Common arbovirus-induced diseases

Viruses and autophagy, apoptosis and unfolded protein response (UPR)

Many viruses hijack host cell responses for their own benefit and use them as complementary mechanisms for replication and infection. Some of the most important host mechanisms that are usually affected by viral infection are pathways involved in cell death and cellular responses against environmental stress. These mechanisms include apoptosis (i.e. programmed cell death I), autophagy (programmed cell death II) and UPR. These mechanisms play essential functions in regulating cell fate and are important for normal cellular functions. In addition, these mechanisms are tightly regulated and can affect each other. They are usually interconnected and also ‘cross-talk’ with each other. We will briefly review the general concepts of apoptosis, autophagy and UPR and explain their cross-talk and regulatory mechanisms. We will then focus on the role of apoptosis, autophagy and UPR in arbovirus replication and infection and then describe different possible therapeutic approaches for arboviruses by discussing the involvement of apoptosis, autophagy and how they may determine therapeutic strategies.

An overview of autophagy, apoptosis and UPR

Autophagy

Lysosomes are the final destination for degradation of long-lived and dysfunctional cellular components through autophagy, a highly regulated catabolic process. This process is essential for maintaining cellular integrity, homeostasis, survival, differentiation and development (Refs Reference Baba92, Reference Levine and Kroemer93, Reference Mrschtik and Ryan94, Reference Fernandez and Lopez-Otin95). In mammals, the role of nutrient deprivation, hormonal stimuli, including glucagon and insulin, and other autophagy activation cues such as temperature, oxygen concentration and cell density have been elucidated (Refs Reference Klionsky and Emr96, Reference Mortimore, Reeta Pösö and Lardeux97, Reference Ghavami98, Reference Papadopoulos and Pfeifer99). There are three different types of autophagy, all of which differ in their mechanisms and functions: chaperone-mediated autophagy (CMA), microautophagy and macroautophagy (Refs Reference Mizushima100, Reference Xilouri and Stefanis101, Reference Napolitano102, Reference Kawamura103, Reference Cuervo and Wong104, Reference Jia and Sowers105). During CMA, specific cytosolic proteins are selectively tagged by the CMA substrate chaperone complex and then moved to the lysosome for degradation (Refs Reference Cuervo and Wong104, Reference Yang106, Reference Bursch107). This is the only form of autophagy in which no vesicular traffic is involved (Ref. Reference Majeski and Fred Dice108). Microautophagy directly targets small proteins and organelles using lysosomes (Refs Reference Mijaljica, Prescott and Devenish109, Reference Voigt and Poggeler110, Reference Lemasters111). However, macroautophagy is the major regulated catabolic mechanism by which the bulk of damaged cytoplasmic proteins and organelles are sequestered within an autophagosome (Refs Reference Yang and Klionsky112, Reference Grasso and Vaccaro113, Reference Klionsky and Schulman114). In this review, we will focus on macroautophagy (referred to herein as ‘autophagy’).

The first step in autophagy (see Fig. 3) involves formation and expansion of a double-membrane structure, which is called the ‘isolation membrane’ or ‘phagophore’. The edges of this membrane eventually fuse to form a new double membrane-bound vacuole, known as the autophagosome that sequesters the cytoplasmic cargo. The autophagolysosome is formed by fusion of the autophagosome with a lysosome and lysosomal contents are degraded by hydrolytic enzymes (Refs Reference Dice115, Reference Tooze116, Reference Pattingre117, Reference Hurley and Schulman118). As a result of degradation, nucleotides, amino acids and free fatty acids (FFAs) are generated and then reused for energy metabolism, macromolecular production and biosynthesis (Refs Reference Levine and Yuan119, Reference Shibutani and Yoshimori120). It is assumed that the different steps in macroautophagy are mediated by autophagy-related genes (ATG), which encode proteins involved in autophagy (Refs Reference Klionsky121, Reference Xie122). These proteins have been classified into five different functional categories: (i) a protein serine/threonine kinase complex that responds to upstream events such as target of rapamycin (TOR) kinase (Atg1/ULK1, Atg13 and Atg17); (ii) a lipid kinase group that controls vesicle nucleation (Atg6/Beclin1, Atg14, Vps34/PI3KC3 and Vps15); (iii) two ubiquitin-like conjugation pathways that stimulate vesicle expansion (the Atg8 and Atg12 conjugation systems); (iv) a recycling pathway that is required for disassembly of Atg proteins (Atg2, Atg9, Atg18); and (v) vacuolar permeases that permit the efflux of amino acids from the degradative compartment (Atg22) (Refs Reference Levine and Kroemer93, Reference Levine and Yuan119). The mammalian TOR (mTOR) kinase acts as a negative regulator of autophagy and is a central controller of cell growth, aging and proliferation (Refs Reference Mizushima123, Reference Hall124). Under starvation conditions, inhibited mTOR induces autophagy through phosphorylation of the Ulk1-Atg13-FIP200-Atg101 complex (Refs Reference Mercer, Kaliappan and Dennis125, Reference Chiarini126), leading to localization of Ulk1/2 and Atg13 to the autophagic isolation membrane (Refs Reference Jung127, Reference Czarny128). During the initiation step of autophagy, Beclin 1 interacts with Vps34, which contributes to Atg protein recruitment and autophagosome nucleation (Refs Reference Simonsen and Tooze129, Reference Morel, Dupont and Codogno130). Interaction with various Beclin-1-interacting proteins facilitates the coordination of these events (Ref. Reference Jordan and Randall131). LC3, the mammalian ortholog of Atg8, is cleaved by Atg4 and then conjugated to the polar head of phosphatidylethanolamine (PE) to generate LC3-II, which is necessary during the elongation step of autophagy (Refs Reference Kabeya132, Reference Slobodkin and Elazar133). Hence, the autophagosome is regulated in response to the Beclin-1/Vps34/UVRAG complex, known as the maturation step (Refs Reference Matsunaga134, Reference Zhong135, Reference Mizushima136). An overview of autophagy is summarised in Figure 3.

