Development and Evaluation of a Dipstick Assay for Detection of \u003cI\u003ePlasmodium falciparum\u003c/I\u003e and \u003cI\u003eP. vivax\u003c/I\u003e sporozoites in Mosquitoes (Diptera:

ARTICLE

Development and Evaluation of a Dipstick Assay for Detection ofPlasmodiumfalciarum and P. vivax sporozoites in Mosquitoes

(Diptera.. Culieidae)

RUSSELL E. COLEMAN, JAMES F. BARTH, MICHAEL J. TURELL, SCOTT W. GORDON,3

JETSUMMON SATTABONGKOT, ROBERT COPELAND,4 AND ROBERT A. WIRTZ5

Department of Vector Assessment, Virology Division, USAMRIID, 1425 Porter Street, Fort Detrick, MD 21702-5011

J. Med. Entomol. 37(4): 581-587 (2000)ABSTRACT We developed a nitrocellulose-based, dipstick circumsporozoite (CS)-enzyme im-

munoassay [ELISA] for the simultaneous detection of Plasmodiumfalciparum and P. vivax-210 CSprotein. The assay had a detection threshold of <250 P. falciparum or 400 P. vivax sporozoites persample, gave results concordant with dissection of salivary glands and CS-ELISA, but was slightlyless sensitive than the CS-ELISA in microtiter plates. The assay consistently detected one infectedmosquito in a pool of 10 or 20 mosquitoes, and was 100% specific in discriminating between speciesof Plasmodium when mosquito suspensions were spiked with sporozoites. The assay could becompleted in h, required no specialized equipment, and therefore was useful for field applications.

KEY WORDS detection, dipstick, circumsporozoite protein, mosquito, malaria

SPOROZOITE INFECTION RATE in mosquitoes, when com-bined with abundance, is a direct measure of theintensity of malaria transmission (Gu 1995). Tradi-tionally, sporozoite infection rates have been based onthe microscopic detection of sporozoites in the sali-vary glands after manual dissection. However, thisprocess is labor intensive and requires highly trainedtechnicians. The use of immunological assays, in par-ticular the circumsporozoite enzyme-linked immuno-assay (CS-ELISA), has widely replaced manual dis-section for determining mosquito sporozoite rates(Zavala et al. 1982, Wirtz et al. 1987a, Wirtz and Burkot1991). The main advantages of the CS-ELISA are thatit is species-specific and can detect sporozoites inpooled samples. Processing pooled specimens is ahighly efficient and economic method for determiningsporozoite rates, particularly when vector infection

Studies using laboratory-reared mosquitoes infected by feedingP. vivax patients were conducted under human protocol ap-proved by the ethical committee of the U.S. Army Medical Researchand Materiel Command (HSSP Log #A-2466; Comparative suscep-tibility of known and suspected species strains of Thai Anopheles toPlasmodium parasites).

Current address: Department ofEntomology, U. S. Army MedicalComponent, Armed Forces Institute of Medical Sciences, 315/6Rajvithi Road, Bangkok 10400, Thailand.

Two whom reprint requests should be sent: Department ofVectorAssessment, Virology Division, USAMRIID, 1425 Porter Street, FortDetrick, MD 21702-5011

Department of Preventive Medicine and Biometrics, UniformedServices University of the Health Sciences, 4301 Jones Bridge Road,Bethesda, MD 20814-4799.

International Center ofInsect Physiology and Ecology, Box 30772,Nairobi, Kenya.

Entomology Branch, Centers for Disease Control and Prevention,4770 Buford Highway -MS F-12, Atlanta, GA 30341-3724.

rates are low (Wirtz et al. 1987a, Gu 1995). The CS-ELISA has been used in numerous epidemiologicalinvestigations on the vectorial capacity of a wide va-

riety of anopheline mosquitoes for Plasmodium para-sites (e.g., Rosenberg et al. 1989, 1990), and has in-creased our understanding of the transmissiondynamics of malaria (Rosenberg et al. 1989). How-ever, the CS-ELISA does not differentiate betweensporozoites in the salivary glands and protein in thethorax, and therefore cannot completely replace dis-section as a means of conclusively incriminating ma-laria vectors. Also, the CS-ELISA is not suitable forsettings outside an equipped laboratory.

