Introduction

Vaccine development has a critical role in inhibiting the spread of influenza A virus (FLUAV) between animals, humans, and from animals to humans. However, FLUAV vaccines are typically effective only on a seasonal basis and against anticipated strains [8, 25]. The circulation of unanticipated and highly pathogenic FLUAV strains has increased in recent years [11]. Furthermore, there has been a recent increase in the incidence of zoonotic transmission of FLUAV, notably highly pathogenic avian influenza (HPAI), from birds to humans and other mammals [26, 36]. The discovery in 2005 of a cross-species transmission of the FLUAV H3N8 subtype from horses to dogs in the United States [10, 47], in addition to the August 2007 spread of H3N8 equine influenza to the previously noninfected continent of Australia [4], has increased concerns over this subtype as well.

FLUAV with the H3 hemagglutanin are also of significance to humans. Currently, the prominent subtype for seasonal influenza is H3N2. In addition, the influenza pandemic of 1890 is presumed to have been caused by an H3N8 subtype virus [16]. Although it is unlikely that an H3N8 influenza virus will soon be the cause of a human epidemic or pandemic, this HA subtype has previously and will continue to be a cause of seasonal morbidity and mortality. The preliminary testing of antiviral drugs against H3N8 strains is therefore of interest.

Small-molecule inhibitors of influenza have been available for several years and have proven effective at reducing the severity and duration of illness [20]. However, some of these drugs, notably the adamantanes, are now considered ineffective against many human FLUAV strains [10]. Recently, a small percentage of FLUAV strains have been shown to be resistant to neuraminidase inhibitors as well [6, 13, 24, 27]. The inadequacy of the current armament of FLUAV therapeutics is evident, and new therapeutics effective against a multiplicity of FLUAV subtypes are clearly desirable.

Phosphorodiamidate morpholino oligomers (PMOs) [42] are oligonucleotide analogs that can act as antisense agents by base pairing with complementary RNA target sequence and forming a steric block [41]. PMOs are water soluble and stable in cells [22, 32]. Cellular uptake of PMOs can be greatly enhanced through conjugation of a cell-penetrating peptide (CPP) [14, 30]. CPP-PMOs (PPMOs) were recently shown to successfully inhibit several subtypes of FLUAV in cell cultures [18]. Here, we show that intranasal delivery of PPMOs at non-toxic doses was able to reduce the replication of Influenza A/Eq/Miami/1/63 (H3N8) in the lungs of mice by over 95% compared to controls. The effective PPMOs were designed against sequences in the polymerase basic 1 (PB1) and nucleoprotein (NP) genes that are highly conserved between FLUAV subtypes. These results indicate that PPMOs may have utility as a novel therapeutic against a multiplicity of FLUAV subtypes.

Materials and methods

Virus and cells

Influenza A/Eq/Miami/1/63 (H3N8) and A/Eq/Prague/56 (H7N7) were obtained from Rocky Baker (Veterinary Diagnostic Laboratory, Oregon State University) and grown in the allantoic cavity of 10-day-old embryonated chicken eggs for 3 days. Virus was harvested and stored at −80°C for future use. Virus was titered by standard plaque assay on Madin–Darby canine kidney (MDCK) cells (see below). MDCK cells were grown in minimum essential media (MEM) (Hyclone) supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin (Gibco), and 10% fetal bovine serum (FBS) (Hyclone).

PMO and PPMO

PMOs were synthesized at AVI BioPharma, Corvallis, OR, by previously described methods [41]. PPMOs were produced with the CPP (RXR)4XB, as previously described [2]. Two oligomer sequences specific to multiple strains of FLUAV RNA, along with a nonsense sequence control having no significant homology to FLUAV or any human mRNAs, were prepared as both PMO and PPMO. The first of the two antisense sequences, PB1-AUG (5′-GACATCCATTCAAATGGTTTG-3′), is a 21mer complementary to the AUG translation start site region of the polymerase subunit PB1 mRNA. The second sequence, NP-v3′ (5′-AGCAAAAGCAGGGTAGATAATC-3′), is complementary to the 22 terminal nucleotides at the 3′ end of the NP virion RNA (vRNA). The nonsense sequence utilized as a negative control for both PMO and PPMO is a 20mer of 50% G/C content, named DScr (5′-AGTCTCGACTTGCTACCTCA-3′). In addition, the DScr sequence was prepared as a PPMO with fluorescein conjugated to its 3′ end (DScr PPMO-Fl).

