Rapid, unplanned growth of urban centers in South-East Asian and South American countries combined with inadequate water supply and sewerage systems have dramatic consequences on the transmission of DENV (16, 17).
The DENVs
Replication cycle
DENV is an enveloped, single-stranded positive-sense RNA virus. The RNA genome consists of approximately 10,700 nucleotides and encodes a 3,411 amino acids long precursor polyprotein containing three structural proteins (capsid [C], precursor membrane [prM], and envelope [E]) and seven non-structural (NS) proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5). The structural proteins are components of the mature virus particle whereas the NS proteins are expressed only in the infected cell and are not packaged to detectable levels into mature particles. The structural proteins are not involved in replication of the viral genome (18-20).
The open reading frame is flanked by two untranslated regions (5’ and 3’ UTR) of approximately 95–135 and 114–650 nucleotides, respectively. The 5’-end contains a type I cap, similar to cellular mRNA, and the viral RNA (vRNA) is translated by a cap-dependent initiation scanning the 5’-UTR. The 3’-end lacks a poly(A) tail but ends in a conserved stem-loop (SL) structure. Both the 5’- and 3’-UTRs are required for efficient translation and replication (21, 22). The UTRs have characteristic secondary structures that confer distinct functions and show high sequence conservation among different DENV serotypes. The 5’-UTR contains a large stem-loop (SLA) that is proposed to act as the promoter for the viral RNA-dependent RNA polymerase (RdRp) NS5 (23). Both the 5’- and the 3’-UTRs contain complementary Upstream AUG Regions (UAR) and cyclization sequences (CS) that hybridize in order to mediate genome cyclization and RNA synthesis (21).
The various steps in the flavivirus life cycle include 1) virions binding to cell-surface attachment molecules and receptors, and are internalized through endocytosis. 2) Due to the low pH of the endosome, viral glycoproteins mediate fusion of viral and cellular membranes, allowing disassembly of the virion and release of vRNA into the cytoplasm. 3) vRNA is translated into a polyprotein that is processed by viral and cellular proteases, and 4) the viral NS proteins replicate the genome RNA. 5) Virus assembly occurs at the endoplasmic reticulum (ER) membrane, where C protein and vRNA are enveloped by the ER membrane and glycoproteins to form immature virus particles. 6) Immature virus particles are transported through the secretory pathway, and in the acidic environment of the trans-Golgi network (TGN), furin-mediated cleavage of prM drives maturation of the virus. 7) Mature virus is released from the cell.
Virus entry
Viral entry into the host cell is mediated by receptor-mediated endocytosis through an as yet unidentified cell-surface receptor. Candidate cellular receptors required for viral entry are various glycoproteins (i.e. heparin sulfates), dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN), or a mannose receptor (24-26). The human C-type lectin-like molecule CLEC5A has been suggested to act as a critical macrophage receptor for DENV and has been described as a proinflammatory receptor for DENV that contributes to lethal disease in mice (27, 28). There is a general consensus that the viral E glycoprotein affects host cell receptor binding, viral entry, and is a major target for humoral immunity. The E protein is composed of three domains: domain I, domain II harboring the fusion peptide at its distal tip, and domain III responsible for receptor-binding activity. In the mature state, E exists as a homodimer with the fusion peptide inaccessible. Low-pH–induced trimerization exposes the hydrophobic fusion peptide in a manner consistent with membrane fusion mediated by class II fusion protein (29). Mutations of residues constituting the ligand pocket at the interface of domain I and II alter the required pH threshold and affect virulence (30). There are two potential asparagine (N)-linked glycosylation sites at positions Asn-67 and Asn-153, whereof the former is unique for DENVs and the latter is conserved in most flaviviruses (31). The glycosylation pattern differs according to DENV serotype and even among different strains, as well as the cells in which the virus is propagated. The degree and position of N-linked glycans affect the antigenic properties of DENV (32-34).