In the current article, we report on the develop-ment and evaluation of a simple, low technology assaysuitable for identifying potential malaria vectors andmeasuring sporozoite infection rates. This assay doesrequire sophisticated laboratory apparatus and is op-erational under field conditions. Although not in-tended to replace the CS-ELISA, the field operabilityand ease of use of the dipstick assay may prove usefulin areas where the CS-ELISA is not available andimmediate results are required.

Materials and Methods

Initial efforts focused on developing an assay thatwas rapid, simple, and had a minimal sensitivity of no>500 Plasmodium falciparum (Pf) or P. vivax (Pv)sporozoites. In addition, the assay had to differentiateaccurately between P. falciparum and P. vivax on asingle dipstick.

Reagents. Monoclonal antibodies (MAB) specific tothe circumsporozoite (CS) protein antigen of P. fal-

582 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 37, no. 4

ciparum (Pf) and P. vivax (Pv)-210 strain were used.The capture antibodies were MABs PF2A10,PVNSV3, and PV22. The development and character-ization of these MABs as well as their use in theCS-ELISA was described by Wirtz et al. (1987b, andunpublished data). Horseradish peroxidase (HRP)conjugates of PF2A10 and PVNSV3 were used as de-tector antibodies, and Moss 1-Step tetramethylbenzi-dene (TMB) was used as the HRP substrate (Moss-l,Hanover, MD). PF2A10 is used widely for both cap-ture and detection (as a HRP-conjugate) of Pf cir-cumsporozoite protein in the CS-ELISA, and provedequally effective in the dipstick format. PVNSV3 alsois used widely in the CS-ELISA and proved effectiveas a detector (HRP-conjugate) antibody. However,PV22 provided much greater sensitivity than didPVNSV3 as a capture antibody and was selected forinclusion in the assay.The Pf recombinant, R32tet32 positive control an-

tigen was described by Wirtz et al. (1987b). Collins etal. (1989) described the Pv recombinant NSIslV20used as the positive control antigen in our study. An-tigen was diluted to 2 ng//xl in distilled water andstored frozen at -70C. Working solutions were pre-pared by diluting this stock solution in blocking buffer.Blocking buffer consisted of 5% (wt:vol) dehydratedskim milk in distilled water. Skim milk blocking buffercan be stored frozen for up to 1 yr or at 4C for severalweeks. For field purposes, 0.5 g of dehydrated skimmilk was stored dry in a 15-ml conical centrifuge tubes.Before use, 10 ml of distilled H20 was added to eachtube, mixed thoroughly, and allowed to stand for 30min.

Unfortunately, the capture antibody (CAB) PV22does not recognize the repeat region of the Pv cir-cumsporozoite protein, so the synthetic peptideNSIslV20 could not be used as a positive control whenusing PV22. We therefore used frozen sporozoitepreparations as a positive control When using PV22.Continued use of PV22 as a CAB will require devel-opment of a synthetic peptide for use as a positivecontrol.

Laboratory-Infected Mosquitoes. The NF54 strainof Plasmodiumfalciparum was cultured as previouslydescribed (Burkot et al. 1984). Gametocytes appearedin the cultures between days 4 and 7 after inoculationand usually were mature, asjudged by appearance andthe ability to exflagellate, 5 or 6 d thereafter. Anophelesstephensi Liston mosquitoes, at least 3 d after emer-

gence, were fed on a mixture ofcultured gametocytes,defibrinated blood, and human serum (Burkot et al.1984). Mosquitoes that had fed on P. falciparum cul-tures were maintained at 25 _ IC, 67 3% RH, anda photoperiod of 16:8 (L:D) h.Approximately 200 colonized An. dirus Peyton &

Harrison mosquitoes were fed on the arms of Plasmo-dium vivax gametocyte-positive infected volunteers atthe government malaria clinic in Kanchanaburi, Thai-land (Sattabongkot et al. 1991), according to an ap-proved human use protocol (Walter Reed Army In-stitute of Research Protocol #44, on file in ourlaboratory). Mosquitoes were held in the laboratory

(Department ofEntomology, Armed Forces ResearchInstitute of Medical Science, Bangkok, Thailand) for14 d, after which a sample was dissected for sporoz0-ites. The remaining mosquitoes were stored frozen at-70C before testing for sporozoites.Assay Procedures. A nitrocellulose membrane (Bi0-

