Quantitative Assessment of Antimalarial Activity in Vitro by a Semiautomated Microdilution Technique ROBERT E

ANTIMICROBIAL AGENTS AND CHEmOTHEzRAY, Dec. 1979, p. 710-718 Vol. 16, No.6 0066-4804/79/12-0710/09$2.00/0

Quantitative Assessment of Antimalarial Activity In Vitro by a Semiautomated Microdilution Technique ROBERT E. DESJARDINS,t* CRAIG J. CANFIELD,' J. DAVID HAYNES,2 AND JEFFREY D. CHULAY2 Department ofPharmacology' and Departnent ofImmunology,2 Walter Reed Army Institute ofResearch, Washington, DC 20012 Received for publication 10 September 1979

A rapid, semiautomated microdilution method was developed for measuring the activity of potential antimalarial drugs against cultured intraerythrocytic asexual forms of the human malaria parasite Plasmodium falciparum. Microti- tration plates were used to prepare serial dilutions of the compounds to be tested. Parasites, obtained from continuous stock cultures, were subcultured in these plates for 42 h. Inhibition of uptake of a radiolabeled nucleic acid precursor by the parasites served as the indicator of antimalarial activity. Results of repeated measurements of activity with chloroquine, quinine, and the investigational new drug mefloquine demonstrated that the method is sensitive and precise. Several additional antimalarial drugs and compounds of interest were tested in vitro, and the results were consistent with available in vivo data. The use of P. falciparum isolates with known susceptibility to antimalarial drugs also permitted evaluation of the cross-resistance potential of each compound tested. The applications and expectations of this new test system within a drug development program are discussed.

