Evaluation of Macaca radiata as a non-human primate model of Dengue virus infection

Dengue virus (DENV) causes a wide range of illnesses in humans, including dengue fever and dengue haemorrhagic fever. Current animal models of DENV infection are limited for understanding infectious diseases in humans. Bonnet monkeys (Macaca radiata), a type of Old World monkey, have been used to study experimental and natural infections by flaviviruses, but Old World monkeys have not yet been used as DENV infection models. In this study, the replication levels of several DENV strains were evaluated using peripheral blood mononuclear cells. Our findings indicated that DENV-4 09-48 strain, isolated from a traveller returning from India in 2009, was a highly replicative virus. Three bonnet monkeys were infected with 09-48 strain and antibody responses were assessed. DENV nonstructural protein 1 antigen was detected and high viraemia was observed. These results indicated that bonnet monkeys and 09-48 strain could be used as a reliable primate model for the study of DENV.

DENV infection would provide a powerful tool for understanding the mechanisms of DENV-induced pathogenesis and for the development of antiviral drugs and vaccines.
Bonnet monkeys (M. radiata) are indigenous to Southern India, where DENV circulates widely. In previous studies, Kyasanur forest disease virus (KFDV), tick-borne encephalitis virus (TBEV), Japanese encephalitis virus (JEV) and West Nile virus (WNV) have been inoculated into bonnet monkeys, resulting in high viraemia [23][24][25] . Furthermore, neutralisation antibodies have been detected after natural infection in bonnet monkeys 26 . However, the experimental inoculation of DENV using bonnet monkeys has not yet been reported.
Some reports have shown that sylvatic DENV is transmitted between NHPs and mosquitoes in the sylvatic environment. It is phylogenetically distinct from urban DENV circulating in humans [27][28][29][30][31] . It is thought that sylvatic DENV is an effective candidate challenge virus for the development of an NHP model.
In this study, to develop a novel NHP model for DENV infection, we evaluated the replication efficiency of five DENV isolates using peripheral blood mononuclear cells (PBMCs) derived from M. radiata. The most replicative DENV, 09-48 strain, was used to infect bonnet monkeys and a biological analysis was performed using plasma samples. We also performed a phylogenetic analysis to determine the relationships between this highly replicative strain and other isolates.
Infection of M. radiata with DENV-4. Three M. radiata were inoculated intravenously with 1 × 10 6 PFU of DENV-4 (09-48 strain). The collection of plasma samples, measurement of rectal temperatures and observations of the clinical presentation of monkeys were performed before inoculation and at 2, 3, 5, 7, 10, 14 and 29 dpi. Viraemia is a major clinical manifestation of DENV infection. The amount of viral RNA was analysed in plasma by quantitative RT-PCR at various time points. The viral load peaked at 2 dpi, reaching 2.2-4.0 × 10 6 copies/mL, followed by a gradual decrease to below the limit of detection and there were no significant differences among M. radiata individuals (Fig. 3a). Furthermore, the NS1 antigen, as examined by ELISA, was detected in all plasma samples derived from DENV-4-infected M. radiata (Fig. 3b). These results suggested that DENV-4 propagated in all M. radiata. However, no clinical symptoms, e.g. haemorrhage or weight loss, were observed and rectal temperatures were not substantially altered at any time point (Fig. 3c and d).
Profiles of leukocytes and platelets in blood. Collected blood samples were analysed to determine their cell properties using an automated haematology analyser. Both leukocytes and platelets were slightly decreased at 2, 3, 5 and 7 dpi ( Fig. 3e and f).
Analysis of antibody responses. IgG and IgM antibodies against DENV antigens were assessed by ELISA.