Rad Pure Nitrocellulose Membrane, 0.45-m poresize, Bio-Rad, Hercules, CA) was used as the solidphase for capturing the sporozoite antigens. A 9 by12-cm sheet of nitrocellulose membrane was cut intosix 9 by 2-cm strips. Each of these strips was used tomake 20 dipsticks (0.45 by 2 cm each). A piece oflabeling tape was folded over the long edge ofthe 9 by2-cm strips of nitrocellulose membrane to precludetouching the membrane and also served as a label.Each 9 by 2-cm strip then was cut into 20 dipsticks witha fine pair of scissors; however, at this stage we did notcut completely through the tape label. This resulted in

a sheet of 20 dipsticks held together by the labels.Keeping the dipsticks attached to each other facili-tated subsequent steps and simplified their long-termstorage.Capture antibody (0.5 pJ) was applied to the nitro-

cellulose membrane with a.10-/xl motorized, micro-capillary pipette (Rainin EDP-Plus pipette, Rainin In-struments, Woburn, MA). This volume of captureantibody produced a distinct circle that did not reachthe edges ofthe dipstick. Up to four capture antibodiescould be applied to a single dipstick (i.e., each dipstickcould be used to detect up to four different antigens);however, in this study, only PF2A10 and PVNSV3 orPV22 were applied. PF2A10 was used at a concentra-tion of 0.2 rng/ml, PVNSV3 at 0.017 rag/ml, and PV22at 0.225 mg/ml (diluted in distilled H20). The captureantibody was allowed to dry for 30 min at room tem-perature, after which each sheet of 20 dipsticks wassubmerged for 30 min in 5% skim-milk blocking buffer.Dipsticks were dried for 30 min, and stored betweentissue paper in a self-sealing plastic bag.Whole mosquitoes (individuals or in pools of up to

20) were triturated in 0.5 ml skim-milk blocking buffercontaining 0.5% (vol:vol) of nonidet P-40 (NP-40,Sigma, St. Louis, MO) using a motorized pestle (Kon-tes, Vineland, NJ) in a 1.5-ml microcentrifuge tube. Aminimum of 1 positive and 1 negative control wereincluded in each assay. Negative controls consisted ofskim-milk blocking buffer alone or containing an un-infectedAn. stephensi mosquito. P. falciparum-positivecontrols consisted ofthe synthetic peptide R32tet39.. P.vivax-positive controls consisted of synthetic peptideNSIslV20 (for use with capture antibody PVNSV3) or

P. vivax sporozoites (for use with capture antibodyPV22). Serial dilutions of the synthetic positive anti-gens were prepared in 0.5 ml of skim-milk blockingbuffer. Antigen concentrations ranged from 16 ng to15 pg synthetic peptide per 0.5 ml skim-milk blockingbuffer. P. falciparum and P. vivax sporozoites wereharvested and prepared as previously described(Burkot et al. 1984, Sattabongkot et al. 1991). Thesporozoite preparations were stored at -70C for lateruse. Serial dilutions of the sporozoites were preparedin skim-milk blocking buffer containing 0.5% (vol:vol)

July 2000 COIMaN ET AI.: DISIcI Assay vorMI SPorozorrs IN MOSQUITOES 583

NP-40. Sporozoite concentrations ranged from a highof 32,000/0.5 ml skim-milk blocking buffer to a low of30/0.5 ml skim-milk blocking buffer. The assay wasperformed and results graded and recorded as de-scribed below.

Dipsticks were incubated for 30 min in the 1.5-mlmierocentrifuge tubes containing the samples, thenrinsed for 10-15 in distilled H20. The dipsticks wereincubated for 30 min in 0.5 ml of the conjugate anti-body solution (0.5 mg/ml stoc! solutions of PF2A10-HRP and PVNSV3-HRP conjugate diluted 1:1500 and1:1000, respectively, in skim-milk blocking buffer) in1.5-ml mierotubes. Dipsticks were rinsed in distilledH20, and 10-15 dipsticks incubated for 6 min in 10 mlof Moss 1-step TMB enzyme substrate in a 15-ml cen-trifuge tube. Dipsticks were removed from the sub-strate, rinsed in distilled H.O, and air-dried. Resultswere recorded 15 min after the dipstick was re-moved from the substrate as follows: 0 (negative), 1(weak positive), 2 (moderate positive), 3 (fairlystrong), or 4 (extremely strong). Dipsticks werestored in a laboratory notebook to provide a perma-nent record of the results.Laboratory Evaluation. To determine the sensitiv-

