Conditions Suitable for Vaccine Development

INFECTION AND IMMUNITY, Nov. 1976, p. 11-17 Copyright 1976 American Society for Microbiology Vol. 14, No. 5 Printed in U.S.A. Isolation of a Temperature-Sensitive Dengue- Virus Under Conditions Suitable for Vaccine Development KENNETH H. ECKELS,* WALTER E. BRANDT, VENTON R. HARRISON, JACK M. McCOWN, AND PHILIP K. RUSSELL Division of Communicable Diseases and Immunology, Walter Reed Army Institute of Research, Washington, D.C. 001 Received for publication 1 May 1976 Dengue virus, type, in viremic human sera and after passage in cell cultures produces mixtures of small and large plaques when assayed in LLC-MK cells. Clones of dengue virus type obtained by plaque selection in primary green monkey kidney cell cultures were tested for temperature sensitivity in vitro and for virulence by intracerebral inoculation of suckling mice. Sublines of a smallplaque clone were found to have lower nonpermissive temperatures than the parent virus by both plaque formation and release of infectious virus into the culture media. Small-plaque sublines were significantly less virulent in suckling mice than was the parent virus. Sublines from a large-plaque clone were not temperature sensitive and closely resembled parent virus mixed-plaque morphology. When small-plaque sublines were serially passaged using undiluted inocula, reversion occurred as evidenced by the appearance of large plaques and return of mouse virulence. Small-plaque virus could be maintained through several serial passages without reversion by using low-input inocula. Desirable passage history as well as temperature-sensitive and attenuation characteristics of the S-1 small-plaque subline make it appear suitable as a vaccine candidate virus. Passage in cell culture has resulted in reduced in vivo virulence for numerous flaviviruses. These include yellow fever (11), West Nile, and Russian spring-summer encephalitis (8), Central European tick-borne encephalitis (4), Japanese encephalitis (3), and Kyasanur forest (7) viruses. Plaque selection has been used as a final step in the purification of many of these viruses; however, plaque size and virulence have been related only rarely. Mayer (4) described a small-plaque variant of Central European tick-borne encephalitis virus obtained from a persistently infected human amnion cell culture. The small-plaque variant had reduced virulence for mice by the peripheral route when it was compared with the parent virus. Paul (7) obtained a spontaneous variant of Kyasanur forest virus by repeated passage in monkey kidney epithelial cell cultures. In addition to small-plaque and lesser in vivo virulence markers, replication of the variant was restricted at 40 C, a permissive temperature for the parent virus. Mixed populations of large and small plaques have been frequently found in wild dengue virus type (DEN-) isolates from humans and mosquitoes in both Asia and the Caribbean region. Previous studies with DEN- demonstrated that large- and small-plaque clones could be readily derived from isolates in first cell culture passage; the clones were identical by neutralization tests, but the small-plaque clone had a markedly reduced virulence for suckling mice (P. K. Russell and P. Sukhavachana, unpublished data). Many naturally occurring and induced temperature-sensitive mutants of other types of viruses have been associated with lower animal virulence (). Indeed, a lower permissive temperature of replication than that of the parent virus has been associated with some viruses currently in use (1) and which have been proposed for use (6, 1, 13) as live-virus vaccines. This paper describes the isolation and purification of small- and large-plaque clones from a 11 human serum isolate of DEN- virus. Plaque size is correlated with temperature sensitivity and with suckling mouse virulence for these clones. MATERIALS AND METHODS Virus. DEN- virus, PR-159 strain, was recovered from serum obtained in Puerto Rico on 19 August 1969 from a 31-year-old male patient with uncomplicated dengue fever. The serum was free of bacteria on culture, and radioimmunoassay for hepatitis B

