BIOPHARMACEUTICS & DRUG DISPOSITIONBiopharm. Drug Dispos. 19: 147–151 (1998)

Pharmacokinetics and Urinary Excretion of DMXBA (GTS-21), aCompound Enhancing Cognition

V. Mahnir, B. Lin, K. Prokai-Tatrai and W.R. Kem*Department of Pharmacology and Therapeutics, University of Florida College of Medicine, Gainesville, FL 32610-0267, USA

ABSTRACT: DMXBA (3-(2, 4-dimethoxybenzylidene)-anabaseine, also known as GTS-21) is currentlybeing tested as a possible pharmacological treatment of cognitive dysfunction in Alzheimer’s disease. Inthis study, plasma and brain pharmacokinetics as well as urinary excretion of this compound have beenevaluated in adult rats. DMXBA concentrations were determined by HPLC. Following a 5 mg kg−1 ivdose, DMXBA plasma concentration declined bi-exponentially with mean (9SE) absorption and elimina-tion half-lives of 0.7190.28 and 3.7191.12 h, respectively. The apparent steady state volume of distribu-tion was 21509433 mL kg−1, total body clearance was 14809273 mL h−1 kg−1, and AUC0-� was37909630 ng h mL−1. Orally administered DMXBA was rapidly absorbed. After oral administration of10 mg kg−1, a peak plasma concentration of 10109212 ng mL−1 was observed at 10 min after dosing.Elimination half-life was 1.74090.34 h, and AUC0-� was 14409358 ng h mL−1. DMXBA peak brainconcentration after oral administration was 6649103 ng g−1 tissue, with an essentially constant brain–plasma concentration ratio of 2.6190.34, which indicates that the drug readily passes across theblood–brain barrier. Serum protein binding was 80.391.1%. Apparent oral bioavailability was 19%.Renal clearance (21.8 mL h−1 kg−1) was less than 2% of the total clearance (14809273 mL h−1 kg−1);urinary excretion of unchanged DMXBA over a 96 h period accounted for only 0.2890.03% of the totalorally administered dose. Our data indicates that DMXBA oral bioavailability is primarily limited byhepatic metabolism. © 1998 John Wiley & Sons, Ltd.

Key words: DMXBA; Alzheimer’s disease; GTS-21; nicotine; nicotinic agonist; pharmacokinetics

Introduction

There is an intense pharmaceutical interest in de-veloping new drugs to delay the loss of cognitivefunction in Alzheimer’s disease. Cholinergic neu-rons in the brain seem particularly susceptible tothe as yet unknown degenerative stimuli ofAlzheimer’s disease, and nicotinic receptor con-centrations may decrease by more than 50% [1,2].Receptors displaying high affinity for nicotine arepreferentially lost during the progression of thisdisease, while alpha-7-subunit-containing receptorsdisplaying low affinity for nicotine but highaffinity for a-bungarotoxin are less affected [3]. Anew approach for alleviating some cognitivedeficits due to cholinergic neuron dysfunctionwould be pharmacological stimulation of these re-maining brain nicotinic receptors [4]. DMXBA(chemical name, 3-(2, 4-dimethoxybenzylidene)-an-abaseine; coded name, GTS-21) is a compoundundergoing clinical tests for treatment of

Alzheimer’s disease [5]. This compound is a syn-thetic analog of anabaseine (Figure 1), a marinenatural product [6]. DMXBA acts as a partial ago-nist upon the a7 nicotinic receptor subtype but asan antagonist upon the a4b2 nicotinic acetyl-cholinic receptor subtype [7–9]. It acts as an an-tagonist upon skeletal muscle and ganglionicnicotinic receptors, but only at very high (\10mM) concentrations [10]. DMXBA enhances learn-ing in aging rabbits [9] and rats [11], and in nu-cleus-basalis-lesioned rats [13]. In addition to itscognition-enhancing actions, DMXBA also has cy-toprotective effects upon cholinergic neurons [12].

The present paper describes the first study ofthe pharmacokinetics and cumulative urinary ex-cretion of DMXBA in rats.

