Abstract
Dengue viruses (DENV) are currently responsible for more human morbidity and mortality than any other known arbovirus, and all four DENV are known to exist in sylvatic cycles that might allow these viruses to persist if the urban (Aedes aegypti) cycle could be controlled. To determine whether DENV were being maintained in a sylvatic cycle in a forested area about 14 km southwest of Iquitos, Peru, a city in which all 4 serotypes of DENV circulate, we placed 20 DENV seronegative Aotus monkeys in cages either in the canopy or near ground level for a total of 125.6 months. Despite capturing >66,000 mosquitoes in traps that collected some of the mosquitoes attracted to these monkeys, blood samples obtained once a month from each animal were tested and found to be negative by an enzyme-linked immunosorbent assay for IgM and IgG antibodies to dengue, yellow fever, Venezuelan equine encephalitis, Oropouche, and Mayaro viruses. Although all four DENV serotypes were endemic in nearby Iquitos, the findings of this study did not support a DENV sylvatic maintenance and transmission cycle in a selected area of the Amazon rainforest in northeastern Peru.
Introduction
Dengue viruses (DENV) comprise four antigenetically and genetically distinct serotypes as members of the genus Flavivirus, family Flaviviridae (Westaway et al. 1985). The human endemic/epidemic transmission cycles of these viruses involve Aedes aegypti as the primary vector and humans as the amplifying host (Gubler 2002). These four DENV are currently responsible for more human morbidity and mortality than any other known arbovirus (Bhatt et al. 2013). Each of the serotypes are also maintained in an enzootic sylvatic transmission cycle involving forest canopy-dwelling Aedes species mosquitoes and nonhuman primates in Southeast Asia, and DENV-2 is maintained in a similar cycle in West Africa (Rudnick 1986, Vasilakis et al. 2011, Hanley et al. 2013).
In Southeast Asia, the sylvatic transmission cycle occurs in the canopy of the forests of Malaysia. This cycle is maintained by mosquito species in the Aedes niveus group and monkeys in the genera Macaca and Presbytis as amplifying hosts. In western Africa, the sylvatic cycle is maintained in the forests and savannahs of Senegal and involves Aedes furcifer, Aedes taylori, and Aedes luteocephalus mosquitoes as the vectors and two monkey species, the African green monkey (Chlorocebus sabaeus) and the Guinea baboon (Papio papio), as amplifying hosts (Diallo et al. 2003). Also, a third primate host of sylvatic DENV, the patas monkey (Erythrocebus patas), has a broad range across Central Africa.
Sporadic human cases have been caused by each of the four serotypes of sylvatic DENV during the last decade, including cases of severe dengue with hemorrhagic fever (Cardosa et al. 2009, Hanley et al. 2013, Liu et al. 2016, Pyke et al. 2016). Some of these cases involved humans who apparently were exposed to sylvatic DENV while visiting or working in the forest in Southeast Asia. In West Africa, a small number of human cases have been caused by sylvatic DENV-2 in Senegal, and an outbreak of febrile illness during the 1960s in Ibadan, Nigeria, was caused by sylvatic DENV, thus demonstrating that sylvatic DENV can cause sporadic cases in rural communities and epidemics in urban communities (Vasilakis et al. 2008, Althouse et al. 2012, Hanley et al. 2013).
Although sylvatic cycles of DENV are maintained in western Africa and Southeast Asia, such a cycle has not been confirmed for the Americas. A single report suggested that a sylvatic cycle may be occurring in Bolivia based on DENV seroconversions among indigenous Ayoreo people in a remote area devoid of Ae. aegypti (Roberts et al. 1984). Further observations on the possibility of a sylvatic cycle in Bolivia have not been reported. Enzootic transmission of DENV-1 or DENV-2 was not detected during surveys of nonhuman primates in Panama (Rosen 1958), and phylogenetic evidence of a sylvatic lineage of DENV in the Americas has not been detected, despite the presence of human dengue in the Americas for centuries (Vasilakis et al. 2011).
Several studies have reported the detection of DENV RNA and/or antibodies against DENV for all four serotypes in mammalian wildlife (notably, bats, rodents, and marsupials), especially in many species of bats in the Neotropics. Because infectious virus was not recovered from the bats, the authors concluded that bats were incapable of sustaining DENV replication and were unlikely to act as amplifying hosts for this virus (Vicente-Santos et al. 2017). However, the detection of DENV-1 in Haemagogus leucocelaenus collected in a rainforest in northeastern Brazil suggested that this virus might be involved in a sylvatic cycle (de Figueiredo et al. 2010).