The recent development of methods for the tem is partially automated for rapid analysis of continuous in vitro culture ofthe human malaria the data. parasite Plasmodium falciparum (4, 21) provides a valuable resource for the investigation of MATERIALS AND METHODS new approaches to malaria chemotherapy. Di- Preparation of parasites. The parasite inocula rect measurement of the antimalarial activity of used in these experiments consisted of two isolates of potential new drugs against parasites grown in P. fakciparum. The African Uganda I strain was orig- vitro is now feasible. inally obtained by Martin and Arnold from a child in Kampala, Uganda, in 1965 (9). It has been widely used Resistance of P. fakciparum to chloroquine is in phase II clinical studies in human volunteers and an important consideration in the treatment and has retained its susceptibility to chloroquine, quinine, prevention ofmalaria in many parts ofthe world, and pyrimethamine. The Vietnam Smith strain, ob- including South and Central America (12), tained in 1969 from an American soldier at the Walter Southeast Asia (25), India (20), and the African Reed Army Medical Center, was first reported by continent (6). It is therefore essential to use well- Craig J. Canfield in 1971 (2). That patient, showing an characterized susceptible and resistant strains of R-I level of resistance to the combination of sulfalene the parasite in the evaluation of potential new and trimethoprim after treatment failures with other chemotherapeutic agents to detect cross-resist- antimalarial drugs, provided the parasite isolate des- standard anti- ignated Smith strain (2). This strain was later adapted ance with chloroquine and other to owl monkeys (24) and has also been used extensively malarial drugs. in human volunteer studies, where its resistance to The method described in this report can pro- chloroquine, quinine, and pyrimethamine has been vide quantitative measurements of the antima- amply documented (10). larial activity of large numbers of compounds, The two strains of parasites were grown continu- based on the inhibition of uptake of a radiola- ously in stock cultures by a modification of the meth- beled nucleic acid precursor by the parasite dur- ods of Trager and Jensen (21) and Haynes et al. (4). A ing short-term cultures in microtitration plates. 6% suspension of human type A+ erythrocytes was The parasites used are continually available prepared in culture medium which consisted of pow- The dered RMPI 1640 (GIBCO Laboratories, Grand Island, from long-tenn maintenance cultures. sys- N.Y.) diluted in sterile water with 25 mM HEPES (N- 2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; t Present address: Wellcome Research Laboratories, Re- Calbiochem, La Jolla, Calif.), 32 mM NaHCO3 search Triangle Park, NC 27709. (GIBCO), and 10% heat-inactivated (40 min at 56°C 710 VOL. 16, 1979 ASSESSMENT OF ANTIMALARIAL ACTIVITY IN VITRO 711 and then clarified by centrifugation) human type A+ i- fresh frozen plasma (in acid-citrate-dextrose antico- agulant). Stock cultures were maintained in 5.0 ml of the 6% erythrocyte suspension in 25-ml tissue culture CQOQO"O0KP""o flasks (Corning Glass Works, Corning, N.Y.). The o ~ 'OK ,>OqBQ1Q flasks were flushed with a gas mixture consisting of 5% 02, 5% C02, and 90% N2 (Air Products Corp., Allen- town, Pa.), sealed, and incubated at 370C. Best results were obtained in individual experiments when the F 88 growth rate in the stock cultures was high as indicated by a doubling of the parasitemia every 24 h. This was accomplished by daily changes of the culture medium and by dilution with fresh erythrocytes every 2 or 3 H QQQQQQQQQQQQ days so that less than 2% of the cells were infected at L any time. For each experiment, samples of the stock Li cultures were further diluted in culture medium con- FIG. 1. Representation of a microtiter plate, with taining sufficient noninfected type A+ human eryth- 96 wells arranged in 8 rows (A-H) and 12 columns rocytes to yield a final hematocrit of 1.5% and parasit- (1-12). When prepared as described in the text, wells emia of 0.25 to 0.5% in preparation for addition to the 1 through 8 of row A serve as a parasite control (no microtitration plates. drug present), and wells 9 through 12 of row A serve Preparation of drugs. Preparation of compounds as an erythrocyte control (no drug and no parasites). in a form suitable for addition to the microtitration Each compound ispresent in duplicate columns, over plate at a known concentration required consideration a 64-fold range, with the highest concentrations in of the aqueous solubility and sterility of each agent. row B and in twofold dilutions to the lowest concen- Many compounds were relatively insoluble in water trations in row H. and had to be dissolved initially in an organic solvent. These problems were overcome by using a combination of sterile water and ethanol as the solvent. Each A constant volume (200 id) of the parasitized eryth- compound was first dissolved in a measured volume of rocyte suspension described above was added to each either water or ethanol, which was then diluted with well of the microtiter plate except the last four wells its counterpart to yield a 70% ethanol-30% water mix- of row A, to each of which 200 A1 of an equivalent ture containing a known concentration of the drug. suspension of nonparasitized type A+ human eryth- This solution was allowed to stand at room tempera- rocytes in culture medium was added. The total vol- ture for 30 min. Further dilutions were then made by ume in every well was then 225 pl. The first eight wells addition of the compound in the 70% ethanol solution of row A, containing no drugs, served as a parasite to a measured volume of culture medium with con- control. The last four wells of row A, containing nei- stant mixing. Plasma proteins in the culture medium ther drugs nor parasites, served as a nonparasitized presumably provided binding sites for compounds erythrocyte controL The parasitized erythrocytes were which might otherwise have precipitated in an exposed to each compound in two columns, with the aqueous solution. The final dilution contained less highest concentration in row B and the lowest concen- than 0.1% ethanol, which had no measurable effect on tration in row H. Two plates were prepared for each the parasites in this system. experiment; the Uganda I strain of the parasite was Preparation of microtitration plates. Microti- added to one, and the Smith strain was added to the tration techniques were used to measure the activity other. of a large number of compounds efficiently. The mi- After preparation as above, the plates were placed crotiter plate (Cooke Laboratory Products, Alexan- in a humidified airtight box (Instrumentation Depart- dria, Va.) used consisted of 96 flat-bottom wells, ar- ment, Walter Reed Army Institute ofResearch, Wash- ranged in a matrix of eight rows (A through H) and 12 ington, D.C.), which was then flushed with a gas columns (1 through 12) (Fig. 1). The plates were mixture of 5% 02,5% C02, and 90% N2 and sealed. The prepared and the parasites were harvested by using box was then placed in an incubator at 37°C for 24 h. strict aseptic techniques inside a laminar flow hood in Preparation of isotope and labeling of para- the following sequence. An Eppendorf pipette or a sites. Uptake of [G-3H]hypoxanthine (Amersham/ Microdrop I (Cooke Laboratory Products) multiple- Searle Corp., Arlington Heights, Ill.) was