Novel, Rapid, and Inexpensive Cell-Based Quantification of Antimalarial Drug Efficacy

We report on the development of a new SYBR Green I-based plateassay for analyzing the activities of antimalarial drugs againstintraerythrocytic Plasmodium falciparum. This assay is considerablyfaster, less labor-intensive, and less expensive than conventionalradiotracer (e.g., [³H]hypoxanthine and [³H]ethanolamine)-basedassays or P. falciparum lactate dehydrogenase activity-basedassays. The assay significantly improves the pace at which antimalarialdrug discovery efforts may proceed.

INTRODUCTION

Top

Introduction

The continued emergence and spread of multidrug-resistant strainsof Plasmodium falciparum and P. vivax are arguably the mostpressing problems in the area of infectious diseases today.Also, although the recent deciphering of the P. falciparum genomereveals many promising new drug targets, the financial costof bringing drugs to the clinic is a major obstacle in the developmentof new antimalarials (6). A faster, less expensive, high-throughputmeans of screening the activities of drugs against a varietyof malarial parasite strains would greatly assist preclinicaldrug development.

Quantitative assessment of the effects of drugs on parasitegrowth and development can be achieved by direct (but extremelytedious) microscopic examination of blood smears. An alternativeassay is measurement of the effect of drug exposure by determinationof the level of incorporation of radiolabeled hypoxanthine.While the latter method can be automated, it requires radioactivematerials and is not convenient for detection of parasite stage-specificeffects. Another assay measures parasite lactate dehydrogenaseactivity by methods that do not require radioisotopes. However,this assay requires multiple processing steps and expensivereagents and is not particularly cost-effective for large-scaledrug screening efforts.

We have thus endeavored to develop more rapid and convenientcell-based assays for quantifying antimalarial drug activities.We have strived to enhance simplicity and reduce cost. In thispaper, we report on the development of one such assay that relieson the fluorophore SYBR Green I.

MATERIALS AND METHODS

Top

Materials and Methods

Cell culture. Asexual culture is routinely performed. Parasite cultures areinitiated from stabilates preserved in liquid nitrogen (thelevel of parasitemia during storage is $>=$ 10%). Followingthe initiation of a fresh culture, at least two full life cycles(96 h) are completed before parasites are used for assays. Ingeneral, cultures are synchronized in the laboratory, and assaysare initiated when the parasites are at the ring stage. However,we find that this assay is equally applicable to asynchronousculture and that similar 50% inhibitory concentrations (IC₅₀s)are calculated from data with asynchronous and synchronous cultures(data not shown). Prior to assay initiation, the level of parasitemiaof an aliquot of a stock culture is measured by light microscopyfollowing Giemsa staining or by fluorescence-activated cellsorter analysis after staining with propidium iodide. In general,stock cultures with 5 to 10% parasitemias are used so that theyare likely to be in the mid-log growth phase. The level of parasitemiais adjusted prior to transfer to incubation wells. The timethat the stock culture is exposed outside a proper gas environment(5% CO₂, 5% O₂, 90% N₂) is minimized ( $<=$ 15 min).

Overall assay design. Positive controls for each assay contained no inhibitor anddifferent levels of initial parasitemia, and negative controlsincluded chloroquine (CQ) at high concentrations (see Results).Typically, wells are run in triplicate and drugs are testedagainst both CQ-resistant (CQR) and CQ-susceptible (CQS) strainsat various concentrations (see Results). Incubations are carriedout for one or two parasite life cycles (48 to 96 h). At theend of the test period, a solution of SYBR Green I (dilutedto a 20x concentration in phenol red-free complete medium) isadded to each well by use of a volume equal to 10% of the finalliquid volume in the well. The resulting solutions are mixedwith a low-retention pipette and allowed to stand for 30 minin the incubator. However, accurate detection can be done assoon as 1 min after dye addition. In our hands, lysis of redblood cells with detergents (in particular, harsh ionic detergents,such as sodium dodecyl sulfate) can compromise the signal-to-noiseratio. The data presented in this paper were obtained in theabsence of any detergents or red blood cell lysis. The SYBRGreen I fluorescence emission (530 ± 4.5 nm) from theplates is collected with a Gemini plate reader (Molecular Devices)at an excitation wavelength of 490 nm with 515-nm long-band-passand 530-nm band-pass emission filters. The Gemini instrumentis fully automated, can read 384 well plates, and has a sensitivityof submicromolar concentrations for fluorescence emitters inthe visible range. With this reader we find that the use of"black plates" reduces fluorescence cross talk between wellsand improves assay reproducibility. To facilitate analysis ofthe data, we used Softmax Pro software (version 4.1, life sciencesedition) and the Excel program.

