Inhibition of rat brain monoamine oxidase enzymes by fluoxetine and norfluoxetine

Naunyn-Schmiedeberg's Arch Pharmacol (1996) 354:17 24 © Springer-Verlag 1996

Andrew H o l t - Glen B. Baker

Inhibition of rat brain monoamine oxidase enzymes by fluoxetine and norfluoxetine

Received: 15 December 1995/Accepted: 8 February 1996

Abstract Fluoxetine and its primary metabolite, norfluoxetine, are inhibitors of neuronal uptake of 5-hydroxytryptamine. While fluoxetine has also been reported to inhibit monoamine oxidase (MAO) in vitro at concentrations much lower than those measured in brain following chronic fluoxetine treatment, neurochemical profiles are not consistent with substantial MAO inhibition in vivo. In an attempt to explain this inconsistency, we have examined the interactions of fluoxetine and norfluoxetine with rat brain MAO-A and -B by a radiochemical assay method.

Fluoxetine and norfluoxetine were competitive inhibitors of MAO-A in vitro, with Ki values of 76.3 pM and 90.5 pM, respectively. Both compounds were non- competitive or uncompetitive inhibitors of MAO-B in vitro. Inhibition of MAO-B was time-dependent and was very slowly reversible by dialysis. ICso values v e r -

s u s metabolism of 50 pM /%phenylethylamine were 17.8 pM (fluoxetine) and 18.5 pM (norfluoxetine). Anal- ysis of the time-dependence of MAO-B inhibition by fluoxetine revealed that an initial competitive interaction between the enzyme and the inhibitor (K i 245 gM) was followed by tight-binding enzyme inactivation (ki~aot 0.071 min-1).

Following administration of fluoxetine (20mg kg-1 day-1) for 7 days, the cortical concentration of fluoxetine + norfluoxetine was estimated by gas-liquid chromatography to be 700 pM. Such drug treatment reduced MAO-A activity by 23% in 1:8 (w/v) cortical homogenates, but not in 1 : 80 homogenates. Inhibition of MAO-B in 1 : 8 homogenates was modest (12%) and was not significantly reduced by homogenate dilution. The concentration of 5-hydroxyindole-3-acetic acid, measured by high pressure liquid chromatography, was reduced by 47% in cortices from drug-treated rats,

A. Holt (I~)- G.B. Baker Neurochemical Research Unit, Department of Psychiatry, 1E7.44 Mackenzie Health Sciences Centre, University of Alberta, Edmonton, Alberta, Canada, T6G 2B7

while concentrations of 5-hydroxytryptamine, noradrenaline, dopamine, 3,4-dihydroxyphenylacetic acid and homovanillic acid were unchanged. These results suggest that, following chronic drug administration leading to relatively high tissue concentrations of fluoxetine and norfluoxetine, inhibition of either form of MAO would be restricted by competition for the enzyme with intraneuronal amine substrates.

Key words Fluoxetine • Norfluoxetine • Monoamine oxidase inhibition • Kinetics

Introduction

Termination of the postsynaptic action of amine neurotransmitters is primarily by neuronal and extraneuronal uptake (Burgen and Iversen 1965). Follow- ing neuronal uptake, monoamine oxidase [amine: oxygen oxidoreductase (deaminating) (flavin containing); EC 1.4.3.4; MAO] acts to regulate intraneuronal concentrations of monoamines and thus to control vesicu- lar amine content (Trendelenburg et al. 1987; Kopin 1994). Many drugs which potentiate central 5-hydroxytryptaminergic neurotransmission by inhibiting one or both of these processes are also efficacious in alleviat- ing the symptoms of depression. However, irreversible inhibitors of MAO-A predispose patients to the poten- tially fatal "cheese effect" (Blackwell 1963; Callingham 1989), while tricyclic uptake inhibitors often have un- pleasant atropine-like side-effects as well as inducing tachycardia and sedation (Bowman and Rand 1980). It is against this background that, in less than a decade, serotonin-selective reuptake inhibitors such as fluoxetine (Prozac ®) have achieved worldwide popularity in the treatment of depression (Gram 1994). In general, adverse reactions to fluoxetine are both less common and less severe when compared with the more tradi- tional antidepressants (Bowden et al. 1993; Kasper et al. 1994).

18

The clinical efficacy of fluoxetine apparently results from potent and selective inhibition of 5-HT uptake (Wong et al. 1974, 1975; Fuller et al. 1991). This may cause desensitisation of some 5-HTI autoreceptor sub- types, resulting in increased release of 5-HT (Briley and Moret 1993; Blier and de Montigny 1994). However, a number of established tricyclic uptake inhibitors are also weak, reversible inhibitors of MAO (Edwards and Burns 1974; Roth 1976; Reid et al. 1988), and these generally show nominal selectivity towards MAO-B. It is possible that such interactions are through the propylamine side chain of these drugs, and the phenylpropylamine structure of fluoxetine may thus allow this molecule also to interact with MAO. Although alterations in brain levels of neurotransmitters and their metabolites ex vivo following administration of fluoxetine do not indicate substantial MAO inhibition (Fuller and Perry 1975, 1981; Fuller and Snoddy 1986), inhibition in vitro of rat MAO-A and human and rat MAO- B by fluoxetine has recently been observed (Reid et al. 1988; Kokotos Leonardi and Azmitia 1994). No aspects of the inhibition other than potency were addressed by these workers, and quoted ICs0 values were not consistent between studies. Nevertheless, it has been suggested that MAO-A inhibition may contribute to the antidepressant efficacy of fluoxetine (Kokotos Leo- nardi and Azmitia 1994).

