Animal models of dengue virus infection and disease

The pathogenesis of DENV infections is, despite intensive research, not well understood. Many fundamental questions in dengue pathogenesis are difficult to address due to the lack of appropriate animal models of infection and disease. No non-human species naturally exhibits the more severe forms of dengue disease that mimic human DF, DHF, and DSS, and this has hampered the development of a suitable animal model.

Non-human primates

NHPs are the only vertebrates apart from humans known to be naturally infected by DENV. However, the viral strains isolated from NHPs are genetically distinct from those infecting humans indicating that these transmission cycles diverged a long time ago (2). In addition, NHPs do not develop hemorrhagic fever, but only a mild infection. Animal models of DENV infection were however initially limited to NHP despite the fact that NHPs only develop mild viremia without severe clinical symptoms. Attempts to induce dengue-like disease in other animals failed and the suboptimal NHP models remained the only available model for a long time. Studies in rhesus macaques revealed no abnormalities in haematocrit or prothrombin time, and only a minority of animals displayed a limited decrease in platelet counts. Thus, DENV infection in macaques has been unable to offer any major insights into the molecular mechanisms of DENV pathogenesis. In contrast, NHPs are widely used in vaccine testing as they are capable of developing neutralizing antibodies in response to DENV infection (220).

Mouse models

Rodents are susceptible to DENV infection although they do not exhibit disease similar to that in human DENV infections. To compensate for these limitations, murine models have relied on mouse-adapted DENVs that appear to be attenuated with respect to human infection. New mouse models have been developed to render the mice more susceptible to infection but none have been able to mimic DENV-specific immune responses.
Mouse models based on intracranial inoculation of DENV have for long been used in parallel to studies in NHPs. Mice usually exhibit a neurovirulent phenotype quite unlike human dengue disease, and succumb to intracranial infection. Multiple efforts have generated a diverse set of mouse models for DENV infection, each with distinct advantages and disadvantages (220). The various approaches can be divided into four groups: 1) Immunocompetent mouse models, 2) Severe combined immunodeficient (SCID)-tumor transplant mouse models, 3) Humanized mouse models, and 4) Interferon-deficient mouse models. Consequently, the various models have contributed in different ways to understanding mechanisms underlying DENV pathogenesis and immunity, and in the development of antiviral drugs and vaccines.
The unimpaired nature of the immune system of immunocompetent mice renders them poorly susceptible for DENV infection. Despite the lack of clinical symptoms, these models have been used extensively to study other aspects of DENV pathogenesis (212, 221-225). In addition, immunocompetent mouse models of DENV infection have been popular for drug and vaccine development studies since the intact immune system is valuable for assessing vaccine immunogenicity. Clinical and neurological disease can be induced by high-dose infection and/or intracranial injection, as well as mouse-adapted DENV strains that render mice more susceptible to DENV infection (226, 227). However, the relevance of observations made with manipulated DENV strains should be interpreted with caution regarding their relevance to wild-type infections.
Measurable viremia is a desired trait when studying DENV infection, and a common strategy to render the mouse more susceptible to DENV infection is to manipulate the mouse in one or several ways. Transplanted human tumor mass that provides a replication site for infectious DENVs have been applied in SCID mice. Since SCID mice lack an adaptive immune system, DENV infection has been successfully established within the transplanted cells and even some dengue disease features have been reproduced (228-230). However, it is unclear how any insights into pathogenesis might apply to human pathogenesis since viral replication is restricted primarily to the transplanted cells.
An improved strategy based on the SCID-tumor model is ‘humanized’ SCID mice that are irradiated to destroy the haematopoietic progenitors in the bone marrow, prior to transplantation with human CD34+ haematopoietic stem cells. The result is an adaptive immune system consisting exclusively of human cells, with certain parts of the innate system being humanized as well (231, 232). The advantage with this model is the increased susceptibility to clinical DENV isolates without adaptation to mice, and the mice display some signs of human disease. The exact cell types in which DENV replicates in humanized mice remains to be identified. In addition, the difficulties with the humanized mouse model are multiple, such as the genetic variation in stem cell donors, the hardship to generate sufficient numbers of humanized mice, variability in the degree of human cell engraftment, and radiation sensitivity of SCID-mice (233). A limitation of crucial importance is the lack of lymph nodes that support a human immune system (234-236) as well as the lack of a robust immune response in these humanized mice; both fluctuations in the elicited immune response as well as poor reproducibility (237-238. Human-specific cytokines required for human cell development and survival are not successfully generated, and there is a low and variable level of T-cell-dependent antibody responses (239, 240). Thus, humanized mouse models are time-consuming and labor-intensive and, therefore, not optimal for most drug and vaccine development purposes. These new models may eventually have the potential to answer important and fundamental questions concerning infection of human cells in vivo, but there are still major issues to be resolved regarding stem cell development in order to obtain a functional and intact adaptive immune system.
Another way to increase the susceptibility of mice to DENV infection is to generate mice deficient in the IFN pathway (241, 242). The IFN system is a potent suppressor of DENV replication, and mice lacking the IFN-α/β and –γ receptors (AG129 mice) can be lethally challenged without intracranial inoculation (241). The AG129 mouse strain exhibits viral replication in relevant peripheral cell types such as macrophages and DCs in spleen and lymph node (243), hepatocytes, and myeloid cells in the bone marrow (154). A variety of DENV isolates have been used and the clinical symptoms are either neurological disease or a hemorrhagic shock-like vascular permeability syndrome depending on the virus isolate and route of inoculation (241, 243-248). Consequently, AG129 mice have been used for numerous studies, including tropism, pathogenesis, immune protection and enhancement, as well as for antiviral drug testing and vaccine development (241-243, 246, 247, 249, 250). However, the genetically deficient background does not make the AG129 mouse suitable for genetic studies since it would be time-consuming to introduce a third knock-out gene.

Conclusions

The DENVs are old viruses that have re-emerged during the latter half of the 20th century. Regarded as a tropical fever disease affecting more than two thirds of the world's population, dengue is also the main cause after malaria of tropical fever among travelers and ranks as the most important mosquito-borne viral disease in the world. The lack of potent antiviral drugs and an effective vaccine results in ∼500,000 individuals, mainly children, being hospitalized with severe dengue every year and causes tremendous economic losses for both households and whole nations. The pathogenesis of the DENVs are not well understood, partly due to the absence of good animal models. Effective vector control measures are the sole weapon against dengue today, while we are hoping for improved diagnostics, clinical treatment, and an effective vaccine.