Comparative sensitivity of three mosquito cell lines for isolation of dengue viruses*

Bulletin ofthe World Health Organization, 63 (2): 279-286 (1985) © World Health Organization 1985 Comparative sensitivity of three mosquito cell lines for isolation of dengue viruses* G. KUNO,1 D. J. GUBLER,2 M. VaLEZ,3 & A. OLIVER4 Comparative studies were carried out on three mosquito cell lines (C6/36 clone of Aedes albopictus, AP-61 from A. pseudoscutellaris, and TRA-284 from Toxorhynchites amboinensis) to determine their sensitivity to dengue virus isolation, growth, and handling characteristicsfor immunofluorescent testing. Virus isolation ratesfrom human sera were the highest in the TRA-284-SF (a line adapted to serum-free medium), followed by the TRA-284 parental line andAP-61. Virus isolation was the lowest in the C6/36 line. All3 cell lines were comparable in terms of ease of handling, but C6/36 cells were preferable for detecting infected cells by the direct fluorescent antibody test (DFA T) because offrequent cell clumping in the AP-61 and TRA-284 lines. Early detection of viral antigen of all 4 serotypes in the infected cells byDFA T was dependent upon the virus titre in the serum. The AP-61 and TRA-284-SF cells were the best for early detection and identification of viral antigen. Similarly, both AP-61 and TRA-284 cells were more resistant than C6/36 cells to toxic effects ofhuman sera. Based on the economy of using the serum-free medium, their higher sensitivity for dengue virus isolation, and their ease of handling, it is recommended that the TRA-284-SF cell line be usedfor routine dengue virus isolation in laboratories with cell culture capability. The high sensitivity of the C6/36 clone of Aedes albopictus cells to dengue viruses has been well documented in recent years (1-3). These cells provide a simple and rapid method for dengue virus isolation and, together with the use of serotype-specific monoclonal antibodies for identification, provide the basis for an effective and economical method for dengue virus surveillance (3). In studies with this cell line, however, we observed that some dengue viruses did not grow well (Kuno, G. & Gubler, D. J., unpub- lished data). We therefore studied two other mosquito cell lines to determine their sensitivity to dengue virus infection. This report describes a com- parative study on the isolation of dengue viruses from human sera using three mosquito cell lines -the C6/36 clone of A. albopictus, AP-61 from A. pseudo- scutellaris, and TRA-284 from Toxorhynchites * From the San Juan Laboratories (Dengue Branch), Division of Vector-borne Viral Diseases, Center for Infectious Diseases, Centers for Disease Control, GPO Box 4532, San Juan, Puerto Rico 00936, USA. ' Research Microbiologist. Requests for reprints should be sent to this author. 2 Chief, Dengue Branch. 3 Medical Technologist. 4 Biological Technician. amboinensis. Virus growth characteristics, the ease with which the cells were processed for immuno- fluorescent testing, and other advantageous traits for the selection of the most suitable cell line for viro- logical surveillance of dengue are described. MATERIALS AND METHODS Cell culture The C6/36 cells were maintained in plastic flasks with Eagle's minimal essential medium (MEM), pH 7.0, containing 10% heat-inactivated fetal bovine serum (FBS), 0.2 mmol/l non-essential amino acids, and 0.2 mmol/l L-glutamine, which was a minor modification of the medium used by Igarashi (1). The AP-61 cells were maintained in 6-oz (170-ml) bot- tles with Mitsuhashi-Maramorosch/Varma-Pudney (MM/VP12) medium, pH 7.0 (4, 5), and the TRA- 284 cells were maintained in plastic flasks with the MM/VP12 medium (6). A subline of the TRA-284 cell line, adapted to a serum-free medium, pH 7.2, that consisted of equal volumes of Leibovitz (L-15) medium and tryptose phosphate broth (TPB), was designated as TRA-284-SF and also maintained in plastic flasks (6). 