Barriers to Preclinical Investigations

Published on December 2016 | Categories: Documents | Downloads: 120 | Comments: 0 | Views: 929
of 7
Download PDF   Embed   Report

Comments

Content

PERSPECTIVES
V E C TO R - B O R N E D I S E A S E S — O P I N I O N

Barriers to preclinical investigations of anti-dengue immunity and dengue pathogenesis
Ashley L. St. John, Soman N. Abraham and Duane J. Gubler

Abstract | Dengue virus (DENV) is a human pathogen that causes severe and potentially fatal disease in millions of individuals each year. Immune-mediated pathology is thought to underlie many of the complications of DENV infection in humans, but the notable limitations of the available animal models have impeded our knowledge of the interactions between DENV and the immune system. In this Opinion article, we discuss some of the controversies in the field of dengue research relating to the interaction between DENV and the mammalian host. We highlight key barriers hindering our understanding of the molecular pathogenesis of DENV and offer suggestions for the most effective ways in which the role of the immune system in the protection from, and pathology of, DENV infection can be addressed experimentally.
Dengue fever has emerged as a major global public-health problem over the past four decades. Closely tied to unprecedented human population growth, urbanization, globalization and the lack of effective vector control, both the causative viruses and the mosquito vectors that transmit them have spread globally in the tropics, resulting in an increased frequency and magnitude of epidemics and the emergence of the severe form of disease, dengue haemorrhagic fever (DHF)1,2. In 2013, an estimated 3.6 billion people, more than half of the world’s population, live in areas that are at risk for dengue virus (DENV) infection3. There are neither vaccines nor specific antiviral therapies available for this disease. The research that has defined our understanding of the pathogenesis and immunology of DENV infection can be predominantly segregated into three types of studies: those examining patients with dengue, to describe the clinical presentation; in vitro studies to define the viral life cycle or investigate direct interactions between DENV and cultured cells; and controlled hypothesis-driven animal experiments to
420 | JUNE 2013 | VOLUME 11 © 2013 Macmillan Publishers Limited. All rights reserved

DENV molecular pathogenesis that have been particularly difficult to understand using the current experimental methodologies. We also discuss the paradox that further substantial advances in the field might be possible only through controlled animal experimentation, which has inherent limitations because animal models do not precisely replicate human disease. We propose that this contradiction can be reconciled with the choice of appropriate animal models to answer specific experimental questions and with cautious interpretation of results, keeping in mind the multifaceted complexity of the interactions between DENV and the human host.
Dengue virus infection in humans DENV is a member of the family Flaviviridae (TABLE 1), which contains several other notable human pathogens. There are four antigenically distinct DENV serotypes (DENV1–DENV4), which have a distribution throughout tropical regions of the world similar to that of their principal mosquito vector, Aedes aegypti 2. The key pathophysio­ logical host response to DENV infection manifests in the form of increased vascular permeability, plasma leakage, microvascular bleeding and reduced functioning of the coagulation cascade. Current estimates suggest that between 1% and 70% of individuals who experience infection have mild haemorrhagic manifestations such as petechiae, purpura, ecchymoses and epistaxis (nosebleed)6,7 (FIG. 1). Severe dengue, known as DHF and dengue shock syndrome (DSS), occurs in <1% of infections2 and can be lifethreatening. During DHF–DSS, bleeding involves multiple organs, frequently including the gastrointestinal tract, and fluid may pool within body cavities (FIG. 1); lesions in the blood vessels or signs of necrosis have not been observed, even in fatal cases8. DSS, the most severe form of DHF, is characterized by a rapid, weak pulse and sudden drop in blood pressure, which is the result of collapse of the vascular system owing to hypovolaemia caused by vascular leakage. It is not fully understood why most individuals resolve DENV infections quickly and without complications, whereas others experience a potentially fatal vascular leak
www.nature.com/reviews/micro

analyse the mechanisms underlying DENVinduced pathologies. Early key studies that elucidated the course of infection, the incidence of pathological complications and the characteristic symptoms were carried out in experimentally infected human subjects4,5. Today, there are limitations on human experimentation because of ethical and regulatory concerns, and because of the expense of undertaking large studies in genetically diverse populations. Furthermore, human studies using patient populations (as opposed to experimental infections) may show a correlation, but do not show unequivocal causation, as these studies lack the ability to measure the responses of patients in comparison to exact controls. For dengue, additional variables could include unidentified factors that might influence the host’s susceptibility to disease in addition to influencing the disease course. Although animal models have supplemented our knowledge, unfortunately a single animal model has not been able to faithfully reproduce all aspects of human DENV infection. In this Opinion article, we highlight the key aspects of anti-dengue immunity and

PERSPECTIVES
Table 1 | Important flavivirus human pathogens
Flavivirus species Prominent pathologies during human infection
Dengue virus (DENV1–DENV4) Fever–arthralgia–rash syndrome, plasma leakage, haemorrhage and severe disease associated with organ failure All four serotypes can cause a spectrum of illness ranging from asymptomatic to severe and fatal disease, as determined by a complex interaction of the virus with the innate immune response of the individual host; the strain of virus, the previous infection history and age of the host, and host genetics all influence disease severity. Yellow fever virus West Nile virus Japanese encephalitis virus Tick-borne encephalitis virus Fever, nausea, abdominal pain, jaundice and haemorrhage Fever, myalgia, encephalitis and long-term neurological defects Fever, encephalitis and long-term neurological defects Meningitis, encephalitis, mild fever and long-term neurological defects

