ORIGINAL PAPER

The use of a prodrug approach to minimize potential CNSexposure of next generation quinoline methanolswhile maintaining efficacy in in vivo animal models

Jason C. Sousa • Erin Milner • Dustin Carroll • William McCalmont • Sean Gardner • Jay Moon •

Jacob D. Johnson • Patricia Lee • Jennifer Auschwitz • Norma Roncal • Diana Caridha • Anchalee Tungteung •

Qiang Zeng • Sean Reyes • Bryan Smith • Qigui Li • Michael P. Kozar • Victor Melendez • Geoffrey Dow

Received: 31 January 2013 / Accepted: 4 December 2013

� Springer-Verlag France (outside the USA) 2013

Abstract The use of mefloquine (MQ) for antimalarial

treatment and prophylaxis has diminished largely in

response to concerns about its neurologic side effects. An

analog campaign designed to maintain the efficacy of MQ

while minimizing blood–brain barrier (BBB) penetration

has resulted in the synthesis of a prodrug with comparable-

to-superior in vivo efficacy versus mefloquine in a P.

berghei mouse model while exhibiting a sixfold reduction

in CNS drug levels. The prodrug, WR319670, performed

poorly compared to MQ in in vitro efficacy assays, but had

promising in vitro permeability in an MDCK–MDR1 cell

line BBB permeability screen. Its metabolite, WR308245,

exhibited high predicted BBB penetration with excellent

in vitro efficacy. Both WR319670 and WR308245 cured

5/5 animals in separate in vivo efficacy studies. The in vivo

efficacy of WR319670 was thought to be due to the for-

mation of a more active metabolite, specifically

WR308245. This was supported by pharmacokinetics

studies in non-infected mice, which showed that both IV

and oral administration of WR319670 produced essentially

identical levels of WR319670 and WR308245 in both

plasma and brain samples at all time points. In these

studies, the levels of WR308245 in the brain were 1/4 and

1/6 that of MQ in similar IV and oral studies, respectively.

These data show that the use of WR319670 as an anti-

malarial prodrug was able to maintain efficacy in in vivo

efficacy screens, while significantly lowering overall pen-

etration of drug and metabolites across the BBB.

Keywords Malaria � Mefloquine � Quinoline methanol �Blood–brain barrier � Central nervous system � Prodrug

1 Introduction

According to current Centers for Disease Control and

Prevention guidelines (Arguin 2012), mefloquine (MQ)

(Fig. 1a) is arguably the most widely applicable drug for

malaria chemoprophylaxis for non-immune individuals

traveling to endemic countries. Like chloroquine (CQ),

MQ can be administered on a weekly basis, a characteristic

that enhances the likelihood of compliance and is more

forgiving in cases of late or missed dosages. Further, both

drugs are deemed safe for children of all ages/sizes and for

This manuscript was reviewed by the Walter Reed Army Institute of

Research and the US Army Medical Research and Materiel Command,

and there is no objection to its publication or dissemination. The

opinions expressed herein are those of the authors and do not

necessarily reflect the views or opinions of the Department of the Army

and the Department of Defense. All animal experiments were

conducted in compliance with the Animal Welfare Act and other

federal statutes and regulations relating to animals and experiments

involving animals and adhere to the principles stated in the Guide for

the Care and Use of Laboratory Animals (National Academy Press,

1996).

J. C. Sousa (&) � E. Milner � D. Carroll � W. McCalmont �S. Gardner � J. Moon � J. D. Johnson � P. Lee � J. Auschwitz �N. Roncal � D. Caridha � Q. Zeng � S. Reyes � B. Smith �Q. Li � M. P. Kozar � V. Melendez � G. Dow

Walter Reed Army Institute of Research, Silver Spring,

MD 20910, USA

e-mail: Jason.sousa@amedd.army.mil

G. Dow

e-mail: geoffrey.dow@us.army.mil

A. Tungteung

United States Army Medical Component, Armed Forces

Research Institute of Medical Sciences, Bangkok, Thailand

G. Dow

United States Army Medical Materiel Development Activity,

1430 Veterans Drive, Frederick, MD 21702, USA

123

Eur J Drug Metab Pharmacokinet

DOI 10.1007/s13318-013-0162-9

women who are pregnant or breastfeeding. Unfortunately,

the widespread development of CQ-resistant strains of

malaria parasites precludes its use in all but a few select

malaria-endemic regions.