In an attempt to explain the apparent ambiguity which exists between ex vivo and in vitro observations, we have examined more closely the interactions in vitro between fluoxetine and its primary metabolite, norfluoxetine, and rat brain MAO enzymes, with particular emphasis on enzyme/inhibitor kinetics. Further ex vivo observations of MAO inhibition and of brain levels of biogenic amines and their metabolites, made following acute or chronic administration of fluoxetine to rats, could then be interpreted with reference to the known kinetic effects of fluoxetine and norfluoxetine on MAO enzymes.

Materials and methods

Radiochemical assay for monoamine oxidase. A modification of the method of Lyles and Callingham (1982) was used to determine MAO-A and -B activities in tissue homogenates. Assays, performed in triplicate in ice-cooled disposable soda glass tubes, contained 25 gl of homogenate, 25 gl of distilled water or inhibitor in aqueous solution and 50 pl of radiolabelled substrate. Tubes were flushed with oxygen, stoppered and incubated at 37°C for an appropriate time period. Incubation times were chosen following preliminary experiments designed to ensure that the rate of metabolite production remained linear for the duration of the incubation.

Enzyme activity was terminated by plunging the tubes into ice and adding HC1 (3 M, 10 gl) to each. Blank tubes had HC1 added before incubation with substrate. Metabolites were extracted into 1 ml of ethyl acetate/toluene (1 : 1, v/v, saturated with water) and 700 gl of the organic phase were added to polyethylene vials containing 4 ml of Ready Safe liquid scintillation fluid, along with 100 gl of glacial acetic acid to reduce chemiluminescence. Samples were counted for radioactivity in a scintillation spectrometer (Beckman LS 6000SC) with quench correction by automatic external standar- disation.

Substrates used were, for MAO-A, 5-hydroxy[G-3H]tryptamine creatinine sulphate (1 gCigmol - t ) and for MAO-B, /~- [ethyl-1- l~C]phenylethylamine hydrochloride (1 gCipmol 1). The extrac- tion efficiencies for metabolites of these amines into ethyl acetate/toluene were determined to be 87% and 94%, respectively.

Time-dependence of MAO inhibition. The effects of preincubation time on inhibitor potencies were examined by preincubating tissue homogenates at 37 ° C with submaximally effective concentrations of fluoxetine and norfluoxetine for a range of times between 0 and 30 rain, before incubation with radiolabelled substrates. The kinetics of the interaction between fluoxetine and MAO-B were further examined by the method of Kitz and Wilson (1962). Cortical homogenates were preincubated at 37 ° C with a range of concentrations of fluoxetine (100-200 gM) or water (controls) for 0, 2, 4, 7 or 10 rain prior to addition of [14C] PEA (50 gM). Following'oxygenation, samples were incubated for 5 min and enzyme activities determined as described above.

Determination of inhibitor constants. To determine kinetic constants for drug interactions with MAO-A, rates of deamination of [3HI 5-HT (50 250 gM) were examined in the absence (controls) or presence of fluoxetine or norfluoxetine (75 250 pM). However, in parallel studies of MAO-B inhibition, homogenates were first preincubated with fluoxetine or norfluoxetine (25 100 gM) for 20 min before addition of [I4C]PEA (6-35 pM). Following oxygenation, samples were incubated for 5 min and enzyme activities determined as described above.

In order to obtain ICso values for inhibitors versus MAO-B, homogenates were incubated with [14C]PEA (50 gM) for 10 min following preincubation with water (controls) or fluoxetine or norfluoxetine (0.1 pM-1 mM) for 20 rain.

Preparation of tissue homogenates. Male Sprague-Dawley rats (180-250 g), supplied by University of Alberta Health Sciences La- boratory Animal Services, were maintained on a 12 h light-dark cycle during which they were allowed free access to drinking water and a standard laboratory rat diet, until required for experimentation. Animals were adapted to handling for a period of at least 7 days prior to experimentation, and were killed by stunning and decapita- tion. For in vitro studies, brains were removed and homogenised, 1:40 (w/v), in ice-cold potassium phosphate buffer (0.2 M, pH 7.8), with an Ultra Turrax T25 mechanical homogeniser. For ex vivo studies, cortices were dissected from whole brains, a portion was homogenised, 1:8 or 1:80 (w/v), in ice-cold potassium phosphate buffer and the remaining tissue was frozen at - 80 ° C for subsequent determination of tissue drug levels (see below). The protein content of each homogenate was estimated by the method of Lowry et al. (1951), with bovine serum albumin as standard.

Reversal of inhibition of MAO-B by dialysis. Tissue homogenates were preincubated with water (controls) or fluoxetine or norfluoxetine (500 pM) at 37 ° C for 20 rain. At this concentration, both compounds caused almost complete inhibition of MAO-B. Aliquants of 2 ml were then dialysed against 1 1 of potassium phosphate buffer (50 mM, pH 7.8) at 37 ° C for 0, 2, 5 or 9 h, with buffer changes at 1, 3 and 6 h. Samples dialysed for 0, 2 or 5 h were transferred to glass test tubes following dialysis and stored at 37 ° C for the remainder of the 9 h period. MAO-B activities were estimated in dialysed homogenates as described above following incubation with [14C] PEA (50 pM) for 2 h and were compared with those of control samples, dialysed for the same time period. A more extensive incubation period was necessary in these experiments to give a measurable degree of substrate turnover, since the activity of rat brain MAO-B declines rapidly during storage or dialysis at 37 ° C (see Fowler et al. 1981). However, control experiments indicated that production of

metabolites did not deviate significantly from linearity during the 2 h incubation period (not shown).