Figure 3.

Graphic representation of autophagy. Autophagy is a process for the degradation and recycling of damaged or unnecessary cellular compartments, which has several tightly regulated steps including induction, nucleation, expansion and completion, fusion and degradation. The mTOR is known as the key regulator of autophagy induction and can be suppressed by ULK1, leading to trigger VPS34-Beclin 1-class III PI3-kinase complex. Several different membrane pools contribute to the formation of the phagophore. The Atg proteins (Atg2, Atg9, Atg18) are essential for phagophore formation. The ATG and LC3 conjugation system also contribute in autophagosome membrane formation and elongation. The autophagolysosome then is formed by fusion of the autophagosome with a lysosome to degrade and reuse the compounds. ATG, autophagy-related genes; mTOR, mammalian target of rapamycin.

Apoptosis

There are two main functionally distinct pathways for apoptosis induction (Fig. 4): the extrinsic and the intrinsic mitochondrial pathways (Refs Reference Call, Eckhardt and Camidge137, Reference Yeganeh138, Reference Rashedi139). Caspases are involved in most of the apoptotic processes and are activated by ligation of death receptors [tumour necrosis factor receptor (TNFR), Fas, TNF-related apoptosis-inducing ligand (TRAIL)] or release of specific proteins from the mitochondria (Refs Reference Fulda and Debatin140, Reference Ghavami141). However, accumulating evidence suggests that the two pathways are intimately intertwined (Refs Reference Yeganeh138, Reference Ghavami142), which will be described in the next sections. The extrinsic apoptosis cascade is stimulated after the binding of cell surface receptors to their ligands, resulting in Fas-associated protein with death domain (FADD)-dependent activation of initiator caspases, namely caspase-8, and subsequently caspase-3 and -7 (Refs Reference Hassan143, Reference Ghavami144). As a consequence, effector caspases (i.e. caspase-3 and caspase-7) are dimerized and activated and, once active they can cause apoptosis (Refs Reference Ghavami141, Reference Shalini145).

Figure 4.

Graphic representation of apoptosis signalling pathways. Apoptosis is initiated via two different routes including extrinsic and intrinsic apoptotic pathways. The extrinsic signals are initiated by cell death ligands (e.g. FasL, APO-2L, TRAIL, TNF) and activate FADD and subsequently cleave pro-caspase-8. Cleavage of pro-caspases-8 and -10 initiate activation of caspases-8 and -10, which later can directly trigger effector caspases including caspases-3, -6 and -7. The intrinsic pathway is stimulated via DNA damage. Once DNA damage occurs, p53 is activated and induces apoptosis in a mitochondria-dependent manner. In this pathway, pro-apoptotic and antiapoptotic proteins are up- and down-regulated, leading to release of cytochrome c. Released cytochrome c later can activate caspase 9 which in turn activates caspase-3. FasL, Fas (Apo-1/CD95) ligand; TNF, tumour necrosis factor receptor TRAIL, TNF, tumour necrosis factor receptor.

The mitochondrial apoptotic death mechanism integrates various extracellular stimuli including drugs, nutrients and radiation and also different intracellular stimuli such as oxidative stress, oncogene expression, endoplasmic reticulum (ER) stress and DNA damage (Refs Reference Fuchs and Steller146, Reference Kubli and Gustafsson147). The apoptotic signals in this pathway converge on the mitochondria to release apoptogenic proteins such as cytochrome c, apoptosis-inducing factor (AIF), Smac/DIABLO, Omi/HtrA2 and mitochondrial endonuclease G (Refs Reference Er148, Reference Crow149, Reference Ghavami150, Reference Ghavami151). The Bcl-2 family of proteins serve as important regulators of the release of these mitochondrial proteins that can be divided into two classes: (i) antiapoptotic members (e.g. Bcl-2 and Bcl-xL); and (ii) proapoptotic members (e.g. Bax, Bak, Bid, Bad, Noxa, Puma and others) (Refs Reference Petros, Olejniczak and Fesik152, Reference Kim153). Up-regulation of proapoptotic proteins or down-regulation of antiapoptotic proteins can cause an increase in permeability of the mitochondrial membrane, which later promotes release of cytochrome c and other proteins into the cytosol (Refs Reference Ghavami151, Reference Ghavami154, Reference Saelens155, Reference Wang156). In the presence of deoxyadenosine triphosphate (dATP), the released cytochrome c interacts with Apaf-1 and caspase-9 and forms a ternary complex, leading to activation of caspase-3 and then apoptosis (Refs Reference Ghavami142, Reference Ghavami157, Reference Wu and Bratton158). In addition, p53 plays a stimulating role in intrinsic apoptosis induction (Refs Reference Moll and Zaika159, Reference Ghavami160, Reference Reed and Quelle161). Thus, the two direct p53 transcriptional targets Noxa and Puma can mediate the pro-apoptotic activity of Bax and Bak, and thereby promote apoptosis (Refs Reference Michalak162, Reference Zhang, Li and Xu163).

It is widely accepted that there is cross-talk between the two extrinsic and intrinsic pathways, such that activity in one pathway interferes with signalling steps in the other pathway (Ref. Reference Ghavami141).

The pro-apoptotic cytochrome c-releasing factor Bid is positioned to serve as a link between the extrinsic death receptor pathway and the intrinsic pathway (Ref. Reference Ghavami154). Cleavage of the BID protein in the cytoplasm by caspase-8 causes Bid to localise in the cytosol while truncated Bid translocates to the mitochondria and activates the mitochondrial pathway after apoptosis induction through death receptors, and can be used to amplify the apoptotic signal (Ref. Reference Peng164). Although Bid is a downstream target of caspase-8 in the extrinsic apoptotic pathway, it also acts as ligand for Bax and Bak, causing caspase-9 activation (Refs Reference Ghavami154, Reference Tilly165). Caspase-9 activation proteolytically activates downstream caspases (e.g. caspases-3,-6,-7), which, in turn, can result in apoptosis (Refs Reference Li166, Reference Pour-Jafari, Ghavami and Maddika167).