syndrome or even severe frank haemorrhaging. Risk factors for severe disease include the presence of heterologous antibodies from a previous infection9, certain strains or subtypes of infecting virus10,11, and the age and genetic background of the human host 2,12–14.
Key questions for dengue virus immunity Dengue remains a disease with many unanswered questions. There are multiple levels of DENV–host interaction from the time virus particles are injected by mos­ quitoes, to the establishment of systemic infection, to its resolution. Each stage requires specialized considerations for effective experimental design to investigate the mechanisms underlying disease. The course of host–virus interactions Initial target cells for infection. Little is understood regarding the early events following DENV inoculation into the skin, including immune responses at the site of inoculation and the impact of the infecting viral strain or subtype, both of which could potentially determine the host’s success in viral clearance. The role of the infecting viral strain is discussed below. Subcutaneously injected virus can potentially encounter several cells of the immune system, including tissue-resident dendritic cells (DCs), macrophages and mast cells15,16 (FIG. 2). The cell types that are first infected in the skin have not been extensively studied, but it is thought that macrophages and DCs are initial targets, and receptors have been identified for DENV uptake by both cell types: the mannose receptor on macrophages and DC‑SIGN (DC-specific ICAM‑3‑grabbing non-integrin 1) on DCs17,18 (FIG. 2). Aside from the demonstration that Langerhans cells are permissive to DENV infection in
NATURE REVIEWS | MICROBIOLOGY

human skin explants15, a detailed analysis of the DC and monocyte subsets that sustain infection within the skin has not been carried out. Most efforts to define the human cell types infected by DENV have sought to identify viral replication at a much later stage of infection: in the blood during peak viraemia. In monkeys and mice, antigen has been detected in cells that are morpho­ logically consistent with mononuclear phagocytes19,20 (which are thought to be one of the prime targets of DENV infection), but this is not well documented in human infection. Phenotypic differences potentially exist among the target cells for infection in the skin (early infection), in the lymphoid organs (during viral dissemination) and in various tissues when systemic infection has been established. Clearance of dengue virus from the injection site. Early in DENV infection, immune cells are targets of infection, but also promote DENV clearance from the skin. Much emphasis is placed on the role of DENVspecific neutralizing antibodies in viral clearance21. However, during primary infection, substantial levels of neutralizing antibodies should not be present until the later stages of disease, as the adaptive immune response requires time to develop. This suggests that in the early days of infection additional, antibody-independent mechanisms of viral clearance from the injection site are required. It is possible that, as occurs in West Nile virus infection, γδ T cells play a part in viral clearance from the skin22. In mice, natural killer (NK) cells and NK T cells (recruitment of which is promoted by activated skinresident mast cells) have been implicated in viral clearance at the site of skin injection and in draining lymph nodes16. Stromal and endothelial cells can also produce cytokines,

particularly when these cells are activated by pro-inflammatory factors derived from neighbouring immune cells23. Although most immune cells release pro-inflammatory mediators when they are infected with DENV, some, such as mast cells16, can directly detect and respond to the virus without being infected. This emphasizes the fact that there are likely to be additional mechanisms of DENV structural recognition by the immune system beyond antibody detection. For cells that are infected, pro-inflammatory responses to DENV replication are initiated by intracellular pathogen recognition receptors, including vesicle-associated Toll-like receptor 3 (TLR3) and the cytosolic sensors retinoic acid-inducible gene I protein (RIG‑I) and melanoma differentiation-associated protein 5 (MDA5; also known as IFIH1)24,25. These pathways lead to the production of pro-inflammatory cytokines, including tumour necrosis factor (TNF), type‑I interferons (IFNs), IFNγ and interleukin‑6 (IL‑6)24,25. Systemic dengue virus infection. As DENV disseminates in a host, it is detected first in draining lymph nodes and then in remote lymph nodes19. In humans and non-human primates, this results in viraemia (which can be detectable 24–48 hours before the onset of clinical symptoms4), and DENV can be isolated from the blood during the acute phase of disease. At this stage, substantial changes occur to the cellularity of the haematopoietic system, as would be expected during systemic viral infection; these changes include leukopenia, neutropenia, thrombocytopenia and, occasionally, eosinophilia26 (FIG. 2). DENV infection is also characterized by altered bone marrow cellularity, but the mechanisms or consequences of these changes are unknown26. Vascular pathology, including microvascular permeability of endothelial cells, occurs as viraemia declines27. Although infected endothelial cells have been observed in human infection28, it is uncertain how frequently this happens, because of the paucity of autopsies on human patients with dengue. Human endothelial cells can be infected in cell culture, and infection of liver sinusoidal cells occurs in immunocompromised animals29. Although controversial, it has been suggested that endothelial cells are not commonly infected during human infection with DENV30,31 and that the breakdown in the endothelial cell barrier is consistent with other mechanisms of vascular permeability26. Owing to both the lack of evidence that endothelial cells are directly targeted by DENV and the massive
VOLUME 11 | JUNE 2013 | 421

© 2013 Macmillan Publishers Limited. All rights reserved

PERSPECTIVES
Headache, fever Bleeding gums, nose and eyes Vascular symptoms: Leukopenia Thrombocytopenia Neutropenia Late eosinophilia Reduced coagulation Skin symptoms: Rash Bruising Petechiae Purpura

Vascular symptoms: Hypovolaemia Low blood pressure Shock Hepatic injury Fluid pooling in body cavities Gall bladder thickening Haemorrhaging within organs Vomiting Intestinal bleeding