Mice

All animal experiments were performed with 6–8-week-old female Balb/C mice (Simonsen Laboratories), which were allowed food and water ad libitum throughout the studies. Infections and treatments of mice were performed under anesthesia with 67 mg/kg ketamine and 4.5 mg/kg xylazine, and mice were euthanized by CO2 asphyxiation for necropsy. All animal experiments were carried out under BSL-2 conditions with protocols approved by the Institutional Animal Care and Use Committee of Oregon State University.

Evaluation of intranasal administration as a method of PPMO delivery into lung tissue

To determine if intranasal (i.n.) instillation would effectively deliver PMO compounds into the lungs, 40 μl of 0.9% saline or 40 μl of 0.9% saline containing 3.75 mg/kg of DScr PPMO-Fl was administered i.n. to noninfected mice. At 16 or 24 h post-treatment, PPMO-Fl and saline-treated animals were euthanized and the lungs removed for imaging. One lung was imaged immediately by fluorescent whole-mount microscopy with a Zeiss Stereo Discovery V8 microscope. The other lung was flash frozen in liquid nitrogen and then sectioned into upper, middle and lower lung samples for examination using a Leica DMLB fluorescent microscope and SPOT CCD camera (Diagnostic Instruments, Inc.).

Detection of pulmonary infiltrates

Noninfected mice were used to examine the effect that a prospective antiviral PPMO dosing regimen could have on inflammation of pulmonary tissue. Mice were administered two i.n. doses, 24 h apart, of a 50/50 mixture of PB1-AUG and NP-v3′ PPMO at concentrations from 0 to 7.5 mg/kg. Mice were weighed at 24, 48 and 72 h after the initial dose and euthanized at 72 h for removal of the lungs and spleen. One lung was formalin fixed, sectioned, and hematoxylin and eosin (H&E) stained for histological examination. All histological examination was performed by pathologists at the Veterinary Diagnostic Lab, College of Veterinary Medicine, Oregon State University. Leukocyte cellularity of the other lung, and of spleen tissue, was measured using flow cytometry with the following methodology: Fresh tissue samples were placed in PBS containing 1% FBS. Single cell suspensions were made by passing tissues through a 70-μm nylon mesh screen cell strainer (BD Falcon). The strainers were rinsed into 50-ml tubes and the suspension washed twice by room-temperature centrifugation (1,500×g) and resuspension of the pellet in 10 ml of DMEM + 1% FBS. Red blood cells were lysed by suspending the pellet in 1 ml of cold 1× RBC lysis buffer (eBioscience) followed by two washes in 10 ml DMEM + 1% FBS. Cell viability and enumeration was determined by trypan blue exclusion using a Vi-cell TM XP cell viability analyzer (Beckman Coulter). A final centrifugation and suspension of the cells (5 × 106/ml) in FACS Buffer (1× PBS + 1% FBS + 0.02% NaN3) was performed prior to subjecting the cells to antibody staining. Individual samples were stained using antibodies: CD4-PE (clone L3T4), CD8-PE (clone Ly-2), CD11b-PE (MAC-1), CD11c-PE (N418) (eBioscience), CD45R-FL (B220) (BD Pharmingen), GR-1 TriColor (Caltag) or manufacturer’s corresponding isotype controls (Fl = fluorescein, PE = phycoerythrin). The stained samples were examined (minimum of 50,000 events) on a Cytomic FC 500 (Beckman Coulter) and data analyzed using FlowJo Software (TreeStar).