UPR

The ER contains an extensive network of tubules, sacs and cisternae, which extend from the cell plasma membrane through the cytoplasm and to the nuclear envelop through a continuous connected network (Refs Reference Saveljeva168, Reference Sovolyova169). The ER is the main sub-cellular compartment involved in proper folding of proteins and their maturation. Approximately one-third of the total proteins are synthesised in the ER. Many different perturbations can alter the function of the ER leading to unfolding or misfolding of proteins in the ER. This condition is referred to as ER stress (Refs Reference Sovolyova169, Reference Hetz, Chevet and Harding170). The ER creates a series of adaptive mechanisms to prevent cell death complications and these together are referred to as the UPR (Refs Reference Hetz, Chevet and Harding170, Reference Logue171). The UPR can be involved in the secretory pathway leading to restoration of protein folding homeostasis. However, if there is too much stress on the ER, and the ER cannot cope with this stress, it will eventually lead to cell death (Ref. Reference Guan172). The UPR also plays an important role in maintaining cellular homeostasis of specialised secretory cells such as pancreatic beta cells, salivary glands and plasma B cells (Ref. Reference Hetz, Chevet and Harding170). It is becoming increasingly evident from animal models that UPR has several functions that are not directly linked to protein folding including inflammation, energy control and lipid and cholesterol metabolism (Ref. Reference Hetz, Chevet and Harding170). The existence of UPR was first reported by Kozatsumi et al. more than 25 years ago (Ref. Reference Kozutsumi173). They showed that glucose regulated proteins (GRPs) that are associated with the ER are up-regulated upon sensing the presence of unfolded or misfolded proteins in the ER (Ref. Reference Kozutsumi173). While the mechanisms and signalling events behind it were not known at the time, today we have a much better understanding of the UPR and how these events are regulated in the ER at the molecular level. ER stress response signals are constantly monitored by three main classes of sensors. These include inositol requiring enzyme 1 alpha (IRE-1α) and IRE-1β, protein kinase RNA like ER kinase (PERK) and activating transcription factor 6 (ATF6; both α and β isoforms) (Fig. 5). In normal healthy cells these sensors are in an inactive state.

Figure 5.

Graphic representation of ER stress and virus replication. ER stress is enhanced in the viral infected cells and activates UPR proteins (e.g. PERK, ATF6, and IRE1). Activated PERK leads to induce ATF4 via phosphorylation of eIF2α, causing attenuation of translation and inducing genes encoding CHOP. Upon IRE1 activation, TRAF2 and sXBPmRNA1 splicing are initiated in the cytoplasm, subsequently leading to activation of UPR target genes. The degradation of ATF6 is increased through recruitment of ATF6, a UPR sensor. ATF6 translocates to the Golgi and is cleaved to a nucleus targeting form that promotes expression of UPR-responsive genes. The consequences of UPR activation are necessary for viral replication and pathogenesis. ATF, activating transcription factor; CHOP, C/EBP homologous protein; ER, endoplasmic reticulum; IRE1, inositol-requiring enzyme; PERK, protein kinase RNA like ER kinase; UPR, unfolded protein response.

IRE1

This is a type I transmembrane protein receptor having an N-terminal ER luminal-sensing domain. The cytoplasmic C-terminal region contains both an endoribonuclease domain and a Ser/Thr kinase domain (Ref. Reference Sovolyova169). There are two homologues of IRE1: IRE1α and IRE1β. Activation of IRE1 involves dissociation from Grp78, followed by dimerization, oligomerization and trans-autophosphorylation, which leads to conformational changes and activation of its RNase domain (Ref. Reference Hetz, Chevet and Harding170). Activated IRE1 excises a 26-nucleotide intron region from mRNA that encodes the transcription factor X-box binding protein 1 (XBP1). Dissociation of this 26-nucleotide intron region from XBP1 leads to a shift in the coding reading frame and produces a more stable form of XBP1 called XBP1 spliced form (XBP1s) (Ref. Reference Hetz, Chevet and Harding170). IRE1-XBP1s signalling axis modulates pro-survival responses by targeting many genes involved in protein folding, maturation and ER-associated degradation (Ref. Reference Sovolyova169). XBP1 also modulates phospholipid synthesis which is required for ER expansion under ER stress (Ref. Reference Hetz174). Some examples of XBP1 target genes include ERdj4, P58^IPK, human ER-associated DNAJ (HEDJ), DnaJ/Hsp-40-like genes and protein disulphide isomerase (PDI) P5 (PDI-P5) and ribosome associated membrane protein 4 (RAMP4) (Ref. Reference Sovolyova169). Different studies have shown that activation of IRE1 signalling is robust at first but as time progresses it diminishes (Refs Reference Sovolyova169, Reference Lin175). However, artificial maintenance of IRE1 signalling is achieved by a chemically-activated mutant form of IRE1, which is positively correlated with enhanced cell survival conditions under ER stress, suggesting that IRE1 signalling mainly plays a role in a pro-survival pathway (Refs Reference Sovolyova169, Reference Lin175, Reference Coelho, Gaspar and Domingos176).