1 ECKELS ET AL. surface antigen was negative. The virus was identified as DEN- by plaque reduction neutralization tests using reference antisera and was found to be antigenically identical to the PR-109 strain from the same epidemic (9). Cells. Primary green monkey kidney (PGMK) cells were obtained from Lederle Laboratories, Pearl River, N.Y., as frozen suspensions. Lots of cells were tested by the supplier for the presence of simian adventitious agents by inoculation of growth fluids in PGMK, Lederle 130 human diploid cells, and BSC-1 cells. Hemadsorption tests using guinea pig erythrocytes were also performed on monolayers of the pretested PGMK cells prior to their certification. Certified PGMK cells were grown in Eagle minimal essential medium containing 5 mm HEPES (N--hydroxyethylpiperazine-N'--ethanesulfonic acid) buffer (Grand Island Biological Co., Grand Island, N.Y.) supplemented with 10% fetal bovine serum (Microbiological Associates, Bethesda, Md.), mm L-glutamine, and 100 jg of streptomycin sulfate and 50 ug of neomycin sulfate per ml. All sera used in cell culture media were thoroughly screened for adventitious agents and bacteriophage prior to use. Upon formation of a monolayer, the PGMK cells were used immediately, or a 1: subpassage was made and secondary cell monolayers were used for virus passage and plaquing. Growth medium with or 5% fetal bovine serum was used to maintain confluent cell monolayers for virus passage. A continuous line of rhesus monkey kidney (LLC- MK) cells was grown in medium 199 containing 0% fetal bovine serum. The LLC-MK cell line was subpassaged weekly and used for plaque assay of DEN- virus. Plaque assays. Monolayers of LLC-MK cells in 5-cm plastic flasks were used for plaquing and characterizing DEN- virus clones. After virus inoculation and adsorption for 1 to h, flasks received 7 ml of medium 199 containing 10% fetal bovine serum, 36 mm NaHCO3, 0.0% diethylaminoethyldextran, 0.5% 100x concentrated Eagle basal medium vitamins and amino acids, 1% purified or Noble agar (Difco), and antibiotics. After incubation at a designated temperature for 6 days, flasks received 4 ml of a second overlay containing 1% Noble agar and 1:6,000 neutral red in Hanks balanced salt solution. Plaque flasks were incubated at 35 C for 4 h and overnight at room temperature before plaques were counted. For plaquing in PGMK cell monolayers, the first overlay medium contained Eagle minimal essential medium supplemented with 10% fetal bovine serum, 36 mm NaHCO3, 0.0% diethylaminoethyl-dextran, 1% Noble agar, and antibiotics. After virus inoculation and adsorption for 1 to h, flasks received 7 ml of the first overlay. After 6 days of incubation at 35 C, a second overlay containing 1% Noble agar and 1:6,;000 neutral red in Hanks balanced salt solution was added. Plaques were counted the following day. Viral replication. Adsorption was carried out for 1.5 h at 35 C, using undiluted virus suspensions in 5-cm plastic flasks (Costar, Cooke Engineering, INFECT. IMMUN. Alexandria, Va.) containing x 106 cells. The inoculum was removed by washing the monolayers three times with Hanks balanced salt solution, and incubation was started at the indicated temperatures h after addition of the virus. Samples of cell culture fluids were removed during the experiments, diluted 1:10 in growth medium, and frozen at -70 C. Animal inoculation. One- to two-day-old suckling mice were inoculated intracerebrally with 0.0-ml volumes of diluted virus. The mice were observed for sickness and death for 1 days. RESULTS Plaque purification of DEN- virus. Six passages of the DEN- isolate were made in PGMK cells using undiluted cell culture fluid for the inoculum at each passage. A large stock of seed virus was prepared at the 6th passage. This seed virus is referred to below as the parent virus. The original human serum isolate and the parent seed virus contained similar mixed populations of large (3 to 5 mm)- and small (1 to mm)-plaque-forming virus (Fig. 1). Virus titers gradually increased from 1.7 x 104 to 3.0 x 106 plaque-forming units/ml during the first six passages in PGMK cells. Plaque assays in LLC-MK cells were used throughout this study to characterize the plaque morphology of all DEN- isolates and clones. All clone purification was performed in certified PGMK cells, where plaque morphology was variable and not