Joint pain

Altered haematopoiesis

Infrequent complications: Encephalitis Acute pancreatitis Renal failure Myocarditis Splenic rupture Pulmonary haemorrhage

not yet explained. Unfortunately, the complexity of the interactions between DENV and the host cannot be adequately modelled by in vitro manipulation of human cells and biological products outside the organs and systems that govern their responses to infection. Not only are the specific target cells of DENV not clearly known, but also in vitro analysis does not capture the responses of cells that do not become infected but might nonetheless respond to DENV and to the inflammatory products in their microenviron­ment in vivo. As a further complication, interactions between DENV and host cells can change over the course of infection, which begins as a localized cutaneous infection with innate immune activation and develops into a systemic infection characterized by viral replication in target organs (including the lymphoid system) and high viral titres in the blood19. The most severe pathologies associated with DENV infection frequently manifest in patients as viraemia resolves and fever begins to subside27, but it is unclear at what stage of infection these pathological processes are initiated.
The role of immunological memory Primary DENV infection results in longlasting immunity to the infecting serotype and, potentially, partial immunity to subsequent infection with other serotypes. Many individuals in dengue-endemic countries have experienced DENV infection previously. Re‑infection with the same serotype as that which caused the primary infection has not been documented, demonstrating that immunity to a homologous DENV strain can be highly protective and lifelong 5. Pre-existing neutralizing antibodies must have a major role in preventing subsequent infection with the same DENV serotype. Secondary infection with a heterologous serotype can cause a broad spectrum of illnesses, ranging from asymptomatic infection to severe haemorrhagic disease. For example, some epidemiological studies have shown that patients with a secondary DENV infection and babies born to mothers who have previously been infected with DENV are more likely to develop severe disease37,38. One theory for why this occurs was first described in reference to other flaviviruses (Murray Valley encephalitis virus, West Nile virus and Japanese encephalitis viruses)39 and is termed antibody-dependent enhancement of infection (ADE). ADE involves the binding of immune complexes of nonneutralizing antibody and infectious virus to the Fc receptors of immune cells29,40,41
www.nature.com/reviews/micro

Figure 1 | Dengue virus pathogenesis in humans.  Systemic infection with dengue virus (DENV) Nature Reviews | Microbiology affects multiple organ systems. This diagram depicts the clinical symptoms and pathogenesis of dengue in humans, across the spectrum of mild to severe disease. Clinical manifestations associated with dengue fever are listed in blue boxes, those associated with dengue haemorrhagic fever are listed in purple boxes, and more rare complications of DENV infection are listed in the green box.

pro-inflammatory response that is detectable in the serum of patients with dengue32, it has been postulated that immune factors act as intermediaries in the response of endothelial cells to DENV. In particular, cytokine storm (elevated levels of many cytokines in the serum) has been identified as a potential underlying mechanism of vascular pathology21. During DENV infection, copious amounts of certain pro-inflammatory and vaso­ active cytokines, including TNF and vascular endothelial growth factor A (VEGFA), are produced32. However, although these factors are frequently elevated in the serum of patients with dengue, some studies have raised questions about their direct role in dengue pathology. For example, in spite of the well-established role of TNF in promoting vascular leakage in other diseases and experimental contexts33,34, there has been no clear correlation shown between levels of TNF and the manifestation of dengue fever
422 | JUNE 2013 | VOLUME 11

versus DHF–DSS in humans. Some evidence has been obtained for there being both lower expression of TNF receptors on granulocytes during DHF–DSS35 and an association between DHF–DSS and certain TNF polymorphisms12. In one study, serum TNF was elevated during both DHF–DSS and dengue fever compared with levels in healthy controls, but a significant difference in TNF levels between DHF–DSS and dengue fever groups was not detected36. Thus, it is possible that TNF has a role in DENV infection, but whether it is (or can be, in some instances) a significant cause of DHF–DSS in humans is still unclear. Multiple pro-inflammatory factors might act in concert to promote vasculopathy. Drawbacks of in vitro analysis. Ultimately, the mechanisms of many of the immune changes that occur in response to DENV infection, and the roles of these immune changes in protection versus pathology, are

© 2013 Macmillan Publishers Limited. All rights reserved

PERSPECTIVES
and has been shown to enhance DENV infectivity in vitro in human cells42. Studies of DENV in vitro and in primates infused with enhancing antibody have shown that ADE can enhance infection of monocytes by promoting viral uptake and intracellular replication40,43. However, it should be noted that this enhancement occurs in only a small fraction of secondary infections and with only certain strains of virus2,44. Another immunological concept that was first identified in the context of another virus — in this case, influenza virus45 — and later shown to be relevant to DENV infection46,47 is original antigenic sin. This theory describes how amplification of pathogenspecific lymphocytes during secondary heterologous challenge can constitute a misallocation of resources for the immune system if those lymphocytes are not as effective in promoting clearance of the secondary challenge strain as they were for the primary strain. It has been reported that T cell responses in children experiencing a secondary DENV infection have a low strain specificity, resulting in T cell anergy and apoptosis, and possibly higher viral titres owing to a lack of optimal viral clearance by T cells46. In other experimental models resulting in original antigenic sin, T cell responses are usually highly specific for the primary challenge strain, and this might also be the case for DENV48. ADE and original antigenic sin have both been offered as explanations for the association of secondary DENV infection with an increased incidence of DHF–DSS21,46, as both theories provide an understanding of how immunological memory could be detrimental to the host when a secondary challenge is highly similar to, but antigenically distinct from, the primary infecting virus. Each theory provides a potential mechanism for achieving high viral titres in vivo. Although viraemia levels have been associated with disease severity 49,50, people with secondary infections frequently do not have higher levels of viraemia or severe disease. Also, contrary to the widely held view, ADE (as a specific mechanism to explain secondary infection as a risk factor for DHF–DSS) has not been demonstrated to occur in vivo in humans, although it can be achieved with human products in an ex vivo experimental setting 51. DHF–DSS is often diagnosed in patients with primary infection, and capillary leakage has been reported to more frequently accompany primary infection than secondary infection in certain studies52. Secondary infection is thus not a requirement for DHF–DSS and should
NATURE REVIEWS | MICROBIOLOGY © 2013 Macmillan Publishers Limited. All rights reserved