ity and specificity of the dipstick assay, we comparedresults obtained with the dipstick assay with thoseobtained by microscopic evaluation of the salivaryglands, and the standard mierotiter plate CS-ELISA.Comparison with Manual Dissection. An intact sali-

vary gland was removed from a mosquito exposed toP. falciparum 21 d previously and examined for sporo-zoites by phase-contrast microscopy (400). Resultswere graded as follows: 0 (negative), 1 (1-10 sporo-zoites), 2 (11-100 sporozoites), 3 (101-1000 sporozo-ires), or 4 (>1,000 sporozoites). The remainder of themosquito (including the other salivary gland) wastriturated in a 1.5-ml microcentrifuge tube, tested bydipstick assay, and results graded and recorded asdescribed previously.Comparison with CS-ELISA. Thirty-six mosquitoes

fed on P. falciparum infected cultures or P. vivax-infected human volunteers were triturated individu-ally in 100/xl of PBS/NP-40. Fifty microliters of thetriturated mosquito sample was diluted in 250/xl ofskim-milk blocking buffer, tested by dipstick, andgraded as previously described. Forty microliters ofthe original triturated mosquito sample was diluted in80 /xl of casein blocking buffer for testing by CS-ELISA. Two 50-/xl samples were tested by CS-ELISAfor Pf and Pv sporozoites (Wirtz et al. 1987a). Thesamples tested by ELISA contained approximately0ne-third as many sporozoites as did the samplestested by dipstick. ELISA results were graded visuallyas follows: 0 (negative), 1 (weak positive), 2 (mod-erate positive), 3 (strong positive), or 4 (extremelystrong positive).These same procedures were used to test 263

anopheline mosquitoes collected in the Suan PhungDistrict, Ratchaburi Province, Thailand, between 25and 26 September 1995. Mosquitoes were sorted tospecies (1 An. kochi Doenitz, two An. jamisii Liston, 42An. minimus Theobald, 19 An. dirus, 181 An. maculatus

Theobald, 1 An. aconitus Doenitz, and 17 An. sawad-wongporni Rattanarithikul & Green), pools of 1-5specimens ofeach species were triturated in skim-milkblocking buffer, and 50 1 of each sample was testedby dipstick assay at the field site. The remaining por-tion of each sample (50 1) was brought to theArmed Forces Research Institute of Medical Sciencesin Bangkok and stored at -70C until tested by CS-ELISA.A second study was conducted at the U.S. Army

Medical Research Unit in Kenya. Mosquitoes from theKisian and Saradidi areas of western Kenya were col-lected, identified to species (An.funestus Giles andAn.gambiae Giles s.1.), and salivary glands were dissectedinto 50 1 ofphosphate buffered saline (PBS). Glandswere examined for sporozoites under a phase-contrastmicroscope and sporozoite positive glands were savedin 100 txl of casein blocking buffer (Wirtz et al. 1987a).Glands from 77 sporozoite-positive and seven sporo-zoite-negative mosquitoes were stored at -70C untiltested. On the day ofthe test, 150 1 of casein blockingbuffer containing 0.5% (vol:vol) NP-40 (Wirtz et al.1987) was added to each tube and the salivary glandswere homogenized. An additional 750/xl of blockingbuffer was added to each tube, bringing the totalvolume to 1 ml. Half of this volume vas used for thedipstick assay and half for the CS-ELISA. Dipstickassay results were graded visually as follows: 0 (neg-ative), 0.5 (very weak), 1.0 (weak), 1.5 (weak plus),2.0 (moderate), 2.5 (strong/moderate), or 3.0 (verystrong).

Effect ofPool Size. Pools of 1, 2, 3, 5, 10, 20, 50, or 100uninfected mosquitoes were spiked with 1,000, 5,000,or 10,000 sporozoites and were assayed. In a subse-quent experiment, a single Pf-positive mosquito wasadded to similar pools of uninfected mosquitoes andassayed.