Viral inhibition experiments in mice

To compare the relative efficacy of non-conjugated PMO to that of PPMO, mice were treated i.n. with 100 μl of saline, or 100 μl of saline containing a 50/50 mixture of PB1-AUG/ NP-v3′ PMO totaling 3.75 mg/kg or 3.75 mg/kg of DScr PMO. Additional groups were treated with saline containing either 3.75 mg/kg of 50/50 PB1-AUG/NP-v3′ PPMO or 3.75 mg/kg of DScr PPMO. Treatments were performed at 4 h before and 20 h after infection with 2.5 × 105 PFU of Influenza A/Eq/Miami/1/63 (H3N8). After infection, the mice were weighed every 24 h, and at 72 h euthanized and lungs harvested for determination of virus titer by plaque assay.

In a subsequent experiment, designed to determine which of the two antisense PPMOs was more effective against H3N8 in vivo, mice were treated i.n. with 100 μl saline, or 100 μl saline containing 3.75 mg/kg of either PB1-AUG, NP-v3′or DScr PPMO. Treatment was again administered at 4 h before infection and 20 h post-infection. Mice were weighed every 24 h, then euthanized at 72 h pi, and lung tissue was harvested for plaque assays.

To determine the effect of PPMO treatment timing in relation to the time of infection on antiviral efficacy, mice were treated with two doses of 3.75 mg/kg of NP-v3′ PPMO under five different regimens: at either 4 h pre-infection and 20 h post-infection, 1, 2, or 4 h post-infection and 20 h post-infection, or 20 h post-infection and 36 h post-infection. Control mice were treated with DScr PPMO or saline on similar schedules. Body weights were recorded as above. All animals were euthanized at 72 h pi. and lungs then collected for plaque assays.

Plaque assays

Plaque assays were performed by grinding lungs in 1.5 ml MEM supplemented with 100 U/ml penicillin and 100 μg/ml streptomycin and subjecting them to two freeze–thaw cycles followed by centrifugation at 2,300×g for 2 min. Supernatant was serially diluted, and standard plaque assays were performed on >90% confluent MDCK cells in 12-well plates. Briefly, MDCK cells in duplicate wells were infected with 100 μl of serially diluted virus for 1 h by incubation at 37°C with periodic shaking. Virus-containing supernatant was removed by aspiration, and the MDCK cell monolayer was overlaid with MEM containing 100 U/ml penicillin, 100 μg/ml streptomycin, 3% BSA (Sigma), and 1 μg/ml TPCK Trypsin (Sigma) with 1% Sea Plaque Agarose (Cambrex). Two days later, agarose was removed and plaques were visualized by staining with 0.75% crystal violet (Alfa Aesar) diluted in 100% methanol and quantified by determining the highest dilution in which duplicate wells contained at least ten plaques.

Sequencing of the PB1 gene

MDCK cells were infected with 0.01 MOI A/Eq/Miami/1/63 (H3N8) or A/Eq/Prague/56 (H7N7), and viral RNA was isolated from cell culture supernatants at 48 h pi by centrifugation at 2,300×g for 5 min (to remove cellular debris), and using Ambion’s MagMAX™ -96 Viral RNA Isolation Kit, according to the manufacturer’s protocol. Using primers PB1-forward (5′-GAATTCAGCTTAGCGAAAGCAG-3′), and PB1-reverse 248 (5′-TCCATCGATTGGATTAAGTTGTG-3′), or PB1-reverse 850 (5′-GGCAATCCTGATTGTTCAAG-3′), the gene of interest was converted to cDNA and amplified via one-step RT-PCR (Superscript III One Step RT-PCR kit, Invitrogen). PCR products were then cloned into pCR8/GW/TOPO (Invitrogen) vector according to the manufacturer’s protocol. Plasmids from five bacterial colonies containing each PCR product were isolated and purified using the PureLink Quick Plasmid Miniprep Kit (Invitrogen). Plasmids were sequenced at Oregon State University’s Center for Genome Research and Biocomputing (CGRB) core facility.

Statistics

All statistical analyses were performed using Prism 4 software (GraphPad Inc.) using One-Way ANOVA with Bonferroni or Dunnett’s multiple comparison test.

