Broad-Spectrum Antiviral Therapeutics

Currently there are relatively few antiviral therapeutics, and most which do exist are highly pathogen-specific or have other disadvantages. We have developed a new broad-spectrum antiviral approach, dubbed Double-stranded RNA (dsRNA) Activated Caspase Oligomerizer (DRACO) that selectively induces apoptosis in cells containing viral dsRNA, rapidly killing infected cells without harming uninfected cells. We have created DRACOs and shown that they are nontoxic in 11 mammalian cell types and effective against 15 different viruses, including dengue flavivirus, Amapari and Tacaribe arenaviruses, Guama bunyavirus, and H1N1 influenza. We have also demonstrated that DRACOs can rescue mice challenged with H1N1 influenza. DRACOs have the potential to be effective therapeutics or prophylactics for numerous clinical and priority viruses, due to the broad-spectrum sensitivity of the dsRNA detection domain, the potent activity of the apoptosis induction domain, and the novel direct linkage between the two which viruses have never encountered.


Introduction
A serious threat is posed by viral pathogens, including clinical viruses (HIV, hepatitis viruses, etc.), natural emerging viruses (avian and swine influenza strains, SARS, etc.), and viruses relevant to potential bioterrorism (Ebola, smallpox, etc.). Unfortunately, there are relatively few prophylactics or therapeutics for these viruses, and most which do exist can be divided into three broad categories [1][2][3]: (1) Specific inhibitors of a virus-associated target (e.g., HIV protease inhibitors, RNAi) generally must be developed for each virus or viral strain, are prone to resistance if a virus mutates the drug target, are not immediately available for emerging or engineered viral threats, and can have unforeseen adverse effects. (2) Vaccines also require a new vaccine to be developed for each virus or viral strain, must be administered before or in some cases soon after exposure to be effective, are not immediately available for emerging or engineered viral threats, can have unforeseen adverse effects, and are difficult to produce for certain pathogens (e.g., HIV). (3) Interferons and other pro-or anti-inflammatories are less virus-specific, but still are only useful against certain viruses, and they can have serious adverse effects through their interactions with the immune and endocrine systems.
To overcome these shortcomings of existing approaches, we have developed and demonstrated a novel antiviral approach that is effective against a very broad spectrum of viruses, nontoxic in vitro and in vivo, and potentially suitable for either prophylactic or therapeutic administration. Our approach, which we call a Double-stranded RNA (dsRNA) Activated Caspase Oligomerizer (DRACO), is designed to selectively and rapidly kill virus-infected cells while not harming uninfected cells.
Our DRACO approach combines two natural cellular processes. The first process involves dsRNA detection in the interferon pathway. Most viruses have double-or single-stranded RNA (ssRNA) genomes and produce long dsRNA helices during transcription and replication; the remainder of viruses have DNA genomes and typically produce long dsRNA via symmetrical transcription [4][5]. In contrast, uninfected mammalian cells generally do not produce long dsRNA (greater than ,21-23 base pairs) [4][5]. Natural cellular defenses exploit this difference in order to detect and to attempt to counter viral infections [6][7]. For example, protein kinase R (PKR) contains an N-terminal domain with two dsRNA binding motifs (dsRBM 1 and 2) and a C-terminal kinase domain [8][9]. Binding of multiple PKR proteins to dsRNA with a length of at least 30-50 base pairs [5] activates the PKRs via transautophosphorylation; activated PKR then phosphorylates eIF-2a, thereby inhibiting translation of viral (and cellular) proteins. Other examples of proteins that detect viral dsRNA include 29,59oligoadenylate (2-5A) synthetases [10], RNase L (activated via dimerization by 2-5A produced by 2-5A synthetases in response to dsRNA [11]), TLR 3 [12], interferon-inducible ADAR1 [13], and RIG-I and Mda-5 [6][7].
The second natural process used by our approach is one of the last steps in the apoptosis pathway [14], in which complexes containing intracellular apoptosis signaling molecules, such as apoptotic protease activating factor 1 (Apaf-1) [15][16] or FLICEactivated death domain (FADD) [17][18], simultaneously bind multiple procaspases. The procaspases transactivate via cleavage, activate additional caspases in the cascade, and cleave a variety of cellular proteins [14], thereby killing the cell.