All four DENV serotypes have been documented as the cause of human disease in Iquitos City, an urban community in the Amazon rainforest with a population of ∼300,000 in northeastern Peru (Liebman et al. 2012, Stoddard et al. 2014). Also, a serosurvey conducted during 1992 detected DENV antibody in 66% of the population of Iquitos and in 32–67% in three surrounding rainforest communities (Hayes et al. 1996). DENV antibody among humans in the rainforest communities was attributed to infection in the Iquitos community. Ae. aegypti is among the most abundant mosquito species in the Iquitos community and considered to be the primary vector of DENV.
Several species of nonhuman primates and canopy-dwelling mosquitoes inhabit the forest adjacent to Iquitos, but studies have not been conducted to consider their possible involvement in a sylvatic cycle of DENV. The nonhuman primate species expected to inhabit the study area include Saguinus fuscicollis, Saimiri sciureus, and Aotus vociferans (Freese et al. 1982, Pacheco et al. 2011) and monkeys tentatively identified as S. fuscicollis were observed in the study area during the study. Therefore, in conjunction with an ongoing longitudinal study on the ecology and epidemiology of DENV and Ae. aegypti, a study was conducted from May 2001 to September 2002 using sentinel Aotus monkeys in the rainforest near Iquitos to determine whether DENV was maintained in a sylvatic cycle involving canopy-dwelling mosquitoes and nonhuman primates.
Materials and Methods
Monkeys
Aotus nancymaae monkeys were chosen as sentinel animals because this species is common in the Amazon rainforest (Aquino and Encarnación 1988), and laboratory studies have shown that Aotus species develop a viremia and antibody response after infection with DENV (Kochel et al. 2000). Young adult captive-born monkeys were purchased from the Peruvian Primate Center located in Iquitos, Peru, and were quarantined for 30 days before use in the study. During the quarantine period, blood samples were obtained, and sera were assayed by an enzyme-linked immunosorbent assay (ELISA) to ensure that the animals were negative for antibodies to dengue or other arboviruses. Only animals that tested negative for antibody to these arboviruses were used in the study.
Because the investigators had not maintained Aotus monkeys in cages in field sites before this study, we initiated the study by housing two monkeys per cage at one field site, including one cage in the canopy and one near the forest floor at the end of May 2001. The health of these animals was closely monitored daily. Because no problems were observed with the first 2 pairs of animals, the next 2 pairs were placed in the field during July 2001, with the final 2 pair of animals placed in the field as selected sites the following month (August 2001) for a total 12 monkeys at 3 different sites. Eight additional Aotus monkeys were added in January 2002, for a total of 20 monkeys at 5 different field sites, with each site consisting of 2 cages, one in the canopy and one near the ground. All animals were removed from the field in August 2002, at the end of the study, and returned to the Peruvian Primate Center.
Study site
The sentinel monkeys were placed at a forested site (grid coordinates: S: 3° 49′ 8.77″, W: 73° 20′ 47.651″), in a Peruvian army camp, located ∼14 km southwest of Iquitos. Populations of wild primates were present at this site. Because Peruvian military personnel do not permit access by unauthorized personnel, and made periodic patrols of the area, there was a reduced risk of the animals being disturbed by humans. The study site is located in a tropical rain forest where rain occurs year round, but heavier rains occur from January to March.
Study design
Paired cages were placed in the forest canopy (∼15 meters above ground level) and near the forest floor (∼1 meter above ground level). Each pair of cages was placed ∼250 meters away from any other pair within the study site. Animals were housed as male/female pairs (i.e., two monkeys/cage). The baseline weight of each animal was taken the day that it was introduced to the field site and the weight of each animal was determined thereafter at monthly intervals for weight loss, which could indicate potential health problems. In addition, all monkeys were checked twice daily by an experienced animal caretaker and checked weekly by a veterinary technician and monthly by a veterinarian from the Peruvian Primate Center for any signs of illness. Sentinel monkeys were housed in custom-built, light-weight, galvanized steel cages that allowed mosquito access to the monkeys, and contained a polyvinyl chloride (PVC) nest box, a PVC perch, a water bottle, and a feeder. The animals were fed daily on a laboratory primate diet provided by the Peruvian Primate Center as well as seasonal fruits and fresh tap water.
To determine if monkeys became exposed to dengue, or other arboviruses, blood samples (2.5 mL) were obtained once a month from each animal, beginning 1 month after placement at the study site. For restraining and collection of blood samples, monkeys were anesthetized with ketamine hydrochloride (10 mg/kg) intramuscularly, using a 27-gauge needle and bled from the femoral vein with a 26-gauge needle attached to a 3-cc syringe. Blood samples were then transported to the Naval Medical Research Center Detachment (NMRCD) laboratory in Iquitos, and aliquots of the sera were frozen at −80°C for transport to our main laboratory facilities in Lima, Peru. Sera were tested for IgG and IgM antibodies to DENV, yellow fever virus (YFV), Venezuelan equine encephalitis virus (VEEV), Oropouche virus (OROV), and Mayaro virus (MAYV) by an ELISA in 96-well microplates by standard methods (Ansari et al. 1993).