Dipstick Optimization. To be truly field-usable, an

assay must not depend upon specialized storage (i.e.,refrigeration or freezing) conditions. We thereforeevaluated the effect of various storage conditions onthe dipsticks, the test reagents, and positive controlmaterial.One thousand dipsticks were prepared, and 200

each were stored at 70, 20, 4, 26C (ambient tem-perature), or at 34C (in an incubator with relativehumidity maintained at 95-100%). PF2A10 and PV22were used as capture antibody on these dipsticks.Dipsticks were held from 1 d to 18 mo before use. Onthe day of the test, 10 dipsticks from each storagetemperature were removed and tested against fourserial dilutions of P. falciparum or P. vivax sporozoites.Two dipsticks were used as negative controls. Resultswere recorded as previously described and comparedwith results obtained when the dipsticks initially wereprepared.

Phosphate buffered saline is the rinsing agent usedin many diagnostic assays, including the CS-ELISA,whereas distilled Hg,O is often used to prepare appro-priate blocking buffers for use in other steps. Becausethis assay is intended for field use, we felt that distilledHg.O or PBS might not always be available. We there-

584 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 37, no. 4

Pg of Synthetic AntigenO0 1000 10000

---- Pf Sporozoites T P -[Pv Sporozoites / /Synthetic Antigen

10 O0 1000 10000

Number of Sporozoites

Fig. l. Sensitivity ofthe malaria dipstick assay for P.falciparum and P. vivax sporozoites and the P.falciparum recombinantantigen.

fore examined the effect of different water sources onthe efficacy ofthe assay. In a series of experiments, weevaluated distilled H20, PBS (pH 7.2), ordinary tapwater, and water with a high organic content (pondwater) as rinsing agents or reagent diluents. A total of200 dipsticks was prepared as previously described. Ineach trial (3 replicates), 40 dipsticks were used, 10 foreach of the four types of wash/rinse solutions. Eightdipsticks were tested against four serial dilutions of P.vivax or P. falciparum sporozoites, with two dipsticksserving as negative controls. Results were recorded aspreviously described.

Results

Sensitivity of the Dipstick Assay. Serial dilutions ofthe P. falciparum recombinant peptide positive con-trol material and frozen sporozoites were evaluated in

a series of blinded trials. No mixtures of P. falciparumand P. vivax were made (i.e., the test was conductedwith either P. falciparum or P. vivax antigen in a singletube, not a mixture of the two). The assay reliablydetected 125 pg of the P. falciparum recombinantantigen and 250 P. falciparum or 400 P. vivax sporo-zoites (Fig. 1). By increasing the antigen incubationtime to 45 min, the assay differentiated between P.falciparum and P. vivax CS protein on a single dipstickand reliably detected as few as 150 sporozoites (Figs.1 and 2). We subsequently conducted a double-blinded test with 25 samples containing 1,000 or 5,000P. falciparum or P. vivax sporozoites (alone or a com-bination of both species), as well as three negative

controls. The dipstick assay correctly identified all 28samples (3 negative controls, 8 Pf, 15 Pv, and 2 mix-tures containing Pf and Pv).Comparison with Manual Dissection. Two separate

tests with a total of 60 P. falciparum laboratory-in-fected mosquitoes were conducted (Table 1). Wedetermined the sporozoite status of 57 of 60 mosqui-toes examined microscopically. Of these 57 mosqui-toes, 18 were negative and 38 were positive by dis-section and dipstick assay; one was positive bymicroscopic examination (only four sporozoites ob-

P. vivax- 1,000 Sporozoites (+4)

P. falciparum- 250 sporozoites (+2)

Negative Control (0)

Fig. . Plasmodium vivax-positive # 162), P. falciparum-positive # 164) and negative #165) malaria dipstick assays.