Results

Our previous study showed that NP-v3′ and/ or PB1-AUG PPMOs were effective at inhibiting the replication of several strains of FLUAV in various cell lines, in a dose-dependent and sequence-specific manner, with minimal attendant cytotoxicity [18]. Encouraged by these results, we sought to evaluate the efficacy of these PPMOs against A/Eq/Miami/1/63 (H3N8) replication in a mouse model. We first evaluated the potential toxicity of these two PPMOs in noninfected mice under conditions similar to those being considered for virus inhibition experiments. No such preliminary assessment was carried out for PMO, as it has been shown to be safe in vivo at high doses [3, 14]. None of the dosing regimens employed resulted in any significant weight loss over the 72-h period of monitoring (Fig. 1a). Histological examination revealed no significant differences in the lung tissue of PPMO-treated compared to saline-treated mice (data not shown). To investigate whether PPMOs induce pulmonary inflammation, we used flow cytometry to detect leukocytic infiltrates in a lung and the cellularity of spleen tissue isolated from each group shown in Fig. 1a. Our results indicate that there were no significant changes in spleen leukocyte populations for any PPMO treatment group compared to the saline-treated group (Fig. 1b). The lungs of mice that received the highest PPMO dose (7.5 mg/kg) showed a statistically significant increase of 2.21 fold in macrophages (95% CI = 2.78–1.64 fold) and 3.38 fold in granulocytes (95% CI = 4.98–1.78 fold) (Fig. 1c). It is noteworthy that the pulmonary infiltrates evident by flow cytometry were not associated with statistically significant weight loss or observable illness in mice over the 3-day duration of this experiment.

Fig. 1
figure 1

Evaluation of PPMO toxicity in mice. Duplicate mice were treated with various doses of 50/50 NP-v3′/PB1-AUG PPMO or 0.9% saline (control) through intranasal administration. a Mice were weighed every 24 h following treatment, for 72 h, to determine weight loss. At 72 h post-treatment, all mice were killed, spleen (b) and lung (c) were removed, and flow cytometry was performed on single cell suspensions using antibodies targeting immune-cell-specific markers. *P < 0.5, **P < 0.01 using One-Way ANOVA with Bonferroni’s multiple comparison test

Full size image

FLUAV has been reported to replicate in both the upper and lower respiratory tracts [34, 38, 46]. Because the lungs are known to be the primary site of viral replication during FLUAV infection, we explored i.n. administration as a way to maximize PPMO delivery to the lungs. We first investigated the distribution of PPMO in the lungs of noninfected mice after i.n. delivery, to determine the feasibility of this route for treatment during an infection. Mice were given either saline or 3.75 mg/kg of PPMO-Fl and necropsied at 16 or 24 h post-administration for examination of lung tissue. A view of the whole lung shows that PPMO-Fl is distributed in a gradated fashion from the upper to the lower lung. The signal from the PPMO-Fl is noticeably stronger in proximity to the major bronchioles (Fig. 2a). Figure 2b includes higher-magnification views of control lung tissue, which received no PPMO-Fl (Fig. 2b, panel 1), and of tissue from the upper (Fig. 2b, panel 2), middle (panel 3) and lower (panel 4) regions of the lung receiving PPMO-Fl. These images confirm that distribution of PPMO after i.n. administration is highest in the upper lung, but that a detectable amount of PPMO-Fl is present in the lower lung region (Fig. 2b, panel 4). Together, these data clearly show that PPMO compounds delivered by the i.n. route enter the lungs efficiently.