Despite its favorable properties, MQ also has drawbacks

that prohibit more universal application for malaria chemo-

prophylaxis and treatment. First, while much less prevalent

than CQ resistance, MQ-resistant strains of malaria parasites

have begun to appear in Southeast Asia. Perhaps more sig-

nificant are MQ’s reported adverse reactions. In addition to

gastrointestinal upset, MQ has been implicated in more

serious neuropathic side effects, including dizziness, vivid

dreams, hallucinations, paranoia, and schizophrenia (CDC

2009; Yelmo et al. 2010). While more common at larger

treatment doses, these effects have been noted at lower

prophylactic doses as well. Further, these symptoms have

been known to persist well beyond the completion of the

normal dosing regimen. As a result, MQ is not recommended

for use in patients with active symptoms of, or a history of,

depression or other psychiatric disorders (Yelmo et al. 2010).

Among the recent drug development programs at the

Walter Reed Army Institute of Research (WRAIR) is an

effort to produce an analog of MQ that is less capable of

penetrating the blood–brain barrier (BBB), while main-

taining levels of efficacy that are equal to or greater than

MQ. Because the precise target responsible for MQ’s

neurological effects is not known (Toovey 2009), the pre-

vailing strategy in this effort has been to prevent overall

drug levels in the brain. To that end, a library of MQ

analogs has been synthesized to explore the relationship

between a range of physiochemical properties and efficacy/

BBB permeability (Milner et al. 2010a, b).

NH

OH

N

F

FF

FF

F

OH

N

F

FF

FF

F

NHH3C

(A) (B)

OH

N

F

FF

FF

F

NH3C NH2

(C)

Fig. 1 a Mefloquine

(WR142490); b WR308245;

and c WR319670

Eur J Drug Metab Pharmacokinet

123

One of the compounds from this library, WR308245

(Fig. 1b), initially generated substantial interest due to its

promising results in in vitro efficacy and cytotoxicity

assays. However, its apparent permeability (Papp) in

in vitro studies using the MDCK cell line transfected with

human MDR1 suggested high BBB penetration. In vivo

mouse PK studies confirmed this finding with drug levels in

the brain approximately six times that of MQ at the Cmax

after IV dosing.

Another drug in this class of analogs, WR319670

(Fig. 1c), was found to have in vivo activity against P.

berghei in a mouse model of antimalarial efficacy while

exhibiting poor activity levels in [3H] hypoxanthine

(Chulay et al. 1983; Milhous et al. 1985) and SYBR Green

(Smilkstein et al. 2004; Johnson et al. 2007) in vitro effi-

cacy screens. These data suggested that the activity of

WR319670 in the in vivo model could be due to drug

metabolism, i.e., that WR319670 acted as a prodrug. Based

on the structure of WR319670 and common Phase I bio-

transformations, it was postulated that one of the metabo-

lites would be WR308245. In vivo mouse PK confirmed

the formation of WR308245 upon IV administration of

WR319670. In addition, drug brain levels for both

WR319670 and WR308245 achieved a Cmax of approxi-

mately 1/3 that of MQ.

While the use of prodrugs in the treatment of malaria is

not new (Chung et al. 2008), we believe that this is the first

reported use of the prodrug approach to minimize overall

penetration of the BBB and thereby reduce complications

due to adverse CNS effects.

2 Materials and methods

2.1 MDCK–MDR1 permeability

All BBB permeability screens using the MDCK cell line

transfected with human MDR1 were performed under

contract by absorption systems (Exton, PA).

2.2 In vitro activity

In vitro activities against several strains of Plasmodium

falciparum were evaluated using both the [3H] hypoxan-

thine method first described by Desjardins et al. (1979) and

later modified by Milhous et al. (1985), and an SYBR

Green I fluorescence assay introduced by Smilkstein et al.