used as an well filling pipette was used to place 25 id ofthe culture index of growth of the parasites. The isotope is sup- medium in each well of the microtiter plate. Twenty- plied as a lyophylate (1,000 mCi/mmol) in ampoules five microliters of the drug solution, prepared as de- containing 1.0 mCi. The contents of a single ampoule scribed above, was then added to each of two adjacent were dissolved in 2.0 ml of 50% ethanol to provide a wells in the second row (B) ofthe plate. Six compounds stock solution which was stored at -20°C. In prepa- were thereby accommodated by each plate. After the ration for addition to the microtiter plates, the ethanol drugs were added to the wells of row B, an Automatic was evaporated from a 0.2-ml sample of the stock Diluter (Cooke Laboratory Products) was used to solution, and 4.9 ml of culture medium was added to make serial twofold dilutions across the plate in each the remaining 0.1 ml of the isotope in water. The final column. When this was complete, row A remained free isotope solution consisted of 20 ,uCi of [G-3H]-hypo- of any drug, and each of the drugs was present in xanthine per ml of culture medium. duplicate columns at seven concentrations over a 64- After the 24-h incubation period described above, fold range in rows B through H. the plates were removed from the box and 25 pl of the 712 DESJARDINS ET AL. ANTIMICROB. AGENTS CHEMOTHER. isotope in culture medium (0.5 ,uCi) was added to each function was used because it provided convenient pa- well. The plates were then returned to the box, which rameters for application to the in vitro test system (3). was again flushed with the 5% 02-5% C02-90% N2 gas Estimates of each parameter were obtained by an mixture, sealed, and incubated at 37°C for an addi- iterative process which also provided a variance-co- tional 18 h. variance matrix from which the 95% confidence limits Harvesting parasites and scintillation count- of the ID50 -were calculated. The computer output ing. At the end of the second incubation period, each included a graph of the data with the fitted curve and plate was harvested on a MASH II automated cell accompanying estimate of the ID50 (see Fig. 2-4). In harvester (Microbiological Associates, Bethesda, Md.). addition, a variety of statistical tests were performed This instrument aspirated and deposited the particu- to determine the goodness of fit of each set of data to late contents of each of the wells onto small disks of the model equation. These included calculations ofthe filter paper (no. 934-AH, Whatman, Inc., Clifton, N.J.), R-square and F -ratio for the regression and a nonpar- which were then washed with copious volumes of ametric sign test on the residual errors. distilled water. Each disk was dried and placed in a glass scintillation vial containing 10 ml of a toluene- based scintifluor for counting. All 96 vials, correspond- RESULTS ing to the 96 wells ofthe microtiter plate, were counted Both the Uganda I and Smith strains of P. in a Searle model Delta 300 liquid scintillation spec- trometer for a sufficient period of time to ensure a falciparum were readily adapted to culture con- counting error of less than 5% for each sample. ditions. They were maintained continuously for Data analysis. The counts and external standard- 3 months while these experiments were con- ization ratio for each vial were recorded on a paper ducted and both were successfully preserved in punch tape for input to a desk top computer, the liquid nitrogen and returned to culture ad libi- Tekronix 4051 Graphic System. The counts were con- tum. verted by standard equations to disintegrations per The uptake of [G-3H]hypoxanthine by paras- minute for each well, which were tabulated in an 8-by- itized erythrocytes in the microtiter plates was 12 matrix corresponding to the 8 rows and 12 columns consistent. The mean values for the parasite of the plate. The mean values for parasite control control wells at the end of the 18-h were uptake and nonparasitized erythrocyte control uptake pulse of [G-3H]hypoxanthine were calculated from the dis- generally between 40,000 and 60,000 dpm and integrations per minute in row A, wells 1 through 8, had a coefficient of variation of less than 5%. and wells 9 through 12, respectively. The compound Nonparasitized control erythrocytes contained identifications and concentrations were entered at the less than 2% of the amount of isotope present in keyboard of the computer. All concentrations were parasitized control cells. expressed as nanograms of the salt form per milliliter. Representative graphs, illustrating the re- The diphosphate salt ofchloroquine was used in every sponse of Uganda I and Smith strain parasites plate. The exact salt forms of most of the compounds to chloroquine in one experiment, are shown in were not identified at the time they were tested, so no Fig. 2. An excellent fit of the data to the regres- attempt was made to calculate base weight or molar concentration. Unknown compounds were routinely sion equation was obtained in each case. The diluted to provide an initial concentration of 125 ng/ ID50 for chloroquine against the Uganda I strain ml and reevaluated at higher or lower concentrations of P. falciparum in this experiment was 10.3 ng/ when necessary. ml with a 95% confidence interval (CI) of 9.9 to The automated system used in these experiments 10.8 ng/ml. The resistance of the Smith strain generated twofold serial dilutions with a 64-fold range to chloroquine was illustrated in this experiment of concentrations for each compound. The concentra- by an ID50 of 156 ng/ml (95% CI of 138 to 193 tion-response curves for active compounds over this ng/ml). The asymmetry of the confidence inter- range were characteristically sigmoidal after logarith- vals is related to the mic transformation of the concentration and were logarithmric transformation interpreted by nonlinear regression analysis. Concen- of the concentration in the model equation. tration-response data for each compound were fit to a The effects of quinine on both strains of the generalized sigmoidal function by the Marquardt al- parasite in this same experiment are illustrated gorithm (1), using'a nonweighted least-squares crite- in Fig. 3. As anticipated, considerable cross-re- rion. The function used was a hyperbolic tangent sistance to chloroquine was exhibited by quinine, function with four parameters: Y, = (U - L)/2 [ 1 + with an ID50 in this experiment of 20.9 ng/ml tanh (,B log C -, log Xi)] + L, where Yi is the (95% CI of 10.5 to 38.5 ng/ml) for the Uganda I disintegrations per minute of the oth sample, X, is the strain compared with 135 ng/ml (95% CI of 111 concentration of compound corresponding to the o1h to 162 ng/ml) for the Smith strain. sample, U is the upper asymptote of the function (approximated by the parasite control mean), L is the Mefloquine (WR 142,490), an investigational lower asymptote of the function (approximated by the new antimalaria drug (22), was not cross-resist- nonparasitized erythrocyte control mean), ,B is a scal- ant with chloroquine or quinine in vitro (Fig. 4). ing parameter, and C is the concentration of agent The slope of response to mefloquine was smilar corresponding to 50% inhibition of the uptake of [G- to that seen with chloroquine. In this experi- 3Hlhypoxanthine (the ID50). This form of the logistic ment, the ID50 for mefloquine was 6.39 ng/ml VOL. 16, 1979 ASSESSMENT OF ANTIMALARIAL ACTIVITY IN VITRO 713