Epidermis DC-SIGN Mannose receptor Viral RNA Infected DC Activated mast cell Dermis Lymphatic vessel NK cell Trafficking to lymph node Cellular recruitment Infected macrophage Cellular cytotoxicity

DENV

T cell

Blood vessel

Figure 2 | Host responses to cutaneous dengue virus injection.  Most infections by dengue virus Nature Reviews | Microbiology (DENV) occur after subcutaneous injection of the virus into the skin. Released viral particles may infect nearby cells (thought to be predominantly monocytes or dendritic cells (DCs)) or activate resident immune cells such as mast cells. A local inflammatory response to DENV in the skin prompts the recruitment of leukocytes from the vasculature, including natural killer (NK) cells and T cells, which promote the killing of virus-infected cells at the site of injection. DENV is thought to then travel to draining lymph nodes via lymphatic vessels to establish systemic infection. These localized inflammatory responses occur many days before there are any signs of severe infection.

not be a requirement for experimental design to understand dengue pathogenesis. Understanding the differential processes that lead to dengue pathogenesis during primary versus secondary infection is an important aim for the field.
The role of immune pathology Some of the emphasis on the role of antibody in dengue pathogenesis comes from the experimental demonstration that antibody alone or post-immune serum is sufficient to enhance the pathology of DENV infection in animal models29,53. Similar studies demonstrating causation are still needed for host responses that, on the basis of human studies, are associated with vascular leak syndrome, such as cytokine storm. Factors that might be correlated with dengue severity include TNF and VEGF, but animal studies have not yet clearly established the role of these host factors in disease progression, the cellular source of these factors, or

whether unidentified factors released along a similar time course could be more consequential to pathology. Recent data show that mast cell proteases such as chymase are released early during acute infection, promoting DENV-induced vascular leakage in mice, and serum levels of chymase have been correlated with the severity of disease in human patients20. This is one example of an immunomodulatory product that is released along with cytokines early during acute DENV infection. Another group of vasoactive lipid products, the leukotrienes, also seem to enhance DENV-induced vascular leakage in the immunocompetent mouse model20. Blockade of TNF can reduce vascular pathology in immunocompromised animals during terminal disease54, but its role in the development of DHF–DSS in humans remains unclear 36. Cytokine storm can also occur as a result of infections of entirely different aetiology without achieving the highly specific symptoms that are attributed
VOLUME 11 | JUNE 2013 | 423

PERSPECTIVES
to DENV infection, suggesting that more information is required if we are to fully understand the mechanisms underlying DENV-induced vascular pathology.
Virus-intrinsic or host-associated risk factors Virus-intrinsic factors are likely to influence the ability of certain strains to replicate robustly or induce severe pathology. For example, mutations in the DENV polymerase non-structural protein 3 (NS3) and envelope protein enhance viral fitness and promote neurovirulence, potentially by altering viral binding to cell surface receptors or increasing the efficiency of NS3 (REF. 55). There are many examples of interactions between DENV proteins and immune products (for example, NS1 binds to complement component C4 (REF. 56), and NS5 binds to signalling intermediary STAT2 (signal transducer and activator or transcription 2) to antagonize the IFN pathway 57). Phylogenetic analyses have shown that the emergence of a new DENV variant can correlate with an increased likelihood of DHF–DSS in humans11. Certain changes within DENV structural proteins could alter antibody binding to viral epitopes58, but some regions that have shown variation do not correspond to surface epitopes of mature DENV virions11. Evolution of non-structural genes has also been associated with increased epidemic potential59 and reduced disease severity 10, supporting the existence of a DENV-intrinsic component to virulence that requires further evaluation. Conversely, some host genetic factors are either associated with or protective against the development of DHF–DSS. In addition to TNF12, certain isoforms of major histocompatibility complex (MHC) class I molecules, MHC class II molecules, immuno­ globulin G Fc region receptor II (FcγRII) molecules, cytotoxic T lymphocyte protein 4 (CTLA4) and transforming growth factor β1 (TGFβ1) have all been linked to DHF–DSS13,60–63, whereas in some cases, particular isoforms seem to be protective61. Certain polymorphisms in TAP1 and TAP2, genes associated with antigen presentation and peptide loading on MHC, also have positive or negative correlations with DHF–DSS64. These studies have identified target host factors that might influence dengue pathogenesis; however, as many of these factors have been identified in an individual population or outbreak, they require corroboration in independent studies and subsequent experimental validation. Virus–host interactions are not the same for all infected patients or strains of virus. This fact is underscored by the broad
424 | JUNE 2013 | VOLUME 11 © 2013 Macmillan Publishers Limited. All rights reserved

spectrum of clinical outcomes for patients infected with DENV, from asymptomatic infection to classical dengue fever and, occasionally, lethal complications of haemorrhage and/or shock. These complex factors and seemingly contradictory observations will probably be reconciled only with controlled experimental investigation, which would require preclinical studies using animal models and a range of unadapted viral strains.
Animal models of dengue virus infection Many animal models have been used in DENV research, including multiple strains of immunocompetent and immunocompromised mice, New World and Old World primates, and even humanized animals65. In the pursuit of an ideal model, however, self-imposed boundaries have emerged with regards to acceptable in vivo experimentation using the tools that are currently available. Non-human primates have been underused, in part because results have been highly variable, depending on the primate species and infecting viral strain used. In some studies, primates have shown mild symptoms of DENV infection or vascular leakage66–68. Primates have also been used to test the safety, immunogenicity and protective capacity of vaccine candidates before these candidates are tested in human clinical trials69. This stage of preclinical testing in primates is virtually inevitable for dengue vaccine candidates and itself warrants further study of the pathophysiological changes caused by DENV in the primate model. Immunocompetent animals (including primates, but especially wild-type mice) are considered a poor choice for DENV research, even in studies addressing immune function directly, owing to their relative resistance to infection compared with humans70,71. It should be noted, however, that each model has distinct benefits and limitations. In addition, many of the accepted limitations of certain models (such as the expectation that immunocompetent mice would not support infections with replicating viruses) are not adequately represented in the literature with data, but rather, are discussed anecdotally 70,71. This makes assessing other potential causes of model ‘failure’, such as the choice of viral strain, difficult. Some of these concepts regarding the resistance of certain species to infection date back to the earliest in vivo studies carried out using DENV72, and scarce supporting data have been published in the intervening time. As new techniques have now become available, more sensitive means of measuring