(2004) and later modified by Johnson et al. (2007). In brief,

the [3H] hypoxanthine assay measures the amount of radio-

labeled hypoxanthine incorporated by P. falciparum para-

sites cultured in erythrocytes in response to exposure to a

range of drug concentrations. Plotting radioactivity (as

counts per minute) of the incorporated isotope versus drug

concentration can be used to determine the IC90 for the test

compound. The SYBR Green I assay involves the culturing

of P. falciparum parasites in erythrocytes while exposed to

a range of drug concentrations, followed by cell lysis and

exposure to the nucleic acid-intercalating dye SYBR Green

I. Fluorescence is plotted versus drug concentration to

determine the IC90 of the test compound.

These analyses were repeated using several strains of P.

falciparum, including W2 (CQ resistant, MQ sensitive), D6

(CQ sensitive, less susceptible to MQ), and C235 (resistant

to CQ and MQ).

2.3 In vitro cytotoxicity

In vitro cytotoxicity assays were performed according to

the method described by Caridha et al. (2008). In brief, the

assay measures a compound’s ability to disrupt calcium

homeostasis in rodent macrophage (RAW-264.7) cells

using a 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazo-

lium bromide (MTT) indicator of cell viability. Plots of

absorbance versus concentration of test compound are used

to determine LC50.

2.4 In vivo efficacy

In vivo efficacy against malaria parasites was evaluated

using a modified Thompson test (Dow et al. 2006),

whereby male ICR mice were infected on Day 0 with P.

berghei (KBG-173 strain) and dosed once orally with test

compound, on Day 3. Five mice were used in each dosing

group. Animals were monitored daily through Day 31 for

clinical signs of infection and periodic malaria smears were

taken. Efficacy was expressed as mean survival time where

cures were defined as survival at 31 days post-treatment.

Typically untreated control mice are usually euthanized

humanely on days 6–10 once parasitemia levels become

too high or symptoms of malaria infection are observed.

Dosages were either 160 or 320 mg/kg and dosing vehicle

was either HECT (0.5 % hydroxyethylcellulose in 0.2 %

Tween 80) or 5 % ethanol/5 % Cremophor EL/90 %

HECT.

2.5 Pharmacokinetic studies

Pharmacokinetic studies were performed using both IV and

PO administration. For each time point to be acquired, four

male FVB mice, aged 5–6 weeks, were dosed at 5 mg/kg

(IV) or 160 mg/kg (PO). Drug vehicle for IV studies was

0.5/5/5 % v/v Cremophor EL/dimethylsulfoxide/glucose

solution in 20 mM citrate buffer (pH 3.0), administered at

50 lL/20 g. For PO dosing, drug vehicle was 5 % ethanol/

5 % Cremophor EL/90 % HECT, administered at 150 lL/

20 g.

Eur J Drug Metab Pharmacokinet

123

At each time point, plasma, whole blood, brain, and

liver samples were collected. Whole blood was collected

by cardiac puncture. Following the separation of appro-

priate aliquots, plasma was obtained from the remaining

whole blood via centrifugation. All liquid and tissue sam-

ples were immediately preserved at -80� C and stored

until analytical work was performed.

2.6 Pharmacokinetics

2.6.1 Extraction

Liquid samples were extracted by adding two volumes of

acetonitrile to one volume of blood/plasma, followed by

vigorous vortexing for 10–15 s. Tissue samples were

weighed and 5 mL of deionized water were added for each

gram of tissue, followed by homogenization using a sonic

dismembrator. From this homogenate, aliquots were

extracted in a manner identical to the liquid samples. All

samples were centrifuged at 13,000 rpm for 10 min

immediately after extraction, and the supernatant liquid

was removed for analysis by LC–MS/MS.

