Viral infection of leukocytes or respiratory epithelium activates a variety of pattern recognition-receptors (PRR), such as the TLRs or the NOD receptors that induce the production of IRF3, NFκB, and other cell activation pathways (26
). Virally induced secretion of IFN-β (largely mediated through IRF3) is especially important because this molecule acts in a paracrine fashion on neighboring cells to up-regulate a group of key antiviral proteins (such as OAS1, MX1, and PKR) that prevent subsequent infection and spread of the virus.
Most viruses have developed sophisticated (and different) mechanisms to try to prevent IRF3/IFN pathway activation (27
). VSV, for example, encodes a Matrix protein that inhibits host cell gene expression by targeting a nucleoporin and blocking nuclear export (28
). Hence, any mRNAs induced by activation of the IRF pathway after viral replication is initiated will not lead to production of IFN or its downstream antiviral effector genes. The NS1 protein of H1N1 influenza virus has also been implicated in a number of regulatory functions during influenza virus infection, including binding of the poly(A) tails of mRNAs (to inhibit their nuclear export) (29
), inhibiting host mRNA polyadenylation, contributing to the virus-induced shutoff of host protein synthesis (31
), and inhibiting pre-mRNA splicing (29
). In addition, binding of NS1 to dsRNA prevents in vitro
activation of PKR (35
Given this ability of IRF3/IFN to prevent viral infection, activation of this pathway as a potentially “antigen-independent” way of controlling disease has been investigated. As early as the 1960s, type I IFNs were reported to be antivirals (36
). Early reports of success using nasal spray preparations of IFN-α by Russian drug companies for prophylaxis and treatment of influenza attracted the interest of Western scientists into this arena (37
). Subsequent studies were conducted that found that the reported beneficial effects were minimal in relationship to the considerable amount of side effects associated with the treatment (38
). The limited efficacy of direct application of IFN was likely related a number of issues, including: (1
) dose-limiting toxicities, such as cognitive dysfunction, dyspnea, fatigue, nausea, and vomiting (39
) a very short half-life due to high renal clearance (38
); and (3
) potential generation of neutralizing antibodies. However, given recent advances in recombinant IFN production that can prolong the half-life of IFN and better nebulization technology, a reinvestigation of this approach may be in order (40
An alternative strategy to harness the innate immune system is to use agents that mimic viral infection and activate the innate immunity through PRR. There are four known TLRs (TLR-3, -7, -8, and -9) that recognize different forms of “foreign” nucleic acids and induce type I IFN to activate antiviral gene function (38
). Compared with using recombinant IFN protein, agonists that activate specific TLRs have been shown to be less toxic, to induce longer IFN exposure, and to be easier to administer and more effective (41
). A TLR-7 agonist, imiquimod, was approved by the FDA in 1997 for the treatment of external genital warts (42
). Similarly, polyinosine–polycytidylic (poly I:C) is a TLR-3 agonist has been studied for the treatment of influenza infection (43
). Although a number of studies reported that polyinosine–polycytidylic can be used for prophylaxis against influenza infection, two major limitations of using nuclear acid–based antiviral agents are their susceptibility to nuclease degradation in vivo
and the fact that these molecules are not cell permeable. The latter point is a critical issue because the TLR3 receptor is intracellular (44
). Attempts of overcome these obstacles have included chemical modifications to the nucleic acid backbones and sugars and packaging with liposomes. Unfortunately, these approaches have been shown to increase toxicity (43
The strategy of activating pattern recognition receptors has recently been applied using an aerosolized lysate of nontypeable haemophilus influenzae (21
). Inhalation of this lysate induced profound inflammation in the lungs with an accompanying increase in a variety of IFNs and cytokines, yet it strongly protected mice against subsequent infection with bacterial, fungal, and viral (including influenza) pathogens. You and colleagues (45
) suggested that this effect could be reproduced with lower toxicity by administration of two specific TLRs (but not by single TLR agonists). Thus, this strategy could be promising if the initial induction of lung inflammation could be minimized.
