neuropsychopharmacology 1994; 10: 231-238, MEDLINE

17 Holt A, Baker GB. Inhibition of rat brain monoamine oxidase enzymes by fluoxetine and norfluoxetine. Naunyn-Schmiedeberg's Arch Pharmac 1996; 354: 17-24,

18 Mukherjee J, Yang Z-Y. Evaluation of monoamine oxidase B inhibition by fluoxetine (Prozac): an in vitro and in vivo study. Eur J Pharmac 1997; 337: 111-114,

19 Buckett WR, Thomas PC, Luscombe GP. The pharmacology of sibutramine hydrochloride (BTS 54 524), a new antidepressant which induces rapid noradrenergic down-regulation. Prog Neuro-Psychopharmac Biol Psychiatr 1988; 12: 575-584,

20 Moller Nielsen I, Dubnick B. Pharmacology of chlorphentermine. In: Costa E, Garattini S (eds). Amphetamines and related compounds.. Raven Press: New York, 1970, 63-73.

21 Seiler K-U, Wasserman O. MAO-inhibitory properties of anorectic drugs. J Pharm Pharmac 1973; 25: 576-578,

22 Haefely W, Burkard WP, Cesura AM, Kettler R, Lorez HP, Martin JR, Richards JG, Scherschlicht R, Da Prada M. Biochemistry and pharmacology of moclobemide, a prototype RIMA. Psychopharmacology 1992; 106: (Suppl): S6-14, MEDLINE

23 Duhault J, Verdavainne C. Modification du taux de sérotonine cérébrale chez le rat par les triflurométhyl-phényl-2 éthyl aminopropane (fenfluramine 768 S). Arch Int Pharmacodyn 1967; 170: 276-286,

24 Morgan D, Löfstrandh S, Costa E. Amphetamine analogues and brain amines. Life Sci 1972; 11: 83-96,

25 Buczko W, De Gaetano G, Garattini S. Effect of fenfluramine on 5-hydroxytryptamine uptake and release by rat blood platelets. Br J Pharmac 1975; 53: 563-568,

26 Lane, EL, Prow, MR, Aspley, S, Kilpatrick IC, Heal DJ. d-Fenfluramine releases 5HT from whole blood of rats as measured by in vitro microdialysis. Br J Pharmac 2000; 129: 155P,

27 Celada P, Martin F, Artigas F. Effects of chronic treatment with dexfenfluramine on serotonin in rat blood, brain and lung tissue. Life Sci 1994; 55: 1237-1243, MEDLINE

28 Redmon B, Raatz S, Bantle JP. Valvular heart disease associated with fenfluramine-phentermine. New Engl J Med 1997; 337: 1773-1774, MEDLINE

29 Rothman RB, Redmon JB, Raatz SK, Kwong CA, Swanson JE, Bantle JP. Chronic treatment with phentermine combined with fenfluramine lowers plasma serotonin. Am J Cardiol 2000; 85: 913-915, MEDLINE

30 Heal DJ, Cheetham SC, Prow MR, Martin KF, Buckett WR. A comparison of the effects on central 5-HT function of sibutramine hydrochloride and other weight-modifying agents. Br J Pharmacol 1998; 125: 301-308, MEDLINE

31 Cheetham SC, Viggers JA, Slater NA, Heal DJ, Buckett WR. [³H]Paroxetine binding in rat frontal cortex strongly correlates with [³H]5-HT uptake: effect of administration of various antidepressant treatments. Neuropharmacology 1993; 32: 737-743, MEDLINE

32 Rothman RB, Ayestas MA, Dersch CM, Baumann MH. Aminorex, fenfluramine and chlorphentermine are serotonin transporter substrates. Implications for primary pulmonary hypertension. Circulation 1999; 100: 869-875, MEDLINE

33 Lancashire B, Viggers J, Prow MR, Aspley S, Martin KF, Cheetham SC, Heal DJ. Monamine release and uptake inhibition profiles of antiobesity agents. J Psychopharmacol 1998; 12(Suppl A): 143P,

Table 1 Drug action (IC₅₀ values) on MAO_A and MAO_B activities

Table 2 Drug action on 5HT release and 5HT uptake inhibition plus ratios for MAO_A or MAO_B inhibition to 5HT uptake inhibition