4524 -279 G. KUNO ET AL. Comparative virus isolation The study to evaluate the sensitivity of the three cell lines for virus isolation was carried out in two experiments. Initially, the TRA-284 cells could not be grown in the glass culture tubes. In the first experi- ment, therefore, comparative tests for virus isolation were conducted with the cultures of TRA-284 cells grown in flasks with 5 ml of growth medium. The C6/36 and AP-61 cells were grown in tubes (16 x 125 mm) with screw-caps in 2 ml of growth medium. Monolayer cultures of the TRA-284 cells were prepared by seeding 4 million cells per flask, whereas monolayers of the C6/36 and AP-61 cells were prepared by seeding 1 million cells per culture tube held in a slant rack. All cultures were incubated for 3 days at 28 'C. In the second experiment, the subline of the TRA- 284 cells adapted to a serum-free medium (TRA-284- SF) was used. This subline was adapted to grow in the screw-cap glass culture tubes (16 x 125 mm), thus permitting comparison of all three cell lines grown under identical conditions. Because of slower growth, the monolayer cultures of the TRA-284-SF cells were prepared by seeding 3 million cells per tube and incu- bating the cultures for 3 days at 28 'C. Cultures of all three cell lines were processed for virus isolation as previously described (3). Briefly, 0.05 ml aliquots of acute-phase human sera were inoculated into tube cultures without removing any growth medium, and into flask cultures after removing 3 ml of growth medium. After adsorption for 1 h at 28 IC, 1 ml of maintenance medium was added to the tube cultures, and 3 ml was added to flask cultures. The inoculated cultures were incubated at 28 'C for 10 days. Temporal growth ofdengue viruses Twenty human sera, from which dengue viruses of all 4 serotypes had been previously isolated, were inoculated into 5 tube cultures each of the three cell lines as described above. The inoculated cultures were maintained at 28 'C, and one tube culture of each cell line was processed for infection by the direct fluores- cent antibody test (DFAT) on days 1, 3, 5, 7, and 9, after inoculation. At the time of inoculation, each serum was titrated by inoculating serial 10-fold dilu- tions into 5 tube cultures of the C6/36 cells. After 7 days of incubation, the cells in each tube were pro- cessed for DFAT. Virus titre was expressed at the 50%0 tissue-culture infectious dose (TCID5o) per ml in the C6/36 cell cultures, calculated by the method of Reed & Muench (7). Immunofluorescence techniques After incubation, cells were harvested and spotted on 12-well Teflon-coated slides (Cel-line Associates, Newfield, NJ). Virus infection in the cells was deter- mined by the direct fluorescent antibody test using a fluorescein isothiocyanate (FITC) conjugate pre- pared from pooled human sera with high haemag- glutination inhibition (HI) antibody titres () 5120). Virus identification was by the indirect fluorescent antibody test (IFAT) using serotype-specific mono- clonal antibodies and an antimouse IgG-FITC con- jugate (Antibodies Incorporated, Davis, CA) (3). Human sera All sera used in the comparative isolation tests were from patients in Puerto Rico who had a dengue-like illness or a febrile illness of unknown origin; serum specimens were collected in 1982 as part of the dengue surveillance system of the San Juan Laboratories. All specimens were processed for virus isolation without regard to the serological result (3). For the temporal growth studies, the DEN-2 and DEN-3 sera were from dengue haemorrhagic fever patients in Indonesia, while DEN-1 and DEN-4 sera were collected in Puerto Rico. Table 1. Comparison of the frequency of dengue virus isolations in the three mosquito cell linesa Cell line Experiment 1 Experiment 2 Totalb C6/36 29/100c (29)d 14/55 (25) 43/155 (28) AP-61 34/100 (34) 17/55 (31) 51/155 (33) TRA-284 33/100 (33) - - 33/100 (33) TRA-284-SF - - 20/55 (36) 20/55 (36) a All cells were grown in tube cultures except the TRA-284, which was grown in flasks. b All isolates were dengue 1 and dengue 4. c Number of isolates/number tested. d Figures in parentheses are percentages. 