Upon internalization, the acidic pH in the endosome triggers a conformational change in the E protein mediating membrane fusion. The viral nucleocapsid is released into the cytoplasm whereupon the virus uncoats and releases the genome (31). The input positive-strand vRNA is translated into a single polyprotein, which is cleaved into the individual structural and NS proteins. The input strand translation is followed by a switch from translation to synthesis of a negative-strand intermediate, which serves as a template for new positive-strand vRNA. Multiple rounds of translation produce high levels of viral proteins that together with vRNA are assembled into progeny virions (35).
DENV infection induces intracellular membrane alterations in the cytosol forming vesicle packets (VPs) or smooth membrane structures (SMS) where the viral replication complex (RC) accumulates (36, 37). The induction of membrane structures may serve as a scaffold for anchoring the viral RC. The C-terminal regions of C, prM, and E contain hydrophobic amino acids that serve as signal sequences for insertion of the remaining protein into the ER membrane (38). An ER signal peptidase together with the viral NS2B-NS3 protease cleaves the structural proteins and NS1 protein into individual membrane-bound proteins (39-41). The NS3 protein acts, together with its cofactor NS2B, as the viral serine protease needed for polyprotein-processing through its N-terminal end (42, 43). This heterodimer protein complex cleaves on the cytoplasmic side of the ER membrane at the junctions between NS2A-NS2B, NS2B-NS3, NS3-NS4A, and NS4B-NS5, as well as on the internal sites within C, NS2A, NS3, and NS4A.
The C-terminal end of the NS3 protein has three enzymatic properties: a 5’ RNA-triphosphatase (RTP), a nucleoside triphosphatase (NTPase), and a helicase. NS3 forms a complex with NS5 and assists in viral replication through unwinding of RNA and dephosphorylation prior to 5’-end capping. The remaining NS proteins are cleaved by the viral serine protease NS3 that requires NS2B as a cofactor for catalytic activity (43). However, a host cell signal peptidase mediates post-translational modifications on the NS4A-4B proteins (44).
The NS1 is a glycoprotein with two glycosylation sites that are conserved among flaviviruses. It is synthesized in the ER as a hydrophilic monomer but exists as a more hydrophobic homodimer. The NS1 dimer is transported to the Golgi apparatus where it undergoes carbohydrate trimming (45). The role of NS1 in virus replication is unknown but is believed to facilitate viral infection and DENV pathogenesis (46). NS1 is in addition secreted from infected cells (sNS1) and has been shown to be immunologically important (47, 48). Antibodies raised against sNS1 proteins have been proposed to cause endothelial dysfunction due to cross-reactivity to host proteins and endothelial cells (49). Data indicate that sNS1 could be an important modulator of the complement pathway and is proposed to protect DENV from complement-dependent neutralization in solution (47).
The small hydrophobic proteins NS2A, NS4A, and NS4B are less well characterized. Recent findings propose an inhibitory role in interferon (IFN)-mediated signal transduction. Their hydrophobic nature potentially implicates them in proper localization of viral proteins and vRNA during replication and virion assembly. Formation of DENV-induced cytoplasmic membrane structures are believed to be an arrangement of the NS4A protein.
The largest NS protein encoded in the DENV genome is the NS5 protein, approximately 103 kDa big. The NS5 protein has three major functional domains: the N-terminal S-adenosyl methionine methyltransferase (MTase), the nuclear localization sequences (NLS), and the RdRp activity in its C-terminal domain. The MTase spans amino acid residues 1 to 239 and is responsible for guanine N-7 and ribose 2’-O-methylations required for the capping of the DENV genome. The cap structure is recognized by the host cell translational machinery. The NLS (residues 320–405) interacts with the NS3 viral helicase and is recognized by cellular factors, allowing protein transport to the nucleus. The NS5 polymerase domain RdRp (residue 273–900) is responsible for synthesizing new vRNA genomes (43).
Prior to secretion of new viral particles, the third structural protein (pr)M is processed into the mature M protein in the TGN by furin host protease (50). It is believed that prM protects the E proteins from pH-induced reorganization and premature fusion during secretion; hence, the maturation event is necessary for infectivity (51-53).