Fig. 2
figure 2

Visualization of PPMO-Fl in lungs after intranasal administration. Three mice were treated with either 3.75 mg/kg fluorescein-conjugated-PPMO or 0.9% saline for determination of PPMO distribution in the lungs. a Fluorescent whole-mount microscopy of left lung of PPMO-Fl and saline-control-treated mice necropsied at 16 h post-treatment. b Fluorescent microscopy of cryo-sectioned right lung necropsied at 24 h post-treatment. Section of saline control mouse lung (1) or PPMO-Fl-treated mouse upper (2), middle (3), and lower (4) lung. Arrows indicate concentrations of fluorescent signal

Full size image

Initially, we compared the efficacy of nonconjugated PMO to that of PPMO, to determine if conjugation to a cell-penetrating peptide had any effect on the ability of PMO to inhibit H3N8 replication in the lungs of mice. Mice were infected with the same strain of influenza, A/Eq/Miami/1/63 (H3N8), that had been used in previous cell culture experiments with PPMO [18]. Based on the possible mild pulmonary inflammation evident from the 7.5 mg/kg PPMO dose (Fig. 1c), we employed 3.75 mg/kg of PPMO per dose in this antiviral trial. Mice were dosed with a 50/50 mixture of PB1-AUG and NP-v3′ in the PMO or PPMO form, or administered equivalent DScr or saline controls. Dosing was performed at 4 h before and 20 h after infection. There was no correlation between weight loss and treatment or weight loss and virus load during the 3-day duration of this experiment (data not shown). Plaque assays performed on lung tissue taken at 72 h post-infection showed a 1.5 log10 (95% CI = 2.17–0.72) reduction in virus growth in mice treated with the PB1-AUG/NP-v3′ PPMO combination in comparison to the DScr PPMO-treated group (Fig. 3a). The PB1-AUG/NP-v3′ PMO combination treatment appeared to produce a moderate decrease in titer compared to the DScr PMO-treated group; however, the difference did not achieve statistical significance. These results indicate that there was a reduction in virus growth in vivo produced by the antisense PPMO combination treatment and that the conjugation of the (RXR)4XB peptide to PMO enhances the ability of PMO to inhibit H3N8 replication in vivo.

Fig. 3
figure 3

Reduction of influenza A H3N8 lung titer in mice treated with PPMO. Lung virus titers at 72 h post-infection were determined by plaque assays for all experiments. a Mice were treated with two doses of PMO or PPMO at 3.75 mg/kg as indicated, or with the same volume of 0.9% saline, 24 h apart, starting at 4 h pre-infection. b Mice were treated with 3.75 mg/kg of the indicated PPMO at 4 h before and 20 h after infection. c Mice were treated with two doses of PPMO NP-v3′, 3.75 mg/kg, at the indicated start times. *P < 0.5, **P < 0.01 using One-Way ANOVA with Dunnett’s multiple comparison test, n = 4–5 mice per group

Full size image

Next, we sought to determine which PPMO would produce the highest efficacy: PB1-AUG alone or NP-v3′ alone. Previous data indicated that certain FLUAV strains (A/PR/8/34 (H1N1), A/Thailand/1(KAN-1)/04 (H5N1), and A/WSN/33 (H1N1)) were inhibited similarly by either PB1-AUG or NP-v3′ PPMO, but that other strains (A/Eq/Miami/1/63 (H3N8), A/Eq/Prague/56 (H7N7), and A/Memphis/8/88 (H3N2)) were not [18]. In addition, we sought to determine if our previous results with the PB1-AUG and NP-v3′ PPMO against A/Eq/Miami/1/63 (H3N8) in cell culture would extend to in vivo conditions. Mice were treated with PB1-AUG or NP-v3′ as well as DScr or saline controls. Treatments were all performed at 4 h before and 20 h after infection with A/Eq/Miami/1/63 (H3N8). There was a statistically significant decrease of 1.0 log10 (95% CI = 1.886–0.06974) in virus titer produced by treatment with NP-v3′ PPMO alone compared to DScr PPMO. In contrast, the PB1-AUG PPMO alone produced no significant reduction in virus titer compared to the DScr PPMO (Fig. 3b). These data are consistent with those previously obtained with A/Eq/Miami/1/63 (H3N8) in cell cultures [18].