Because PKR activation and caspase activation function in similar ways and involve proteins that have separate domains with well-defined functions, these two processes can be combined to circumvent most viral blockades [26][27]. In its simplest form, a DRACO is a chimeric protein with one domain that binds to viral dsRNA and a second domain (e.g., a procaspase-binding domain or a procaspase) that induces apoptosis when two or more DRACOs crosslink on the same dsRNA. If viral dsRNA is present inside a cell, DRACOs will bind to the dsRNA and induce apoptosis of that cell. If viral dsRNA is not present inside the cell, DRACOs will not crosslink and apoptosis will not occur.
For delivery into cells in vitro or in vivo, DRACOs can be fused with proven protein transduction tags, including a sequence from the HIV TAT protein [28], the related protein transduction domain 4 (PTD) [29], and polyarginine (ARG) [30]. These tags have been shown to carry large cargo molecules into both the cytoplasm and the nucleus of all cell types in vitro and in vivo, even across the blood-brain barrier.
coli. An empty expression vector was transformed into the E. coli and the same purification protocol was followed, resulting in control extract without DRACOs.
DRACO rapidly entered cells, persisted within cells for days, and mediated apoptosis in cells transfected with dsRNA. PKR-Apaf DRACO with PTD or TAT tags entered cells efficiently, whereas DRACO without a transduction tag did not ( Figure 2A). DRACO entered cells within 10 minutes, reached a maximum after approximately 1.5 hours ( Figures 2B, S1), and persisted inside cells for at least 8 days ( Figure 2C). L929 cells transfected with both DRACO and poly(I):poly(C) dsRNA exhibited greatly increased apoptosis within 24 hours, whereas cells that received only DRACO did not ( Figure 3). Pan-caspase and caspase-9 inhibitors eliminated DRACO-mediated apoptosis in the presence of dsRNA.
We measured the viability of normal human lung fibroblast (NHLF) cells that had been treated with PKR-Apaf DRACOs or negative controls and then challenged with 130 plaque forming units (pfu) per well rhinovirus 1B (Figures 4, S2, S3). Untreated cell populations succumbed to virus within days, indicating that any innate cellular responses were ineffective against the virus or blocked by the virus. DRACOs with PTD, TAT, and ARG tags prevented significant cytopathic effects (CPE) in virus-challenged cell populations by rapidly killing any initially infected cells, thereby terminating the infection in its earliest stages. DRACOs had no apparent toxicity in unchallenged cells. Isolated PKR  and Apaf-1 1-97 domains were nontoxic but not antiviral, even when added simultaneously (but not covalently linked). DRACO with deleterious amino acid changes also had little efficacy. Likewise, an amount of purified bacterial extract (without DRACOs) approximately 10-fold greater than the average volume of DRACOs added to cells was nontoxic and not efficacious, demonstrating that any remaining bacterial contaminants such as lipopolysaccharide did not affect the cells or produce antiviral activity. Thus the antiviral efficacy appears to require intact functional DRACOs. Tests using DRACOs with protein transduction tags on the N terminus, C terminus, or both termini indicated that N-terminal tags generally worked the best (data not shown). DRACOs with transduction tags penetrated cells and were antiviral when administered by themselves (Figures 2, S2A), but efficacy was enhanced by co-administration with Roche FuGene 6 to maximize uptake ( Figure S2B), so FuGene was used in experiments unless otherwise noted. Cell viability measured 7 days post infection (dpi) showed little difference if DRACOcontaining medium was removed 3 dpi after untreated cells had widespread CPE; there was no relapse of viral CPE in treated cells after DRACOs were withdrawn ( Figure 4B).
DRACOs were added approximately 24 hours before virus unless otherwise noted, but other dosing times were tested ( Figure  4C). One dose of PTD-PKR-Apaf DRACO was efficacious against rhinovirus 1B in NHLF cells when added up to 6 days before infection, supporting the western data ( Figure 2C) that DRACO persisted inside cells for at least 8 days. Up to 3 days after infection, one DRACO dose could still rescue a significant percentage of the cell population. After 3 days, virtually all of the cells had already been killed or at least infected by the virus.
Additional DRACO designs exhibited efficacy against rhinovirus ( Figure 5A). Other effective dsRNA detection domains included NTE3L, CTE3L, 26E3L, and RNaseL 1-335 . Other effective apoptotic domains included FADD 1-90 , mApaf1 1-97 , and procaspases [26][27]. Although the initial performance of these alternate DRACOs was generally inferior to that of PKR-Apaf human DRACO in these experiments, better performance might be achieved with further optimization. These results demonstrate that the alternate DRACO designs are nontoxic and efficacious against virus, and they support the DRACO mechanism of action.