ATF6

ATF6 is a type II transmembrane protein that contains a basic leucine zipper (bZIP) transcription factor domain in its cytosolic terminus (Refs Reference Sovolyova169, Reference Yu177). The ATF6 family of ER transducers include ATF6 α, ATF6 β, old astrocyte specifically induced substance (OASIS), LUMAN (also called CREB3), BBF2 human homolog on chromosome 7 (BBF2H7), cyclic-AMP responsive element binding protein hepatocyte (CREBH) and CREBP4 (Ref. Reference Hetz174). Unlike IRE1, ATF6 does not undergo oligomerization, dimerization and autophosphorylation. Under ER stress conditions, Grp78 dissociate from ATF6 thus uncovering the Golgi localisation signal of ATF6. Activated ATF6 translocates into the Golgi complex where it undergoes cleavage by site-1 and site-2 proteases (Ref. Reference Yu177). Thus, the ATF6 N-terminal cleavage product translocates to the nucleus and regulates the expression of genes that are associated with the ER-associated protein degradation pathway. Some of the ATF6 target genes include Grp78, PDI and ER-degradation enhancing-a-mannosidase-like protein 1 (EDEM1). All these proteins work closely to reduce unfolded proteins in the ER lumen (Ref. Reference Sovolyova169). ATF6 also activates pro-survival transcription factor and IRE1 target gene XBP1 (Refs Reference Yoshida178, Reference Guo179). Similar to that of IRE1 signalling, ATF6 is activated by the UPR but is not sustained throughout the UPR response. ATF6 signalling is primarily for pro-survival but in some cases, ATF6 signalling activates the pro-apoptotic transcription factor C/EBP homologous protein (CHOP) during prolonged ER stress (Ref. Reference Yoshida178).

PERK

This is a type I ER transmembrane protein having an ER luminar sensor domain and a cytoplasmic domain. The cytoplasmic domain contains Ser/Thr kinase activity. Upon activation by UPR, PERK dissociates itself from Grp78 and undergoes dimerization and trans-auto phosphorylation (Refs Reference Sovolyova169, Reference Guan172, Reference Doyle180). Activated PERK phosphorylates eukaryotic translation initiation factor 2α (eIF2α). PERK-mediated phosphorylation of Ser51 in eIF2α reduces the activity of eIF2α complex and leads to the inhibition of protein synthesis. This rapidly reduces the number of proteins entering the ER and this can lead to a pro-survival effect on the cell (Refs Reference Hetz, Chevet and Harding170, Reference Guan172, Reference Donnelly181). Phosphorylation of eIF2α also allows translation of mRNAs containing short open reading frames in their 5′ UTR regions. Such translated proteins include activating transcription factor 4 (ATF4) (Ref. Reference Hetz, Chevet and Harding170). ATF4 controls expression of many proteins involved in redox processes and amino acid metabolism, and it modulates the expression of ER chaperones and foldases (Ref. Reference Hetz, Chevet and Harding170). ATF4 also regulates important genes involved in ER apoptosis such as CHOP and growth arrest and DNA damage inducible 34 (GADD34) (Ref. Reference Hetz, Chevet and Harding170). GADD34 is involved in a feedback loop to dephosphorylate eIF2α by protein phosphatase IC (PPIC) to restore protein synthesis (Refs Reference Hetz, Chevet and Harding170, Reference Novoa182). Another substrate for activated PERK is nuclear factor (erythroid-derived 2 factor)-related factor (Nrf2). In normal cells, Nrf2 is present in the cytoplasm in association with cytoskeletal anchor kelch-like Ech-associated protein (KEAP1). Upon activation PERK phosphorylates Nrf2 and this helps Nrf2 to dissociate from KEAP1 and translocate into the nucleus (Refs Reference Sovolyova169, Reference Cullinan183). Upon translocation into the nucleus Nrf2 induces the expression of genes that have an anti-oxidant response element (ARE) within their promoter such as heme oxygenase 1 (HO-1), aiding in protein folding and helping to restore ER homeostasis (Refs Reference Sovolyova169, Reference Cullinan183). The role of Nrf2 as a pro-survival factor is further shown by the fact that cells devoid of Nrf2 display increased sensitivity to cell death via apoptosis after ER stress (Refs Reference Sovolyova169, Reference Cullinan183). The overall UPR signalling pathway is shown in Figure 5.

The role of autophagy in arbovirus replication

Although autophagy was initially proposed as a physiological cellular response to environmental stress followed by virus amplification, increasing evidence now indicates that several viruses may use autophagy as a survival strategy to support their life cycle, which is known as ‘pro-viral autophagy’ (Refs Reference Jordan and Randall131, Reference Yeganeh138, Reference Shelly184, Reference Orvedahl185) (Fig. 6). Virus-induced induction of autophagy seems to be associated with replication/translation of many arboviruses like DENV, JEV, CHIKV, rotavirus, and epizootic haemorrhagic disease virus (EHDV, an orbivirus) (Refs Reference Lee186, Reference Li187, Reference Krejbich-Trotot188, Reference Meng189, Reference Shai190, Reference Arnoldi191). The results that were obtained by monitoring LC3 lipidation in JEV-infected NT-2 cells, a pluripotent human testicular embryonal carcinoma cell line treated with Rapamycin and 3-methyladenine, revealed that there was a direct relationship between autophagy and viral replication The results were confirmed using an Atg5/Beclin-1 knock down model (Ref. Reference Li187). Most commonly, in many eukaryotic cells, it is apparent that the initiation of autophagy can be enhanced in infected DENV cells; in addition, the replication of DENV is positively linked to autophagy induction (Ref. Reference Heaton and Randall192). However, DENV viral replication has been shown to be limited in monocytes, which suggests a possible cell-specific relationship between activated autophagy and DENV production (Ref. Reference Panyasrivanit193). WNV induces autophagy even though its replication is autophagy independent (Ref. Reference Vandergaast and Fredericksen194). The importance of virus-induced autophagy and up-regulation of viral replication has also been shown in CHIKV-infected cells (Ref. Reference Krejbich-Trotot188). The Orbivirus EHDV induces autophagy, apoptosis and c-Jun N-terminal kinase (JNK) activation, and phosphorylates c-Jun, all of which seem to benefit viral replication (Ref. Reference Shai190). JEV also induces autophagy in the early stage of infection and the inoculated viral particles traffic to autophagosomes for subsequent steps of viral infection (Ref. Reference Li187). In vivo studies showed that autophagy played a supporting role in DENV-2 replication and pathogenesis (Ref. Reference Lee195).