viral and immune outcomes might also be useful. Examples that are currently being applied in some studies include quantitative real-time reverse transcription PCR (qRT-PCR), which can be used as a sensitive substitute for plaque assays (although this has the caveat that qRT-PCR does not accurately measure infectious DENV), and flow cytometry, which can be used in place of immunohistochemistry to quantify rare infected populations. Therefore, it is worthwhile revisiting lesser-used models to determine whether additional information about dengue pathogenesis and the host immune response can be garnered using modern techniques and a broader range of low-passage viral strains. In response to the difficulty of assessing DENV virology in immunocompetent animals, and in an effort to achieve uniform severity of infection and clearly interpretable results, most recent animal studies have gravitated towards the use of immunocompromised models (such as the AG129 IFN-deficient mouse model) and ‘antibody-enhanced’ infection (often using monoclonal antibodies to DENV, along with mouse-adapted DENV). These systems have clear advantages for answering certain questions. For example, investigations with the aim of identifying antiviral drugs might be most definitive in immunocompromised mice, in which almost unrestricted DENV replication can occur and there is uniform disease severity 73. However, we must be careful not to overextend the conclusions from such experiments. In many aspects, particularly relating to host immunity and potentially also to immune-mediated path­ ology, there might be differences between the mouse immune response and the response that develops during human infection. Moreover, we do not know whether the underlying mechanisms of key pathological events such as vascular leakage would be the same in mice as during human disease or whether the same cell types that become infected in humans are infected in immunocompromised animals. The absence of key inflammatory pathways could allow otherwise resistant mouse cells to be permissive for DENV replication. In our opinion, the animal models that have lent themselves to studying the basic virology of DENV, such as immunocompromised mice, might not be the most appropriate for studying immunity or immune pathogenesis. We do not advocate the adoption of any one existing model as the gold standard because they each have their pros, cons and caveats. Rather, as outlined below, researchers should select the
www.nature.com/reviews/micro

PERSPECTIVES
most appropriate existing model, with these considerations in mind, to address a unique experimental question. Investigations that aim to understand the development of effective anti-dengue immunity and to use that understanding as a basis for rational vaccine design require an intact immune system to fully encompass the immune response. Immunocompetent mice are thought to be relatively resistant to DENV compared with humans, but this fact also affords us the opportunity to study effective anti-dengue immunity in a host that can respond to, and clear, the infection. Using selective knockout mice could provide a mechanistic understanding of how unique cell types or inflammatory factors contribute to immunity and pathogenesis. There is likely to be a complex interplay between the processes of anti-dengue immunity and DENV-induced pathology, as supported by the association of secondary infection with variable disease severity during dengue outbreaks27,74–76. Studying immune pathology mechanistically has proved difficult because mice with functional immune systems rarely fully recapitulate the severity of human DHF. However, we challenge the premise that full recapitulation of the disease time course or pathological phenotype is a necessity. Changes that occur in vivo in response to DENV infection need not be equally detrimental to the survival of animals and humans to reveal information regarding the progression of disease, and such information could be crucial for identifying a therapeutic or prophylactic strategy. Using animal models that have fallen out of favour for dengue research, such as wild-type immunocompetent mice and non-human primates, to study those physiological changes will require careful experimental design, a realistic assessment of the likelihood that conclusions can be extended to human disease, the use of selected unpassaged viral strains and subtypes, and the use of more sensitive measures of pathology than the extreme outcomes of haemorrhaging, prolonged viraemia or death. Some potential surrogates for end-stage pathology include more sensitive methods to objectively measure vascular permeability or immune cell signalling, statistically significant changes in the number of haematopoietic cells (for example, loss of platelets or proliferation of lymphocytes), the levels of immune factors upstream of vascular leakage (such as heightened levels of cytokines or other vasoactive factors), and alternate time points of assessment to those thought to be analogous to the human course of infection. Thus, individual animal
NATURE REVIEWS | MICROBIOLOGY © 2013 Macmillan Publishers Limited. All rights reserved

models for dengue might have limitations, but should not be dismissed as inherently flawed. It should be noted that severe dengue does not manifest uniformly in humans, so lack of severity in a model should not prevent the study of dengue in varied models. Conversely, as the limitations of animal models for dengue are well known, the results of these studies should be used to inform our collective knowledge of DENV virology, pathogenesis and immunity in order to introduce new theories and concepts to the field, but not to limit the potential of similar investigations in humans. For example, there seem to be differences in the ability of humans and mice to develop neutralizing antibodies to certain epitopes of DENV proteins77. For immunocompromised mice, drugs might reduce dengue pathology but never allow long-term survival of animals that cannot clear the infection. Thus, limitations in the degree to which we can extend conclusions from animal studies should not inhibit our pursuit of understanding this pathogen in a range of animal models as well as in reasoned human trials.
Conclusions The field of dengue research has made great progress over the past 50 years, particularly in terms of defining dengue clinically, determining how to identify and treat dengue with supportive care, and understanding the epidemiology and basic biology of DENV, including viral structure and replication. Animal models have been developed for antiviral drug testing, and many groups have designed promising vaccine candidates. Importantly, many of the remaining questions about DENV as a pathogen centre on how the immune system, both innate and acquired, perceives and responds to DENV infection, how genetic variation in both the host and the virus influence this perception, and where the fine line lies for protection versus pathology. These questions cannot be definitively answered in the context of an immunocompromised system alone, but must be directly addressed with bold experimental design of preclinical studies, including in immunocompetent animals, to fully define anti-dengue immunity. We believe that meeting this challenge will be a prerequisite to identifying the most effective targeted therapies and vaccine strategies against dengue.
Ashley L. St. John, Soman N. Abraham and Duane J. Gubler are at the Program in Emerging Infectious Diseases, Graduate Medical School, Duke-National University of Singapore, 8 College Road, 169857, Singapore.