July 2000 CoIs_nn vx n_t.. DIr'ScI Assnr vorMSvorozorrs i MOSQUITOES 585

Table 1. Comparison of results obtained with the dipstick assay and results obtained when salivary glands manually dissectedand microscopically examined for P. falciparum sporozoites tested by CS-ELISA

P. falciparum P. vivax

Dipstick value Manual dissection value CS-ELISA Value CS-ELISA Value

0 2 3 4 Total 0 2 3 4 Total 0 2 3 4 Total

0 18 0 0 0 19 13 0 0 0 0 13 13 2 0 0 0 150 5 0 0 0 5 0 0 0 0 0 0 0 2

2 0 0 6 4 11 0 0 2 0 0 2 0 0 2 0 33 0 0 3 5 3 11 0 0 0 0 0 0 0 0 0 04 0 0 3 4 4 11 0 0 0 2 9 11 0 0 0 0Total 18 6 12 13 8 57 13 2 2 10 28 13 3 2 2 21

Values in each column/row intersection indicate the number of samples with that specific dipstick value and that specific manual dissectionor CS-ELISA value. Dipstick, manual dissection, and CS-ELISA results were visually graded from 0 (negative) to +4 (strong positive reactionby dipstick and CS-ELISA or >1,000 sporozoites observed in the salivary glands).

served in the salivary gland) but negative by the dip-stick assay. Overall, the sensitivity and specificity ofthe dipstick assay were comparable with that achievedby manual dissection, with only one (1/19) false neg-ative and no (0/39) false positive results. Somewhatsurprisingly, based on the sensitivity of the assay withprepared sporozoite preparations, the dipstick assayaccurately detected five of six mosquitoes that weregraded as a'T' by dissection (indicating that only 1-10sporozoites were observed in the gland).Comparison with ELtSA. Of36 laboratory-exposed

mosquitoes tested by dipstick assay and ELISA, sixwere P. vivax positive, 15 were P. falciparum positive,and 13 were negative by both the dipstick assay andthe CS-ELISA (Table 1). Two mosquitoes that werepositive for P. vivax by ELISA + 1) were negative bythe dipstick assay.A total of 80 mosquito pools containing from 1-5

mosquitoes each was prepared from the 263 field-collected mosquitoes from Thailand. One pool of fiveAn. maculatus was positive for P. vivax by both dipstickassay and the CS-ELISA. The remaining 79 pools werenegative by both dipstick assay and ELISA for both P.falciparum and P. vivax.

Eighty-four field-collected mosquitoes from Kenyawere tested by dipstick assay and CS-ELISA. TheCS-ELISA data were normalized with a square-roottransformation ofthe data, which were compared withthe dipstick values (Table 2). Results of the dipstick

assay and the CS-ELISA were correlated strongly,explaining 81% of the variation. Although samplepreparation procedures used in this test differedslightly from those used in other tests (see Materialsand Methods), the results were essentially the same.All seven negative controls were negative by both theCS-ELISA and the dipstick assay. In addition, 10 otherdissection-positive samples were negative by both theCS-ELISA and the dipstick assay, whereas 66 sampleswere positive by both tests. The 10 dissection-positivesamples that were negative by both dipstick and CS-ELISA may have been P. malariae or P. ovale sporo-zoites. One dissection-positive sample was positive bythe dipstick assay but negative by the CS-ELISA.

Effect ofPool Size. The assay readily detected 1,000,5,000 or 10,000 sporozoites in pools of 1-20 mosqui-toes. The strength of the reaction was slightly weakerin pools of 20 mosquitoes than in smaller pools; how-ever, positives were still easily visible. The assay de-tected 10,000 (but not 1,000 or 5,000) sporozoites inpools of50 or 100 mosquitoes. However, the signal wasweak and difficult to discern. Essentially identical re-suits were obtained when a single infected mosquitowas placed in a pool with varying numbers of unin-fected mosquitoes. The assay reliably detected CSprotein in pools of -<20 mosquitoes, but not in pools of50 or 100 mosquitoes.

Effect of Storage Conditions on the Dipsticks. Dip-sticks held at -70 or -20C retained their efficacy for

Table 2. Comparison of results obtained with the dipstick assay and the CS-ELISA using the salivary glands of A. funestus andA. gambiae mosquitoes collected in Kenya

Dipstick valueSquare root of CS-ELISA OD value

0 0.2 0.2- 0.4 0.4 0.6 0.6 0.8 0.8 -1 1-1.2 1.2-1.4 Total

0.0 17 0 0 0 0 0 0 170.5 2 0 0 0 0 0 0 21.0 4 3 0 0 0 0 0 71.5 0 2 0 0 0 0 32.0 0 5 23 4 0 342.5 0 0 3 5 4 2 153.0 0 0 0 0 2 2 2 6Total 23 5 8 26 11 7 4 84

Values in each column/row intersection indicate the number of samples with that specific dipstick value and that specific CS-ELISA value.Dipstick results were recorded 0 (negative), 0.5 (weak-), 1.0 (weak), 1.5 (weak +), 2.0 (moderate), 2.5 (moderate +) 3.0 (strong).CS-ELISA OD values were normalized using square-root transformation.