Figure 6.

Graphic proviral functions of autophagy. There are five possible mechanisms for modulating viral replication by autophagy. Amphisome formation is thought to be beneficial for viral cellular entry and replication. Induction of autophagosome formation is also important for some virus’ replication. Furthermore, viruses initiate autophagy to benefit from lipid droplets as an energy source during viral replication. Free fatty acids are liberated from lipid droplets during autophagy to produce ATP. Viruses also stimulate autophagy to subvert immune responses by selectively degrading key regulatory molecules. Another mechanism is that viruses promote their replication by prolonging cell survival and suppressing cell death. The mechanistic details related to proviral functions of autophagy are discussed in the text.

Although the function or functions of autophagy in promoting virus replication are not completely understood, experimental evidence suggests that there are multiple autophagy pro-viral mechanisms, including serving as a scaffold for viral replication, contributing to viral entry, regulation of lipid metabolism, suppressing innate immune responses and preventing cell death (Ref. Reference Shi and Luo196). A group of arboviruses including DENV, and JEV may need to invoke autophagy components such as the autophagosome, amphisome and autolysosome to: (i) serve as a scaffold for viral replication; and (ii) escape from the immune system (Refs Reference Li187, Reference Chi197, Reference den Boon and Ahlquist198, Reference Panyasrivanit199). The amphisomes play major roles in DENV entry and localisation of viral translation/replication constituents (Ref. Reference Panyasrivanit199). DENV-2 needs pre-lysosomal fusion vacuoles (amphisomes) while DENV-3 interacts with both amphisomes and autophagolysosomes as the sites for their viral translation/replication complexes (composed of viral RNA and proteins) (Ref. Reference Panyasrivanit199). Poliovirus and CHIKV also stimulate autophagosome formation as a site for aggregation of viral translation/replication complexes (Refs Reference Krejbich-Trotot188, Reference Meng189, Reference Jackson200). After DENV and JEV induce autophagy, the presence of viral replication/translation complexes in both the autophagosome and the endosome suggests an auxiliary role for autophagosome–endosome fusion in viral entry (Refs Reference Li187, Reference Panyasrivanit201). Autophagy can regulate lipid metabolism (lipophagy) through modulating the degradation of triglycerides that have accumulated in cytosolic lipid droplets (Ref. Reference Singh and Cuervo202). Lipid droplet usage as an energy source is another autophagy-mediated pro-viral mechanism that is used for DENV replication (Ref. Reference Heaton and Randall203). Thus, lipid droplets are sequestered in autophagosomes and delivered to lysosomes for degradation to generate FFAs from triglycerides (Ref. Reference Heaton and Randall203). The released FFAs are imported to mitochondria and they undergo β-oxidation to produce ATP for viral replication (Ref. Reference Heaton and Randall203).

The innate antiviral immune response

The innate antiviral response is initiated by binding of pattern recognition receptors (PRRs), retinoic acid-inducible gene (RIG) and Melanoma differentiation-associated protein 5 (MDA5) to intracellular viral pathogen associated molecular patterns (PAMPs) (Ref. Reference Jensen and Thomsen204). The interaction of PRR-PAMP with mitochondrial antiviral-signaling protein (MAVS) through Caspase activation and recruitment domain (CARD)–CARD homotypic reaction leads to signalling cascades that ultimately activate nuclear factor-κB (NF-κB) and interferon regulatory factors (IRF-3) (Refs Reference Loo205, Reference Hou206). Inhibiting interferon (IFN) production followed directly from interaction of Atg5-Atg12 with the CARD of RIG, and MDA5 can promote vesicular stomatitis virus (VSV) replication (Ref. Reference Jounai207). Although the exact mechanism of autophagosome accumulation in JEV replication is still unclear, several studies have demonstrated the importance of fusion between autophagosomes and lysosomes and also autophagy in reducing MAVS-IRF3 activation to facilitate virus replication (Ref. Reference Jin208). Additionally, it has been suggested that autophagy promotes cell survival by delivering damaged mitochondria to lysosomes during JEV infection (Ref. Reference Jin208).

Optimal Flavivirus (e.g. DENV2) replication/translation is associated with the nonstructural viral protein NS4A in up-regulating PI3-K-dependent autophagy, and preventing cell death (Ref. Reference McLean209). Recently, NS4A has been characterised as a main component of the membrane-bound DENV2 replication complexes (Ref. Reference Miller210). With attention to the cross-talk between autophagy and apoptosis, it is becoming apparent that autophagy postpones apoptosis and promotes CHIKV propagation by inducing the IRE1α–XBP-1 pathway in conjunction with ROS-mediated mTOR inhibition (Ref. Reference Joubert211). A schematic representation of autophagy and arbovirus replication is summarised in Figure 6.