280 MOSQUITO CELL LINES FOR DENGUE VIRUS ISOLATION RESULTS Cell line characteristics for growth and comparative virus isolation and identification All cell lines (C6/36, AP-61, and TRA-284-SF) were judged equal with respect to the ease of handling and preparation of tube cultures. The slower growth of the TRA-284-SF cells was compensated for by seeding the culture tubes with 3 million cells of this line. Infected cells could easily be observed by DFAT in all 3 cell lines. While cytopathic effects were more often observed in the AP-61 cells, we found this to be an unreliable method of detecting virus infection. The AP-61 cells were more difficult to read by DFAT because there tended to be more disrupted cells. This may be one reason why attachment of the AP-61 cells on spot slides was inferior to that with both C6/36 and the TRA-284 lines. In the first experiment, 100 acute sera were inocu- lated into the C6/36, AP-61, and TRA-284 cell lines simultaneously. Results are presented in Table 1. Dengue virus isolation was similar in the AP-61 cells (34 isolates) and the TRA-284 cells (33 isolates). Only 29 isolates were made in the C6/36 cells. Statistical analysis (Cochran's Q-test), however, revealed that there was no significant difference in virus isolation rate among the three cell lines (Q = 4.35; P > 0.05). All viruses were identified as DEN 1 or 4. There were 6 viruses that could not be identified in the initial cultures because few cells were infected and because there was insufficient amount of viral antigen to bind the monoclonal antibodies; 4 of these were in the C6/36 cells, and 2 in the AP-61 cells. All the viruses isolated in the TRA-284 cells were identified in the initial cultures. In this experiment, 33 out of 100 human sera were toxic to the C6/36 cells, resulting in detachment of patches of cells or death of a large number of cells. Only 5 of these sera were toxic to the TRA-284 cells destroying approximately 10-40% of the cells; all 5 were haemolysed. AP-61 cells were the most resistant, since none of the sera were toxic to these cells. In the second experiment of the virus isolation study, all cell lines were grown in tube cultures. Of 55 sera tested, 20 isolates were obtained with the TRA- 284-SF cells, compared with 17 in the AP-61 and 14 in the C6/36 cells (Table 1). All isolates were identified as DEN 4. The virus isolation rate in the TRA-284-SF cell line was significantly higher than that in the C6/36 cell line, as revealed by Cochran's Q-test (Q= 6.75; P < 0.05) and by MacNemar's chi-square test (X2 = 4.2; P < 0.05). Overall, the dengue virus isolation rate was highest in the TRA-284-SF cells (360o), followed by the parental TRA-284 line and the AP-61 cells (330/o). The lowest isolation rate was in the C6/36 cells (2807o). The superiority of the TRA- 284-F cell cultures was further underscored in the second experiment with 3 viruses isolated only in that cell line. No viruses were isolated only in the AP-61 or C6/36 cell cultures. To economize further the cost of cell culture, a third experiment was carried out to test the possibility of growing mosquito cells in 1-dram (3.5-ml) vials for virus isolation. In addition to tube cultures, the TRA- 284-SF cells were grown in the 1-dram vials by seeding each vial with one million cells in 2.0 ml of medium. Both tube and vial cultures were inoculated with 72 sera and processed for virus isolation as