Materials and Methods

Chemicals

DMXBA dihydrochloride was synthesized by re-action of 2, 4-dimethoxybenzaldehyde with anaba-seine dihydrochloride, essentially as previously

* Correspondence to: Department of Pharmacology and Therapeutics, Uni-versity of Florida College of Medicine, Gainesville, FL 32610-0267, USA.Tel.: +1 352 3923541; e-mail: Kem@qm.server.ufl.edu

Contract grant sponsor: Taiho Pharmaceutical Company

Received 18 November 1996Accepted 10 February 1997

CCC 0142–2782/98/030147–05$17.50© 1998 John Wiley & Sons, Ltd.

V. MAHNIR ET AL.148

described for related compounds [13]. Its puritywas established by thin-layer chromatography,melting point, mass spectral and NMR analyses.Acetonitrile was HPLC grade. All other solventsand reagents used were of analytical reagentgrade.

Animals

Male Sprague–Dawley rats weighing 400–450 gwere purchased from Charles River Breeding Labo-ratories (Boston, MA) and maintained on a 12 hday–12 h night cycle. They were provided ab libitumaccess to food (Purina rat chow) and water. Allanimals were maintained for at least 1 week underthese conditions before they were used in experi-ments. The rats were fasted overnight before anexperiment.

Drug Administration and Collection of BloodSamples

In vivo experiments were performed under condi-tions similar to those described by Sastry et al. [14].Four animals were anesthetized with Innovar-Vet(80 mL kg−1 ip) consisting of 2 mg mL−1 droperidoland 0.4 mg mL−1 fentanyl. Thirty minutes later5 mg kg−1 DMXBA was administered iv (10 mgmL−1 in saline, pH 5.0) by injection into the leftfemoral vein. Animals remained conscious butslightly sedated during the entire experiment. Bloodsamples (0.15–0.3 mL) were collected by tail veinbleeding in heparinized plastic microcentrifugetubes at 0.083, 0.17, 0.33, 0.5, 0.75, 1.0, 2.0, 3.0, 4.0,6.0 and 8.0 h after dosing. Plasma was immediatelyseparated by centrifugation at 2500×g for 10 min at4°C.

Using the same anesthesia as in the iv experi-ment, four rats were orally administered DMXBA(10 mg mL−1 in saline, pH 5.0) at a dose 10 mgkg−1, using a syringe fitted with a 38 mm longfeeding needle. Blood samples were collected by tailvein bleeding. Preliminary oral experiments mea-suring the plasma concentrations as a function oftime were performed without anesthesia. Compari-son of the oral administration data obtained withanesthetized animals with data obtained in the pre-liminary experiment without anesthesia did not re-veal any apparent influence of the anesthesia onDMXBA pharmacokinetics. In a separate experi-ment three rats orally received 10 mg kg−1 DMXBAand blood as well as brain samples were obtainedafter decapitation at the specified times. Plasma wasseparated by centrifugation. After oral administra-tion of DMXBA urine was collected over a 96 hperiod from rats housed individually in metaboliccages. All samples were stored at −40°C untilanalysis.

Analytical Procedures

Collected plasma was deproteinated with acetoni-trile (1:2 v/v), and centrifuged at 15 000×g for 10min and the resulting supernatant evaporated todryness with a stream of nitrogen at 40°C. Afteradding NaOH to 0.1 M, urine samples were ex-tracted with 2 mL hexane; the organic phase wasthen transferred to a clean tube and evaporatedunder a stream of nitrogen. For DMXBA extractionbrain samples homogenized in four parts (v/w) ofacetonitrile were left overnight at 4°C. Insolublematerial was then separated by a 10 min centrifuga-tion at 15 000×g, then the supernatant was evapo-rated with a stream of nitrogen. All samples werekept at −20°C before further analysis by HPLC.Dry samples were reconstituted in 80 mL 0.1% tri-fluoroacetic acid, centrifuged for 10 min at 15 000×g and stored overnight in the dark (DMXBA issensitive to light) at 4°C. On the next day 10–65 mLof each sample was injected into a Vydac C4column (4 mm×250 mm) which was eluted with32% acetonitrile–0.1% trifluoroacetic acid (v/v) at 1mL min−1. Absorbance was measured at 400 nm.The approximate retention time of DMXBA was 6.2min, the lower limit of its determination was 15 ngmL−1 and the SE of three determinations was lessthan 3%.