In cases of DENV infection, IgM antibodies can be detected immediately and IgG antibodies subsequently PBMCs derived from M. radiata were infected with DENV at an MOI of 0.1. At 2 dpi, the viral titres in culture supernatants were determined by plaque assays using BHK-21 cells and three independent tests were performed. Results were compared using the Student's t-test against DENV-4 and P < 0.05 (*) was considered statistically significant.
increase. We detected IgM antibodies after 3 or 5 dpi, with a peak at 11 dpi. We detected IgG antibodies at 11 or 14 dpi and their levels increased thereafter ( Fig. 4a and b). These results indicated that DENV-4 propagated and induced specific IgM and IgG antibodies in M. radiata. Neutralisation antibodies. Neutralisation antibodies were measured by plaque reduction assays.
Neutralisation antibodies were observed at 5 or 7 dpi. At 29 dpi, neutralisation antibody titres for all three monkeys were increased to over 1:160 (Table 1).

Discussion
In this study, five clinically isolated DENV strains were evaluated to determine viral replication in PBMCs derived from bonnet monkeys. DENV-4 09-48 strain was isolated from a traveller returning to Japan from India and showed the most efficient replication in PBMCs, suggesting that this strain may be a suitable challenge virus in (e and f) Blood samples were collected on the indicated days. The absolute number of leukocytes (e) and platelets (f) in blood were determined using an automated haematology analyser. Results were compared using the paired t-test against pre-inoculation levels; P < 0.05 (*) and P < 0.01 (**) were considered statistically significant. bonnet monkeys. An evaluation of growth kinetics in PBMCs before in vivo analyses may provide insights into the appropriate growth level in vivo for this challenge virus.
Sylvatic DENV, which is distinguished from urban DENV circulating in humans, has been evaluated in various studies [27][28][29][30][31] . Sylvatic DENV cycles between NHPs and mosquitoes in the sylvatic environment and studies have evaluated its genome and isolation. If sylvatic DENV is acclimatised to monkeys, it may be an effective candidate challenge virus for the development of an NHP model. The sequence of the 09-48 strain indicated that it belongs to the urban DENV cluster and is divergent from the sylvatic DENV cluster. Some sequence differences may affect viral replication in PBMCs; further analyses are needed.
In this study, 09-48 strain was intravenously injected into M. radiata. The results should be interpreted with caution because intravenous inoculation is a different route compared with natural DENV infection. In cases of experimental infection in Old World monkeys, viraemia levels are typically very low. However, following the infection of M. radiata with DENV-4 09-48 strain, viraemia levels were very high 19,32,33 , even soon after infection. The peak viral titre was observed within 2 dpi. The NS1 antigen, which is a product of viral infection and replication, was also observed, as evidence of DENV replication in M. radiata in vivo and inoculated DENV stock reached a titre of 5.0 × 10 8 copies/mL. The monkeys used in this study weighed 6-7 kg and the total serum quantity was approximately 250 mL. After inoculation, there was an estimated 2.0 × 10 6 copies/mL and viral titres at 2 dpi were similar to or greater than the inoculation titre, suggesting that much of the detected viral genome was derived from in vivo replication. Additionally, M. radiata showed lower viraemia levels than those of marmosets, but higher viraemia levels than those of many other Old World monkeys.
In typical DENV infection in humans, IgM antibodies are detected 3-5 days after onset and IgG antibodies are detected 10-14 days after onset 34,35 . Human-like antibody responses against DENV infection are important for the establishment of animal infection models. However, in humans or marmosets, infection with DENV resulted in higher IgG or IgM antibody responses compared with those in M. radiata. Notably, M. radiata could be applied for antiviral drug or antibody medical treatment for DENV infection owing to the similarities in genetic backgrounds between M. radiata and humans.
The platelet and leukocyte counts were slightly and transiently decreased by about 24% and 54%, respectively, similar to the counts in previous marmoset experiments 14 . Furthermore, the recovery times of leukocytes and platelets were similar to those of humans following DENV infection 36 . In this study, the body weight and the amount of blood collected for M. radiata were similar to those for other experiments using Old World monkeys 6,37 , suggesting that blood collection did not affect the hemogram properties caused by DENV-4 infection.  Neutralisation antibodies were also detected by plaque reduction assays in this study. In comparison with DENV infection results in primate models, such as marmosets and rhesus macaques, neutralisation titres in DENV-4-infected M. radiata were similar 19,38 . Compared with human infection, neutralisation antibody levels were significantly lower; however, the timing of the increases in IgM, IgG and neutralisation antibodies was similar, suggesting that this model was useful for the evaluation of neutralisation levels induced by vaccines or viral infection.