2.6.2 LC–MS/MS

Chromatography was performed using a surveyor pump

(Thermo Scientific, Waltham, MA) with Waters XTerra

MS C18 50 mm 9 2.1 mm id, 3.5 lm particle size col-

umns (Waters Corp., Milford, MA). Mobile phase con-

sisted of a water/0.1 % formic acid (Solvent A)/

acetonitrile/0.1 % formic acid (Solvent B) gradient. The

gradient began at 2 % B, rose to 98 % B from 1 to 3.5 min,

held steady for 2 min, then returned immediately to its

starting composition and allowed to equilibrate for

1.5 min. Flow rate was 300 lL/min. Samples were injected

using an HTC PAL autosampler (LEAP Technologies,

Carrboro, NC). Tandem mass spectrometry was performed

using a TSQ Quantum AM (Thermo Scientific). Monitored

transitions were: WR319670 (380.11–321.08, ESI?);

WR308245 (339.08–270.03, ESI?); and mefloquine

(379.02–361.42, ESI?).

3 Results

In vitro permeability with MDR1-transfected MDCK cells

was used to rank order potential MQ analogs in terms of

decreasing likelihood to penetrate the BBB. Table 1 sum-

marizes the permeability, in vitro activity, in vitro cyto-

toxicity and in vivo efficacy results for WR308245 and

WR319670 along with that of MQ, used as a baseline

comparison for rank ordering purposes. While the in vitro

permeability data suggested greater BBB penetration than

MQ, WR308245 generated sufficient interest due to its

comparable in vitro efficacy coupled with its superior

in vitro toxicity results. WR319670, while suggesting a

promising decrease in BBB permeability and a comparable

cytotoxicity profile, did not produce impressive in vitro

efficacy results. However, in the in vivo efficacy model

using P. berghei, WR319670 cured 5/5 mice after a single

oral dose, compared to 4/5 for MQ and 5/5 for WR308245

after a 3-dose regimen.

Figure 2 shows the brain (solid line) and plasma (dashed

line) drug levels for WR308245 and MQ in an IV PK

studies for both drugs using non-infected mice. As sug-

gested by the in vitro screen, WR308245 has much higher

BBB permeability than MQ, peaking at about six times the

concentration at Cmax. In Fig. 3, a comparable PK study

with IV dosing of WR319670 confirms the in vitro pre-

diction of lower BBB penetration (about 1/4 that of MQ in

Fig. 2).

In an effort to explain the discrepancy between

WR319670’s in vitro and in vivo efficacy, the PK samples

from WR319670 IV dosing were also monitored for the

formation of WR308245, which is a potential metabolite of

WR319670 based on common Phase I biotransformations.

These data are also included in Fig. 3. Throughout the

course of the study, WR308245 and WR319670 were

present in similar concentrations in both plasma and brain.

Figure 4 summarizes the total brain concentration of

MQ, WR319670, and WR308245 after IV dosing of 5

mg/kg, as well as the appearance of WR308245 in the brain

after the dosing of WR319670.

Table 1 Summary of results for MQ, WR308245, and WR319670

from in vitro MDCK–MDR1 permeability, in vitro efficacy by both

the [3H] hypoxanthine (HX) and SYBR Green I (SG) methods,

in vitro toxicity (MTT), and in vivo efficacy against P. berghei in a

murine model of malaria infection

MQ WR308245 WR319670

MDCK–MDR1 (910-6

cm/s)

9.3 24 5.0

HX W2 (ng/mL) 6.2 ± 2.8

(532)

17 180

HX D6 (ng/mL) 17 ± 11

(536)

45 290

HX C235 (ng/mL) 52 ± 30

(367)

58 710

SG D6 (ng/mL) 12 ± 2 (6) NT 340

SG C235 (ng/mL) 43 ± 12 (6) NT 700

LC50 MTT (ng/mL) 4,000 31,000 3,800

Cures at 160 mg/kg 91

PO

4/5 5/5 5/5

Numbers in parentheses indicate the number of runs used in deter-

mining the reported values

NT not tested

Eur J Drug Metab Pharmacokinet

123

Figure 5 is an overlay of brain and plasma drug levels

for oral PK studies for MQ and WR319670. The formation

of WR308245 following dosing of WR319670 is also

shown. As seen in the IV studies, the levels of WR319670

and WR308245 correlate throughout the lifetime of the

study, with both being significantly lower than MQ in the

brain. This effect is even more significant than the IV

study, as levels of each compound are about 1/6 that of MQ

at Cmax.