described above. Tube cultures were superior to the vial cultures with 24 and 15 viruses isolated respectively. Again, all isolates were DEN 4. In the vial cultures, patches of cells were often detached from the glass, and localized cell death was evident during the 10-day incubation. Minimum incubation periodfor virus detection and identification This study was designed to evaluate the minimum incubation period (MIP) necessary for virus isolation using the 3 cell lines. Five sera containing DEN- 1 virus were inoculated into all 3 cell lines simultaneously, and cultures were harvested on the days shown in Table 2. In 4 of the 5 sera tested, viruses were detected in the C6/36 cells one day after inoculation, and all 5 viruses were detected in the 3 cell lines by day 3 after inoculation. While the percentage of infected cells increased to more than 600o in both the AP-61 and TRA-284-SF cell lines by day 9, it remained around 2%7o or less and did not increase noticeably in the C6/36 cells with 3 of the 5 sera. The virus titres of the sera inoculated ranged from 102.7 to 104.8 TCIDso/ml. Viruses were isolated from all 5 sera containing DEN-2 viruses in the AP-61 and TRA-284-SF cells, whereas only 3 viruses could be isolated in the C6/36 cells (Table 3). Furthermore, viruses were detected earlier in the AP-61 and TRA-284-SF cells than in the C6/36 cells. When the titre of virus was above 102 TCID50 per ml, viruses were generally detected and identified by day 5 in the AP-61 or TRA-284-SF cells, whereas 7-day incubation was necessary when virus titres were below 102 TCID50 per ml. Except for one virus which was detected on day 5, DEN-2 viruses either did not replicate in the C6/36 ceIls or could be detected only on day 9 after inoculation. As with DEN-1 infection, the percentage of infected C6/36 cells by one virus (specimen IN-4870) did not signifi- cantly increase during the 9-day incubation period. The results with sera containing DEN-3 virus were similar to those for DEN 2. Thus, when the virus titre was below 102 TCID50/ml, viruses replicated only in the AP-61 and TRA-284-SF cells, and were detected by days 7-9 after inoculation (Table 4). However, 281 G. KUNO ET AL. Table 2. Temporal study of dengue-1 virus replication in the three mosquito cell lines Immunofluorescence scorea Minimum incubation (days afternoculation)period necessary forVirus Titre (days after inoculation) virus detection specimen (log TCIDso/mI) Cell line 1 3 5 7 9 (days) PR-23334 4.8 C6/36 (+) + + + + 1 (or less) AP-61 - + + + + 3 TRA-284-SF - (+) + + + 3 PR-25850 3.7 C6/36 (+) 1+) + I+) (+) 1 AP-61 (+1 + + + + 1 TRA-284-SF - 1+ + + + 3 PR-25872 4.7 C6/36 1+1 1+) (+1 1+) (+) 1 AP-61 - (+) + + + 3 TRA-284-SF - (+ + + + 3 PR-25960 2.7 C6/36 - 1+1 (+) (+) + 3 AP-61 - + + + + 3 TRA-284-SF - (+) + + + 3 PR-28032 4.7 C6/36 1+) + + + + 1 AP-61 - + + + + 3 TRA-284-SF - + + + + 3 Scoring system: -, no infection; (+), less than 2% cells infected; +, many cells infected. Table 3. Temporal study of dengue-2 virus replication in the three mosquito cell lines Immunofluorescence scorea Minimum incubation (days afternoculation)period necessary forVirus Titre (days after inoculation) virus detection specimen (log TCIDso/ml) Cell line 1 3 5 7 9 (days) IN-8485 < 2.0 C6/36 - - - - - NAb AP-61 - - - (+) - 7 (or less) TRA-284-SF - - - + - 7 IN-7214 < 2.0 C6/36 - - - - - NA" AP-61 - - - + + 7 TRA-284-SF - - - + + 7 IN-4870 2.5 C6/36 - - (+) (+) (+) 5 AP-61 - 1+) + + + 3 TRA-284-SF - - (+ + + 5 IN-5138 2.9 C6/36 - - - - 1+) 9 AP-61 - - + + + 5 TRA-284-SF - - 1+) + 1+) 5 IN-5226 < 2.0 C6/36 - - - - (+1 9 AP-61 - - + + 1+) 5 TRA-284-SF - - - + + 7 a Scoring system: -, no infection; (+), less than 2% cells infected; +, many cells infected. h NA: not applicable. 