Laboratory diagnosis
Since laboratory-based dengue diagnosis is often unavailable at the time of care, the preliminary diagnosis relies on a combination of travel history and clinical symptoms. Travel history provides key information that can rule out other potentially life-threatening diseases since the incubation period of DENV is less than 2 weeks (54). A confirmed diagnosis for a DENV infection is established by culture of the virus, polymerase-chain reaction (PCR), or serologic assays. There are, however, limitations with each test, and detection is based on different virological markers, namely infectious virus, vRNA, and DENV-specific antibodies, respectively.
Culturing the virus requires an acute patient serum with sufficient levels of virus, and the period when DENV can be successfully isolated in patient serum is short. Viremia peaks before the onset of symptoms, hence virus levels might drop significantly once the patient seeks medical care. Furthermore, rising levels of antibody interfere with virus culture already within a day or two after the subsidence of fever. Apart from sample collection limitations, practical considerations limit the use of this method. Culture of the virus is both time- and labor intensive; infectious patient material must be kept cold, and a bio-safety level 3 laboratory is required, necessitating professional training of the personnel. These requirements limit the use of this diagnostic tool, especially in rural areas (4).
The Ae. albopictus cell line C6/36 (CRL 1660, ATCC) is commonly used to isolate DENVs from patient material. Specimens that may be suitable for virus isolation include acute phase serum; plasma or washed buffy coat from the patient; autopsy tissues from fatal cases, especially liver, spleen, lymph nodes, and thymus; and mosquitoes collected in nature (55). Detection of vRNA from serum, plasma, or cells with PCR is based on DENV-specific oligonucleotide primers, and is fast and robust, although sensitive only in the very early stages of disease (54). PCR is particularly useful in situations when virus culture has not been successful but nevertheless depends on sample collection during the symptomatic phase.
The third laboratory diagnostic option is not based on direct detection but on the presence of anti-DENV antibodies. Thus, it is not hindered by the limitations of virus culture and PCR, and the timing of sample collection can be more flexible. The acute anti-DENV IgM antibody response lasts for a couple of weeks after infection and the IgG antibodies for several years. The immunoglobulins (Ig) are not easily inactivated and do not have the same strict requirements for low temperature as infectious virus specimen. The assay techniques are relatively simple and there are commercial diagnostic kits available, whereof the assays based on IgM detection are the most commonly used in routine diagnostics (4). The major drawback with serological tests is the considerable risk for false-positive results due to potential cross-reactivity with other flaviviruses, for example, vaccination against Yellow fever virus (YFV) (56).
Due to the drawbacks of serological methods to reliably diagnose acute infections, alternative methods based on the detection of the viral NS1 protein have been developed. NS1 can be found both membrane-associated inside the host cell and in a soluble, secreted form. The amount of secreted NS1 in patient serum correlates with viremia and DENV pathogenesis (46, 57-60), and the NS1 protein is detectable in serum by enzyme-linked immunosorbent assay (ELISA) from the first day of fever up to 9 days post-infection (6, 61-63). NS1-based ELISAs have become an important diagnostic tool for acute samples in which IgM is not detectable and where PCR is not available. Several commercial NS1 antigen kits are available and are widely used in endemic as well as non-endemic countries. The sensitivity varies from 63% to 94% (58, 62, 64, 65), and it depends on sample time-point, DENV serotype, and if it is a primary or secondary DENV infection 66).
Vaccines
Unlike flaviviruses such as YFV, Japanese encephalitis virus (JEV), and Tick-borne encephalitis virus (TBEV), no licensed vaccine exists for dengue. Vaccination must protect against all four serotypes without predisposing for antibody-dependent enhancement (ADE) and has proven difficult to design. Nearly 80 years of vaccine-related research and development have passed, and over 25 unique DENV vaccine candidates have been tested in clinical trials during the past decade.
To be safe, a dengue vaccine must be functionally tetravalent, eliciting simultaneous protection against all four DENV serotypes. Hence, vaccination cannot proceed in an analogous sequential manner, and herein lies the greatest obstacle (67-69). Live attenuated vaccines can induce durable humoral and cellular immune responses that mimic natural infection (70). However, the viral replication must be discrete to preclude the development of significant illness. A reasonable range of viremia for a live attenuated vaccine is believed to be approximately 101-102 infectious units/mL (71) compared to high levels of viremia upon natural infection that can be 105-107 infectious units/mL (59).