distinctive. Plaque picking in PGMK cells was therefore necessarily followed by characterization of the selected clones in the LLC-MK plaquing system. Passage history of the clones is summarized in Table 1. Initial plaque purification was performed by locating a single, well-isolated plaque after neutral red staining, removing a plug of agar over the plaque, and growing the isolate in PGMK cells under fluid culture media. After two of these cycles, clones designated 5 and 6 were selected for further purification. Sublines of these two clones at the 10th passage level were carried to the 14th passage level by three cycles of plaque purification, the progeny of one plaque being directly plaqued again under agar for each cycle. Table 1 lists the results of the S-1 subline, one of eight sublines of clone 6, and the L-1 subline, one of five sublines derived from clone 5. Although all of the small-plaque sublines retained their small-plaque morphology, this was not true for the progeny of the largeplaque clone. All five sublines of clone 5 after three cycles of plaque purification contained mixed viral populations similar to the parent, passage 6 virus. Whereas small-plaque populations could not be eliminated from these prepa-

VOL. 14, 1976 TEMPERATURE-SENSITIVE DENGUE- VIRUS 13 HUMAN PARENT ISOLATE FIG. 1. Original DEN- human isolate and passage 6, parent virus, plaqued in LLC-MK cells. rations, the percentage of large plaques ranged from 73 to 85, whereas the parent virus pool contained approximately 70o large-plaque virus. In Fig. the homogeneous character of the S-1 subline at the 14th passage level is compared to the plaque mixture found in the parent preparation and in the b-1 subline preparation. The S-1 subline at the 14th passage level was further purified by direct plaque picking to the 18th passage level. Virus pools prepared at the 18th and 19th passage levels, as listed in Table 1, still retained homogeneous small-plaque populations, with plaques not exceeding a diameter of mm. In vivo virulence of the purified clonal sublines. Small-plaque sublines had reduced intracerebral virulence for suckling mice when they were compared with the parent virus. Virulence ratios listed in Table measure the infectivity in cell culture compared to the infectivity in suckling mice for the small- and large-plaque sublines and parent virus. Although significant differences in virulence were not seen between small- and large-plaque sublines, the smallplaque sublines were clearly less virulent for suckling mice than the parent virus. Also, the Fe, mean survival time for mice inoculated with the small-plaque viruses was at least days greater than for mice inoculated with the parent virus. A higher plaquing efficiency for the small-plaque sublines may account for the lower virulence ratios. However, a longer mean survival time in suckling mice would indicate a real difference in lethality for these viruses as compared with the parent virus. Effects of temperature on plaque formation. The first four small-plaque sublines (S-1 to S-4) at the 14th passage level had plaqueforming titers in LLC-MK cells similar to those of a large-plaque subline (L-1) and the parent virus (6th passage level) when assayed at 35 C; however, only the parent virus and the large-plaque subline formed plaques at 39 C (Table 3). The S-1 subline was selected for the remaining experiments. Plaque formation by this virus was compared with the parent virus at increasing temperatures from 35 to 39 C (Table 4). The nonpermissive temperature for the small-plaque subline was between 38 and 39 C in LLC-MK cells. An effect of increasing temperature on plaque formation by the parent virus was to gradually decrease the size of both