DMXBA HPLC Determinations

We were unable to find a satisfactory internal stan-dard for DMXBA determination. External standardswere prepared by adding DMXBA to 0.2 mL drug-free plasma or urine, or 0.9 g brain homogenate, toachieve concentrations of 0.25, 0.5 and 1.0 ng mL−1

for plasma and urine, or 0.5 and 1.0 ng g−1 forbrain tissue. Samples were prepared in triplicateand processed as described above. Analytical recov-ery of DMXBA was independent of its concentra-tion. The analytical recovery of DMXBA was8092.8% from plasma, 6094.5% from brain, and9193.5% from urine when compared to directcolumn injection of known DMXBA aliquots.

Serum Protein Binding

Pooled trunk blood from three rats was allowed tocoagulate at room temperature for 30 min; after clotremoval, the serum fraction was centrifuged at2000×g to obtain undiluted serum. A 1 mL sampleof rat serum was placed inside 1000 mw cut-offdialysis tubing, which was then immersed in astoppered centrifuge bottle containing DMXBA di-hydrochloride dissolved in 50 mL 0.9% NaCl con-taining 10 mM sodium phosphate, pH 7.4. After 8 hequilibration on a shaker at room temperature, both

© 1998 John Wiley & Sons, Ltd. Biopharm. Drug Dispos. 19: 147–151 (1998)

DMXBA (GTS-21) PHARMACOKINETICS 149

phases were diluted with 20% trichloroacetic acid(1:1 v/v), and centrifuged for 10 min at 10 000×gand the DMXBA concentration determined spec-trophotometrically at 400 nm.

Octanol/Water Partition Coefficient

A 50 mL sample of 0.1 M DMXBA dihydrochloridewas added to a centrifuge tube containing 3.0 mLsaline-saturated octanol and 3.0 mL octanol-satu-rated saline (0.9% NaCl containing 10 mM sodiumphosphate, pH 7.4). The tube was then stopperedand shaken for 8 h at room temperature. Subse-quently 1.0 mL glacial acetic acid and 1.0 mLmethanol were added to a 1.0 mL aliquot of eachphase in order to spectrophotometrically measureDMXBA at 400 nm. The partition coefficient of theunionized form of DMXBA was calculated assum-ing a pKa of 7.6.

Pharmacokinetic Analysis

Plasma concentration of DMXBA was fitted to thebi-exponential equation Cp=A exp(−at)+B exp(−bt) using a nonlinear least-squares pro-gram Kinetic (Biosoft, UK) [15]. Secondarypharmacokinetic parameters were calculated as fol-lows: t1/2a=0.693/a ; t1/2b=0.693/b ; AUC0–�=A/a+B/b ; Vc=D/(A+B); Cltot=D/AUC0–�;k21= (Ab+Ba)/(A+B); k10=ab/k21; k12=a+b−k21−k10; Vt=Vc ·k12/k21; Vss/Vc+Vt; where D is ivadministered dose, Vc and Vt are apparent volumesof distribution for central and peripheral compart-ments, Vss is the steady state volume of distribution,and k12, k21, and k10 are the elimination rate con-stants between the compartments. Renal clearance(Clr) of DMXBA was calculated from the amount ofthe drug excreted in the urine divided by AUC(o-ral). Bioavailability was calculated as the dose-nor-malized ratio of oral and iv AUCs.