586 JotmNnL OF MEDICAL ENTOMOLOGY Vol. 37, no. 4

18 mo (the longest period the dipsticks were evalu-ated). Dipsticks held at four or 26C retained theirefficacy for at least 12 mo, whereas those stored in anincubator at 34C and 95-100% RH were effective foronly 5-6 mo. Interestingly, in a separate experiment,several packets of dipsticks stored in the incubatorwere opened and closed several times during a 6-moperiod. Heavy mold began growing on these dipsticks;however, these still retained their efficacy for at least6 mo (although with a fainter signal than the packetsthat remained sealed for the entire 6 mo) despite themold.Choice of Wash/Rinse Solutions. The assay per-

formed equally well whether PBS, distilled water, tapwater, or tap water with a high organic content (watercontaining high populations of larval mosquitoes andmosquito food) was used as the rinsing agent, as thediluent for the skim milk blocking buffer used to trit-urate the samples, or as the diluent for the detectorantibody.

Discussion

The malaria sporozoite dipstick assay was simple toperform, giving results in <1 h under field conditions.When compared with manual dissection, the assayappeared sensitive, and it reliably detected mosqui-toes containing as few as 5-10 sporozoites in a singlesalivary gland. Presumably, this heightened sensitivity(compared with results obtained with purified sporo-zoites) was caused by the presence of circulating CSprotein within the mosquito or sporozoites that hadnot yet penetrated the salivary glands (Golenda et al.1990).

In addition to qualitatively determining the sporo-zoite status of mosquitoes, the dipstick assay also pro-vided quantitative information about the antigen load.The intensity ofthe signal was correlated strongly withthe number of sporozoites in the mosquito. Mosqui-toes with higher sporozoite loads gave much more

intense signals than mosquitoes with lower sporozoiteloads. Although this system is somewhat subjective,we successfully used a hand-held portable densitom-eter to quantitate sporozoite loads with the dipstickassay (R.E.C., unpublished data). Use of the densi-tometer will allow the user to calculate standard de-viations and more accurately determine sporozoiteloads.Although the dipstick assay as reported here was

only capable of the simultaneous detection of P. fal-ciparum and P. vivax-210, we have developed a pro-totype assay that detected and differentiated four par-asites: P. falciparum, P. vivax-210, P. vivax-247, and P.malariae. However, this multiple system was ham-pered by nonspecific binding between some of theantibody/antigen combinations, resulting in a highincidence of false positives. We used four differentdetector antibodies (i.e., one for each species of par-asite) in this system; however, use of a Plasmodiumgenetic detector antibody would presumably simplifythe entire system. This would make it much easier tooptimize the assay and would decrease the number of

false positives by minimizing the amount of nonspe-cifi binding. When using four different detector an-tibodies, it was necessary to use extremely dilute con-

centrations of each antibody (i.e., 1:3000 instead of1:1000 or 1:1500 used in this study), lowering thesensitivity of the assay.There is clearly a need for a field assay capable of

rapidly detecting Plasmodium-infected mosquitoes. In

many epidemiological studies it is acceptable to delaytesting of mosquito specimens until they can bebrought to a laboratory equipped with the CS-ELISA;however, in humanitarian or military operations animmediate answer is required, even though a well-equipped laboratory is not available. Although thedipstick assay that we have developed shows promise,it is hampered by the fact that it is a multi-step processthat requires 1 h to complete. Current efforts in ourlaboratory focus on the development of a single-stepassay that can be completed in 15 min or less. Inaddition, alternative methods of preparation and pro-cessing of samples for testing are being explored, as isthe development of panel assays capable of detectingother arthropod-borne pathogens. These tools, whencommercially available, will be highly advantageousfor military use, epidemiological studies, and directedcontrol efforts.