Ex vivo enzyme inhibition. Rats were administered vehicle (saline containing dimethyl sulphoxide, 10% v/v) or fluoxetine (20 mgkg-1) intraperitoneally, once daily for 1 or 7 days, and were killed 2 h after the last injection. Cortical homogenates were prepared as described above and activities of MAO-A and -B estimated radiochemically.

Fluoxetine and norfluoxetine concentrations in brain cortex. A modification of the acetylation procedure of Drebit et al. (1988) was used for simultaneous analysis of fluoxetine and norfluoxetine in cortical homogenates. Following administration of fluoxetine as described above, samples of cortex were homogenised in 5 volumes of distilled water and a 1 ml aliquant was used for analysis. Maprotiline (1 gg) was added to the homogenate as an internal standard and carried through the procedure. Homogenates were basified with 250 mg K2CO 3 and extracted with 4 ml of ethyl acetate. The organic layer was retained and evaporated to dryness under nitrogen. The residue was rcdissolved in 2 ml of distilled water and acetylated with acetic anhydride according to the procedure of Martin and Baker (1977). Acetylated drugs were extracted by shaking with ethyl acetate (4 ml) for 10 min on a Vibrax mixer and, after centrifugation (5 rain, 1000 x g), the organic phase was removed and evaporated to dryness under nitrogen. Samples were reconstituted by adding toluene (100 gl) and a 2 gl aliquot was injected onto a gas chromatograph equipped with a fused silica capillary DB-5 column (15 m x 0.25 mm i.d.; polymethyl [5% phenyl] siloxane stationary phase; 0.25 gm film thickness) and a nitrogen-phosphorus detector.

Levels of neurotransmitters and metabolites in brain cortex. In rats which had received chronic fluoxetine, cortical levels of 5-HT, 5- HIAA, DA, DOPAC, HVA and NA were determined by high pressure liquid chromatography (h.p.l.c.) with electrochemical detection, the protocol being modified from that of Baker et al. (1987). Samples of cortex were homogenised in 5 volumes of water containing perchloric acid (0.1 M), ascorbic acid (50gM) and EDTA (270 gM). Following centrifugation at 4 ° C (12000 x g, 10 min), 15 ~tl of the supernatant were used for h.p.l.c, analysis. The system was comprised of a Waters model 510 pump, a Waters Intelligent Sample Processor (model 710B) and a Waters M460 electrochemical detector coupled to a Hewlett-Packard HP3392A integrator. The glassy carbon electrode was set at 0.88 V against a Ag/Ag C1 reference electrode. The aqueous mobile phase, which contained Na2HPO 4 (55 mM), EDTA (37 mM), octanesulphonic acid (73 mM) and acetonitrile (10%, v/v), was pumped at a flow rate of 1 mlmin 1 through a Spherisorb ODS (2) column (4.6 mm x 250mm, 5 gm particle size; Phenomenex, Torrance Calif., USA) with a gBondapak C18 pre-column (2mmx 150ram, 10gm particle size; Waters Chromatography, Milford, Mass., USA). A set of standards for each compound was prepared in perchloric acid (0.1 M) and carried through in parallel with each assay run; the resultant standard curves were linear, with r 2 > 0.99 in all cases.

Materials. Fluoxetine hydrochloride and norfluoxetine hydrochloride were generously donated by Lilly Research Laboratories, In- dianapolis, Ind., USA.

5-Hydroxy[G-aH]tryptamine creatinine sulphate was obtained from Amersham International plc, Bucks, UK and fl-[ethyl-1- 14C]phenylethylamine hydrochloride from Du Pont Canada Inc, Markham, ON. 5-Hydroxytryptamine creatinine sulphate, fi-phenylethylamine hydrochloride and maprotiline hydrochloride were pur- chased from Sigma Chemical Co., St. Louis, Mo., USA. Ready Safe ® liquid scintillation cocktail was obtained from Beckman Instru- ments Inc., Fullerton, Calif., USA. Spectra-Por dialysis membrane (MWCO 6-8000) was obtained from Spectrum Laboratory Prod- ucts, Houston, TX, USA. All other materials were of analytical grade, where possible.

19

Statistical analysis of results. Where appropriate, straight lines and sigmoid curves were fitted to the data using the linear and non-linear regression facilities of GraphPad Prism, version 1.03 (GraphPad Software, San Diego, Calif., USA). Values, shown as mean + SEM, were compared statistically with controls by Student's unpaired t-test.