To determine which species of mosquitoes were present and might be involved in maintaining a sylvatic dengue transmission cycle, mosquitoes were collected using modified mosquito Centers for Disease Control and Prevention (CDC) light traps (John W. Hock Co., Gainesville, FL). To enhance collection of mosquitoes attracted to the monkeys, a CDC trap was attached to a cloth covering constructed around and over each monkey cage. Mosquitoes attracted to the monkeys that flew upward were “funneled” into the CDC trap and collected. No attractant (e.g., light or CO2), other than the monkeys, was utilized. A rainproof covering was provided to protect the entire cage and trap from rain. Traps were set weekly from Sunday morning at 06:00, until Thursday evening at 18:00, providing 5 full days and 5 full nights of collections each week. Mosquitoes were collected twice daily from each trap, at ∼06:00 and 18:00, and transported to the NMRCD laboratory in Iquitos for identification. The mosquitoes were identified to species, pooled by species (up to 25/pool), and stored at −70°C until shipped to the U.S. Army Medical Research Institute for Infectious Diseases at Fort Detrick, MD, for testing for arboviruses.
Ethics statement
The monkeys were housed and cared for according to the “Guide for the Care and Use of Laboratory Animals” (Institute for Laboratory Animal Research 2011), the Animal Welfare Act, and Animal Welfare Regulations. The study was approved by the U.S. Naval Medical Research Center Detachment Institutional Animal Care and Use Committee.
Results
Mosquitoes captured
During the course of this study, a total of 66,097 mosquitoes were captured in the monkey-baited CDC light traps (Andrews et al. 2014). These included 50 distinct species in 12 genera, with 94% of the specimens belonging to 5 genera, Culex (32.7%), Aedes (23.3%), Psorophora (17.5%), Mansonia (10.4%), and Coquillettidia (10.1%). In addition, smaller numbers of Haemagogus janthinomys and Sabethes species were collected, 62 and 16 specimens, respectively. Individuals in both genera were captured in the traps associated with the monkeys located both at 1 and 15 meters above ground level.
Monkey exposure, serological study, and lack of evidence of seroconversion
The 20 Aotus monkeys were exposed to forest mosquitoes for a total of 125.6 months, with 10 monkeys each being exposed for 62.8 months either in the canopy or near the forest floor.
All animals gained weight, or maintained a body weight close to their baseline, until the end of the study, with the exception of a single monkey who had a weight loss of 4.8%, but had no signs of illness. All animals appeared healthy and were returned to the Peruvian Primate Center at the conclusion of the study.
Sera samples from all of these monkeys were negative for antibody to DENV, OROV, VEEV, and MAYV. Therefore, we did not detect evidence of natural infection with any of these viruses in the sentinel monkeys.
Discussion
Despite circulation of all four serotypes of DENV in the nearby city of Iquitos, we did not observe any evidence of sylvatic transmission of DENV to Aotus monkeys despite the equivalent of >10 monkey years of exposure to sylvatic mosquitoes. These monkeys readily attracted mosquitoes as evidenced by the >66,000 captured in CDC traps for which these monkeys were the only attractant provided, and there is an abundant nonhuman primate population in the Amazon rainforest surrounding Iquitos (Freese et al. 1982, Pacheco et al. 2011). This is in contrast to the existence of sylvatic transmission cycles of DENV in Southeast Asia and West Africa (Rudnick et al. 1967, Rudnick 1986, Diallo et al. 2003, Cardosa et al. 2009, Pyke et al. 2016).
Evidence of a DENV sylvatic cycle involving nonhuman primates and mosquitoes as documented in West Africa and Southeast Asia has not been reported in the Americas (Hanley et al. 2013), with the possible exception of two free-ranging black howler monkeys (Alouatta caraya) of northeastern Argentina with a very low neutralizing titer (Morales et al. 2017). However, these may have been the result of spillover from human cases. Several studies have reported possible evidence of DENV infection in mammalian wildlife in the Americas (notably, bats, rodents, and marsupials), especially in many species of bats in the Neotropics (Platt et al. 2000, Aguilar-Setién et al. 2008, Sotomayor-Bonilla et al. 2014). However, because the results of field and experimental studies showed bats to be incapable of sustaining DENV replication (Perea-Martínez, et al. 2013, Cabrera-Romo et al. 2014, Vicente-Santos et al. 2017), it is unlikely that bats serve as amplifying hosts. It has also been suggested that some of the sera considered positive for DENV might have been from an infection with a previously unidentified flavivirus (de Thoisy et al. 2004, Machain-Williams et al. 2013). Therefore, the conclusion based on existing knowledge is that the critical evidence required to support a sylvatic DENV cycle is lacking in the Americas. This includes the inability to document an enzootic cycle based on the isolation of a sylvatic serotype and/or strain of DENV from infected nonhuman primates and/or mammals as well as infected forest-dwelling mosquitoes and other parameters as described previously (Scott 2001).