In addition to improving survival of the cell population, DRACOs reduced viral titers from virus-challenged cells ( Figures  5B, S4). One dose of PKR-Apaf DRACO administered to NHLF cells 24 hours before 300 pfu/well rhinovirus 1B eliminated any measurable viral titer in cell supernatant samples collected 4 dpi.
The median effective concentration for DRACOs with PTD, TAT, and ARG tags against a variety of viruses was 2-3 nM, as illustrated for PTD-PKR-Apaf DRACO against rhinovirus 1B, murine encephalomyelitis, and murine adenovirus (Figures 5C).
DRACOs were effective against a broad spectrum of other viruses in a variety of cell types (Tables 1-2). DRACOs were effective against rhinoviruses 2 and 30 in NHLF cells (data not shown) and rhinovirus 14 in HeLa cells ( Figure S4). DRACOs     were effective against murine adenovirus in L929 cells if added before or up to at least 72 hours after virus (Figures 6, S5), demonstrating efficacy against a DNA virus ( Figures 6A, S5), in murine cells (using human apoptotic DRACO domains to recruit endogenous murine procaspases), when treatment is delayed until significantly after infection ( Figure 6B), and with a variety of DRACO designs ( Figure 6C). DRACOs were effective against murine encephalomyelitis in L929 cells regardless of whether the DRACO-containing medium was removed 3 dpi ( Figure 7A), whether DRACOs were added before or after infection ( Figure  7B), and which DRACOs were used (Figures 7C, S6). DRACOs were effective in Vero E6 cells against Amapari and Tacaribe, arenaviruses that are closely related to lymphocytic choriomeningitis virus (LCMV), Lassa, and Junin viruses ( Figures 8A, S7, S8). Likewise, DRACOs were effective against Guama strain Be An 277 ( Figures 8B, S9); comparable results were obtained for Guama strain Be Ar 12590 (data not shown). Guama virus is a significant human pathogen and is closely related to other bunyaviruses such as Rift Valley fever, hantavirus, and Crimean-Congo virus. DRACOs were similarly effective against dengue type 2 (New Guinea C) hemorrhagic fever virus, a major human pathogen that is very closely related to other flaviviruses such as West Nile virus, Yellow fever virus, and Omsk virus ( Figures 8C, S10, S11). DRACOs were also effective against H1N1 influenza A/PR/8/34 in normal human hepatocytes ( Figure S12  DRACOs appeared promising in proof-of-concept trials with adult BALB/c mice. Intraperitoneal (i.p.) PKR-Apaf DRACO penetrated the liver, kidney, and lungs and persisted at least 24-48 hours ( Figure 9A). Live mice and harvested mouse organs showed no apparent toxicity. PTD-PKR-Apaf and TAT-PKR-Apaf DRACOs administered i.p. from day -1 through day 3 greatly reduced the morbidity in mice challenged intranasally (i.n.)     Figure 9B). Similarly, PTD-RNaseL-Apaf, TAT-RNaseL-Apaf, and ARG-RNaseL-Apaf DRACOs administered i.p. from day -1 through day 3 prevented morbidity in mice challenged i.n. with 0.3 LD 50 influenza and reduced the day-2 viral titers by an order of magnitude or more ( Figure 9C). PKR-Apaf DRACO administered i.n. to mice penetrated the lungs and persisted over 24 hours ( Figure 10A). PTD-PKR-Apaf, TAT-PKR-Apaf, and ARG-PKR-Apaf DRACOs administered i.n. on day 0 reduced the morbidity in mice challenged i.n. with 1 LD 50 influenza ( Figure 10B). Based on these encouraging initial animal trials, future work should be done to test and optimize antiviral efficacy, pharmacokinetics, and absence of toxicity in vitro and in vivo. Future experiments can further characterize and optimize dsRNA binding, apoptosis induction, cellular transduction, and other DRACO properties. More extensive trials are also needed to determine how long after infection DRACOs can be used successfully, or if DRACOs are useful against chronic viral infections without producing unacceptable levels of cell death in vivo.
DRACOs should be effective against numerous clinical and NIAID priority viruses, due to the broad-spectrum sensitivity of the dsRNA detection domain, the potent activity of the apoptosis induction domain, and the novel direct linkage between the two which viruses have never encountered. We have demonstrated that DRACOs are effective against viruses with DNA, dsRNA, positivesense ssRNA, and negative-sense ssRNA genomes; enveloped and non-enveloped viruses; viruses that replicate in the cytoplasm and viruses that replicate in the nucleus; human, bat, and rodent viruses; and viruses that use a variety of cellular receptors (Table 1).