CHLOROQUINE ve UGANDA I

D p M

IBMss

8 I Is too 88 ID-50 - 11.2SCng/ml) CONCENTRATION Cng/mi)

C1HLOROQUINE ve SMITH

50000

40m D p M

18888

ID-SO - 156 l8Cng/m) CONCENTRATION (ng/ml) FIG. 2. Comparative activity of chloroquine diphosphate against the African Uganda I strain and the Vietnam Smith strain ofP. falciparum in vitro. The range of concentration of the drug was 1.95 to 125 ng/ml for the Uganda I strain and 19.5 to 1,250 ng/ml for the Smith strain. The disintegrations per minute represent the uptake of[G-3HJhypoxanthine by theparasites at each concentration ofthe drug. The ID50 was 10.3 ng/ml (95% CI = 9.9 to 10.8 ng/ml) for the Uganda I strain and 156 ng/ml (95% CI = 138 to 193 ng/ml) for the Smith strain. The graphs shown are part of the computer output described in the text.

(95% CI of 5.82 to 6.91 ng/ml) for the Uganda I ng/ml). Considerable cross-resistance to quinine strain and 6.66 ng/ml (95% CI of 6.43 to 7.19 ng/ was indicated by the mean ID50 values of 109 ml) for the Smith strain. ng/ml for the Smith strain and 26.1 ng/ml for Multiple determinations of the ID50 for each the Uganda I strain. The lack of cross-resistance of these three antimalarial drugs demonstrated to mefloquine was also apparent, with a mean the precision of the test system. The results of ID50 of 7.8 ng/ml for the Smith strain and 6.7 12 separate experiments performed over a period ng/ml for the Uganda I strain. of 3 months are shown in Table 1. As indicated, The three antimalarial drugs, amodiaquin, the mean ID50 for chloroquine was nearly 20 primaquine, and pyrimethamine, were evaluated times higher with the Smith strain (182 ng/ml) in one experiment (Table 2). Amodiaquin, a 4- than with the susceptible Uganda I strain (9.5 aminoquinoline similar to chloroquine, was 714 DESJARDINS ET AL. ANTIMICROB. AGENTS CHEMOTHER. QUININE vs UGANDA I

D p M

18888

0 1888 ID-SO -2 CmCCNRTN18 188 ID-Q-28.8QCg/mj)CONCENTRATION Cng/m I)