It is expected that a live attenuated vaccine would be successful and require only a single dose since the vaccine against YFV is based on a live attenuated virus. However, it is more likely that booster immunizations will be required based on results from clinical trials using tetravalent formulations of live vaccine candidates aimed at eliciting neutralizing antibodies (72-75) The obvious challenge is when and how to boost; infectivity and immunogenicity in NHP models have not always clearly predicted the outcome of human trials (76, 77). Vaccination compliance may also be lower with a multi-dose vaccination strategy, especially in regions where resources are scarce, and at the same time where the need for a vaccine often is the most acute.
Currently, there are several dengue vaccine candidates at different stages of preclinical or clinical development. The most advanced clinical development stage is a candidate developed by Sanofi Pasteur (CYD-TDV), which is under evaluation in phase II and phase III clinical studies. Phase III efficacy studies of CYD-TDV are currently underway in 31,000 children and adolescents in 10 countries in Asia and Latin America. These multi-center studies in a variety of epidemiological settings will be important to obtain data regarding efficacy and safety, and will shed further light on the relationship between vaccine-induced immune responses and protection against clinical dengue disease (73).
There are in addition other live-attenuated, subunit, and DNA vaccine candidates at earlier stages of clinical development. Other technological approaches include viral-vectored and virus-like particle vaccines, which currently are being tested in preclinical studies. It is hoped that clinical trials evaluating novel recombinant subunit proteins, DNA, and vectored vaccines would be initiated in the coming years. These approaches could be part of a prime-boost strategy, or stand-alone (78). The use of different types of vaccines depends on the purpose of vaccination and target group reflecting the disease setting. In endemic areas, there is an urgent need for routine immunization against dengue for infants and young children aged 1–3 years. A dengue vaccine would be coordinated with current childhood immunization schedules. Due to the socioeconomic status of many endemic countries, this type of vaccine ought to be inexpensive. In contrast, a protective vaccine for international travel, seasonal work personnel, and military staff that visit or work in DENV endemic areas are more tolerant to increased cost. Vaccination in this case will need to be rapid.
Hence, the different requirements for a dengue vaccine vary according to target group and their specific needs (life-long immunity or temporal protection), and efficient antiviral drugs would be a useful complement for protection and/or treatment. In addition, antiviral drugs would be more potent in an outbreak situation than a vaccine when there is no time to complete a multi-dose immunization schedule spanning 6 months or more.
Treatment and therapeutic approaches
Currently, vector control, regarded as both expensive and ineffective, is the only method for disease prevention (79, 80). In the absence of available vaccines and antiviral drugs against DENV infection, specific treatment for dengue patients consist primarily of supportive care including bed rest, antipyretics, and analgesics. Urgent resuscitation with intravenous fluids to replace lost intravascular volume in DSS patients is a prerequisite; Ringer's lactate has been shown to be effective in moderately severe dengue, and starch or dextran have been suggested for more severe cases (81). Aspirin and other salicylates should be avoided due to plasma leakage (6).
The design of novel therapeutic approaches for dengue disease has focused on the various stages of the viral replication cycle. The conformational changes of the E protein and its interaction with prM or M have been a major interest. These transition states present opportunities for antiviral targeting of the entry, assembly, or maturation steps of the virus life cycle. Antiviral peptides have been designed and tested for blocking of both DENV and West Nile virus (WNV) entry with positive results, indicating that antiviral peptides could be a promising form of DENV therapy. Targeting of mature virus entry into host cells is an extremely promising candidate since delivery of target compounds into the host cell during stages of fusion and maturation is significantly more challenging.
Another approach to inhibit the structural changes of the E-prM protein interactions has been to synthesize peptides mimicking the pr peptide of the M protein, thereby preventing membrane fusion and release of newly synthesized virions. The viral protease is another interesting target for antiviral discovery, since proteases are common to most viruses and generally important for efficient replication. Protease inhibitors for hepatitis C virus (HCV) may eventually be further developed to inhibit the DENV protease NS2B-NS3.