AcknowledgmentsWe thank T. Kollars, J. Ryan, G. Ludwig, K. Kenyon, and

T. Klein for critically reviewing the manuscript and providingadvice and guidance. This study used reagents produced withsupport from the UNDP/World Bank/WHO Special Pro-gram for Research and Training in Tropical Diseases. ThePf2A10 and PvNSV#3 cell lines were developed at New YorkUniversity and the Naval Medical Research Institute, respec-tively. The recombinant positive controls were a gift fromSmith Kline Beecham. The work was supported by fundsfrom the U.S. Army Medical Research and Materiel Com-mand.

References Cited

Beier, J. C., P. V. Perkins, F. K. Onyango, T. P. Gargan, C. N.Oster, R. E. Whitmier, D. K. Koech, and C. R. Roberts.1990. Characterization ofmalaria transmission byAnoph-eles in western Kenya in preparation for malaria vaccinetrials. J. Med. Entomol. 27: 570-577.

Burkot, T. R., J. L. Williams, and I. Schneider. 1984. Infec-tivity to mosquitoes of Plasmodium falciparum clonesgrown in vitro from the same isolate. Trans. R. Soc. Trop.Med. Hyg. 78: 339-341.

Collins, W. E., R. S. Nussenzweig, W. R. Ballou, T. K. Rue-bush II, E. H. Nardin, J. D. Chulay, W. R. Majarian, J. F.Young, G. F. Wasserman, I. Bathurst, and others. 1989.Immunization of Saimiri sciureus boliviensis with recom-binant vaccines based on the circumsporozoite protein ofPlasmodium vivax. Am. J. Trop. Med. Hyg. 40: 455-464.

Golenda, C. F., W. H. Starkweather, and R. A. Wirtz. 1990.The distribution of circumsporozoite protein (CS) inAnopheles stephensi mosquitoes infected with Plasmo-dium falciparum malaria. J. Histochem. Cytochem. 38:475-481.

July 2000 COLEMAN ET AL.: DIPSTICI,: ASSAY FOR MALARIA SPOROZOITES IN MOSQUITOES 587

Gu, W.-D. 1995. Estimating sporozoite rates by examiningpooled samples of mosquitoes. Trans. R. Soc. Trop. Med.Hyg. 89: 359-360.

Rosenberg, R., R. Wirtz, D. Lanar, J. Sattabongkok, T. Hall,A. Waiters, and C. Prasittisuk. 1989. Circumsporozoiteprotein heterogeneity in the human malaria parasite Plas-modium vivax. Science 245: 973-976.

Rosenberg, R., R. G. Andre, and L. Somchit. 1990. Highlyefficient dry season transmission of malaria in Thailand.Trans. R. Soc. Trop. Med. Hyg. 84: 22-28.

Sattabongkot, J., N. Maneechai, and R. Rosenberg. 1991.Plasmodium vivax: gametocyte infectivity of naturallyinfected Thai adults. Parasitology 102: 27-31.

Wirtz, R. A. and T. R. Burkot. 1991. Detection of malarialparasites in mosquitoes. In Advances in disease vectorresearch, vol 8. Springer, New York.

Wirtz, R. A., T. R. Burkot, P. M. Graves, and R. G. Andre.1987a. Field evaluation of enzyme-linked immunosor-

bent assays for Plasmodium falciparum and Plasmodiumvivax sporozoites in mosquitoes (Diptera: Culicidae)from Papua New Guinea. J. Med. Entomol. 24: 433-437.

Wirtz, R. A., F. Zavala, Y. Charoenvit, G. H. Campbell,T. R. Burkot, I. Schneider, K. M. Esser, R. L. Beaudoin,and R. G. Anders. 1987b. Comparative testing ofmonoclonal antibodies against Plasmodiumfalciparumsporozoites for ELISA development. Bull. WorldHealth Org. 65: 39-45.

Zavala, F., R. W. Gwadz, F. H. Collins, R. S. Nussenzweig,and V. Nusenzweig. 1982. Monoclonal antibodies tocircumsporozoite proteins identify the species of ma-laria parasite in infected mosquitoes. Nature (Lond.)299: 737-738.

Receivedfor publication 23 March 1999; accepted 24 March2000.