The data presented in this report suggest that there are other agents, like DMXAA, that have the ability to activate antiviral pathways after systemic or intranasal instillation without inducing strong lung inflammatory responses. In this study, we showed that DMXAA directly stimulated IFN-β production in respiratory epithelial culture (Figures 1A, 1B, 3A, and 4B) as well as inducing the downstream IFN-stimulated antiviral genes PKR and OAS-1 (, and ). This activation of the IFN pathway by DMXAA correlated with protection of C10 bronchial epithelium from VSV-mediated cytopathic effects in culture (). We also showed activation of the IFN pathway after administering DMXAA to mice intranasally or intraperitoneally. There was a significant increase in OAS1 protein in epithelial cells along the respiratory tract, including the nasal mucosa, the trachea mucosa, and the lung epithelium (). Although intraperitoneal administration of DMXAA showed suboptimal protection against influenza virus challenge (), intranasal administration was able to prolong the survival of mice () and to decrease the viral titer () while inducing only minor amounts of inflammation, as assessed by changes in the bronchoalveolar lavage fluid.
These activities of DMXAA are potentially important for a number of reasons. First, the drug appears to be highly permeable, unlike agents such as poly I:C, and thus does not require liposomal encapsulation for administration. Second, DMXAA has been shown to simultaneously activate multiple antiviral pathways in macrophages. These pathways include (1
) NFκB activation (12
) IFN-β through TBK1/IRF3 activation (11
) the NOD pathway (12
), and (4
) the MAP kinase pathways (p38, ERK, and Jun) (13
). A recent study by Sabbah and colleagues (46
) showed that NOD2 can also function as a cytoplasmic viral PRR by triggering activation of IRF3 and production of IFN-β. Our data indicate that the same pathways seen in macrophages are activated in epithelium. Third, DMXAA does not appear to induce strong inflammatory responses in the lungs (unlike bacterial lysates) and was well tolerated by the mice. In fact, DMXAA is in Phase III clinical trials for use in lung cancer and has not induced serious toxicities. One major unanswered question is how a small organic compound, like DMXAA, is able to trigger such a broad response. Despite many years of work in multiple labs, the “DMXAA receptor” has yet to be discovered. However, it is known that DMXAA does not require MyD88 or the RIG-I–like receptors (9
). Further research into this question is ongoing.
There are a number of caveats to this study. DMXAA had a relatively narrow window of therapeutic efficacy and worked best when administered close to the time of viral infection; treating mice with DMXAA 24 hours after infection was not effective. This may not be surprising considering the range of virally produced proteins that inactivate IFN responses. The value of agents like DMXAA would thus appear to be primarily in prophylaxis rather than therapy (i.e., to be used when a known exposure would be likely to occur). Another limitation of DMXAA is that there appears to be a refractory period that lasts 3 to 5 days (data not shown). This suggests that daily administration may not be effective. Finally, the effects of DMXAA in human (versus murine) bronchial epithelial cells are attenuated. We studied human immortalized (but not transformed) bronchial epithelial cells and primary cultures of human tracheal epithelial cells and found lower IFN responses after DMXAA administration (Figure E3) accompanied by minimal anti-VSV protection (data not shown). This species difference has been previously observed in human leukocytes (47
). For these reasons, it is unlikely that DMXAA itself will be useful clinically in patients; however, related compounds with higher activity in human cells may be valuable.
In summary, this study provides additional “proof of principal” data that pharmacologic activation of the IRF3/type I IFN pathway is an attractive strategy in the development of prophylactic antiviral therapies. This can be achieved in mice in a nontoxic and efficient manner by a small, cell-permeable, xanthone-like molecule (DMXAA) that activates multiple antiviral pathways by an undiscovered mechanism. Understanding the molecular targets of DMXAA could be important because this would likely aid development of related compounds with high activity in humans. Even if the effects are relatively short lived, the prophylactic use of a DMXAA-like drug or other types of IRF-3 activators (such as TLR agonists) could be highly useful for the protection of “front-line” health professionals or military professionals who might need to enter environments where exposure to pandemic viruses or viral bioterror agents could be expected.