QUININE vs SMITH

68888

58800

D 408 pP 30088

20808

18888

I to le I ego18 CONCENTRATION (ng/m I) ID-SQ - 134. 7OCng/mI ) FIG. 3. Comparative activity of quinine sulfate against the African Uganda I strain and the Vietnam Smith strain of P. falciparum in vitro. The range of concentration of the drug was 8.7 to 555 ng/ml in both cases. The IDro was 20.9 ng/ml (95% CI = 10.5 to 38.5 ng/ml) for the Uganda I strain and 135 ng/ml (95% CI

- 111 to 162 ng/ml) for the Smith strain.

somewhat more effective than the latter drug an IDro of 5.1 ng/ml. Against the Smith strain, against the Smith strain, with an ID50 of 23.7 which is known to be resistant to pyrimethamine ng/ml. Primaquine, an 8-aminoquinoline used in vivo, this drug was inactive at the highest clinically for its effect on exoerythrocytic schi- concentration tested, 1,500 ng/ml. zonts of Plasmodium vivax, was relatively inef- Several of the many other compounds evalu- fective in this test system, which specifically ated to date are presented in Table 3. Each of measures activity against only the asexual eryth- these compounds emerged from the U.S. Army rocytic forms of P. fakciparum. Pyrimethamine, Antimalarial Program and is a candidate for a dihydrofolate reductase inhibitor with selec- clinical studies. These results are generally con- tive activity against the protozoan enzyme, was sistent with existing in vivo data in animals and very effective against the Uganda I strain, with in some cases with clinical data in humans. VOL. 16, 1979 ASSESSMENT OF ANTIMALARIAL ACTIVITY IN VITRO 715 MEFLOQUINE ve UGANDA I

40M88

38888 D p M 28888

188

8 1 18 l18 1888 ID-50 - 8. 39(n/mI ) CONCENTRATION Cng/mI)

MEFLOQUINE ve SMITH

Bes88

s8888

D 40888 p M 30800

18888

a to too t808

CONCENTRATION Cnn/ml1S., ID-58 - 6. 6(nc/m I ) FIG. 4. Comparative activity of mefloquine hydrochloride (WR 142,490) against the African Uganda I strain and the Vietnam Smith strain of P. fakiparum in vitro. The range of concentration of the drug was 1.95 to 125 ng/ml in both cases. The ID50 was 6.39 ng/ml (95% CI = 5.82 to 6.91 ng/ml) for the Uganda I strain and 6.66 ng/ml (95% CI 6.43 to 7.19 ng/ml) for the Smith strain.