Results

Time-dependence of MAO inhibition

Inhibition of brain MAO-B was time-dependent in the presence of submaximally effective (100 ~tM) concentrations of inhibitors, whereas the effects upon MAO-A of preincubating with inhibitors at 200 gM were independent of preincubation time (Fig. 1). Inhibition of MAO- A after preincubation with fluoxetine for 3 h was the same as that at 30 min (not shown). Since dialysis studies had indicated that the inhibition of MAO-B by fluoxetine was essentially irreversible (see below), the time-dependence of this interaction was studied by the method of Kitz and Wilson (1962). The initial binding of fluoxetine to MAO-B showed pseudo-first-order be- haviour (Fig. 2, upper panel; see Tipton 1980). Linear regression of the slope replot (Fig. 2, lower panel) gave a straight line which did not pass through the origin (see Discussion). Replots of data obtained from 3 animals yielded an estimated inhibitor dissociation constant (Ki) of 245 gM for the initial, reversible interaction between fluoxetine and MAO-B. The intercept with the ordinate axis has the value logo 2/kinac,, where kinac t is the rate constant for onset of irreversible inhibition. From experiments on 3 separate homogenates, kinac t was estimated to be 0.071 min-1

Competitive inhibition of MAO-A

Figure 3 shows representative Lineweaver-Burk plots illustrating the inhibition of 5-HT deamination by fluoxetine (a) and norfluoxetine (b). The common inter- cepts with the ordinate axes are indicative of competitive inhibition of MAO-A. Ki values obtained by linear regression of slope replots (not shown) are 76.3 laM (fluoxetine) and 90.5 gM (norfluoxetine).

Non-competitive/uncompetitive inhibition of MAO-B

Figure 4 shows representative Lineweaver-Burk plots illustrating the inhibition of /~-PEA deamination by fluoxetine (a) and norfluoxetine (b). Results at the lowest drug concentrations suggest non-competitive inhibition, while higher concentrations of both drugs seem also to reduce the affinity of the enzyme for /~-PEA, yielding uncompetitive kinetic plots. IC5o values, obtained by non-linear regression of inhibitor

2.(

100

80

8 6(

"~ 4(

2c

20

6 5 1'0 1'5 2% 2'5 3'0

Preincubation time (min)

Fig. 1 Effects of preincubation time on the effects of submaximal concentrations of inhibitors on the activity of rat brain MAO. Homogenates were preincubated with ftuoxetine (open symbols) or norfluoxetine (closed symbols) at 200 pM before the addition of [aH] 5-HT (250 pM) to assay remaining MAO-A activity (circles) or at 100 oM before the addition of [14C]PEA (50 pM) to assay remaining MAO-B activity (squares). Results, expressed as a percentage of activities in controls, show the means _+ SEMs from 5 experiments, where error bars exceed symbol size

%

"~ 1.8

1.~

1.4

1.2

_o 1.0

Preincubation time (rain)

2C

. . . . i . . . . i 0 )0 0.005 0.010

1/[FLU] (~I,M -1)

Fig. 2 Upper panel: Effect offluoxetine concentration on the rate of onset of MAO-B inhibition. Rat brain homogenates were preincubated with fluoxetine at 100 pM (O), 150 pM (A) or 200 gM (V) before incubation with [14C]PEA (50 I-tM) to assay remaining MAO-B activity. Results, expressed as a percentage of activity in controls, show the means ± SEMs from 3 experiments, where error bars exceed symbol size. Lower panel: Slope replot of the data presented in the upper panel. The intercept with the abscissa yielded a K i of 245 ± 153 gM and the intercept with the ordinate axis yielded a kin,c t of 0.071 ± 0.014 min- 1

plot data (Fig. 5), are 17.8 _+ 2.4 gM (fluoxetine) and 18.5 _+ 2.8 pM (norfluoxetine).

!

1~0 1'5 2'0

1/S (1/mM)

0.06

0.04

0.02

-10 -5

b

0.06 •

0 . 0 4 ~ /

-10 -5 0 5 10 15 20

I/s 0/rrN)

Fig. 3 Representative Lineweaver-Burk plots of the inhibition of rat brain MAO-A by fluoxetine (a) and norfluoxetine (b). S represents substrate concentration and v represents reaction velocity. Homogenates (n = 5) were incubated with [3H] 5-HT (50-250/aM) in the absence (O) or presence of fluoxetine or norfluoxetine at 75gM (A), 150pM (V) or 250pM (I~) for 5min at 37°C and MAO-A activities determined as described in the text

Slow reversal of MAO-B inhibition by dialysis

The dissociation of either fluoxetine or norfluoxetine from MAO-B during dialysis was very slow, even at 37 ° C, with enzyme activities recovering to less than 20% of those in control tissues after dialysis for 9 h. MAO-B activities in control homogenates were (pmolh-Xmg -1 protein __+ SEM; n = 5): 5.81 _+ 0.85 (undialysed), 0.86 _+ 0.08 (2 h), 1.01 ± 0.16 (5 h) and 0.99 __+ 0.15 (9h). MAO-B activities in drug-treated homogenates, expressed as a percentage of the corres- ponding control activity _+ SEM were, for fluoxetine: 0.6% _+ 0.4% (undialysed), 1.9% _+ 0.9% (2 h), 8.9% _+ 4.8% (5 h) and 12.3% _+ 4.1% (9 h), and for nor-

fluoxetine: 2.9% __ 1.1% (undialysed), 11.3% _+ 2.8% (2h), 11.3% __ 3.2% (5h) and 18.0% __+ 3.3% (9h).