Various experiments indicate that antiviral PPMOs act through steric blockage of particular viral RNA regions, such as the AUG start site or critical regulatory motifs in noncoding regions [14, 21, 33, 44, 45, 48]. Sequence analysis has indicated that FLUAV resistant to PPMO treatment can contain sequence mismatches between the PPMO and the RNA target. For example, the NP-v3′ PPMO-targeted region of A/Memphis/8/88 (H3N2) has two known mismatches and was resistant to this PPMO in cell culture [18] (see Table 1). As shown here in vivo, and in previous work in cell cultures, PB1-AUG PPMO is not effective against the A/Eq/Miami/1/63 (H3N8) strain. In an effort to gain insight into a possible explanation for the inactivity of PB1-AUG PPMO against this strain, we sequenced the first 850 nt of the PB1 gene, which was not previously available in the Influenza Sequence Database [31]. We compared the PB1-AUG PPMO target sequence of the A/Eq/Miami/1/63 (H3N8) strain used in this experiment to several other FLUAV strains used in previous experiments [18] (Table 1). The sixth nucleotide in the PB1-AUG PPMO target region of A/Eq/Miami/1/63 (H3N8) has a C-to-U mutation compared to other strains of FLUAV. The identical mutation was also found to be present in A/Eq/Prague/56 (H7N7), a strain against which PB1-AUG PPMO was also found to be ineffective in cell culture testing [18]. In contrast, no mutations in the PPMO target sequence were found in any of the viral strains for which PB1-AUG or NP-v3′ PPMO have previously been shown to be effective (Table 1).

Table 1 Comparison of PPMO target sequences in various strains of FLUAV
Full size table

Finally, we tested the effect of the timing of PPMO treatments, in relation to the time of viral infection, on antiviral efficacy in mice. One group of mice was treated initially at 4 h before infection, while the other four groups were treated initially at 1, 2, 4 or 20 h after infection, respectively, with NP-v3′ PPMO. Infections and treatments were performed as before, with 3.75 mg/kg of PPMO per treatment. Figure 3c shows that treatment initiated at 4 h pre-infection produced a 1.5 log10 (95% CI = 1.884–1.050) reduction in virus growth, similar to that observed in previous experiments (Fig. 3a, b). With initial PPMO treatment occurring at 1 or 2 h after infection, we observed a significant decreases of 0.83 log10 (95% CI = 1.248–0.4136) and 0.57 log10 (95% CI = 0.9871–0.1528), respectively (Fig. 3c), while treatment initiated at 4 or 20 h post-infection produced no significant decrease in virus growth compared to mice treated with the DScr control PPMO on the same schedule. These results show that PPMO treatment before or early after infection with a non-lethal influenza strain is capable of significantly reducing virus growth in the lungs of mice.

Discussion

There is a pressing need for new antiviral drugs against FLUAV [29]. PPMOs were previously shown to be effective at inhibiting the replication of multiple FLUAV subtypes in cell cultures, with high specificity and low cytotoxicity [18]. Here, we show that the PB1-AUG and NP-v3′ PPMO in combination or NP-v3′ alone, but not PB1-AUG alone, effectively reduced the growth of the A/Eq/Miami/1/63 strain of H3N8 by as much as 1.5 log10 when delivered into lung tissue of mice through an intranasal route of administration (Fig. 3a–c).

The results show that the PPMOs in this study were effective if administered before or shortly after viral infection, producing reductions in viral titer similar to those reported for other antisense- [1] or siRNA-based experiments [17, 43]. However, if treatment was delayed until 4 h after viral infection, a statistically significant reduction in viral titer was not achieved (Fig. 3c). It is possible that the PPMO is only effective if administered early during the course of an infection, before virus replication is fulminant in the lungs. However, we note that other antiviral studies with T-705 [40], RWJ-270201 [39], and Oseltamivir [39, 40] have shown that although drug treatment did not reduce virus growth significantly early in infection, viral titers were indeed reduced later in infection compared to controls, and protection from lethal infection was observed. In addition, the efficacy of anti-FLUAV drugs appears to be variable depending on the viral strain used for challenge [19, 35].