DISCUSSION on morphological assessment of the activity of chloroquine against parasites obtained from in- The adaptation of automated microtitration fected owl monkeys or humans was described equipment described here provides a rapid and and adapted for field use in epidemiological sur- quantitative measurement of antimalarial activ- veys by Rieckmann et al. (14). Although this ity for large numbers of compounds against P. method appears to provide reliable results with falciparum cultivated in vitro. The potential sufficient accuracy for certain applications, the value of in vitro drug susceptibility data for more precise and rapid method using asexual epidemiological purposes and in support of new parasites maintained in culture as described in drug development was discussed in a review by this report offers many advantages for a new Trigg in 1976 (23). A microtitration system based drug development program. 716 DE-SJARDINS ET AL. ANTIMICROB. AGENTS CHEMOTHER. TABLE 1. In vitro antimalarial activity of The use of a radiolabeled nucleic acid precur- choloroquine diphosphate, quinine sulfate, and sor as an indicator of parasite growth in an in mefloquine hydrochloride (WR 142,490) against two vitro antimalarial drug screen was reported by strains ofP. falciparum McCormick and Canfield (11). By measuring ID50 (ng/ml) suppression of the incorporation of [6-'4C]orotic Drug African Vietnam acid into deoxyribonucleic acid of Plasmodium (Uganda I) (Smith) knowlesi obtained from infected rhesus monkeys, the activity of several known antimalarial Chloroquine 9.5 ± 0.78a 182 ± 23.4 (7.8-11.2)" (130-233) drugs was demonstrated. Suppression of the up- Quinine 26.1 ± 5.57 109 ± 8.5 take and incorporation of [G-3H]hypoxanthine (23.8-48.4) (90.3-128) into nucleic acids by Babesia parasites in vitro Mefloquine 6.7 ± 1.00 7.8 ± 1.41 has also been used as an indicator of drug activ- (WR 142,490) (4.5-8.9) (4.7-10.9) ity (5). Though the feasibility ofusing the uptake a All values are the mean ± a standard error of 12 of [G-3H]hypoxanthine to measure antimalarial separate determimations. activity in vitro has been demonstrated by the b Value in parentheses are 95% confidence limits results in this report, the optimum conditions based on 12 separate determinations. with respect to medium composition, level of parasitemia, duration of culture, and pulse time have not been fully evaluated. However, in a TABLE 2. In vitro antimalarial activity of single experiment performed in this laboratory, amodiaquin, primaquine, andpyrimethamine there was good correlation between the ID50 against two strains ofP. falciparum values for chloroquine and mefloquine as deter- ID,0 (ng/mil) mined by morphological assessment and the iso- Drug African Vietnam topic method described here. Hypoxanthine, un- like thymidine, is capable ofcrossing the malaria (Uganda I) (Smith) parasite membrane (8). It is ultimately incorpo- Amodiaquin 9.8 23.7 rated into both ribonucleic acid and deoxyribo- Primaquine 543 1,136 nucleic acid and therefore provides a reasonably Pyrimethamine 5.1 >1,500 broad index of parasite metabolism. The simul- taneous use of a second label (["4C]-leucine) is being evaluated to extend the utility of the TABLE 3. In vitro antimalarial activity ofseveral screen. potential new antimalarial drugs against two The results presented in Table 1 clearly illus- strains ofP. falciparum trate the ability of the test system to quantitate IDso (ng/mil) the extent of cross-resistance among antimalar- Compound' Af.rican Vietnam ial drugs. The excellent activity of mefloquine in (Uganda I) (Smith) vitro against a chloroquine-resistant strain of P. falciparum is consistent with its efficacy against WR 184,806 10.0 7.7 infections with the same strain in owl WR 180,409 35.4 48.5 monkeys WR 172,435 7.9 3.2 (18) and with similar strains in clinical studies WR 171,669 2.5 3.9 (22). Phannacokinetic studies with this drug in WR 194,965 4.6 6.5 humans show a peak blood concentration after WR 99,210 1.4 1.8 a single 1,500-mg therapeutic dose of 1.2 to 2.0 WR 158,122 52.0 171 ,ug/ml and a mean half-life of 14 days (R. E. WR 225,448 >1,500 >1,500 Desjardins, C. L. Pamplin, J. von Bredow, K. G. a WR 184,806; DL-2,8-bis(trifluoromethyl)-4-[l- Barry, and C. J. Canfield, Clin. Phannacol. hydroxy-3-(N-t-butylamino)propyl]quinoline phos- Ther., in press). Based on these considerations phate. WR 180,409: DL-threo;-a-(2-piperidyl)-2-tri- and on the information in the graphs of Fig. 4, fluoromethyl-6-(4-trifluoromethylphenyl)-4-pyridine- the concentration of mefloquine in the blood is methanol phosphate. WR 172,435: 3-Di-n-butylamino- well above that required for complete inhibition 1-[2,6-bis(4-trifluoromethylphenyl)4-pyridyl] pro- of the parasite (0.1 Ag/ml) for several days after panol methanesulfonate. WR 171,669: 1-[1,3-di- a single oral dose of 1,500 