282 MOSQUITO CELL LINES FOR DENGUE VIRUS ISOLATION Table 4. Temporal study of dengue-3 virus replication in the three mosquito cell lines Immunofluorescence score' Minimum incubation Virus Titre (days after inoculation) period necessary for specimen (log TCID5o/ml) Cell line 1 3 5 7 9 (days) IN-10023 < 2.0 C6/36 - - - - - NAb AP-61 - - - + - 7 (or less) TRA-284-SF - - - + - 7 IN-9054 < 2.0 C6/36 - - - - - NAb AP-61 - - - - + 9 TRA-284-SF - - - - + 9 IN-9580 3.0 C6/36 - + + + NTc 3 AP-61 - + + + NT 3 TRA-284-SF - + + + NT 3 IN-813 3.4 C6/36 - - (+) (+) + 5 AP-61 - (+) + + + 3 TRA-284-SF (+) (+) + + + 1 IN-839 3.3 C6/36 (+) - (+) (+) + 1 AP-61 (+) + + + + 1 TRA-284-SF (+) (+) + + + 1 a Scoring system: -, no infection; (+ b NA: not applicable. c NT: not tested. ), less than 2% cells infected; +, many cells infected. when the virus titre was more than 103 TCID5o/ml, viruses were detected and identified either on day 1 or 3 after inoculation. Virus was not isolated from the two sera with low titre in the C6/36 cells, whereas isolations were made from all sera with the AP-61 and TRA-284-SF cell lines. Of interest was the serum (specimen IN-10023), which had only a few infected cells in both of these cell lines on day 7, but the day-9 cultures were negative. Four of 5 DEN-4 viruses were isolated in all 3 cell lines (Table 5). In one serum, the virus was detected only in the AP-61 cells 7 days after inoculation. This serum had a virus titre below 102 TCID5o per ml. When the virus titres of sera were higher, the mini- mum incubation period for all cell lines was between 3 and 5 days. DISCUSSION Several continuous cell lines have been derived from both haematophagous and nonhaemato- phagous mosquitos. Three of these have been shown to be highly sensitive for isolation and propagation of dengue viruses (1, 4, 6). However, little is known as to which of these cell lines is best for overall use in a dengue laboratory. This study was designed to answer that question. All three cell lines were comparable with regard to ease of handling and preparation of tube cultures. The increased sensitivity of the TRA-284-SF subline through adaptation to the serum-free medium is exceptional, since the adaptation of the C6/36 and AP-61 cells to the same serum-free medium resulted in decreased sensitivity to either all or some of the dengue serotypes (8). The TRA-284-SF cell line should be advantageous to those laboratories where acquisition of bovine sera of good quality is either expensive or difficult. A previous study showed that many human sera were more toxic to the C6/36 cells than to the TRA- 284 cells, when the volume ratio of human sera and growth medium was 1/50 (6). Varma et al. could inoculate as much as 0.4 ml of human serum per 4 ml of growth medium/flask (volume ratio of 1/10) into cultures of AP-61 cells without significant toxicity (4). In the present study, the C6/36 cells were the most susceptible to serum toxicity. A similar result was obtained elsewhere (9). The higher tolerance of 283 G. KUNO ET AL. Table 5. Temporal study of dengue-4 virus replication in the three mosquito cell lines Immunofluorescence score' Minimum incubation period necessary for Virus Titre (days after inoculation) virus detection specimen (log TCID5o/ml) Cell line 1 3 5 7 9 (days) PR-62860 4.7 C6/36 - (+) + + + 3 (or less) AP-61 - + + + + 3 TRA-284-SF - - + + + 5 PR-63066 3.7 C6/36 - - + + + 5 AP-61 - + + + + 3 TRA-284-SF - + + + + 3 PR-63069 3.7 C6/36 - - + - - 5 AP-61 - - + + + 5 TRA-284-SF - - + + + 5 PR-63522 3.7 C6/36 - - + + + 5 AP-61 - - + + + 5 TRA-284-SF - - + + + 5 PR-64470 < 2.0 C6/36 - - - - - NAb AP-61 - - - + - 7 TRA-284-SF - - - - - NA " Scoring system: -, no infection; b NA: not applicable. (+ ), less than 2% cells infected; +, many cells infected. the AP-61 and TRA-284-SF cell cultures is advan- tageous when virus isolation from specimens with a very low virus titre is