Assay refinements. We refined the assay to reduce expense and increase both thesensitivity and the reproducibility of the assay. Incubationsare typically carried out in 200-µl final volumes in 96-wellplates (384-well plates can also be used if automated liquid-handlinginstrumentation is used). The preferred medium is phenol red-freeRPMI to reduce the levels of autofluorescence and backgroundabsorbance. We use plates with coated wells to reduce fluorescencecross talk. Measurements depend on the completion of at leastone full life cycle (48 to 54 h), which necessitates low initiallevels of parasitemia (0.5 to 1.0%) and low levels of hematocrit(1 to 2%). Overall growth inhibition is assessed by comparisonof the growth in the treated wells with that in the controlwells, to which no drug was added. Several wells on each plateare reserved for use for generation of a control standard curve.To generate the standard curve, a culture with a known levelof parasitemia is introduced into these wells immediately priorto dye staining and is stained along with the rest of the plate,so that the absolute SYBR Green I intensities in the test wellscan be extrapolated to a defined level of parasitemia. Importantly,such standard curves are linear over a range of 0.2 to 15% parasitemiaand overlap for the two strains examined in this study (datanot shown). As an additional control, several wells are reservedfor initial plating in the absence of drug but at differentinitial levels of parasitemia (i.e. 0.5, 0.75, and 1.0%). Whenthese controls are stained 48 to 96 h later, they reveal thegeneral health of the culture (e.g., the linearity of growth)within the plate format. If, for example, the control well fluorescencefor 1.0% initial parasitemia is not twice that for 0.5% initialparasitemia, the assay is redone.

RESULTS

Top

Results

SYBR Green I (Fig. 1, inset) is a DNA-intercalating dye thatprefers G and C base pairs. It can be prepared by heating 2-chloro-4-[2,3-dihydro-3-methyl-(benzo-1,3-thiazol-2-yl)-methylidene]-1-phenylquinoliniumiodide with N-(3-dimethylaminopropyl)-N-propylamine in 1,2-dichloroethanefor 2 h. The first compound can be rapidly made from 2-chloro-4-methyl-1-phenylquinoliniumchloride, which is easily synthesized from 1,2-dihydro-4-methyl-1-phenyl-2-quinolone(9). The dye is highly fluorescent when it is intercalated intoDNA but is poorly fluorescent when it is not (Fig. 1). Sincetypical laboratory growth of malarial parasites requires propagationin human red blood cells and since red blood cells do not containDNA, the use of SYBR Green I and its derivatives thus providesattractive advantages for malarial parasite growth detectionassays. Also, the excitation and emission spectra of DNA-intercalatedSYBR Green I are broad (Fig. 1), allowing considerable versatilityin the choice of detection wavelengths.

View larger version (19K):
[in this window]
[in a new window]
FIG. 1. Fluorescence excitation (390 to 505 nm) and emission (505 to 615 nm) spectra for 7.9 nM SYBR Green I (bottom trace) and 7.9 nM SYBR Green I in the presence of 4.2 nM plasmid DNA base pairs. At this dye:base pair molar ratio (1.9), fluorescence quantum efficiency increases 25- to 30-fold. The increase in SYBR Green I fluorescence quantum yield is linear with molar ratios of 7.0 to 0.5 (assuming approximately equal proportions of GC and AT base pairs; data not shown). RFU, relative fluorescence units. (Inset) Chemical structure of SYBR Green I.

Indeed, by laser confocal microscopy with the 488-nm line ofan argon laser, SYBR Green I provides brilliant resolution ofinfected versus noninfected erythrocytes in asynchronous culture(Fig. 2), even though P. falciparum has an AT-rich genome. Importantly,detection is via illumination at convenient visible wavelengths(Fig. 1) and not UV wavelengths, as is the case with many otherDNA-staining dyes. Also, very importantly, in Fig. 2 the dyeis uniformly exposed to the entire cell culture, but it fluorescesappreciably only when it binds to parasite DNA. Moreover, ahigh level of fluorescence is seen within seconds after additionof the dye (data not shown). Background staining is minimaland no washing steps are necessary, making use of this dye ina 96-well plate format ideal. That is, in theory, parasitesmay be grown with or without drug in a traditional red bloodcell culture environment, stained by one simple addition ofSYBR Green I, incubated for 5 min, and then rapidly read onan appropriate fluorescence-activated plate reader. Initially,results such as those shown in Fig. 2 were combined with thoseof fluorescence-activated cell sorter analysis to give a robust,quantitative measure of parasite growth and development. Wevalidated this procedure by direct comparison of the resultsobtained by the procedure with those obtained by light microscopyand have used it extensively for routine measurement of inhibitionof merozoite invasion of erythrocytes.

View larger version (141K):
[in this window]
[in a new window]
FIG. 2. SYBR Green I staining (in green) of intraerythrocytic malarial parasites attached to thin glass with polylysine and imaged by laser confocal microscopy with an Olympus Fluoview microscope in z-series mode. Excitation is at 488 nm, and emission at 510 to 550 nm is collected with a 505-nm dichroic filter and a combination of appropriate long- and short-band-pass filters. The entire field is bathed in SYBR Green I, but as evidenced by comparison of the differential interference contrast (A), fluorescence (B), and overlay (C), images only parasites are highly fluorescent.