The role of apoptosis in arbovirus replication

To date, several investigations have been carried out on the importance of apoptosis in different virus infections, pathogenesis and replication, but many issues are still unclear and under debate (Refs Reference Collins212, Reference Clarke213, Reference Turpin214). As summarised in Figure 7, a number of arboviruses such as Sindbis virus, WNV and JEV seem to use apoptosis as a virulence factor to promote their own pathogenesis (Reference Lewis215, Reference Samuel, Morrey and Diamond216, Reference Sun, Yu and Deubel217). Each of these viruses has specific targets and biochemical-induced mechanisms during virus-induced programmed cell death. The observations suggest that Sindbis virus-induced apoptosis plays an important role in Sindbis virus pathogenesis and mortality (Ref. Reference Lewis215). After entry of Sindbis virus into the host cell and subsequent formation of Sindbis virus double-stranded RNA intermediates, dsRNA-dependent protein kinase (PKR) recognises these particles (Refs Reference Stollar, Shenk and Stollar218, Reference Stollar and Stollar219, Reference Domingo-Gil220). PKR blocks cellular translation through eIF2a phosphorylation, which later can inhibit Mcl-1 (anti-apoptotic Bcl2 family protein) biosynthesis (Ref. Reference Akgul221). PKR also controls c-Jun N-terminal kinases (JNK) through IRS1 phosphorylation and later activates 14-3-3 (Ref. Reference Venticinque and Meruelo222). Thus, 14-3-3 affects the accessibility of substrates (e.g., Bad) to kinases and serves to localise kinases to their substrates, thereby leading to release of Bad and disruption of the complex between anti-apoptotic Bcl2 family proteins, Bcl-xl and Bak. Both Bad and Bik can displace Bak from Mcl-1, which results in Bak oligomerization and cytochrome c release, and subsequent induction of apoptosis (Ref. Reference Venticinque and Meruelo222). CHIKV triggers the apoptosis machinery and uses apoptotic blebs to evade immune responses and facilitate its dissemination by infecting neighboring cells (Ref. Reference Krejbich-Trotot223). CHIKV infection can induce apoptotic cell death via at least two apoptotic pathways: the intrinsic pathway, which has been reported to be involved in virus replication and results in activation of caspase-9, and the extracellular pathway, which is dependent on the induction of cell surface or soluble death effector ligands that activate caspase-8. Thus, both pathways activate caspase-3 and finally induce cell death, and this facilitates virus release and spread (Ref. Reference Joubert211). The replication of Crimean-Congo haemorrhagic fever virus (CCHFV), an arbovirus from the family Bunyaviridae, is associated with the death receptor pathway of apoptosis. Up-regulation of pro-apoptotic proteins (i.e. BAX and HRK) and novel components of the ER stress-induced apoptotic pathways (i.e. PUMA and Noxa) have also been shown in a CCHFV-infected hepatocyte cell line, which suggests a link between CCHFV replication, ER stress and apoptotic pathways. Notably, differential high levels of transcription factors, such as CHOP, which are activated through ER stress, are present in hepatocytes following CCHFV replication (Ref. Reference Rodrigues224). In this study, it was shown that the over-expression of IL-8, an apoptosis inhibitor, during CCHFV infection was independent from apoptotic pathways. However, in other studies, a positive correlation was detected between IL-8 induction and DENV infection (Refs Reference Rodrigues224, Reference Medin and Rothman225, Reference Li226). In contrast to Sindbis virus, CHIKV and CCHFV replication in infected cells have been proposed to be necessary for apoptosis induction, as demonstrated by the use of UV-inactivated viral particles (Refs Reference Lee, Liao and Lin227, Reference van Marle228, Reference Jan and Griffin229). The replication of Flaviviruses (e.g. WNV, JEV and DENV) can be limited by virus-induced programmed cell death at the early stage of virus infection. These viruses might block or delay apoptosis via activating several cell survival pathways, such as PI3K/Akt signalling, to improve their replication rate (Refs Reference Lee, Liao and Lin227, Reference Urbanowski and Hobman230). Blocking PI3K (using LY294002 and wortmannin) showed that the induction of apoptosis might be a result of p38 MAPK activation and did not affect JEV and DENV viral particle production (Ref. Reference Lee, Liao and Lin227). In 2001, del Carmen Parquet et al. demonstrated that WNV-induced cytopathic effect was caused during induction of apoptosis and that viral replication is an essential event for virus-induced cell death (Ref. Reference del Carmen Parquet231). WNV capsid protein has an anti-apoptotic role, ensuring that it can block or delay apoptosis by suppression of the phosphatidylinositol (PI) 3-kinase-dependent process at the early stage of infection (Ref. Reference Urbanowski and Hobman230). In addition, Akt is a downstream target of PI3-kinase and can directly phosphorylate the pro-apoptotic protein Bad at position Ser 136 (Ref. Reference Fujikura232). WNV can initiate apoptosis through caspases-3 and -12 and p53 after several rounds of replication and it is noteworthy that initial viral dose exerts an influence on kinetics of WNV-induced cell death (Refs Reference van Marle228, Reference Medigeshi233, Reference Chu and Ng234, Reference Yang235). After some RNA virus infections, expression of multiple miRNAs in host cells might have either a positive or negative effect on virus replication. One such cellular miRNA, Hs_154, limits WNV replication by inducing apoptosis through inhibition of two anti-apoptotic proteins like CCCTC binding factor (CTCF) and EGFR-co-amplified and overexpressed protein (ECOP) (Refs Reference Lee, Liao and Lin227, Reference Smith236). JEV, an RNA virus, may induce ROS-mediated ASK1-ERK/p38 MAPK activation and thus lead to initiation of apoptosis (Ref. Reference Yang237). In mouse neuroblastoma cells (line N18) infected with ultraviolet-inactivated JEV (UV-JEV), replication-incompetent JEV virions induced cell death through a ROS-dependent and NF-kB-mediated pathway (Ref. Reference Lin, Liao and Lin238). Initial suppression of UV-JEV-induced cell death, followed by co-infection with active or inactive JEV, showed that JEV may trigger cell survival signalling to modify the cell environment for timely virus production (Ref. Reference Lin, Liao and Lin238). NS1′ protein, a neuroinvasiveness factor that is only produced by the JEV serogroup of Flaviviruses during their replication, was introduced as a caspase substrate in virus-induced apoptosis; however, use of a caspase inhibitor had no effect on virus replication (Ref. Reference Melian239). Empirical evidence showed that JEV can affect Bcl-2 expression to increase anti-apoptotic response rather than anti-viral effect to enhance virus persistence and reach equilibrium between replication and cell death (Ref. Reference Liao240).

Figure 7.