Brain drug levels and ex vivo MAO inhibition

Following i.p. administration of fluoxetine to rats, both fluoxetine and norfluoxetine were detectable in cortical homogenates (Table 1). Higher concentrations of both compounds were present in brains of animals which

0.15

0.10

0.05

-50 0

t

v

5'0 160 ' 1~0

1/s 0/mM)

0'201 /

01/ / b ~ 0.10

oo5 l

-50 0 50 100 150 1/s 0/mM)

Fig. 4 Representative Lineweaver-Burk plots of the inhibition of rat brain MAO-B by fluoxetine (a) and norfluoxetine (b). Homogenates (n = 5) were incubated with fluoxetine or norfluoxetine at 0 ~tM (O), 25 gM (A), 50 gM (V) or 100 ~tM (0) before incubation with [u~C]PEA (6-35/aM) for 5 min at 37°C and determination of MAO-B activities

120

100 o

i~ 60

40

logl0[lnhibitor], M

Fig. 5 ICso plots for the inhibition of rat brain MAO-B by fluoxetine and norfluoxetine. Homogenates were preincubated with fluoxetine (©) or norfluoxetine (0) at 37 ° C for 20 rain before incubation with [~4C]PEA (50 ~tM) for 10 min to assay remaining MAO-B activity. Results, expressed as a percentage of activity in controls, show means _+ SEMs from 5 animals. ICso values, obtained by nonlinear regression of the data, were 17.8 _+ 2.4 ~tM (fluoxetine) and 18.5 _+ 2.8 ~tM (norfluoxetine). Mean MAO-B activity in controls was 74.1 _+ 3.8 nmolh-~ mg -1 protein

had received ftuoxetine for 7 days when compared with those which had received the drug acutely. Further- more, the ratio of norfluoxetine:fluoxetine was con- siderably higher in animals which had received chronic

21

Table 1 Drug levels measured ex vivo in rat cortex following administration of fluoxetine

Group (n) Fluoxetine Norfluoxetine (ggg i tissue) (pgg-1 tissue)

Control (10) 0 0 Fluoxetine (1 day) (4) 12.10 _+ 1.53 6.36 _+ 0.23 Fluoxetine (7 days) (6) 68.2 _+ 15.4 163.0 _+ 25.9

Rats were administered saline (control) or fluoxetine (20 mgkg 1) intraperitoneatly, once daily for 1 or 7 days and were killed 2 h after the last injection. Levels of fluoxetine and norfluoxetine in cortical homogenates were measured by gas chromatography following an acetylation procedure. In each case, the mean _+ SEM has been given

120

100

3 s0

.~ 60

40

20

1 day, 1:80 7 day, I:80 7day, 1:8

Fig. 6 MAO activities measured ex vivo in rat brain homogenates following acute or chronic administration of fluoxetine. Rats were administered vehicle or fluoxetine (20mgkg 1day- l ) for 1 or 7 days and MAO activities were estimated radiochemically in cortical homogenates prepared 1 : 8 (w/v) or 1 : 80 (w/v). Column shading styles indicate the following measurements: open, MAO-A (vehicle); closed, MAO-A (fluoxetine); cross-hatched, MAO-B (vehicle); stip- pled, MAO-B (fluoxetine). Results show means _+ SEMs from 4 animals (1 day) or 6 animals (7 day). Mean amine oxidase activities in controls (n=10) were 161.6 _+17.4 nmolh i mg -1 protein (MAO-A) and 99.6_+12.Snmolh-1mg 1 protein (MAO-B). * P < 0.05, Student's unpaired t-test, compared with appropriate vehicle

fluoxetine, reflecting the longer half-life of the metabolite (Lemberger et al. 1985; Torok-Both et al. 1992).

When inhibition of MAO-A and -B were measured ex vivo in 1 : 80 (w/v) homogenates of the same tissues, no differences were seen between enzyme activities in controls and in animals which had received fluoxetine for one day (Fig. 6). Following chronic drug administration, both MAO-A and MAO-B activities were unchanged when compared with activities in vehicle- treated rats, although a modest reduction in MAO-B activity just failed to reach statistical significance. How- ever, when enzyme activities were estimated in 1:8 (w/v) homogenates from rats which had received chronic fluoxetine, it was found that MAO-A activity had been reduced by 23% when compared with controls, while MAO-B activity had been further reduced by 4%

22

.9 o

o

5-HT 5-HIAA DA DOPAC HVA

Fig. 7 Levels of neurotransmitters and neurotransmitter metabolites in rat cortex following chronic administration of fluoxetine. Rats received vehicle (open columns) or fluoxetine (20 mgkg -1 day- l; closed columns) for 7 days and concentrations of neurotransmitters and their metabolites in cortical homogenates were determined by h.p.l.c. Results show means +_ SEMs from 6 animals. * P < 0.05, Student's unpaired t-test, compared with vehicle

by increasing the tissue:buffer ratio (Fig. 6) such that the total decrease of 12% was now statistically different from controls.

Levels of neurotransmitters and metabolites in brain cortex

Chronic administration of fluoxetine (20mgkg -1 day -1) reduced cortical levels of 5-HIAA by 47%, compared with vehicle-treated animals (Fig. 7). Levels of all other neurotransmitters and metabolites examined remained unaltered by treatment with fluoxetine.

Discussion

The principal metabolic pathway for fluoxetine identi- fied to date in human subjects and rats is N-demethyla- tion to the primary monoamine, norfluoxetine, and this metabolite is also a potent inhibitor of 5-HT uptake (Fuller 1984). Steady state plasma and tissue levels of norfluoxetine in these species are significantly higher than those of the parent drug since the half-life for elimination of norfluoxetine is significantly longer in both cases (Lemberger et al. 1985; Torok-Both et al. 1992; Altamura et al. 1994). Thus, it was important to consider the potential effects on MAO of both fluoxetine and norfluoxetine. Both compounds were without effects (results not shown) on the metabolism of ben- zylamine (2 jiM) by a third amine oxidase enzyme, semicarbazide-sensitive amine oxidase from rat brown adipose tissue (EC 1.4.3.6; Callingham et al. 1991; Holt et al. 1992) and this class of enzyme was not considered further here.