Ashley L. St. John and Soman N. Abraham are also at the Department of Pathology, Duke University Medical Center. Soman N. Abraham is also at the Department of Immunology and the Department of Molecular Genetics and Microbiology, Duke University Medical Center, Durham, North Carolina 27710, USA. Correspondence to D.J.G.  e‑mail: [email protected] doi:10.1038/nrmicro3030 Published online 8 May 2013
1. Gubler, D. J. Dengue, urbanization and globalization: the unholy trinity of the 21st century. Trop. Med. Health 39, 3–11 (2011). 2. Gubler, D. J. Dengue and dengue hemorrhagic fever. Clin. Microbiol. Rev. 11, 480–496 (1998). 3. Wilder-Smith, A. et al. DengueTools: innovative tools and strategies for the surveillance and control of dengue. Glob. Health Action 5, 17273 (2012). 4. Siler, J. F., Hall, M. W. & Hitchens, A. P. Results obtained in the transmission of dengue fever. JAMA 84, 1163–1172 (1925). 5. Sabin, A. B. Research on dengue during World War II. Am. J. Trop. Med. Hyg. 1, 30–50 (1952). 6. Aggarwal, A., Chandra, J., Aneja, S., Patwari, A. K. & Dutta, A. K. An epidemic of dengue hemorrhagic fever and dengue shock syndrome in children in Delhi. Indian Pediatr. 35, 727–732 (1998). 7. Vaughn, D. W. et al. Dengue viremia titer, antibody response pattern, and virus serotype correlate with disease severity. J. Infect. Dis. 181, 2–9 (2000). 8. Bhamarapravati, N., Tuchinda, P. & Boonyapaknavik, V. Pathology of Thailand haemorrhagic fever: a study of 100 autopsy cases. Ann. Trop. Med. Parasitol. 61, 500–510 (1967). 9. Zompi, S., Montoya, M., Pohl, M. O., Balmaseda, A. & Harris, E. Dominant cross-reactive B cell response during secondary acute dengue virus infection in humans. PLoS Negl. Trop. Dis. 6, e1568 (2012). 10. Steel, A., Gubler, D. J. & Bennett, S. N. Natural attenuation of dengue virus type‑2 after a series of island outbreaks: a retrospective phylogenetic study of events in the South Pacific three decades ago. Virology 405, 505–512 (2010). 11. Messer, W. B., Gubler, D. J., Harris, E., Sivananthan, K. & de Silva, A. M. Emergence and global spread of a dengue serotype 3, subtype III virus. Emerg. Infect. Dis. 9, 800–809 (2003). 12. Vejbaesya, S. et al. TNF and LTA gene, allele, and extended HLA haplotype associations with severe dengue virus infection in ethnic Thais. J. Infect. Dis. 199, 1442–1448 (2009). 13. Chen, R. F. et al. Combination of CTLA‑4 and TGFβ1 gene polymorphisms associated with dengue hemorrhagic fever and virus load in a dengue‑2 outbreak. Clin. Immunol. 131, 404–409 (2009). 14. de la, C. S. B., Kouri, G. & Guzman, M. G. Race: a risk factor for dengue hemorrhagic fever. Arch. Virol. 152, 533–542 (2007). 15. Wu, S. J. et al. Human skin Langerhans cells are targets of dengue virus infection. Nature Med. 6, 816–820 (2000). 16. St John, A. L. et al. Immune surveillance by mast cells during dengue infection promotes natural killer (NK) and NKT-cell recruitment and viral clearance. Proc. Natl Acad. Sci. USA 108, 9190–9195 (2011). 17. Tassaneetrithep, B. et al. DC-SIGN (CD209) mediates dengue virus infection of human dendritic cells. J. Exp. Med. 197, 823–829 (2003). 18. Miller, J. L. et al. The mannose receptor mediates dengue virus infection of macrophages. PLoS Pathog. 4, e17 (2008). 19. Marchette, N. J. et al. Studies on the pathogenesis of dengue infection in monkeys. 3. Sequential distribution of virus in primary and heterologous infections. J. Infect. Dis. 128, 23–30 (1973). 20. St John, A. L., Rathore, A. P., Raghavan, B., Ng, M. L. & Abraham, S. N. Contributions of mast cells and vasoactive products, leukotrienes and chymase, to dengue virus-induced vascular leakage. eLife (in the press). 21. Rothman, A. L. Immunity to dengue virus: a tale of original antigenic sin and tropical cytokine storms. Nature Rev. Immunol. 11, 532–543 (2011). 22. Wang, T. et al. IFN-γ-producing γδ T cells help control murine West Nile virus infection. J. Immunol. 171, 2524–2531 (2003).