Nucleoside analogues are usually prodrugs that need to be converted to their antiviral nucleotide metabolite forms. Ribavirin (1-β-d-ribofuranosyl-1H-1, 2, 4-triazole-3-carboxamide) possesses broad spectrum antiviral activity and is used in combination with IFN to combat HCV infection. Ribavirin depletes the nucleotide pool and thereby indirectly affects capping and polymerase activities of both cellular and viral proteins. In addition, ribavirin causes a more error-prone replication of several viral genomes. Despite successful in vivo results with several RNA viruses, ribavirin has a cytostatic effect in DENV-infected cells and has not been effective in animal models.
Nucleic acid–based therapies offer various alternatives. RNA interference (RNAi) is thought to protect the host from viral infections by degrading the extraneous genetic material such as vRNA. It has been used in therapeutic approaches for several infectious diseases, tumors, and metabolic disorders. Small interfering RNA (siRNA) treatment reduces viral load of WNV in mice, but there are several obstacles yet to overcome; the RNA of flaviviruses are resistant to RNAi since replication occurs in reorganized ER membrane packets. In addition, HCV replication was found to be stimulated by the RNAi machinery.
Another nucleic acid–based antiviral approach is antisense DNA or RNA decoys, for example, phosphorodiamidate morpholino oligomers (PMOs). These compounds act by forming a stable, sequence-specific duplex with RNA, thereby blocking access to target RNA by biomolecules required for replication. PMOs targeting the translation initiation site of DENV RNA, the 3’ UTR, 5'SL, and 3'CS were effective in reducing the viral load in various cell lines. These compounds meet most of the requirements for an anti-DENV therapeutic; non-toxic, cheap, easy to administer, stable for months at variable temperatures, but remain to be tested in animal models.
Sulfated polysaccharides have been investigated for anti-DENV activity, although inconsistency in the activity results indicates that they need to be further tested both in vitro and in vivo (24).
The processing of N-linked oligosaccharides in the ER is important for viral glycoprotein maturation, and inhibition of glucosidase-mediated trimming affects the replication cycle of several enveloped viruses. DENV production was inhibited in mouse neuronal cells by two ER α-glucosidase inhibitors, castanospermine (CST) and deoxynojirimycin. CST was effective against all four serotypes in human hepatoma cells and prevented mortality in DENV-2–infected mice. This effect was restricted to DENV, not being observed against other flaviviruses such as WNV and YFV. A third ER α-glucosidase inhibitor, N-nonyl-deoxynojirimycin, inhibits DENV-2 infection in BHK-cells. These results with α-glucosidase inhibitors are encouraging and should be investigated further in vivo.
Nitric oxide (NO) is generated by macrophages, monocytes, dendritic cells (DCs), and neutrophils; the same cells that are supposed to be the main sites of replication for DENV. In vitro assays have revealed that NO specifically affects the viral RdRp activity, suggesting possible viral targets of NO during DENV infection (82).
Hence, there are multiple options for designing novel therapeutics for dengue disease. However, the main concern with most therapeutic approaches is that they are not validated for inhibitory effects on all four DENV serotypes. In addition, several studies have not been examined in an animal model, and several reported antivirals have been tested at only one time point, pre- or post-infection in tissue culture systems, and therefore need to be subjected to more diverse regimes, and different cell types.
Risk factors for severe dengue
DENV pathogenesis remains a challenging jigsaw puzzle with many pieces missing to understand the complex interplay of viral and host factors. Despite intensive research, it is not well understood. The severity of DENV infection is modulated by multiple risk factors such as age (83, 84), the genetic background of the host (85, 86), viral serotype (83, 87) and genotype (88, 89), and secondary DENV infection by a heterologous serotype (85, 90-93). Finally, the virus serotype and genotype also influence the symptomatic picture of disease and outcome (Fig. 2). These observations were initially based on epidemiological findings, but accumulating laboratory and experimental data have contributed to the recognition of DENV virulence as an important risk factor.