The present studies have shown that fluoxetine and norfluoxetine are competitive, reversible inhibitors of

rat brain MAO-A, inhibition being independent of preincubation time. Neither compound would appear to be a substrate for MAO-A since no hydrogen perox- ide was produced during incubation of these compounds with rat brain mitochondria, at concentrations three times higher than their Ki values reported here (A. Holt, unpublished observations). Thus, fluoxetine and norfluoxetine probably act as substrate analogues for MAO-A, reducing metabolism of endogenous substrates by competition for the enzyme active site.

It is very difficult to demonstrate competitive, reversible inhibition ex vivo since preparation of a homogenate in buffer will effectively dialyse the inhibitor from the enzyme as a result of inhibitor dilution (Green 1984). While no inhibition of MAO-A was apparent in a 1:80 (w/v) cortical homogenate, reducing the buffer volume and thus increasing the inhibitor concentration ten-fold reduced metabolism of [ 3 H ]

5-HT by 23 %. In whole cortex, the combined concentration of fluoxetine and norfluoxetine following chronic drug administration was 231 gg g-1 and this ap- proximates to 700 gM inhibitor if it is assumed that lg = 1 ml. Although this concentration is very much higher than the Ki value for either compound, the Ki is the inhibitor dissociation constant in the absence of a competing ligand and inhibition of MAO-A in situ will depend very much upon the intraneuronal concentrations of competing amine substrates and their rela- tive affinities for MAO-A, compared with those of fluoxetine and norfluoxetine. Thus, while it is tempting to extrapolate 23% inhibition of metabolism of 250 gM 5-HT in a 1:8 homogenate back to the whole tissue, this would provide no meaningful estimate of the extent of inhibition in vivo.

Both fluoxetine and norfluoxetine were found to be time-dependent, slowly reversible inhibitors of MAO-B in vitro, although kinetic plots were ambiguous and did not fully reveal the true nature of the interaction. An examination of this time-dependent inhibition of MAO-B by fluoxetine, by the method of Kitz and Wilson (1962), yielded a slope replot which did not pass through the origin. This suggests that initial formation of a readily-reversible complex is followed by further metabolism to a tightly-bound transition-state ana- logue. This intermediate was not covalently bound to MAO-B since activity could be recovered, albeit slowly, by dialysis. It seems, at least in vitro therefore, that fluoxetine (and probably norftuoxetine) are mechanism-based inhibitors of MAO-B (Tipton 1994) and follow the reaction mechanism:

Ki k~nl~cT

_ _ I ~ E + I ~ E - I ~- E (1)

where EI* represents the tightly-bound, slowly-reversible enzyme-inhibitor complex. The presence of both competitive and non-competitive components of inhibition (Eq. 1) may explain the ambiguity seen in

23

Lineweaver-Burk plots. The ICs0 values for fluoxetine and norfluoxetine were determined in vitro to be 17.8 pM and 18.5 pM, respectively. These values are substantially lower than the estimated combined inhibitor concentration of 700 gM present in cortex following chronic administration of fluoxetine and one might therefore expect to see substantial inhibition of MAO-B ex vivo. However, this was not the case, with /3-PEA deamination reduced by less than 10%, compared with controls. This apparent lack of inhibition was not a result of inhibitor dilution since the level of inhibition in a 1 : 8 (w/v) homogenate was only margin- ally higher than that in a 1:80 (w/v) homogenate. Rather, it is likely that the initial formation of a reversible complex between fluoxetine (or norfluoxetine) and MAO-B is inhibited competitively by high concentrations of endogenous amine substrates.

Measurements of changes in tissue content of amines and their metabolites may provide evidence of MAO inhibition in vivo. While such changes are often specific to a particular brain region, an overall trend is apparent for each compound when results from various studies in rodents are compared (Campbell et al. 1979; Waldmeier et al. 1981; Kato et al. 1986; O'Regan et al. 1987; Colzi et al. 1990; McKenna et al. 1991; Berry et al. 1994). Substantial ( > 95 %) chronic inhibition of MAO- A is necessary to double tissue NA content, while partial (80%) inhibition may increase NA levels only by 30%. Similarly, substantial MAO-A inhibition increases DA levels only by around 25% while DOPAC may, under similar conditions, be reduced by 60% and HVA by 50 80%. Concurrent inhibition of MAO-B may further deplete tissue DOPAC and HVA but effects on DA are marginal in most studies. Again, partial inhibition of MAO-A greatly reduces the magnitude of the changes in DA metabolite levels. Finally, substantial inhibition of MAO-A, but not MAO-B, initiates a 50-100% increase in 5-HT and a 30-40% decrease in 5-HIAA levels, with partial inhibition having only marginal effects on tissue levels of these compounds.

From the measurements of brain amines and their metabolites made in this study, it is apparent that the degree of inhibition of either form of MAO by fluoxetine and norfluoxetine was insufficient to initiate measurable changes in cortical levels of these compounds. The significant decrease in 5-HIAA with little effect on 5-HT is consistent with inhibition of 5-HT uptake (Fuller et al. 1974). These results suggest that, with the exception of the small, non-competitive com- ponent of MAO-B inhibition which can be measured directly, the initial, competitive interactions between fluoxetine, norfluoxetine and either form of MAO result in, at best, moderate inhibition of biogenic amine turnover in vivo.