VOLUME 11 | JUNE 2013 | 425

PERSPECTIVES
23. Starnes, T. et al. Cutting edge: IL‑17F, a novel cytokine selectively expressed in activated T cells and monocytes, regulates angiogenesis and endothelial cell cytokine production. J. Immunol. 167, 4137–4140 (2001). 24. Tsai, Y. T., Chang, S. Y., Lee, C. N. & Kao, C. L. Human TLR3 recognizes dengue virus and modulates viral replication in vitro. Cell. Microbiol. 11, 604–615 (2009). 25. Loo, Y. M. et al. Distinct RIG‑I and MDA5 signaling by RNA viruses in innate immunity. J. Virol. 82, 335–345 (2008). 26. Halstead, S. B. Antibody, macrophages, dengue virus infection, shock, and hemorrhage: a pathogenetic cascade. Rev. Infect. Dis. 11 (Suppl. 4), S830–S839 (1989). 27. Halstead, S. B. Dengue. Lancet 370, 1644–1652 (2007). 28. Gubler, D. J. & Zaki, S. in Pathology of Emerging Infections 2 Ch. 3 (eds Nelson, A. M. & Horsburgh, C. R.) 43–71 (American Society for Microbiology Press, 1998). 29. Zellweger, R. M., Prestwood, T. R. & Shresta, S. Enhanced infection of liver sinusoidal endothelial cells in a mouse model of antibody-induced severe dengue disease. Cell Host Microbe 7, 128–139 (2010). 30. Jessie, K., Fong, M. Y., Devi, S., Lam, S. K. & Wong, K. T. Localization of dengue virus in naturally infected human tissues, by immunohistochemistry and in situ hybridization. J. Infect. Dis. 189, 1411–1418 (2004). 31. Halstead, S. B. Controversies in dengue pathogenesis. Paediatr. Int. Child Health 32 (Suppl. 1), 5–9 (2012). 32. Hober, D. et al. Serum levels of tumor necrosis factoralpha (TNF-α), interleukin‑6 (IL‑6), and interleukin‑1β (IL‑1β) in dengue-infected patients. Am. J. Trop. Med. Hyg. 48, 324–331 (1993). 33. Edwards, M. J., Heniford, B. T. & Miller, F. N. Tumor necrosis factor mediates disseminated intravascular inflammation (DII) in the genesis of multiple organ edema. J. Surg. Res. 54, 140–144 (1993). 34. Kunder, C. A., St John, A. L. & Abraham, S. N. Mast cell modulation of the vascular and lymphatic endothelium. Blood 118, 5383–5393 (2011). 35. Wang, L. et al. Implications of dynamic changes among tumor necrosis factor-α (TNF-α), membrane TNF receptor, and soluble TNF receptor levels in regard to the severity of dengue infection. Am. J. Trop. Med. Hyg. 77, 297–302 (2007). 36. Chakravarti, A. & Kumaria, R. Circulating levels of tumour necrosis factor-α and interferon-γ in patients with dengue and dengue haemorrhagic fever during an outbreak. Indian J. Med. Res. 123, 25–30 (2006). 37. Halstead, S. B. et al. Dengue hemorrhagic fever in infants: research opportunities ignored. Emerg. Infect. Dis. 8, 1474–1479 (2002). 38. Kliks, S. C., Nimmanitya, S., Nisalak, A. & Burke, D. S. Evidence that maternal dengue antibodies are important in the development of dengue hemorrhagic fever in infants. Am. J. Trop. Med. Hyg. 38, 411–419 (1988). 39. Hawkes, R. A. Enhancement of the infectivity of Arboviruses by specific antisera produced in domestic fowls. Aust. J. Exp. Biol. Med. Sci. 42, 465–482 (1964). 40. Goncalvez, A. P., Engle, R. E., St Claire, M., Purcell, R. H. & Lai, C. J. Monoclonal antibodymediated enhancement of dengue virus infection in vitro and in vivo and strategies for prevention. Proc. Natl Acad. Sci. USA 104, 9422–9427 (2007). 41. Daughaday, C. C., Brandt, W. E., McCown, J. M. & Russell, P. K. Evidence for two mechanisms of dengue virus infection of adherent human monocytes: trypsinsensitive virus receptors and trypsin-resistant immune complex receptors. Infect. Immun. 32, 469–473 (1981). 42. Halstead, S. B., Chow, J. S. & Marchette, N. J. Immunological enhancement of dengue virus replication. Nature New Biol. 243, 24–26 (1973). 43. Halstead, S. B. & O’Rourke, E. J. Antibody-enhanced dengue virus infection in primate leukocytes. Nature 265, 739–741 (1977). 44. Halstead, S. B., Venkateshan, C. N., Gentry, M. K. & Larsen, L. K. Heterogeneity of infection enhancement of dengue 2 strains by monoclonal antibodies. J. Immunol. 132, 1529–1532 (1984). 45. Webster, R. G. Original antigenic sin in ferrets: the response to sequential infections with influenza viruses. J. Immunol. 97, 177–183 (1966). 46. Mongkolsapaya, J. et al. Original antigenic sin and apoptosis in the pathogenesis of dengue hemorrhagic fever. Nature Med. 9, 921–927 (2003). 47. Halstead, S. B., Rojanasuphot, S. & Sangkawibha, N. Original antigenic sin in dengue. Am. J. Trop. Med. Hyg. 32, 154–156 (1983). 48. Singh, R. A., Rodgers, J. R. & Barry, M. A. The role of T cell antagonism and original antigenic sin in genetic immunization. J. Immunol. 169, 6779–6786 (2002). 49. Gubler, D. J., Reed, D., Rosen, L. & Hitchcock, J. R. Jr. Epidemiologic, clinical, and virologic observations on dengue in the Kingdom of Tonga. Am. J. Trop. Med. Hyg. 27, 581–589 (1978). 50. Vaughn, D. W. et al. Dengue in the early febrile phase: viremia and antibody responses. J. Infect. Dis. 176, 322–330 (1997). 51. Kliks, S. C., Nisalak, A., Brandt, W. E., Wahl, L. & Burke, D. S. Antibody-dependent enhancement of dengue virus growth in human monocytes as a risk factor for dengue hemorrhagic fever. Am. J. Trop. Med. Hyg. 40, 444–451 (1989). 