14 ECKELS ET AL. the small and large plaques until (by assumption) only the large plaques were visible at 3900. In a separate test, 4000 was found to be nonpermissive for plaque formation by the parent virus. Effects of temperature on viral replication in cells in fluid culture. The S-1 subline was compared to the parent virus at 35 and 3900 in TABLE 1. Plaque size and titer of DEN-, PR-159 virus after passage and clone selection in PGMK cell culture Plaque Virus sage Pas- LLC-MK size in ml) Titer at (PFU/ 350Ca cells (mm) Human serum 0 1-, 3-5 1 x 103 Parent, passage 6b 6 1-, 3-5 3 x 106 Clone 5c 10 1-, 3-5 1.1 X 104 Clone 6c 10 1-5 x 104 L-1 (from clone 5)d 14 1-, 3-5 9 x 104 S-1 (from clone 6)d 14 1-9 x 105 S-le 18 1- x 105 S-i 19 1-1 x 105 a PFU, Plaque-forming units. b First six passages made using undiluted cell culture fluid as the inoculum for each successive passage. c Clones derived by plaque selection at passages 7 and 9. dclonal sublines derived by plaque selection at passages 11, 1, and 13; five large-plaque sublines derived from clone 5 and eight small-plaque sublines derived from clone 6 in identical fashion. e S-1 subline further plaque purified at passages 15, 16, and 17. A. r;a t/o;f*!.v esx 4 INFECT. IMMUN. LLC-MK cells and at 37, 39, and 400C in PGMK cells (Fig. 3A, B). The parent virus replicated well in the primary cells at 390C, whereas in the continuous cells infected at the same time its replication was substantially depressed at the same temperature. Replication of the S-1 subline in primary cells was significantly depressed at 390C, but in continuous cells replication was totally nonpermissive at 390C. When infected primary cells were incubated at 400C (Fig. 3B), replication of the parent virus was further depressed, whereas replication of the small-plaque subline was totally restricted at this temperature, 10C higher than in the continuous cells. The effect of temperature on the S-1 subline at 10C intervals was tested in LLC-MK cells (Fig. 4). Whereas the nonpermissive temperature was again 390C, there was also a substantial depression of replication from 37 to 38 C. Stability of small-plaque sublines on passage. It was observed that genetic stability of small-plaque sublines on serial passage in PGMK cells was influenced by the multiplicity of infection (MOI). Parallel lines of S-1 and S-6 viruses were passed three times using either diluted inocula (MOI, approximately 10-4) from passage 14 to 17 or undiluted inocula (MOI, approximately 10-1) from passage 14 to 17 (Table 5). With both S-1 and S-6 sublines, serial passage with a relatively high MOI resulted in PARENT.i -. I FIG.. Parent DEN- virus and clonal sublines S-1 and L-1 plaqued in LLC-MK cells.

VOL. 14, 1976 TABLE. TEMPERATURE-SENSITIVE DENGUE- VIRUS 15 Infectivity of selected small- and large-plaque sublines ofden- virus in cell culture and suckling micea Virus Suckling mice log LLC-MK cell culture PFU/ Ratio: LD5/PFU LD,dml ml at 35TC S-1 (passage 14) 3.4 9 x 105 0.003 S- 3.6 1. x 106 0.003 S-3 4.4 9 X 105 0.03 S-4 3.5 1 X 106 0.003 S-5 4.3 6 x 105 0.03 S-6 3.3 1. x 106 0.00 S-7 4.6 1 x 106 0.04 S-8 4.6 1.1 X 106 0.04 X = 0.019 ± 0.018 L-1 (passage 14) 4.3 4 X 105 0.05 L- 4.6 4 x 105 0.10 L-3 4.8 3 x 105 0.1 L-4 4.8 4 X 105 0.16 L-5 3.8 3 x 105 0.0 X = 0.108 ± 0.078 Parent (passage 6) 5.6 3 x 106 0.13 x = 0.168 0.06b a LD54), 50% lethal dose; PFU, plaque-forming units. b Mean + a- LD5,/PFU ratio for five determinations. TABLE 3. Plaque formation by small- and largeplaque sublines and the parent DEN- virus in LLC-MK cells at 35 and 39 C Plaque-forming units/ml Virus 350C 39 C S-i 9.0 X 104 <5 x 10 S- 7.0 x 104 <5 x 10' S-3.9 x 10; <5 x 10' S-4 1.7 x 10; <5 x 101 L-1 9.0 X 104.8 x 104 Parent (passage 6) 1.6 x 10;1 1.7 x 104 TABLE 4. Plaque formation by the S-1 and parent DEN- viruses in LLC-MK cells at increasing temperatures Plaque-forming units/ml Temp ( C) S-i Parent 35 1.7 x 104 1.5 x 105 37 4.5 x 103 1.4 x 105 38 5.0 x 103 7.5 x 104 39 <1.0 x 10'.9 x 104 reversion of the homogeneous small-plaque virus to a population containing large plaques. The appearance of large plaques was accompanied by loss of temperature sensitivity and increased virulence for suckling mice. The smallplaque characteristic as well as reduced mouse virulence and temperature sensitivity could be maintained by passage of diluted virus at a low MOI for at least six passages. Yields of infectious virus produced at 7 days were approximately the same for cells inoculated with either a high or low MOI. In vivo stability of the S-1 subline was tested after one passage in suckling mice. Brain homogenates from mice dying after intracerebral inoculation of the S-1 and parent viruses were assayed for the presence of small and large plaques in LLC-MK cells. Only virus capable