Us and West Nile and Sindbis Viruses

70 JounNer, oF THE AMERTCANMoseurro CoNTRoLAssocrATroN VoL.5, No. 1

COMPARISONOF THREE METHODS FOR DETERMINING TRANSMISSION RATES IN VECTOR COMPETENCESTUDIES WITH CULEX UNIVITTA?US AND WEST NILE AND SINDBIS VIRUSES

A. J. CORNEL exo P. G. JUPP

Department of Virology, Uniuersity of the Witwatersrand,and Arbouirus Unit, National Institute for Virology, Priuate Bag X4, Sandringham2131, South Africa

ABSTRACT. Culex uniuittatus mosquitoeswere infected by ingesting West Nile and Sindbis viruses, respectively.After 13-19 days, comparative viral transmission rates were determined using 3 methods: 1) transmission to hamstersby bite, 2) in vitro transmission while feeding on suspensionsin capillary tubes and 3) transmission while feeding on droplets of suspension.Transmission rates were 59-100%, exceptfor Sindbis virus by droplet feedingwhen the rate was 9%. Except for the latter, rates determined by the different methods were very similar in the case of each virus. It was concluded that capillary feeding is an acceptablealternative to feeding on hamsters for determining transmission rates for Cr. uniuittatus with West Nile and Sindbis viruses.

INTRODUCTION Aitken 1979).The droplet methoddeveloped by Gubler and Rosen(1976) was usedby them for laboratory In our a well established method the transmission of dengue-2 virus by Aedes for determining the transmission rates of mos- albopictus(Skuse). They found that the per- quito vectors with arboviruses of is the feeding centageof mosquitoestransmitting was related infected individual mosquitoes on individual to the extent of the infection in their salivarv susceptible animals. The choice of animal de- glands. pends on the virus and vector in question and includeschickens (Jupp 1974),mice or hamsters (Mclntosh et al. 1980)and monkeys (Jupp et al. METHODS 1981).For our purposeshamsters have been the most useful laboratory animals as they can be A laboratory colony of Cx. uniuittatus was used with a number of viruses including Rift usedwhich originated from mosquitoescollected Valley fever (RVF), West Nile (WN), Sindbis in Johannesburg-F1, Fz or Fz adult females, (SIN) and chikungunya viruses. Successful aged 1-18 days, were given infective feeds. All transmission is measuredbyvirus isolation from the experiments were undertaken in an insec- a hamster if the virus is pathogenic, such as tary where the mosquitoeswere maintained at RVF and WN viruses, or by the subsequent 75-85% relative humidity and 25-26"C. detection of hemagglutinating antibodies in the The viruses usedwere the H442 strain of WN animal's blood, as in the caseof SIN and chi- virus (Kokernot and Mclntosh, 1959) and the kungunya viruses. AR86 strain of SIN virus (Weinbren1955). Each Although the use of animals is thought to give virus had beenpassaged 5 times in mice (intra- an assessmentof transmissionthat is morecom- cerebral)and 4 times in Vero cells. parable to the real situation in the field than Infective meals of high titer were provided by that measuredby an in uitro method, it is never- intramuscular inoculation of chicks. For SIN theless cumbersome,time-consuming and ex- virus, l-day-old chicks were inoculatedand ex- pensivein a project requiring many transmis- posed to the mosquitoes 24 hr later, while for sion tests. Such was our problem when we WN virus 2-day-oldchicks were inoculatedand wished to undertake a study on the effect of exposedto the mosquitoes36 hr later.The mos- temperature on the vector competence of Cr. quitoes were starved for 24 hr preceding each uniuittatus Theobald with WN and SIN viruses. infective meal, and feedingtook place for 2-3 hr It was therefore decidedto evaluate two in uitro in the dark in the insectary. Immediately before transmission methods against our existing ham- exposure,a blood samplewas collectedto deter- ster rnethodto seewhether one or both methods mine the concentration of virrs present. This gave comparableresults and could therefore be was titrated in Vero cells, and the 5O% cyto- used as an alternative. pathic dose (CPDbo)end points were calculated "capillary" The methods chosenwere the and by the method of Reedand Muench (1938).Viral "droplet" methods. The former was developed titers were expressedas logroCPDs,o/mlof chick by Aitken (1977)and has been used successfully blood. to determine transmission rates for Aedes ae- Transmissions were attempted 13-19 days eypti (Linn.) with yellow fever virus (Beaty and after the infective feedwith WN and SIN viruses MARCH 1989 Vmus TnlNstr,tlssloN RATEStN Cx. vt'uwrrtrus