Significant increases in NA levels and decreased turnover have been measured in the dorsomedial hy- pothalamic nucleus, dorsal raph6 and motor parietal cortex, but not in other nuclei, following administra-

tion of ftuoxetine to rats for four days (Frankfurt et al. 1994). It is thus possible that moderate MAO-A inhibition by fluoxetine and norfluoxetine may cause regio- specific reductions in amine turnover, which may not be evident when measurements are made in homogenates of larger brain regions, such as whole cortex. It remains to be seen whether antidepressant efficacy in general is associated with altered neurotransmitter turnover in discrete brain nuclei.

In summary, while in vitro experiments suggest that both fluoxetine and norfluoxetine are reasonably potent MAO inhibitors with moderate selectivity for MAO-B, their potencies in vivo are markedly reduced, probably by competing endogenous amines. Further- more, since fluoxetine has a higher affinity for MAO-A than for MAO-B when formation of an initial reversible complex is considered, some selectivity for MAO- A may exist in vivo. While any effects on transmitter metabolism which could be attributed solely to MAO inhibition were not evident in cortical homogenates, it is possible that alterations in NA turnover in brain nuclei observed by other workers are a consequence of MAO-A inhibition. Finally, the present results do not preclude the possibility that moderate inhibition of MAO-A by fluoxetine in 5-hydroxytryptaminergic neurones may augment the increase in neurotransmission which results from inhibition of 5-HT uptake.

Acknowledgements The authors wish to thank Eli Lilly Research Laboratories for their gift of fluoxetine and norfluoxetine and Gail Rauw for her technical assistance with h.p.l.c, and g.l.c, techniques. Funding was provided by the Alberta Mental Health Research Fund and the Medical Research Council of Canada.

References

Altamura AC, Moro AR, Percudani M (1994) Clinical pharmaco- kinetics of fluoxetine. Clin Pharmacokinet 26:201 214

Baker GB, Coutts RT, Rao TS (1987) Neuropharmacological and neurochemical properties of N-(2-cyanoethyl)-2-phenylethylamine, a prodrug of 2-phenylethylamine. Br J Pharmacol 92: 243-255

Berry MD, Scarr E, Zhu M-Y, Paterson IA, Juorio AV (1994) The effects of administration of monoamine oxidase-B inhibitors on rat striatal neurone responses to dopamine. Br J Pharmacol 113:1159 1166

Blackwell B (1963) Hypertensive crisis due to monoamine-oxidase inhibitors. Lancet ii: 849-851

Blier P, De Montigny C (1994) Current advances and trends in the treatment of depression. Trends Pharmacol Sci 15:220 226

Briley M, Moret C (1993) Neurobiological mechanisms involved in antidepressant therapies. Clin Neuropharmacol 16:387-400

Bowden CL, Schatzberg AF, Rosenbaum A, Contreras SA, Samson JA, Dessain E, Sayler M (1993) Fluoxetine and desipramine in major depressive disorder. J Clin Psychopharmacol 13:305 311

Bowman WC, Rand MJ (1980) In: Textbook of pharmacology, 2nd edn. Blackwell Scientific Publications, Oxford, pp 15.14 15.18

Burgen ASV, Iversen LL (1965) The inhibition of noradrenaline uptake by sympathomimetic amines in the rat isolated heart. Br J Pharmacol Chemother 25:34 49

24

Callingham BA (1989) Biochemical aspects of the pharmacology of moclobemide. The implications of animal studies. Br J Psychia- try 155 [Suppl 6]:53-60

Callingham BA, Holt A, Elliott J (1991) Properties and functions of the semicarbazide-sensitive amine oxidases. Biochem Soc Trans 19:228-233

Campbell IC, Robinson DS, Lovenberg W, Murphy DL (1979) The effects of chronic regimens of clorgyline and pargyline on monoamine metabolism in the rat brain. J Neurochem 32:49-55

Colzi A, D'Agostini F, Kettler R, Borroni E, Da Prada M (1990) Effect of selective and reversible MAO inhibitors on dopamine outflow in rat striatum: a microdialysis study. J Neural Transm 32 [Suppl]:79 84

Drebit R, Baker GB, Dewhurst WG (1988) Determination of maprotiline in plasma and urine by gas-liquid chromatography with nitrogen-phosphorus detection. J Chromatogr Biomed Appl 432:334-339

Edwards DJ, Burns MO (1974) Effects of tricyclic antidepressants upon human platelet monoamine oxidase. Life Sci 15:2045-2058

Fowler C J, Oreland L, Callingham BA (1981) The acetylenic monoamine oxidase inhibitors clorgyline, deprenyl, pargyline and J- 508: their properties and applications. J Pharm Pharmacol 33:341-347

Frankfurt M, McKittrick CR, Luine VN (1994) Short-term fluoxetine treatment alters monoamine levels and turnover in discrete brain nuclei. Brain Res 650:127-132

Fuller RW (1984) Specificity in the inhibition of monoamine uptake. Clin Neuropharmacol 7 [Suppl]: 310-311

Fuller RW, Perry KW (1975) Inability of an inhibitor of amine uptake (Lilly 110140) to block depletion of brain 5-hydroxytryptamine by L-dopa. J Pharm Pharmacol 27:618-620

Fuller RW, Perry KW (1981) Elevation of 3,4-dihydroxyphenylacetic acid concentration by L-5-hydroxytryptophan in control and fluoxetine-pretreated rats. J Pharm Pharmacol 33:406 407