of producing small plaques was recovered from mice inoculated with S-1, whereas the parent virus-inoculated mice yielded progeny virus that produced a typical plaque mixture. Tests for adventitious agents in the S-1 virus pool. The S-1 virus pool prepared at the 19th passage level was tested for adventitious bacterial and viral agents. Tests for bacterial and mycoplasmal contaminants were negative after inoculation of the S-1 pool in Trypticase soy, thioglycolate, and mycoplasma broths. Tests for adventitious viral agents were done in adult and suckling mice and in four types of cell cultures. Young adult mice remained healthy after receiving a combination intracerebral and intraperitoneal inoculation of the S-1 subline. Suckling mice inoculated by the same route using neutralized S-1 virus also remained healthy over a period of 3 weeks. Tests after two passages of the S-1 subline at 14-day intervals in WI-38, primary rabbit kidney, primary rhesus monkey kidney, and PGMK cell cultures were also negative for cytopathic or morphological changes in the cell monolayer. DISCUSSION The plaque size variants observed in the unpassaged DEN-, PR-159 isolate and in the passaged parent virus appear to be due to viral

. 16 ECKELS ET AL. INFECT. IMMUN. PARENT 6 6 _35C S-l 35C / a<+ -- --* ~PARENT 5 5 /10 /S- 0-1 39C.a... PARENT. 4 4,. 40C CL 3 3...-------...---...- D... oo El -1 40C 3 4 5 3 4 5 DAY POST INFECTION DAY POST INFECTION FIG. 3. Growth ofden- parent and S-i viruses at 35 and 39 C in LLC-MK cells (A) and at 35, 39, and 400C in PGMK cells (B). 6 36 C suppression of viral replication processes at the body temperature of the host. The degree of suppression at a given temperature is deter- 5 37 C mined in part by the virus and in part by the host cell. We have shown in this study that viral replication is suppressed at different restrictive temperatures in continuous and pri- 4 mary monkey kidney cells. Virulence or attenuation might be a reflection of which cells in a a- 0~ 3 certain host are effectively replicating the vi- 0 38 C rus. Thus, measurement of temperature sensitivity in vitro cannot generally be used to predict attenuation for a given host, but it is a O 39 C practical screening method for mutants or variants in many virus-cell systems. In the case of a DEN- small-plaque clone, temperature sen- 3 4 5 6 sitivity appears to be a very useful laboratory DAY POST INFECTION marker, which is associated thus far with ature on the tenuation. FIG. 4. Effect of increasing temperat replication of the S-i subline in LLC-MKL cells. The small-plaque clonal subline S-1, at the 19th passage level, which has been tested more subpopulations with differing viru temperature sensitivity characterrlence and extensively than other DEN- clones, has sevistics. A eral characteristics that indicate its potential small-plaque clone was derived from the parent as a vaccine candidate. These include temperavirus and multiple sublines were is( purified. Homogeneity of the original[oated and ture sensitivity, reduced neurovirulence for substantiated by the similar chariaclone. was suckling mice, and, in other studies to be re- *acteristics ported (Harrison et al., in preparation), refailure to duced virulence for primates and low neurovir- demonstrated by the sublines. The isolate pure large-plaque clones in s peated attempts strongly pite suggests of re- ulence for rhesus monkeys. In addition, the S-1 a high-frequency mutation from large ponsistento subline is free of detectable adventitious agents plaques to and has only been passaged in cell cultures a mixture of small plaques and large plaques. acceptable for human vaccine development. There appears to be a much-lower. -frequency Passage experiments indicate that the genetic mutation of small-plaque to large-pllaque morpurity of the S-1 subline can be maintained phology. There is a strong positive during serial passage by using a low MOI. corredlation be- ACKNOWLEDGMENTS tween small-plaque morphology, teimperature sensitivity, and reduced suckling mlouse viru- We thank C. Hampton for technical assistance and J. lence. The reason for reduced in vivc ) virulence Lowenthal, S. Berman, and D. Dubois for helpful discus- partial of the small-plaque clones is probalbly sions.