Graphic representation of apoptosis and viral replication. Viral infection, in general, can induce both intrinsic and extrinsic apoptotic pathways. Viruses like CHIKV, CCHFV and RVFV initiate extrinsic signals through cell death ligands (e.g. FasL, APO-2L, TRAIL, TNF), causing caspases-8 activation which then triggers caspases-3, -6 and -7). AHSV and WNV directly trigger caspase 3; however, CHIKV targets caspase 9. DENV and WNV affect the intrinsic pathway of apoptosis through stimulation of P53. Once P53 is activated, mitochondria-dependent apoptosis can be activated. Viral infection can also induce PKR and this kinase can affect eIF2a, resulting in activation of effector caspases and initiation of apoptosis. Viruses can also have anti-apoptotic activity. DENV, WNV and JEV trigger survival signalling through PI3K-AKT signalling pathway. PKR can be initiated by Sindbis virus which leads to inhibition of cellular translation through eIF2a phosphorylation, suppressing Mcl-1 biosynthesis. Sindbis virus can regulate 14-3-3 through activation of JNK followed by induction of PKR (for other details see text). AHSV, African horse sickness virus; CHIKV, Chikungunya virus; CCHF, Crimean–Congo haemorrhagic fever virus; DENV, Dengue virus; FasL, Fas (Apo-1/CD95) ligand; JEV, Japanese encephalitis virus; JNK, c-Jun N-terminal kinases; TNF, tumour necrosis factor receptor; TRAIL, TNF-related apoptosis-inducing ligand; PKR, (dsRNA)-activated protein kinase; RVFV, Rift valley fever virus; WNV, West Nile virus.

Numerous in vitro studies have confirmed that DENV can induce apoptosis in a wide variety of mammalian cells including endothelial cells, hepatocytes, mast cells, monocytes, dendritic cells and neuroblastoma cells, but the mechanisms are not completely understood. Dendritic cells are believed to be the primary DENV targets that play central roles in supporting active replication during virus pathogenesis. However, a recent study reported that DENV replication in monocyte-derived dendritic cells (mdDCs) was positively correlated with pro-inflammatory cytokine secretion such as TNFα and apoptosis (Ref. Reference Silveira241). To achieve high replication in macrophages, hepatoma and dendritic cells, DENV may subvert apoptosis by inhibiting NF-kB in response to TNFα stimulation (Refs Reference Palmer242, Reference Li243). Interaction between DENV capsid protein and the hepatoma cell line (Huh7) calcium modulating cyclophilin-binding ligand (CAML) also positively affected viral replication by inhibiting apoptosis (Ref. Reference Li243). Activation of p53-dependent apoptosis by DENV may also contribute to inhibition of inflammation and reduce immune responses to efficiently disseminate viral progeny (Ref. Reference Nasirudeen and Liu244). Microarray analysis following DENV infection in p53-positive and -deficient cell lines revealed that activation of the pro-apoptotic gene caspase-1 played a basic role in p53-mediated apoptotic pathway and was necessary for up-regulation of numerous immune response genes (Ref. Reference Nasirudeen and Liu244). As mentioned, apoptosis serves as a critical and final step in viral infectious cycles that may favour virus propagation. The pro-apoptotic NSs and anti-apoptotic NSm proteins of the Phelebovirus genus of the family Bunyaviridae (e.g. RVFV) delayed apoptosis to efficiently replicate by regulating p53 (Refs Reference Yang235, Reference Austin245). The RVFV protein inhibits either caspase-8 activity or the death receptor-mediated apoptotic pathway to regulate pro-apoptotic p53 signalling (Ref. Reference Won246). NSs can facilitate viral translation through inhibition of PKR/eIF2a pathway and IFN production at early stages of infection (Ref. Reference Ikegami247). Members of the Orthobunyavirus genus, family Bunyaviridae, delay apoptosis through anti-apoptotic effects of NSs nonstructural protein on IRF-3 activity (Ref. Reference Kohl248).

Apoptosis has also been extensively linked to reovirus replication. BTV induces apoptosis in three mammalian cell lines but not in insect cell lines that were tested. BTV-mediated apoptosis involved activation of NF-kB and required virus uncoating and exposure to both outer capsid proteins VP2 and VP5 (Ref. Reference Mortola, Noad and Roy249). Apoptosis was mediated by both intrinsic and caspase-dependent extrinsic pathways (Ref. Reference Nagaleekar250). African horse sickness virus (AHSV), another orbivirus, also induced apoptosis in mammalian BHK-21 cells but not in insect KC cells, through activation of caspase-3 (Ref. Reference Stassen, Huismans and Theron251).

When apoptotic programmed cell death acts as a barrier against viral replication, previous research has revealed that some arboviruses can delay or block apoptosis to elevate their replication and dissemination. Moreover, viral replication of some arboviruses occurs following the presence of viral-induced apoptosis. However, the exact mechanisms whereby viruses modulate apoptosis in different mammalian cells need to be more extensively studied.