attempted, because larger volumes of sera can be inoculated undiluted without adverse effects on the cells. The results of this study showed that both the parental TRA-284 cell line and the TRA-284-SF subline were equal to, or more sensitive than, the AP- 61 cell line for isolation of dengue viruses from human sera. The C6/36 cells were consistently less sensitive than the other two cell lines. Tesh found the C6/36 cells slightly more sensitive than the AP-61 cells, but only 6 viruses were compared (2). The results of our study, however, are consistent with previous reports, which have shown the C6/36 clone ofAedes albopictus cells to be inferior to either TRA- 284-SF or AP-61 cells (6, 9, 10). It is possible that a subpopulation of the C6/36 cells with lower sensi- tivity to dengue viruses was selected in our laboratory with increased passage, since an earlier study showed that the C6/36 and AP-61 cells were equally sensitive to dengue viruses (11). A similar decrease in sensi- tivity of the C6/36 cells was observed in another laboratory (D. Trent, personal communication, 1983). It appears likely that changes in the cloned C6/36 cells have occurred because of differences in manipulation procedures or media in various labora- tories. The minimum incubation period necessary for detecting dengue virus in cell culture was, in general, a function of the virus titre in the original serum speci- men. Thus, when titres of virus inoculated were more than 103 TCID5o per ml, most of the viruses were detected in about 3 days, but sometimes as early as one day after inoculation. When the titre of virus inoculated was 102 TCID5o per ml or less, however, viruses were usually not detected until day 7 after inoculation. The MIP for DEN-4 viruses was longer than that for DEN 1 or DEN 3, when specimens with comparable virus titres were used. The MIP for DEN- 2 viruses was generally 5 days or longer, which reflected low virus titres in those serum specimens. Similarly, lack of growth and spread of virus, as happened with all 4 serotypes, appeared to be associated with low virus titres in the serum. This phenomenon was more pronounced with DEN- 1 viruses in the C6/36 cells where infection was detected as early as 24 hours after inoculation, but many viruses failed to spread to other cells. It is not known whether this was due to the characteristics of the virus strains, or of the cells, or a combination of both. Both the AP-61 and TRA-284-SF lines had a higher 284 MOSQUITO CELL LINES FOR DENGUE VIRUS ISOLATION 285 proportion of infected cells than the C6/36 line. An earlier study, on the other hand, showed that the per- centage of infected cells was higher in C6/36 cell cultures than in AP-61 cell cultures (2). The C6/36 cells were generally best in terms of uniform cell dispersion on spot slides. Suspensions of the AP-61, and to some extent the TRA-284 cells, tended to clump despite repeated aspiration to break them up. Furthermore, detachment of the AP-61 cells from the slides was a problem during the immuno- fluorescence procedures. The cause of cell detach- ment of the AP-61 cells experienced in this study is not known. No such problem was reported in earlier studies in which clean glass slides (2) or Lab-Teka b slides (10) were used. It has been suggested that the lack of L-15 medium in the AP-61 cell cultures might be a source of the problem (M. G. R. Varma, personal communication, 1983). Multi-well culture vessels, such as Lab-Tek slides and plastic plates, have been used for virus isolation, but the major drawbacks are expense and the need for a CO2 incubator. Culture vessels with a screw cap, on ° Miles Laboratories, Naperville, IL, USA. b Use of trade names is for identification only and does not constitute the endorsement by the Public Health Service or by the US Department of Health and