respectively,by each ofthe 3 methods described quently tested for virus by inoculation of a group below. In eachexperiment the transmission rate of 8 infant mice. The mosquitoeswere judged to was subsequentlydetermined, i.e., the propor- have been infected if all 8 inoculated mice died tion of infected mosquitoes which transmitted within the expecteddeath time of the respective virus. Fisher exact tests in an Epistat statistical viruses. The infection rate, i.e., the proportion computer program were used to compare the of mosquitoeswhich had becomeinfected after rates obtained by the different methods to as- feeding on the infective meal, could then be certain whether they were significantly differ- determined for each group of mosquitoes. ent. In the hamster method, individual infected RESULTS mosquitoeswere fed on individual Syrian ham- sters.Successful transmission was demonstrated AII of the mosquitoes became infected after by the development of hemagglutinating anti- ingesting WN virus (Table 1). Infection rates bodies against SIN virus in the hamsters blood were also hish (57-89%) after ingestion of SIN 21 days later. In the caseof WN virus, death of virus. The transmission rates were slightly a hamster indicatedtransmission, and the ani- higher for both viruseswith the hamster method mal's liver was harvestedand tested for virus to as compared to rates determined with the cap- confirm this. illary method. These differences were statisti- With the droplet method (Gubler and Rosen cally significant,i.e., P = 0.1-0.2(SIN) and P: 1976),an infected mosquito was placedin a tube 0.3 (WN) atthe 5% significancelevel but not at coveredwith fine nylon mesh material so that it the 1.0%Ievel. Comparison of the transmission could feed from a droplet of fluid placed on the rates obtained with the droplet method to those material.In this method,0.025 ml of suspension obtained with hamsters showedthat the differ- consisting of sucrose, fetal calf serum and ence was again significant in the case of WN washed gooseerythroc5rtes were placed on the virus (P : 0.01) at the 5% but not at the lUVo nylon netting. After the mosquito had engorged, level; however the rate obtained for SIN virus the remainder of the drop was diluted l0-fold in with the droplet method was very much lower 0.25ml of Liebovitz'smedium before inoculation (9%). Someadditional tests weretherefore car- into cultures of Vero cells to test for virus. ried out to try to explain this low transmission Transmission was defined to have occurred if rate. SIN virus-infected mosquitoeswere fed on virus was recoveredfrom the unimbibed portion a droplet in the usual way, exceptthat the goose of the droplet. erythrocytes were replaced by sheep erythro- For the capillarymethod (Aitken 1977),a tube cy'tes.The transmission rate still remained be- was employed with inner and outer diameters low l0%. Secondly, infected mosquitoes were about 0.5 and 1.0mm, respectively.It was drawn allowed to feed on a droplet but, when half- out to a fine point and filled with 0.004 ml of engorged,were removedand allowedto complete feeding suspensionconsisting of equal parts of their meals by the capillary method. Both the 10% sucroseand fetal calf serum.The point of droplet and capillary fluids were then titrated the tube was insertedover the proboscisof the for virus in Vero cells. No virus was detectedin mosquito which was then allowed to feed to any of the droplet fluids, but 4.5-5.4 logroCPDso/ repletion. In this method 0.0025ml of the feed- ml of SIN virus was present in all the capillary ing suspensionremained in the tube after feed- suspensions. ing. Subsequentlya 100-fold dilution in 0.25 ml permit of Liebovitz's medium was necessaryto DISCUSSION inoculation of the specimeninto cultures of Vero cells. The results of the transmission tests with the After each mosquito had completed its trans- 3 different methods indicate that the capillary mission feed, it was stored at -70'C and subse- tube method would be an acceptablealternative

Table 1. Comparative transmission rates bv three different methods.