1

10

100

1000

0 4 8 12 16 20 24

Con

cent

rati

on (

ng/g

Bra

in, n

g/m

L P

lasm

a)

Time (Hours)

WR319670 Plasma

WR319670 Brain

WR308245 Plasma

WR308245 Brain

Fig. 3 Brain and plasma levels of WR319670 and the appearance of

WR308245 in both matrices after 5 mg/kg IV dosing of WR319670

10.0

100.0

1000.0

10000.0

0 4 8 12 16 20 24

Con

cent

rati

on (

ng/g

)

Time (Hours)

WR308245

Mefloquine

WR308245-M

WR319670

Fig. 4 Brain levels of MQ, WR319670 and WR308245 after IV

dosing of each compound, as well as the appearance of WR308245 in

the brain after WR319670 dosing. At their respective Cmax values,

both WR319670 and WR308245 are less than a third of the Cmax of

MQ. Combined, they are about one-half that of mefloquine at Cmax

1

10

100

1000

10000

100000

0 24 48 72 96 120 144 168

Con

cent

rati

on (

ng/g

Bra

in, n

g/m

L

Pla

sma)

Time (Hours)

Mefloquine Plasma

Mefloquine Brain

WR319670 Plasma

WR319670 Brain

WR308245 Plasma

WR308245 Brain

Fig. 5 Brain and plasma PK profile of mefloquine and WR319670

and the formation of WR308245 in both matrices after 160 mg/kg PO

dosing. WR319670 and WR308245 each have Cmax values about one-

sixth of that of mefloquine at its Cmax. Combined, they are about one-

third that of mefloquine at Cmax

10.0

100.0

1000.0

10000.0

0 4 8 12 16 20 24

Con

cent

rati

on (

ng/g

Bra

in, n

g/m

L

Pla

sma)

Time (Hours)

WR308245 Plasma

Mefloquine Plasma

WR308245 Brain

Mefloquine Brain

Fig. 2 Brain (solid line) and plasma (dashed line) levels of

WR308245 and MQ after 5 mg/kg IV dosing of each compound.

Although WR308245 has lower plasma levels, at their respective

Cmax values, WR308245 has brain levels six times that of MQ

Eur J Drug Metab Pharmacokinet

123

4 Discussion

The library of compounds from which this data is derived

was designed with the goal of reducing BBB penetration

compared to MQ through the manipulation of a variety of

physicochemical properties (Milner et al. 2010a, b). It was

an unexpected observation in the combined in vivo and

in vitro efficacy data that exposed the prodrug nature of

WR319670. While prodrugs are commonly used to

improve bioavailability, reduce toxicity by directing

metabolism and activation near the target, or to enhance

penetration and transport across biological membranes, the

use of this approach to maintain therapeutic activity while

minimizing drug levels in the CNS is unusual.

In a PK comparison against MQ with IV dosing,

WR308245 exhibited brain concentrations nearly six times

that of MQ at Cmax. This is consistent with the observed

results in the in vitro MDCK–MDR1 assay, which sug-

gested permeability across the BBB 2–3 times that of MQ.

When MQ and WR319670 were each dosed orally, both

WR319670 and its metabolite WR308245 were present in

the brain at about 1/3 the level of MQ at Cmax. Because

WR319670 has little intrinsic activity against the malaria

parasite, as demonstrated in the [3H] hypoxanthine and

SYBR Green I in vitro assays, the drug’s activity in an

in vivo mouse model can most likely be attributed to the

formation of active metabolites, such as WR308245. Fur-

ther, a net 18-fold change in brain levels of the active

WR308245 metabolite was not associated with a demon-

strable change in efficacy, providing vindication of this

prodrug approach. Since the levels of WR308245 were

essentially equal to those of WR319670 after dosing, it is

safe to assume that it is the major metabolite contributing

to antimalarial activity. It is not contended here that the

biotransformation of WR319670 to WR308245 in any way

impacts the mechanism of BBB penetration of the latter

compound. Rather, it is simply shown that the use of

WR319670 as a prodrug maintains antimalarial efficacy

while significantly diminishing the overall drug levels in

the CNS.