Figure 2. Mean plasma concentrations of DMXBA after 5 mgkg−1 iv (— — ) and 10 mg kg−1 oral (—�—) administrationto the rat. Each data point represents the mean9SE obtainedusing four animals

Results and Discussion

This is the first scientific report of the basic pharm-cokinetic properties of DMXBA in the adult malerat. HPLC proved to be a sensitive method fordetecting and measuring DMXBA in biologicalfluids. As little as 10 ng DMXBA per peak could bedetected by measuring its 400 nm absorbance. Wewere able to measure DMXBA plasma and brainconcentrations for 8 h after administration. Plasmaconcentration–time curves obtained after iv admin-istration of DMXBA were well described by thetwo-compartment model. The mean plasma concen-tration–time curve after 5 mg kg−1 iv dosing isshown in Figure 2. The plasma level of DMXBArapidly declined during the distribution phase andmuch more slowly during the elimination phase.The initial pharmacokinetic parameters (mean9SE)obtained as a result of fitting experimental data tothe bi-exponential equation were as follows: A=24709437 ng mL−1; a=1.9190.57 h−1; B=5989186 ng mL−1; and b=0.2490.06 h−1. Secondarypharmacokinetic parameters (Table 1) were calcu-lated as described in the Methods section.

Ten minutes following oral administration at adose of 10 mg kg−1, a peak plasma concentration of10109212 ng mL DMXBA was measured (Figure2). This indicates that the drug was rapidly ab-sorbed into the systemic circulation. The eliminationrate obtained as a result of computer analysis of theconcentration-time curve was 0.4490.08 h−1, corre-sponding to a half-life of 1.7490.34 h (Table 1). Theoral bioavailability was calculated to be 19%. It isunlikely that this modest bioavailability is due to alimited absorption, as the time required for reachingthe peak DMXBA plasma concentration was very

Figure 1. Chemical structures of anabaseine and DMXBA, shownas free bases

© 1998 John Wiley & Sons, Ltd. Biopharm. Drug Dispos. 19: 147–151 (1998)

V. MAHNIR ET AL.150

Table 1. Pharmacokinetic parameters of DMXBAafter iv (5 mg kg−1) and oral (10 mg kg−1) admin-istration in the rat (mean9SE, n=4)

Parameter Intravenous Oral

— 10109212Cmax (ng mL−1)—24709437A (ng mL−1)

5989186B (ng mL−1) 4959881.9190.57a (h−1) —0.2490.06 0.4490.08b (h−1)0.7190.28t1/2a (h) —

1.7490.343.7191.12t1/2b (h)AUC0–� (ng h 144093537909630

mL−1)—21509433Vss (mL kg−1)

14809273Cltot (mL kg−1 h−1) ——21.8Clr (mL kg−1 h−1)

Figure 4. Cumulative urinary excretion of DMXBA after a single10 mg mL−1 oral dose. Each data point represents the mean9S.E. obtained using five rats

short. A more likely basis for the low bioavailabilityis extensive first-pass metabolism and biliary excre-tion. From the chemical structure of DMXBA (Fig-ure 1) we predicted that the dimethoxybenzylidenemoiety would be an excellent target for first-passmetabolism, yielding monohydroxy-monomethoxy-derivatives as major putative primary metabolitestogether with the dihydroxy-derivative. Indeed, wehave detected both monohydroxy-derivatives in invitro experiments with rat liver homogenate [16].

The higher elimination rate observed after oraladministration may suggest nonlinear pharmacoki-netics, since the peak plasma concentration of thedrug was much less than that observed after ivadministration. Alternatively, it may be caused by agreater overlap of the initial distribution phase withthe metabolic phase of elimination when DMXBA isorally administered.

The DMXBA concentration in rat brain after 10mg kg−1 oral administration was measured in aseparate experiment (Figure 3). Brain and bloodsamples were collected after decapitation of animals

(three rats per time point). A peak brain concentra-tion of 6649103 ng g−1 tissue was detected (datafor blood concentrations are not shown) with anessentially constant brain–plasma concentration ra-tio of 2.6190.34. This suggests that DMXBA readilypasses across the blood–brain barrier. The meancumulative excretion curve after 10 mg kg−1 oraladministration is shown in Figure 4. The urinaryexcretion of unchanged DMXBA accounted for only0.2890.03% of the total orally administered dose(2.6% of the amount of DMXBA absorbed) over 96 h.The renal clearance was only 21.8 mL h−1 kg−1.This indicates that the major route of elimination ofDMXBA is not renal.