Fuller RW, Snoddy HD (1986) Effect of fluoxetine pretreatment on the neurochemical changes induced by amfonelic acid combined with spiperone in rats. J Pharm Pharmacol 38:776 777

Fuller RW, Perry KW, Molloy BB (1974) Effect of an uptake inhibitor on serotonin metabolism in rat brain: studies with 3-(p-trifluoromethylphenoxy)-N-methyl- 3-phenylpropylamine (Lilly 110140). Life Sci 15:1161-1171

Fuller RW, Wong DT, Robertson DW (1991) Fluoxetine, a selective inhibitor of serotonin uptake. Med Res Rev 11:17 34

Gram LF (1994) Fluoxetine. N Engl J Med 331:1354-1361 Green AL (1984) Assessment of the potency of reversible MAO

inhibitors in vivo. In: Tipton KF, Dostert P, Strolin-Benedetti M (eds) Monoamine oxidase and disease. Prospects for therapy with reversible inhibitors. Academic Press, London, pp 73-81

Holt A, Sharman DF, Callingham BA, Kettler R (1992) Character- istics of procarbazine as an inhibitor in-vitro of rat semicarbazide-sensitive amine oxidase. J Pharm Pharmacol 44:487 493

Kasper S, H6flich G, Scholl H-P, M611er H-J (1994) Safety and antidepressant efficacy of selective serotonin re-uptake inhibitors. Hum Psychopharmacol 9:1-12

Kato T, Dong B, lshii K, Kinemuchi H (1986) Brain dialysis: in vivo metabolism of dopamine and serotonin by monoamine oxidase A but not B in the striatum of unrestrained rats. J Neurochem 46:1277-1282

Kitz R, Wilson IB (1962) Esters of methanesulfonic acid as irreversible inhibitors of acetylcholinesterase. J Biol Chem 237: 3245-3249

Kokotos Leonardi ET, Azmitia EC (1994) MDMA (ecstasy) inhibition of MAO type A and type B: comparisons with fenfluramine and fluoxetine (Prozac). Neuropsychopharmacology 10:231-238

Kopin IJ (1994) Monoamine oxidase and catecholamine metabolism. J Neural Transm 41 [Suppl]:57-67

Lemberger L, Bergstrom RF, Wolen RL, Farid NA, Enas GG, Aronoff GR (1985) Fluoxetine: clinical pharmacology and physiologic disposition. J Clin Psychiatry 46:14 19

Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951) Protein measurement with the Folin phenol reagent. J Biol Chem 193:265 275

Lyles GA, Callingham BA (1982) In vitro and in vivo inhibition by benserazide of clorgyline-resistant amine oxidases in rat cardio- vascular tissues. Biochem Pharmacol 31:1417-1424

Martin IL, Baker GB (1977) A gas-liquid chromatographic method for the estimation of 2-phenylethylamine in rat brain tissue. Biochem Pharmacol 26:1513-1516

McKenna KF, Baker GB, Coutts RT (1991) N2-Acetylphenelzine: effects on rat brain GABA, alanine and biogenic amines. Naunyn-Schmiedeberg's Arch Pharmacol 343:478 482

O'Regan D, Kwok RPS, Yu PH, Bailey BA, Greenshaw AJ, Boulton AA (1987) A behavioural and neurochemical analysis of chronic and selective monoamine oxidase inhibition. Psychopharmacol- ogy 92:42-47

Reid AA, Hill JL, Murphy DL (1988) Interactions of tricyclic antidepressant drugs with human and rat monoamine oxidase type B. Naunyn-Schmiedeberg's Arch Pharmacol 338:678 683

Roth JA (1976) Multiple forms of monoamine oxidase and their interaction with tricyclic psychomimetic drugs. Gen Pharmacol 7:381-386

Tipton KF (1980) Kinetics and enzyme inhibition studies. In: Sand- ler M (ed) Enzyme inhibitors as drugs. Macmillan Press, London, pp 1 23

Tipton KF (1994) Monoamine oxidase inhibition. Biochem Soc Trans 22:764 768

Torok-Both GA, Baker GB, Coutts RT, McKenna KF, Aspeslet LJ (1992) Simultaneous determination of fluoxetine and norfluoxetine enantiomers in biological samples by gas chromatography with electron-capture detection. J Chromatogr 579:99-106

Trendelenburg U, Cassis L, Grohmann M, Langeloh A (1987) The functional coupling of neuronal and extraneuronal transport with intracellular monoamine oxidase. J Neural Transm 23 [Suppl]: 91-101

Waldmeier PC, Felner AE, Maitre L (1981) Long-term effects of selective MAO inhibitors on MAO activity and amine metabolism. In: Youdim MBH, Paykel ES (eds) Monoamine oxidase inhibitors - the state of the art. John Wiley, New York, pp 87 102

Wong DT, Horng JS, Bymaster FP, Hauser KL, Molloy BB (1974) A selective inhibitor of serotonin uptake: Lilly 110140, 3-(p- trifluoromethylphenoxy)-N-methyl-3-phenylpropylamine. Life Sci 15:471 479

Wong DT, Bymaster FP, Horng JS, Molloy BB (1975) A new selective inhibitor for uptake of serotonin into synaptosomes of rat brain: 3-(p-trifluoromethylphenoxy)-N-methyl-3-phenylpropylamine. J Pharmacol Exp Ther 193:804-811

Documents

Inhibition of rat brain monoamine oxidase enzymes by fluoxetine and norfluoxetine