Because the exact target causing adverse CNS side

effects with MQ is not known, the next generation quino-

line methanol project team has set as a goal a fivefold

decrease in overall brain concentration for any potential

lead candidates. As WR319670 does not provide an ade-

quate window under this criterion, it is not a likely can-

didate for further development and a fuller characterization

of its metabolic profile has not been deemed a priority.

However, the discovery of this prodrug strategy as a means

of maintaining efficacy in in vivo models while minimizing

CNS drug levels may prove to be an important tool in the

future design of quinoline methanol analogs.

References

Arguin PM, Mali S (2012) Malaria. CDC Yellow Book. http://

wwwnc.cdc.gov/travel/yellowbook/2012/chapter-3-infectious-

diseases-related-to-travel/malaria.htm. Accessed 12 Dec 2012

Caridha D, Yourick D, Cabezas M, Wolf L, Hudson TH, Dow GS

(2008) Mefloquine-induced disruption of calcium homeostasis in

mammalian cells is similar to that induced by ionomycin.

Antimicrob Agents Chemother 52:684–693

Chulay JD, Haynes JD, Diggs CL (1983) Plasmodium falciparum—

assessment of in vitro growth by [H-3]-labeled hypoxanthine

incorporation. Exp Parasitol 55:138–146

Chung MC, Ferreira EI et al (2008) Prodrugs for the treatment of

neglected diseases. Molecules 13:616–677

Desjardins RE, Canfield CJ, Haynes JD, Chulay JD (1979) Quanti-

tative assessment of antimalarial activity in vitro by a semiau-

tomated microdilution technique. Antimicrob Agents Chemother

16:710–718

Dow GS, Heady TN et al (2006) Utility of alkylaminoquinolinyl

methanols as new antimalarial drugs. Antimicrob Agents Che-

mother 50:4132–4143

Dow GS, Magill AJ, Ohrt C (2008) Clinical development of new

prophylactic antimalarial drugs after the 5th Amendment to the

Declaration of Helsinki. Ther Clin Risk Manag 4:803–819

Johnson JD, Dennull RA, Gerena L, Lopez-Sanchez M, Roncal NE,

Waters NC (2007) Assessment and continued validation of the

malaria SYBR green I-based fluorescence assay for use in

malaria drug screening. Antimicrob Agents Chemother 51:

1926–1933

Milhous WK, Weatherly NF, Bowdre JH, Desjardins RE (1985)

In vitro activities of and mechanisms of resistance to antifol

antimalarial drugs. Antimicrob Agents Chemother 27:525–530

Milner E, McCalmont W et al (2010a) Structure-activity relationships

amongst 4-position quinoline methanol antimalarials that inhibit

the growth of drug sensitive and resistant strains of Plasmodium

falciparum. Bioorg Med Chem Lett 20:1347–1351

Milner E, McCalmont W et al (2010b) Anti-malarial activity of a non-

piperidine library of next-generation quinoline methanols. Malar

J 9:51–60

Smilkstein M, Sriwilaijaroen N, Kelly JX, Wilairat P, Riscoe M

(2004) Simple and inexpensive fluorescence-based technique for

high-throughput antimalarial drug screening. Antimicrob Agents

Chemother 48:1803–1806

Toovey S (2009) Mefloquine neurotoxicity: a literature review. Travel

Med Infect Dis 7:2–6

Yelmo S, Morera-Fumero AL, Henry M, Renshaw A, Gracia-Marco

R (2010) Mania Associated With Mefloquine Prophylaxis. J Clin

Psychopharmacol 30:339–341

Eur J Drug Metab Pharmacokinet

123