mg. chloro-6-trifluoromethyl-9-phenanthryl]-3-di(n-bu- Amodiaquin is to tyl) aminopropanol hydrochloride. WR 194,965: 4-(t- thought possess potentially butyl)-2-(t-butylaminomethyl)-6-(4-chlorophenyl)- useful activity against infections with chloro- phenol phosphate. WR 99,210: 4,6-diamino-1,2-dihy- quine-resistant P. falciparum, based on clinical dro-2,2-dimethyl-1-[y-(2', 4', 5'-trichlorophenoxy)- and experimental observations (19). The data in propyloxy]-s-triazine hydrochloride. WR 158,122: 2,4- Table 2 support this impression and have led to diamino-6-(2-naphthylsulfonyl)quinazoline. WR interest in the development of new amodiaquin 225,448: proprietary compound. analogs. Also illustrated by the data in Table 2 VOL. 16, 1979 ASSESSMENT OF ANTIMALARIAL ACTIVITY IN VITRO 717 is the known resistance of the Smith strain to composition ofthe culture medium. The activity pyrimethamine and the relatively poor blood of sulfonamide drugs was not detected by the schizonticidal activity of primaquine. present system, presumably because of the The compounds presented in Table 3 are of abundance of para-aminobenzoic acid in stan- interest as potential new antimalarial drugs in dard RPMI 1640 culture medium. various stages ofinvestigation. One ofthese, WR In addition to its potential value both as a 171,669, a '9-phenanthrenemethanol, has been screen for new antimalarial drugs and for inves- shown to be effe,ctive against experimental infec- tigating the comparative activity of several an- tions with chloroquine-resistant P. falciparum alogs within a class of compounds, a number of malaria in human volunteers (15). The rest have other uses ofthe system are under consideration. been shown to be effective against chloroquine- The use of microtitration techniques in an 8-by- resistant strains of P. berghei in mice (13) and 12 matrix of wells presents an excellent oppor- against chloroquine-resistant strains of P. falci- tunity to evaluate potential synergy or antago- parum in owl monkeys. (17). nism among compounds. Combinations of com- The compound WR 225,448 was quite active pounds over a broad range of ratios of their in vivo in owl monkeys against the Smith strain respective concentrations can be generated by of P. falciparum (R. N. Rossan, unpublished sequential dilution in two directions (across rows data). It is therefore of interest that no activity and columns). The ability ofthe system to assess could be detected in vitro at very high concen- the antimalarial activity in blood specimens ob- trations. The compQund did not appear to pre- tained after administration of a drug to human cipitate either during preparation for testing or volunteers is also being explored. It would be of in the wells of the microtiter plate. It is charac- considerable value, in conjunction with early teristic ofan in vitro test system that compounds cliHnical studies with a potential new antimnalarial which require in vivo metabolism to an active drug, to determine the kinetics of antimalarial form will appear inactive in vitro. This possibil- activity in the blood of noninfected volunteers ity is being explored by in vitro metabolic studies after drug administration. with WR 225, 448. The experimental approach to measuring the Another compound of interest, WR 99,210, is activity of potential antimalarial drugs against a triazine derivative. The results with this com- P. falciparum described in this report provides pound (Table 3) are consistent with in vivo quantitative data not previously available and is results in owl monkeys infected with pyrimeth- sufficiently rapid and efficient to serve as a pri- amine-resistant P. falciparum (16). The lack of mary screen. It does not replace the very effi- cross-resistance between WR 99,210 and pyri- cient primary mouse screen (7) or the secondary methamine is remarkable because of the pre- screen in owl monkeys (16). Rather, it provides sumed mechanism of action of both drugs as supplementary information with respect to ac- inhibitors of the enzyme dihydrofolate reduc- tivity against the organism of interest, P. falci- tase. Another inhibitor of this enzyme, WR parum, and allows more critical selection of the 158,122, a quinazoline, did show some cross-re- most active compounds within a given class. It sistance with pyrimethamine in vitro (Table 3) has, in fact, become an integral part of the U.S. and in vivo (16). The remaining compounds in Army Antimalarial Drug Development Pro- Table 3, WR 184,806 (a quinolinemethanol), WR gram. 180,409 and WR 172,435 (pyridine derivatives), and WR 194,965 (a phenylphenol) have all been LITERATURE CITED for and tolerance in humans evaluated safety 1. Bard, Y. 1974. The Marquardt method, p. 94-96. In and are ready for phase II clinical efficacy stud- Nonlinear parametef estimation. Academic Press, Inc. ies. New York. When used as described, the present in vitro 2. Canfield, C. J., E. G. Whiting, W. H. Hall, and B. of antimalaial activ- MacDonald. 