Hamster method Capillary method Droplet method Virus (Titer of infective Days Infection Transmission Infection Transmission Infection Transmission meal) postinfection rate rate rate rate rate rate Sindbis 13-19 47/54 (87%) 34/47(74%) 66/74 (8e%) 39/66 (597o)17/re (8e%) 3/35(e%) (7.0-8.5.) West Nile 13-14 20/20 (100%)20/20 (r00%) 28/28 (r00%) 22/28 (78%) 17/17 (ro0%) L2/I7 (7r7o) (5.0-5.5) * Means LogroCPDso/ml. 72 JoURNALoF THE AMERTcANMosquno CoNrnor, Assocreuon VoL. 5, No. 1

to feeding infected Cx. uniuittatus mosquitoes uate it against an animal method first for the on hamstersin the caseof SIN and WN viruses. particular virus-vector combination under Although the transmission rates obtained for study. both viruseswith the capillary method wereonly just significantly lower by statistics than the ACKNOWLEDGMENTS rates obtained with hamsters, the rates were very similar. Hence our results indicate that this We wish to acknowledge a grant from the method has valid applications as it can give a PoliomyelitisResearch Foundation which con- goodapproximation as to what the transmission tributed towards this work. We also thank Mrs. rate would have been in animals. This finding G. Meenehanfor technical assistanceand the has enabled us to use this method for a long Director-Generalof National Health for permis- series of transmission tests to investigate the sion to publish. effects of temperature on the vector competence of Cx. uniuittatuswith the 2 viruses(Cornel and Jupp, unpublished data). More recently, we suc- REFERENCES CITED cessfully used this artificial method to evaluate Aitken, T. H. 1977.An in uitro feeding technique for the competenceof severalmosquitoes as vectors artificially demonstrating virus transmission by of RVF virus (Jupp and Cornel 1988). mosquitoes.Mosq. News37:130-133. The primary disadvantage of the capillary Beaty, B. J. and T. H. G. Aitken. 1979. In uitro method is that the mosquitoes die after the transmission of yellow fever virus by geographic transmission feed due to the prior removal of strains of Aedesaegypti. Mosq. News 39:232-238. some of their legs. This makes them no longer Gubler, D. J. and L. Rosen. 1976.A simple technique availablefor subsequenttransmission tests. Sec- for demonstrating transmission of Denguevirus by without ondly, it was found necessary mosquitoes the use of vertebratehosts. Am. to test the capil- J. Trop. Med. Hyg. 25:146-150. lary fluid for virus immediately after the mos- Jupp, P. G.1974. Laboratory studieson the transmis- quito had fed since no virus could be isolated sion of West Nile virus by Culex (Culex)uniuittatw after a few days storageat -70"C. Perhaps this Theobald; factors influencing the transmission rate. problem could be overcome by adding a virus J. Med. Entomol. 11:455-458. preservativesuch as dimethyl sulphoxide to the Jupp,P. G. andA. J. Cornel.1988. Vectorcompetence fluids before they are stored or by storing them tests with Rift Valley fever virus and five South at a lower temperature in liquid nitrogen. Ad- African speciesof mosquito. J. Am. Mosq. Control vantagesof the method are that determinations Assoc.4:4-8. of the transmission rate Jupp, P. G., B. M. Mclntosh, I. dos Santosand P. De can be made without Moor. 1981.Laboratory vector studies on six mos- the use of animals and can provide transmission quito and one tick specieswith chikungr.rnyavirus. data for speciesof mosquitoes which will not Trans. R. Soc.Trop. Med. Hyg. 75:15-19. readily refeed on laboratory animals. Further- Kokernot, R. H. and B. M. Mclntosh. 1959.Isolation more, the method is a convenient one and the of West Nile virus from a naturally infected human capillary fluid can be readily titrated in Vero being and from a bird, SyLuiettarufescens (Veillot). cells to determine the concentrations of virus S. Afr. Med. J. 33:987-989. secretedby individual mosquitoes. Mclntosh, B. M., P. G. Jupp, I. dos Santosand B. J. Vector Gubler and Rosen (1976) used the droplet H. Barnard. 1980. studies on Rift Valley fever in South Africa. S. Afr. Med. J.58:127-132. method successfullyfor transmission of dengue- Reed,L. J. and H. Muench. 1938.A simple method of 2 virus by Aedes albopictus,but in view of our estimating fifty per cent endpoints. Am. J. Hyg. very low transmission rate of SIN virus by Cr. 27:493-497. uniuittatus, this method should be used with Weinbren,M. P. 1955.The occurrenceof West Nile caution. Certainly it would be advisableto eval- virus in South Africa. S. Afr. Med. J.29:1092-1097.