52. Meltzer, E., Heyman, Z., Bin, H. & Schwartz, E. Capillary leakage in travelers with dengue infection: implications for pathogenesis. Am. J. Trop. Med. Hyg. 86, 536–539 (2012). 53. Balsitis, S. J. et al. Lethal antibody enhancement of dengue disease in mice is prevented by Fc modification. PLoS Pathog. 6, e1000790 (2010). 54. Shresta, S., Sharar, K. L., Prigozhin, D. M., Beatty, P. R. & Harris, E. Murine model for dengue virus-induced lethal disease with increased vascular permeability. J. Virol. 80, 10208–10217 (2006). 55. de Borba, L., Strottmann, D. M., de Noronha, L., Mason, P. W. & Dos Santos, C. N. Synergistic interactions between the NS3hel and E proteins contribute to the virulence of dengue virus type 1. PLoS Negl Trop. Dis. 6, e1624 (2012). 56. Avirutnan, P. et al. Antagonism of the complement component C4 by flavivirus nonstructural protein NS1. J. Exp. Med. 207, 793–806 (2010). 57. Ashour, J. et al. Mouse STAT2 restricts early dengue virus replication. Cell Host Microbe 8, 410–421 (2010). 58. Kochel, T. J. et al. Effect of dengue‑1 antibodies on American dengue‑2 viral infection and dengue haemorrhagic fever. Lancet 360, 310–312 (2002). 59. Bennett, S. N. et al. Selection-driven evolution of emergent dengue virus. Mol. Biol. Evol. 20, 1650–1658 (2003). 60. Zivna, I. et al. T cell responses to an HLA‑B*07‑ restricted epitope on the dengue NS3 protein correlate with disease severity. J. Immunol. 168, 5959–5965 (2002). 61. Chiewsilp, P., Scott, R. M. & Bhamarapravati, N. Histocompatibility antigens and dengue hemorrhagic fever. Am. J. Trop. Med. Hyg. 30, 1100–1105 (1981). 62. Loke, H. et al. Strong HLA class I–restricted T cell responses in dengue hemorrhagic fever: a doubleedged sword? J. Infect. Dis. 184, 1369–1373 (2001). 63. Fernandez-Mestre, M. T., Gendzekhadze, K., RivasVetencourt, P. & Layrisse, Z. TNF-α‑308A allele, a possible severity risk factor of hemorrhagic manifestation in dengue fever patients. Tissue Antigens 64, 469–472 (2004). 64. Soundravally, R. & Hoti, S. L. Polymorphisms of the TAP 1 and 2 gene may influence clinical outcome of primary dengue viral infection. Scand. J. Immunol. 67, 618–625 (2008). 65. Cassetti, M. C. et al. Report of an NIAID workshop on dengue animal models. Vaccine 28, 4229–4234 (2010). 66. Onlamoon, N. et al. Dengue virus-induced hemorrhage in a nonhuman primate model. Blood 115, 1823–1834 (2010). 67. Halstead, S. B., Shotwell, H. & Casals, J. Studies on the pathogenesis of dengue infection in monkeys. I. Clinical laboratory responses to primary infection. J. Infect. Dis. 128, 7–14 (1973). 68. Omatsu, T. et al. Common marmoset (Callithrix jacchus) as a primate model of dengue virus infection: development of high levels of viraemia and demonstration of protective immunity. J. Gen. Virol. 92, 2272–2280 (2011). 69. Guirakhoo, F. et al. A single amino acid substitution in the envelope protein of chimeric yellow fever-dengue 1 vaccine virus reduces neurovirulence for suckling mice and viremia/viscerotropism for monkeys. J. Virol. 78, 9998–10008 (2004). 70. Bente, D. A. & Rico-Hesse, R. Models of dengue virus infection. Drug Discov. Today Dis. Models 3, 97–103 (2006). 71. Yauch, L. E. & Shresta, S. Mouse models of dengue virus infection and disease. Antiviral Res. 80, 87–93 (2008). 72. Simmons, J. S., St. John, J. H., & Reynolds, H. K. Experimental Studies of Dengue (Bureau of Printing, Manila, 1931). 73. Rathore, A. P. et al. Celgosivir treatment misfolds dengue virus NS1 protein, induces cellular pro-survival genes and protects against lethal challenge mouse model. Antiviral Res. 92, 453–460 (2011). 74. Watts, D. M. et al. Failure of secondary infection with American genotype dengue 2 to cause dengue haemorrhagic fever. Lancet 354, 1431–1434 (1999). 75. Thein, S. et al. Risk factors in dengue shock syndrome. Am. J. Trop. Med. Hyg. 56, 566–572 (1997). 76. Vitarana, T., de Silva, H., Withana, N. & Gunasekera, C. Elevated tumour necrosis factor in dengue fever and dengue haemorrhagic fever. Ceylon Med. J. 36, 63–65 (1991). 77. Wahala, W. M., Huang, C., Butrapet, S., White, L. J. & de Silva, A. M. Recombinant dengue type 2 viruses with altered e protein domain III epitopes are efficiently neutralized by human immune sera. J. Virol. 86, 4019–4023 (2012).

Acknowledgements

The authors thank A. P. S. Rathore for critical reading of the manuscript. This work was supported by the National Medical Research Council of Singapore (grant NIG/1053/2011) and by the Duke-National University of Singapore Signature Research Program, funded by the Ministry of Health, Singapore.

Competing interests statement

The authors declare no competing financial interests.

FURTHER INFORMATION
Duane J. Gubler’s homepage: https://www.duke-nus.edu.sg/content/gubler-duane-j
ALL LINKS ARE ACTIVE IN THE ONLINE PDF

426 | JUNE 2013 | VOLUME 11 © 2013 Macmillan Publishers Limited. All rights reserved

www.nature.com/reviews/micro

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close