In the rat, we observed that DMXBA serum pro-tein binding (80.391.1%) was directly proportionalto its free concentration, at least up to 38 000 ngmL−1, the highest free concentration investigated.

Since both DMXBA and nicotine are 3-substitutedpyrdine nicotinic agonists under consideration forAlzheimer’s disease therapy, we compare some ofthe pharmacokinetic properties of these two com-

Figure 3. Mean concentration of DMXBA in brain following oraladministration at the dose of 10 mg kg−1 to the rat. Each datapoint represents the mean9SE of three animals

Table 2. A comparison of pharmacokinetic properties ofDMXBA with those of nicotine (References for nicotinedata are indicated)

NicotineProperty DMXBA

Ionization at pH 7.4 (%) 61 70 [18]Partition coefficient 71 15 [18]Serum protein binding (%) 80 B10 [17]Vss (L kg−1) 5.2 [19]2.2

3.7t1/2b (h) 0.9 [19]Oral bioavailability (%) 19 45 [17]Cltot (L h−1 kg−1) 1.48 2.90 [19]Clr (L h−1 kg−1) 0.022 0.20 [19]Brain–plasma ratio 2.1 [17,19]2.61

© 1998 John Wiley & Sons, Ltd. Biopharm. Drug Dispos. 19: 147–151 (1998)

DMXBA (GTS-21) PHARMACOKINETICS 151

pounds, as shown in Table 2. DMXBA is less basic,but more lipophilic than nicotine. These differencesare thought to contribute to its significantly greaterplasma protein binding relative to nicotine. Thesmaller apparent volume of distribution of DMXBAis also probably related to its higher plasma proteinbinding. The oral bioavailability of DMXBA is sig-nificantly less than that of nicotine. The plasmahalf-life calculated for unchanged DMXBA is aboutfour times longer than for nicotine; this property isdesirable for maintaining a relatively constantplasma concentration of the compound. The totalclearance of both compounds is primarily due tometabolic transformation, rather than renal clear-ance. A major difference in the elimination mecha-nisms for DMXBA and nicotine is that the primarymetabolites of the former compound pharmacologi-cally resemble the parent compound, whereas theprimary metabolite of nicotine, cotinine, lacks nico-tinic agonist activity [16,17].

Acknowledgements

We thank Dr H. Derendorf for comments on themanuscript and Ms. Judy Adams for assistance inword-processing. This work was partially sup-ported by the Taiho Pharmaceutical Company,Tokushima, Japan.

References

1. P.J. Whitehouse, A.M. Martino, P.G. Antuono, P.R. Lowen-stein, J.T. Coyle, D.L. Price and K.J. Kellar, Nicotinic acetyl-choline binding in Alzheimer’s disease. Brain Res., 371,146–151 (1986).

2. D.M. Araujo, P. A. Lapchak, Y. Robitaille, S. Gauthier and R.Quirion, Differential alternation of various cholinergic mark-ers in cortical and subcortical regions of human brain inAlzheimer’s disease. J. Neurochem., 50, 1914–1923 (1988).

3. K. Sugaya, E. Giacobini, and V.A. Chiappinelli, Nicotinicacetylcholine receptor subtypes in human frontal cortex:changes in Alzheimer’s disease. J. Neurosci. Res., 27, 349–359(1990).

4. P.A. Newhouse and A. Potter, The role of nicotinic systemsin the cognitive disorder of Alzheimer’s disease. In AlzheimerDisease: Therapeutic Strategies, E. Giacobini and R. Becker,(Eds), Birkhauser, Boston, pp. 191–195 (1994).

5. W.R. Kem, V.M. Mahnir and B. Lin, Interaction of DMXBA

(GTS-21), a cognition-enhancing compound, with cholinergicreceptors. Soc. Neurosci. Abstr., 20, 1134 (1994).