A previous report has shown that clinical symptoms, such as rash, were induced by DENV infection in NHPs 19 . In this study, although clear clinical symptoms were not observed, a rash-like bruise was present in MR02 (data not shown). However, this bruise was not analysed using pathological methods, such as immunostaining. Furthermore, viraemia and other responses in MR02 were not different from those in MR01 and MR03. Therefore, this bruise may not a result of DENV infection. In animal model inoculation studies, observations of human-like clinical symptoms are essential. However, in this study, these symptoms of DENV infection were not found and further analyses are needed to explain this result.
In Plaque reduction assay. Sera were diluted with FBS by 10-160-fold and mixed with 50 PFU of virus at a ratio of 1:1. The mixed samples were then incubated at 37 °C for 30 min. BHK cells were seeded in 12-well plates (5 × 10 5 cells/well) and inoculated with samples for 1 h. After inoculation, E-MEM containing 2% FBS and 1% methyl cellulose was overlaid and incubated for 5-6 days. Fixation and staining were applied as described for plaque assays. Samples with 50% fewer plaques than those of the negative control were considered positive.
Real-time RT-PCR. Viral RNA was extracted from plasma using a High Pure Viral RNA Kit (Roche, Basel, Switzerland) according to the manufacturer's instructions. A quantitative real-time RT-PCR analysis of the envelope region of the DENV-4 genome was performed using RNA-direct Real-time PCR Master Mix (Toyobo, Osaka, Japan) and a StepOnePlus System (Applied Biosystems, Foster City, CA, USA). DENV-4-specific primer-probe sets (D4Ten711s forward: GGTGACRTTYAARGTHCCTCAT, D4Ten786c reverse; WGARTGCATRGCTCCYTCCTG and TaqMan probe: D4Ten734p probe, CCAAGAGACAGGATGTGACAGTGCTRGGATC), developed by Ito et al. 39 , were used in this study. Note that in the primer sequences, W indicates a mix of A and T, R indicates A and G, Y indicates C and T and H indicates A, T and C. The real-time RT-PCR conditions were as follows: denaturation (90 °C for 30 s), reverse transcription (61 °C for 20 min), denaturation (95 °C for 1 min) and 40 cycles of amplification and quantification (95 °C for 15 s, 57 °C for 1 min). The copy number was determined by the relative quantification method using the synthesized RNA template. Determination of DENV-4 09-48 strain sequences. DENV RNA was isolated using a High Pure viral RNA Kit (Roche) according to the manufacturer's instructions. cDNA was synthesised from viral RNA and three fragments were amplified by polymerase chain reaction (PCR) using Q5 High-Fidelity 2× Master Mix (NEB, Ipswich, MA, USA) and the following primers (fragment1: D4.001 f, AGTTGTTAGTCTGTCTGGACCG; D4.4620r, TCCCAAACAACCCTCTTTGCAT; fragment2: D4.4210 f, TCCCTTTAGCTGGCCCAATGGT; D4.9427r, TGGCGGATGAGTTGTACTTCCAT; fragment3: D4.8976 f, CGAGCAATCTGGTATATGTGG; D4.10649r, AGAACCTGTTGGATCAACAACAC). Purified PCR products were sequenced using primers (Supplementary Table S1) and analysed using a Big Dye Terminator v3.1 Kit and an ABI 3500 sequencer (Applied Biosystems).

Phylogenetic analysis.
A phylogenetic tree using the full sequence of the 09-48 strain was constructed by the maximum-likelihood method 42 .