Ethics statement for mouse trials
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Committee on Animal Care of MIT (Assurance Number: A-3125-01). Guidelines to minimize suffering were followed, and avertin anesthetic was used for intranasal procedures.

Cell assays
Contact-inhibited cells were grown to 50-80% confluence and non-contact inhibited cells to 20-50% confluence in 96-well plates with 100 ml/well medium. DRACOs or controls were added to columns of wells, 8 wells/column. Except in Figs. 2 and S2, 0.4-1% (vol/vol) Roche FuGene 6 was co-administered with DRACOs and controls to optimize cellular uptake. Wells received virus approximately 24 hours after DRACO unless otherwise noted. On selected days, cell viability in each plate was measured using CellTiter 96 (Promega). Assay schedules, viral doses, and other parameters were optimized for different cell/virus systems. Micrographs were taken in 24-well plates under similar conditions.

DRACO cell penetration assays
Cells in 24-well plates were incubated with DRACOs for varying lengths of time, then trypsinized, washed thoroughly in PBS, and lysed. Lysate from approximately 10 5 cells was loaded in each lane. DRACOs were detected via westerns using mouse anti-His 6 (Invitrogen) and goat anti-mouse IgG HRP (Jackson).

Viral titers
Titers were determined by serial dilutions onto 96-well NHLF (for rhinovirus 1B) or H1-HeLa (for rhinovirus 14) plates, with 8 wells per 10-fold dilution and with the number of wells exhibiting CPE measured 5 dpi. Reed-Muench titers were calculated from the results (1 TCID 50 <0.7 pfu). Error bars indicate s.e.m. from 3 trials.

Statistical analysis
CellTiter 96 cell viabilities were normalized to 100% for untreated uninfected and 0% for untreated virus-killed cells. Mice were challenged i.n. with 0.3-1.3 LD 50 influenza H1N1 A/ PR/8/34. Mice received DRACO i.p. once daily on days -1 and 1-3 and twice on day 0, or just one i.n. DRACO dose simultaneously with virus. Lungs were harvested on day 2 and viral titers determined by serial dilutions onto 96-well MDCK plates. For pharmacokinetics, organs were harvested at designated times, then sonicated into 1 ml PBS with 1% Triton X-100. 1 mg organ solution was mixed with 26Laemmeli buffer, boiled 5 min., and run on a 10-20% SDS PAGE gel with a standard curve of purified DRACO, followed by western blots with anti-Apaf (Millipore) and HEP-labeled anti-rabbit IgG (Jackson Immunoresearch). Blots were developed with Pierce luminescent reagent and exposed to film. DRACO bands were quantitated by Gel Doc densitometry vs. the standards.  cells. Photographs were taken 11 days after challenge with 30 pfu/ well reovirus 3. Right four photos: 200 nM PTD-26E3L-Apaf DRACO was effective against adenovirus 5 in human embryonic kidney AD293 cells. Fluorescent microscope photographs were taken 4 days after challenge with 25 pfu/well adenovirus 5 expressing enhanced green fluorescent protein (EGFP). Scale bars = 50 mm. (TIF)