VOL. 14, 1976 TABLE 5. TEMPERATURE-SENSITIVE DENGUE- VIRUS 17 Characteristics of the S-1 and S-6 sublines passaged at relatively high and low MOI in PGMK cells PFU/mla Virus, passage MOI LD5,,I/PFU Plaque morphology at 35 C (P) 35 C 39 C S-1, p-14 9 x 10) ND" 0.003 Small S-1, p-17 10-4 5 x 10) <5 x 10 0.008 Small S-1, p-19 10-4 1 x 10 <5 x 10' 0.040 Small S-1, p-17 10-' 3 x 10` 5 x 103 0.1 Large and small S-6, p-14 x 106 ND 0.0008 Small S-6, p-17 10-4 1 x 10" ND 0.006 Small S-6, p-0 10-4 1 x 10 ND ND Small S-6, p-17 10-1 1 x 10) ND 0. Large and small a PFU, Plaque-forming units. b Suckling mouse intracerebral 50% lethal doses (LD5,,). " ND, Not determined. LITERATURE CITED 1. Food and Drug Administration. 1974. Tests for in vitro markers. Code of Federal Regulations 1:630.16 (b) (3). U.S. Government Printing Office, Washington, D.C.. Ghendon, Y. Z., A. T. Marchenko, S. G. Markushin, D. B. Ghenkina, A. V. Mikhajeva, and E. E. Rozina. 1973. Correlation between ts phenotype and pathogenicity of some animal viruses. Arch. Gesamte Virusforsch. 4:154-159. 3. Hammon, W. McD., S. Rohitaydhin, and J. S. Rhim. 1963. Studies on Japanese B encephalitis virus vaccines from tissue culture. IV. Preparation and characterization of a pool of attenuated Oct-541 line for human vaccine trial. J. Immunol. 91:95-305. 4. Mayer, V. 1963. Two variants of tick-borne encephalitis showing different plaque morphology. Virology 0:37-390. 5. Mills, J. V., and R. M. Chanock. 1971. Temperaturesensitive mutants of influenza virus. I. Behavior in tissue culture and in experimental animals. J. Infect. Dis. 13:145-159. 6. Mills, J. V., I. van Kirk, D. A. Hill, and R. M. Chanock. 1969. Evaluation of influenza virus mutants for possible use in a live virus vaccine. Bull. W.H.O. 41:599-606. 7. Paul, S. D. 1966. Some biological properties of two variants of Kyasanur Forest disease virus. Indian J. Med. Res. 54:419-44. 8. Price, W. H., R. W. Lee, W. F. Gunkel, and W. O'Leary. 1963. The virulence of West Nile virus and TP 1 virus and their application to a group B arbovirus vaccine. Am. J. Trop. Med. Hyg. 10:403-4. 9. Russell, P. K., and J. M. McCown. 197. Comparison of dengue- and dengue-3 virus strains by neutralization tests and identification of a subtype of dengue-3. Am. J. Trop. Med. Hyg. 1:97-99. 10. Schluter, B., B. Bellomy, and A. Brown. 1974. Pathogenesis of temperature sensitive mutants of Sindbis virus in the embryonated egg. I. Characterization and kinetics of viral multiplication. Infect. Immun. 9:68-75. 11. Theiler, M. 1951. The virus, p. 39-136. In G. K. Strode (ed.), Yellow fever. McGraw-Hill, New York. 1. Wagner, R. R. 1974. Pathogenicity and immunogenicity for mice of temperature-sensitive mutants of vesicular stomatitis virus. Infect. Immun. 10:309-315. 13. Wright, P. F., W. G. Woodend, and R. M. Chanock. 1970. Temperature-sensitive mutants of respiratory syncytial virus: in-vivo studies in hamsters. J. Infect. Dis. 1:501-51.