6. W.R. Kem, B.C. Abbott and R.M. Coates, Isolation and struc-ture of hoplonemertine toxin. Toxicon, 9, 15–22 (1971).

7. R.L. Papke, C.M. de Fiebre and W.R. Kem, The subunitspecific effects of novel anabaseine-derived nicotinic agents.In Alzheimer disease: Therapeutic Strategies, E. Giacobini and R.Becker (Eds), Birkhauser, Boston, pp. 206–211 (1994).

8. C.M. de Fiebre, E.M. Meyer, J.C. Hentry, S.I. Muraskin, W.R.Kem and R.L. Papke, Characterization of a series of anaba-seine-derived compounds reveals that the 3-(4)-dimethy-laminocinnamylidine derivative (DMAC) is a selectiveagonist at neuronal nicotinic a7 [125I]a-bungarotoxin receptorsubtypes. Mol. Pharmacol., 47, 164–171 (1995).

9. D.S. Woodruff-Pak, Y.-T. Li and W.R. Kem, A nicotinicagonist (GTS-21), eyeblink classical conditioning, and nico-tinic receptor binding in rabbit brain. Brain Res., 645, 309–317(1994).

10. G.W. Arendash, G.J. Sengstock, P.R. Sanberg and W.R. Kem.Improved learning and memory in aged rats with chronicadministration of the nicotinic receptor agonist GTS-21. BrainRes., 674, 252–259.

11. E.M. Meyer, C.M. de Fiebre, B.E. Hunter, C.E. Simpkins, N.Frauworth and N.E.C. de Fiebre, Effects of anabaseine-re-lated analogs on rat brain nicotinic receptor binding and onavoidance behaviors. Drug Dev. Res., 31, 127–134 (1994).

12. E.J. Martin, D.S. Panickar, M.A. King, M. Deyrup, B.E.Hunter, G. Wang and E.M. Meyer, Cytoprotective actions of2, 4-dimethoxybenzylidene anabaseine in differentiated PC12cells and septal cholinergic neurons. Drug Dev. Res., 31,135–141 (1994).

13. J.A. Zoltewicz, K. Prokai-Tatrai, L.B. Bloom and W.R. Kem,Long range transmission of polar effects in cholinergic 3-arylidene anabaseines. Conformations calculated by molecu-lar modelling. Heterocyclics, 35, 171–179.

14. B.V.R. Sastry, M.B. Chance, G. Singh, J.L. Horn and V.E.Janson, Distribution and retention of nicotine and its metabo-lite, cotinine, in the rat as a function of time. Pharmacology,50, 128–136 (1995).

15. G.A. McPherson, Analysis of radioligand binding experi-ments: a collection of computer programs for the IBM PC. J.Pharmacol. Methods, 14, 213–228 (1985).

16. W.R. Kem, V.M. Mahnir, B. Lin and K. Prokai-Tatrai, Twoprimary GTS-21 metabolites are potent partial agonists atalpha7 nicotinic receptors expressed in the Xenopus oocyte.Soc. Neurosci. Abstr., 22, 268 (1996).

17. N.L. Benowitz, H. Porchet and P. Jacob. Pharmacokinetics,metabolism, and pharmacodynamics of nicotine. In NicotinePsychopharmacology: Molecular, Cellular, and Behavioral Aspects,S. Wonnacott, M.A.H. Russell and I.P. Stolerman (Eds), Ox-ford University Press, Oxford, UK, pp. 112–157 (1990).

18. A. Leo, C. Hansch and D. Elkins, Partition coefficients andtheir uses. Chem. Rev., 71, 525–616 (1971).

19. D.R. Plowchalk, M.E. Anderson and J.D. deBethizy, A physi-ologically based pharmacokinetic model for nicotine disposi-tion in the Sprague–Dawley rat. Toxicol. Appl. Pharmacol.,116, 177–188 (1992).

.

© 1998 John Wiley & Sons, Ltd. Biopharm. Drug Dispos. 19: 147–151 (1998)

.