1971. Treatment of acute falciparum ma- system is capable detecting lana from Vietnam with trimethoprim and sulfalene. ity for many compounds with various modes of Am. J. Trop. Med. Hyg. 20:524-526. action. However, it is likely that the absolute 3. Finney, D. J. 1971. Quantitative dose-response relation- value of the ID50 for a given compound in vitro ships, p. 58-98. In Statistical methods in biological will vary somewhat among different laboratories assay. Griffon, London. 4. Haynes, J. D., C. L Diggs, F. A. Hines, and R. E. because of differences in technique. It is there- Desjardins, 1976. Culture of human malaria parasites fore important when testing and reporting the Plasmodium falciparum. Nature (London) 263:767- results for a given compound to report also the 769. results obtained simultaneously for known effec- 5. Irvin, A. D., and E. R. Young. 1977. Possible in vitro test for screening drugs for activity against Babesia and tive antimalarial drugs. Another important con- other blood protozoa. Nature (London) 29:407-409. sideration with regard to the mechanism of ac- 6. Kean, B. H. 1979. Chloroquine-resistant falciparum ma- tion of a candidate antimalarial compound is the laria from Africa. J. Am. Med. Assoc. 241:395. 718 DESJARDINS ET AL. ANTIMICROB. AGENTS CHEMOTHER. 7. Kinnamon, K. E., and W. E. Rothe. 1975. Biological 27:718-737. screening in the U.S. Anny antimalarial drug develop- 17. Schmidt, L H. 1979. Experimental infections with human ment program. Am. J. Trop. Med. Hyg. 24:174-178. plasmodia in owl monkeys-their contributions to de- 8. Manandhar, M. S. P., and K. Van Dyke. 1975. Detailed velopment of new broadly active blood schizonticidal purine salvage metabolism in and outside the free ma- drugs. p. 79-90. In M. Adolphe (ed.), Advances in phar- laria parasite. Exp. Parasitol. 37:138-146. macology and therapeutics, vol. 10. Pergamon Press, 9. Martin, D. C., and J. D. Arnold. 1967. Studies on an New York. East African strain of Pwsmodium fakxparum. Trans. 18. Schmidt, L H., R. Crosby, J. Rasco, and D. Vaughn. R. Soc. Trop. Med. Hyg. 61:331-339. 1978. Antimalarial activities of various 4-quinoline- 10. Martin, D. C., J. D. Arnold, D. F. Clyde, A. Ibrahim, methanols with special attention to WR 142,490 (meflo- P. E. Carson, K. E. Reman, and D. Willerson, quine). Antimicrob. Agents Chemother. 13:1011-1030. Jr. 1973. A quinolinemethanol (WR 30,090) for treat- 19. Schmidt, L H., V. Dennis, D. Meuller, R. Crosby, and ment of acute malaria. Antimicrob. Agents Chemother. R. Hamilton. 1977. Activities of various 4-aminoqui- 3:214-219. nolines against infections with chloroquine-resistant 11. McCormick, G. J., and C. J. Canfield. 1972. In vitro strains of PLasmodium fa1cyparum. Antimicrob. Agents evaluation of antimalarial drug combinations. Proc. Chemother. 11:826-843. Helminthol. Soc. Wash. (special isue) 39:292- 297. 20. Sehgal, P. N., M. L D. Sharma, and S. Gogai. 1973. 12. Moore, D. V., and J. E. Lanier. 1961. Observations on Resistance to chloroquine in falciparum malaria from two Plakmodium faklparum infections with abnormal Assam State, India. J. Commun. Dis. 5:175-180. response to chloroquine. Am. J. Trop. Med. Hyg. 10:5- 21. Trager, W., and J. B. Jensen. 1976. Human malaria 9. parasites in continuous culture. Science 193:673-675. 13. Peters, W., J. H. Portus, and B. L Robinson. 1975. 22. Trenholme, G. M, R. L Willims, R. E. Desjardins, The chemotherapy of rodent malaria. XXVII. The H. Frischer, P. E. Carson, K. H.RLcmann, and value ofdrug-resistant strains ofP. berghei in screening C. J. Canfield. 1975. Mefloquine (WR 142,490) in the for blood schizonticidal activity. Ann. Trop. Med. Par- treatment of human malaria. Science 190:792-794. asitol. 69:155-171. 23. Trigg, P. I. 1976. Parasite cultivation in relation to re- 14. Rieckmann, K. H., G. H. CampbelL L J. Sax, and J. search on the chemotherapy of malaria. Bull. W.H.O. E. Mrema. 1978. Drug sensitivity of Pwsmodium falci- 53:399406. parum. An in vitro microtechnique. Lancet i:22-23. 24. Ward, R. A., D. E. Hayes, S. C. Hembree, L C. 15. Rinehart, J. J., J. D. Arnold, and C. J. Canfield. 1976.' Rutledge, S. J. Anderson, and A. J. Johnson. 1972. Evaluation of two phenanthrenemethanols for antima- Infectivity of Plasmodium falcarum gametocytes larial activity in man: WR 122,455 and WR 171,669. Am. from Aotus trivirgatus to anopheline mosquitoes. Proc. J. Trop. Med. Hyg. 26:769-774. Helminthol. Soc. Wash. (special issue) 39:33-46. 16. Schmidt, L H. 1978. Pwsmodium fakciparum and Plas- 25. Young, M. D., P. G. Contacos, J. E. Stitcher, and J. modium vivax infections in the owl monkey (Aotus W. Millar. 1963. Drug resistance in Pkhnndium falci- trivirgatus). III. Methods employed in the search for parum from Thailand. Am. J. Trop. Med. Hyg. 12:306- new blood schizonticidal drugs. Am. J. Trop. Med. Hyg. 314.