Arboviruses and UPR

The scientific literature related to the role of UPR in arbovirus pathogenesis is limited. Here, we review some of the arboviruses and the UPR pathways they elicit to aid replication. WNV is a neurotropic arbovirus that emerged as a pathogen of serious concern in the North American population. People infected with WNV are affected by severe neurological diseases such as meningitis, encephalitis and poliomyelitis (Ref. Reference Medigeshi233). WNV activates multiple UPR pathways leading to transcriptional and translational activation of several UPR target genes (Ref. Reference Medigeshi233). Of the three UPR pathways, the XBP1 pathway was shown to be non-essential for WNV replication and it was replaced by other pathways. ATF6 was degraded by the proteasome and PERK transiently phosphorylated eIF2α and induced the pro-apoptotic protein CHOP (Ref. Reference Medigeshi233). WNV-infected cells showed signs of apoptotic cell death including induction of growth arrest, activation of caspase-3 and activation of poly (ADPribose) polymerase (PARP). WNV titer levels were also significantly increased when grown in a CHOP^−/− deficient mouse embryo fibroblast (MEF) cell line but not in wild type MEF cells (Ref. Reference Medigeshi233). This evidence showed that WNV activates the UPR, and a host mechanism to counteract WNV infection involved activation of CHOP-dependent cell death (Ref. Reference Medigeshi233). In another study, the WNV Kunjin strain activated UPR signalling upon infection in mammalian cells (Ref. Reference Ambrose and Mackenzie252). UPR ATF6/IRE1 pathways were activated by this strain. However, there was no significant phosphorylation of eIF2α indicating that the UPR PERK pathway was not activated (Ref. Reference Ambrose and Mackenzie252). The Kunjin strain nonstructural proteins, NS4A and NA4B, were potent inducers of UPR. Moreover, sequential removal of NS4A hydrophobic domains decreased UPR activation but increased interferon gamma-mediated signalling (Ref. Reference Ambrose and Mackenzie252). These results show that WNV Kunjin strain activates UPR signalling and hydrophobic residues of WNV nonstructural proteins regulate the UPR signalling cascade. The role of ATF6 signalling in WNV replication is poorly understood. Results from the same group showed that ATF6 signalling is required for WNV replication by promoting cell survival and inhibition of the innate immune response (Ref. Reference Ambrose and Mackenzie253). ATF6-deficient cells showed a decrease in protein and virion production when infected with WNV Kunjin strain. These cells also demonstrated increased eIF2a phosphorylation and CHOP transcription, but these events were absent in infected control cells (Ref. Reference Ambrose and Mackenzie253). In contrast, IRE I-deficient cells do not show any discernible differences when compared with IRE I-positive cells upon infection (Ref. Reference Ambrose and Mackenzie253). These results also demonstrate that, in the absence of ATF6, other UPR signalling cascades such as PERK and IRE1 pathways cannot activate or enhance virus production, indicating that ATF6 is required for viral replication. However, it has also been shown that both ATF6 and IRE I are required for signal transducer and activator of transcription (STAT) I phosphorylation, showing that ATF6 is required for inhibition of innate immune response (Ref. Reference Ambrose and Mackenzie253). The arboviruses CHIKV and Sindbis also cause frequent epidemics of febrile illness and long-term arthralgic sequelae that affect the lives of millions of people each year (Ref. Reference Rathore, Ng and Vasudevan254). These viruses replicate in infected patients and also in mammalian cells indicating that they have certain control over the UPR of the host system. Analysis of these viral infections in mammalian cells shows that CHIKV specifically activates the ATF-6 and IRE-1 branches of the UPR pathway and suppresses the PERK pathway (Ref. Reference Rathore, Ng and Vasudevan254). CHIKV nonstructural protein 4 (nsp4) expression in mammalian cells suppresses eIF2α phosphorylation that regulates the PERK pathway (Ref. Reference Rathore, Ng and Vasudevan254). These results provide insight on the replication of CHIKV in mammalian cells by regulating the host UPR mechanism. However, experimental findings with Sindbis virus show that it induced uncontrolled UPR, which is reflected by failure to induce synthesis of ER chaperones, followed by increased phosphorylation of eIF2α and activation of CHOP leading to premature cell death (Ref. Reference Rathore, Ng and Vasudevan254). In another study, it was reported that the UPR XBP1 pathway was activated when neuoroblastoma N18 cells were infected with the arboviruses JEV and DENV (Ref. Reference Yu255). This was evidenced by splicing of XBP1 mRNA and activation of downstream genes ERDJ4, EDEM1 and p58. Reduction of XBP1 by small interfering RNA had no effect on cellular susceptibility to the two viruses but enhanced cellular apoptosis (Ref. Reference Yu255). Overall, these results suggest that both encephalitis and DENV trigger the XBPI signalling pathway and take advantage of this cellular response to alleviate virus induced cytotoxicity (Ref. Reference Yu255). According to another group, DENV infection of A547 ovarian cancer cells elicited the UPR signalling response (Ref. Reference Umareddy256). This was demonstrated by phosphorylation of eIF2α. It was also shown that different serotypes of DENV, such as ATF6 and IRE1, activate other UPR pathways. These results show that different DENV serotypes have the capacity to modulate different UPR pathways. They also demonstrated that de-phosphorylation of eIF2α by a drug called solubrinal reduced virus infection. This unique report showed that the same virus could activate all three UPR pathways (Refs Reference Umareddy256, Reference Fraser257).

Initiation of UPR signalling is critical for cell survival and also for viral replication. All the above results show that arboviruses induce UPR signalling upon infection in mammalian cells. However, the UPR pathways that are activated upon infection with various arboviruses are not the same. Even different strains of the same virus activate different UPR pathways. These results suggest that specific virus-induced UPR pathway usage depends on the type of viral strain used. In vitro studies using ectopically-expressed arbovirus nonstructural proteins alone in mammalian cells showed that the proteins themselves can elicit the UPR response. Mutations of certain hydrophobic residues in nonstructural proteins reduced the UPR signalling response. These results indicate that composition of viral nonstructural proteins can determine the type of UPR pathway to be elicited and the extent of UPR response. Viral nonstructural proteins often undergo mutation; thus, more studies are needed to understand the role of arbovirus nonstructural proteins in inducing UPR. The role these viruses may play in UPR in the invertebrate insect cells is even less defined. Thus, induction of UPR signalling by viruses is one important facet, and equally important is how these viruses respond to anti-viral therapy. Do these viruses use the UPR pathways to decrease the effectiveness of anti-viral therapies? This is also one of the main questions to be answered. Thus, in conclusion, a significant amount of research is needed to investigate the pathogenesis of arboviruses and their relationship with UPR signalling. These studies can provide us with better antiviral therapeutics to control arbovirus replication by addressing various mechanisms of virus propagation.