Our findings should not be interpreted as the absence of a sylvatic DENV cycle in the Americas. As experienced in Africa, and perhaps Asia, only a very limited part of these regions has been surveyed for sylvatic DENV. Thus, enzootic foci could go unrecognized (Vasilakis et al. 2011). Another possible confounder has been the inability to apply reliable DENV diagnostic tools to identify sylvatic strains of DENV. Furthermore, with the exception of the present study that focused on using nonhuman primates as sentinels to detect evidence of a sylvatic DENV cycle, there does not appear to have been any well-designed surveillance of wild populations of nonhuman primates for DENV in the Americas. However, the present study had some limitations that may have interfered with the recognition of evidence of a sylvatic transmission cycle. Wild populations of nonhuman primates are free ranging and may travel over a wide range of habitat and thus be exposed to many species of mosquitoes and viruses (Harvey and Clutton-Brock 1981). In contrast, the confinement of the animals to cages could have prevented them from entering areas where vector mosquitoes were more common. Also, studies by Althouse et al. (2012) indicate that outbreaks of sylvatic dengue might be cyclic in wild primates, and our studies, conducted over a period of just over 1 year, may have missed sufficient dengue transmission to be detected.
The reasons for the lack of evidence of a sylvatic cycle of DENV among wild nonhuman primates in the Western Hemisphere are unknown. However, most nonhuman primates in the Americas live in families/groups smaller than those in Africa or Southeast Asia (Buchanan-Smith 1990, Mitchell et al. 1991, Dunbar 1992, Janson and Goldsmith 1995), making sustained transmission in a sylvatic cycle more difficult. Larger troops that involve multiple females giving birth throughout the year may provide a better source of susceptible monkeys helping to maintain the virus cycle. Also, while DENV viremias tend to be low in New World nonhuman primates, making the initiation of a sylvatic cycle more difficult, viremias were high enough to sustain such a cycle, if a sylvatic strain of DENV were introduced (Hanley et al. 2013). However, because clinical signs of DENV infection are extremely rare in American nonhuman primates, if a sylvatic outbreak were to occur, it would be difficult to detect as it would not likely cause a wave of deaths in nonhuman primate as occurs with YFV (Hanley et al. 2013). The ecological and epidemiological interactions between sylvatic DENV cycles and human populations and the dynamics of human infection and the burden of disease associated with sylvatic DENV serotypes and/or strains remain poorly understood (Vasilakis et al. 2011).
Conclusion
These preliminary results suggest that DENV may not circulate in sylvatic cycles in this area of northeastern Peru. Because the study was conducted in only one location in a forested area of Peru, it is not possible to conclude that DENV do not circulate in sylvatic cycles in South America. Additional studies should be conducted, in other areas of South America, to establish if sylvatic cycles of dengue do, or do not, exist in South or Central America.
Footnotes
Acknowledgments
We convey our sincere thanks and appreciation to Clever Donayre and Miguel Vasquez for assisting in collecting the mosquitoes and transporting them to the laboratory; Alfredo Cetraro for providing outstanding animal husbandry and care for the monkeys; Dr. Enrique Montoya and Arnulfo Romaina, Centro de Reproducción y Conservación de Primates No Humanos (CRCP)—Iquitos, Perú, for veterinary care and collecting the blood samples; and to Alfredo Huaman for his excellent technical support for processing the sera obtained from the monkeys and for performing the serological testing of the sera samples. Financial support: these studies were funded by the Military Infectious Disease Research Program (A60001_02_LI). This research was supported, in part, by the Intramural Research Program of the National Institutes of Health, National Institute of Allergy and Infectious Diseases, and the Comparative Medicine Branch.
Disclaimer
The views expressed in this article are those of the authors and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, or the U.S. Government. The use of any specific product does not constitute endorsement of that product, and the opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the U.S. Army or the U.S. Navy. All authors are current or former employees of the U.S. Government. This work was prepared as part of their official duties.
Author Disclosure Statement
No conflicting financial interests exist.