Human Services. the other hand, do not require CO2 and are economi- cal. Although they are also economical, the one-dram vials with screw caps were not as effective as the tube cultures for virus isolation. The lower virus isolation rate in the vials may have been due to the small space for gas exchange and the small volume of growth medium, both of which could have contributed to more rapid deterioration of the cell cultures. On the basis of the results of this study, we have selected the TRA-284-SF cells, grown in a serum-free medium, for routine virus isolation in our laboratory. Mosquito cell cultures, although they are slightly less sensitive than intrathoracic inoculation of adult mosquitos for isolation of dengue viruses (10, 12), are suitable for routine virological surveillance because of the ease of processing large numbers of specimens for virus isolation in a short period of time (3). This is an important factor for consideration in those laboratories where maintenance of a mosquito colony is difficult or expertise in mosquito inoculation is not readily available. Furthermore, the use of the serum- free medium with the TRA-284-SF cells offers advantages for those laboratories where animal sera for cell culture are either too expensive or difficult to obtain. ACKNOWLEDGEMENTS We appreciate the assistance in statistical analysis given by Mr Raymond E. Bailey of the Centers for Disease Control, Fort Collins, CO, USA. RESUME SENSIBILITE COMPAREE DE TROIS LIGNEES CELLULAIRES DE MOUSTIQUES POUR L'ISOLEMENT DES VIRUS DE LA DENGUE Trois lignees cellulaires de moustiques que l'on sait sensibles a l'infection par le virus de la dengue ont et comparees en vue de leur utilisation dans un nouveau programme de surveillance virologique a Porto Rico. Les lignees cellulaires, le clone C6/36 d'Aedes albopictus, la lign6e AP-61 d'A. pseudoscutellaris et les lignees TRA-284 et TRA-284-SF de Toxorhynchites amboinensis ont e cultivees soit en boites de plastique soit en tubes de verre munis de bouchons a vis. Les cultures cellulaires ont e inoculees avec des fractions de 0,05 ml de serum humain de stade aigu non dilue. Apres incubation pendant 10 jours a 28 IC, les cellules ont e remises en suspension, etalees sur des lames et traitees a la recherche d'une infection a flavivirus par immunofluorescence directe. Les virus de la dengue ont ete identifies par immunofluorescence indirecte au moyen d'anticorps monoclonaux sp&cifiques de sero- type. Toutes les lignees cellulaires ont e jugees equiva- lentes en ce qui concerne la facilite de manipulation et de preparation des cultures. Bien que les cellules infectees soient faciles a observer en immunofluorescence directe dans toutes les lignees cellulaires, la fluorescence etait plus difficile a lire dans les cellules de la lignee AP-61, qui avaient tendance a etre plus souvent elatees. De plus, de nom- breuses cellules de cette lignee tendaient a se detacher de la lame au cours de l'epreuve d'immunofluorescence. Apres inoculation de 100 serums de stade aigu, on a isole 34 virus 286 G. KUNO ET AL. dans la lignee AP-61, 33 virus dans la lignee TRA-284 et seulement 29 virus dans les cellules C6/36. Tous les virus isoles dans les cellules TRA-284 ont et identifies par immunofluorescence indirecte a l'aide d'antieorps mono- clonaux. En revanche, quatre souches isolees sur les cellules C6/36 et deux souches isol6es sur les cellules AP-61 n'ont pu etre identifiees dans la suspension cellulaire d'origine. Toutefois, apres un seul passage, on a pu parvenir a une identification satisfaisante. Une sous-lignee des cellules TRA-284, adapt6e a un milieu exempt de serum (TRA-284- SF), s'est revelee plus sensible que les autres cellules pour l'isolement des virus de la dengue. Ainsi, sur 55 autres serums 6tudies, on a isole 20 souches virales dans la lignee