Figure 3 shows that staining with SYBR Green I is similarlyintense regardless of the stage of intraerythrocytic parasitedevelopment. The intensity per parasite is similar for bothrings (Fig. 3A) and trophozoites (Fig. 3B), and multiple SYBRGreen I spots for the multiple nuclei present at the schizontstage are readily visible (Fig. 3C, in which 12 new nuclei canbe discerned). Since each malarial parasite typically produces8 to 24 progeny upon completion of intraerythrocytic growth,there should be significant amplification of the net SYBR GreenI intensity over time as a culture expands.

View larger version (69K):
[in this window]
[in a new window]
FIG. 3. Stage independence of SYBR Green I staining. By following the procedures described herein, to the best of our ability to ascertain SYBR Green I staining, SYBR Green I stains the ring (A), trophozoite (B), and schizont (C) stages of intraerythrocytic P. falciparum equally well. Thus, importantly, this assay is applicable to assessments of the stage-specific effects of drugs with synchronized cultures.

Figure 4A to C shows the results of typical growth inhibitionassays with CQ for strains HB3 (CQS) and Dd2 (CQR). Data arefit to sigmoidal curves, and the computed IC₅₀s (125 and 13nM for Dd2 [Fig. 4A] and HB3 [Fig. 4B], respectively) are nearthose computed by other assays in other laboratories (7). Moreimportantly, the ratio of these IC₅₀s (that is, the fold resistanceto CQ for Dd2 versus that for HB3) are nearly identical to theratios calculated by the use of other assays (e.g., see reference7 and the references therein). We do not expect exact agreementof the IC₅₀s determined by the different methods, since thedifferent reporters in various assays can have stage- and time-dependenteffects. In Fig. 4C we show CQ growth inhibition data for Dd2with and without verapamil (VPL), which is a well-known chloroquineresistance reversal agent. Importantly, the level of resistancereversal in this assay (approximately sixfold) compares verywell with the results from other laboratories in which moretraditional assays were used.

View larger version (12K):
[in this window]
[in a new window]
FIG. 4. Representative CQ growth inhibition data for CQR strain Dd2 (A), CQS strain HB3 (B), and CQR strain Dd2 in the absence (closed circles) and the presence (open inverted triangles) of 5 µM VPL (C). The computed IC₅₀s are 125 nM (A), 13 nM (B), and 125 and 21 nM, respectively (C). Thus, importantly, the fold CQ resistance for Dd2 versus that for HB3 obtained by this method agrees very well with many previously published ratios of the IC₅₀s for these strains, as does the fold CQ resistance reversal for strain Dd2 in the presence of VPL.

DISCUSSION

Top

Discussion

SYBR Green I and related compounds have been used to eithermonitor the appearance of DNA in biochemical matrices or theDNA concentration versus time. PCR-based assays for gene detection(1, 5, 8) and detection of DNA repair (3), gel mobility assays(4), and bacterial detection assays (2) have all been developed.Because red blood cells do not harbor DNA and because most antimalarialdrugs target the intraerythrocytic stage of parasite development,assays based on similar principles should be helpful for high-throughputscreening for new antimalarial therapies.

By the new assay that we have described, staining, quantificationof growth inhibition, and calculation of the IC₅₀ can be completedwithin 15 min. The activities of about 40 compounds can be preliminarilyscreened on a single 96-well plate. With automated pipetting,this expands to 200 compounds/384-well plate. By including thetime for cell culture, one person in a typical laboratory cantest 20 plates per week. This allows significantly acceleratedand less expensive assessments of candidate drugs or drug combinations.After preliminary screening, for example, the activities of8 to 10 drug pairs can then be rapidly assessed against twodifferent strains (e.g., CQS and CQR strains) per plate. Theassay is exceedingly fast because only one reagent is addedto the plate after parasite growth and no washing or filteringsteps are required. Comparative IC₅₀s are used in preliminaryscreens of antiparasitic compounds; however, an in vitro therapeuticindex is required to describe the selective toxicity for theparasite. Accordingly, we note that the assay described herecan also be easily applied to mammalian cell culture using similarprinciples.

ACKNOWLEDGMENTS

This work was supported by NIH grants AI45957 and AI056312 (toP.D.R.). We thank M. Riscoe (Oregon Health Sciences University)for helpful discussions.

T.N.B. and M.P. contributed equal efforts to this work.

FOOTNOTES

* Corresponding author. Mailing address: Department of Chemistry, Georgetown University, 37th and O St. NW, Washington, DC 20057. Phone: (202) 687-7300. Fax: (202) 687-6209. E-mail roepep{at}georgetown.edu.

REFERENCES

Top