TRA-284-SF contre 17 souches dans les cellules AP-61 et 14 souches dans les cellules C6/36. Trois virus n'ont e isoles que dans les cellules TRA-284-SF. Le temps d'incubation minimal necessaire pour la detection des virus par immuno- fluorescence directe etait en general fonction du titre de virus dans l'echantillon de serum. Ainsi, lorsque le titre de virus inocule etait superieur a 103 DICT50 par ml, la plupart des virus etaient en general deceles dans les trois lignees cellulaires au bout de trois jours, et parfois des le lendemain de l'inoculation. Lorsque le titre du virus inocule etait de 102 DICT50 par ml ou moins, les virus n'etaient en general pas deceles avant le septieme jour suivant l'inoculation. Alors que la proportion de cellules infectees augmentait en general progressivement pendant la periode d'incubation dans les cultures cellulaires AP-61, TRA-284 et TRA-284-SF, tel n'etait pas toujours le cas avec les cellules C6/36. Avec certaines souches de virus DEN- 1, on n'observait pas de propagation dans les cellules C6/36, d'ou la difficulte d'identifier ces virus par immunofluorescence indirecte. Compte tenu de la facilite de manipulation des cultures cellulaires, de leur sensibilite a l'infection virale, de la facilite de l'identification par immunofluorescence indirecte et du prix de revient des epreuves nous avons choisi les cultures cellulaires TRA-284-SF pour la surveillance virologique de la dengue en routine a Porto Rico. REFERENCES 1. IGARASHI, A. Isolation of Singh's Aedes albopictus cell clone sensitive to dengue and chikungunya viruses, Journal ofgeneral virology, 40: 530-544 (1978). 2. TESH, R. B. A method for the isolation and identifi- cation of dengue viruses using mosquito cell cultures, American journal oftropical medicine and hygiene, 28: 1053-1059 (1979). 3. GUBLER, D. J. ET AL. Use of mosquito cell cultures and specific nmonoclonal antibodies for routine surveillance of dengue viruses. American journal of tropical medicine and hygiene, 33: 158-165 (1984). 4. VARMA, M. G. R. ETAL. Cell lines from larvae of Aedes (Stegomyia) malayensis Colless, and Aedes (S.) pseudo- scutellaris (Theobald) and their infection with some arboviruses. Transactions of the Royal Society of Tropical Medicine and Hygiene, 68: 374-382 (1974). 5. VARMA, M. G. R. & PUDNEY, M. The growth and serial passage of cell lines from Aedes aegypti (L.) larvae in different media. Journal of medical entomology, 6: 432-439 (1969). 6. KUNO, G. Dengue virus replication in a polyploid mosquito cell culture grown in serum-free medium. Journal of clinical microbiology, 16: 851-855 (1982). 7. REED, L. J. & MUENCH, H. A simple method of esti- mating fifty per cent end-points. American journal of hygiene, 27: 493-497 (1938). 8. KUNO, G. Cultivation of mosquito cell lines in serum- free media and their effects on dengue virus replication. In Vitro, 19: 707-713 (1983). 9. LEAKE, C. J. ET AL. Comparative isolation of dengue viruses from DHF patients by mosquito inoculation and on three mosquito cell lines. In: Pang, T. & Pathmanathan, R., ed. Proceedings ofthe International Conference on Dengue Fever and Dengue Haemor- rhagic Fever, Kuala Lumpur, Malaysia, 1-3 September 1983. Kuala Lumpur, 1983, pp. 437-445. 10. DE MATTOS, 1. ET AL. Comparative infectivity assays of dengue virus in mosquitoes and mosquito cell lines. In: Kurstak, E. et al., ed., Invertebrate systems in vitro, Amsterdam, North Holland, 1980, pp. 341-346. 11. KUNO, G. Replication of dengue, yellow fever, St. Louis encephalitis and vesicular stomatitis viruses in a cell line (TRA-171) derived from Toxorhynchites amboinensis. In Vitro, 17: 1011-1015 (1981). 12. KUNO, G. ET AL. Factors affecting syncytial develop- ment in Aedes pseudoscutellaris cells by dengue viruses. American journal of tropical medicine and hygiene, 30: 870-875 (1981).