The A/Eq/Miami/1/63 (H3N8) strain used in these experiments does not produce a lethal infection in mice and propagates only to ∼105 PFU/g of lung tissue, as compared to a lethal infection with A/PR/8/34 (H1N1), from which we have observed a titer of ∼107 PFU/g of lung (data not shown). Furthermore, mice infected with A/Eq/Miami/1/63 begin to recover by day 5 post-infection without any treatment (data not shown). Further testing in a mouse model using a lethal influenza infection is necessary to determine if treatment in a therapeutic manner is capable of both increasing survivorship and of reducing virus load over a period longer than three days.

PPMO produced antiviral effects at concentrations that did not cause weight loss or induce a host immune response detectable in the lungs or spleen (Fig. 1a–c). However, some abnormal infiltration of the lungs by immune-system cells was produced by a dose of PPMO higher than those employed in the anti-viral experiments. This potential toxicity was presumably due to the CPP, as previous studies have shown no toxicity for nonconjugated PMO [3, 15]. It has been shown that administration of high-dose poly-arginine via intratracheal injection results in a marked increase in eosinophil and neutrophil infiltration into the lungs of guinea pigs [5]. Mechanistically, this has been attributed to the ability of poly-l-arginine to mimic the effect of constituents stored in the secretory granules of eosinophils, especially major basic protein (MBP) [5]. MBP is a 177-amino-acid-residue polypeptide rich in arginine. When MBP is released from granules, it mediates migration of inflammatory cells via the induction of ion flux and prostaglandin synthesis in airway epithelium as well as histamine release and airway smooth muscle contraction in the bronchoalveolar compartment [5]. It is possible that at high doses the arginine-rich CPP mimics MBP and has an effect on lung epithelium similar to that observed with high doses of poly-l-arginine.

Sequencing of the A/Eq/Miami/63 strain of H3N8 used here showed that it has a single base mismatch with PB1-AUG PPMO at the target site. The PB1-AUG PPMO has effectively inhibited other FLUAV strains lacking the mismatch [18] (see Table 1). The NP-v3′ PPMO sequence has perfect complementarity in A/Eq/Miami/63 (H3N8) (Table 1), and was highly effective. Taken together, these observations indicate that the C-to-U mutation causing the mismatch may be the cause for failure of the PB1-AUG PPMO to inhibit A/Eq/Miami/63 (H3N8). However, that conjecture is not consistent with previously reported results of cell-free translation experiments utilizing in vitro-transcribed RNAs to investigate the activity of a PPMO against RNAs with which it had a various number of mismatches [18, 49]. In those experiments, one mismatch between a PPMO and RNA sequence had little impact on PPMO-mediated inhibition of translation. The discrepancy between the results of those in vitro translations compared to our cell culture or in vivo results with PB1-AUG PPMO and FLUAV suggest that PPMO and RNA can interact quite differently in a cell-free assay compared to intact biological systems. It seems unlikely, however, that the C-to-U mutation would be the sole reason why the PB1-AUG PPMO alone was ineffective, as the resulting U in the viral mRNA would still have two hydrogen bonds with the corresponding G in the PPMO and therefore represents a minor penalty to hybridization potential.

The evolution of drug-resistant strains of FLUAV has resulted in the number of anti-FLUAV drugs recommended for use in humans being cut in half [37]. The use of antivirals administered in combination as a strategy to inhibit the development of drug resistance has been suggested previously [17, 18, 23]. As shown here, although resistant to the PB1-AUG PPMO alone, A/Eq/Miami/1/63 (H3N8) was susceptible to NP-v3′ PPMO or a combination of NP-v3′ and PB1-AUG PPMO (Fig. 3). These results suggest that a combination of PPMOs may provide superior protection against naturally occurring mutants as well as reduce the likelihood of the development of viral escape mutants. The choice of the NP and PB1 genes as targets for PPMOs is supported by other reports showing that these same genes provided highly productive targets for antisense- and siRNA-based efforts at inhibiting FLUAV replication [1, 17, 43]. PPMOs targeting other RNA viruses have been shown to be effective when delivered intraperitoneally [9, 15] or intravenously [48] as well, and it may become important to determine if routes of PPMO administration other than intranasal delivery are effective, as some FLUAV strains have been shown to produce systemic infection in humans [7, 12].

The testing of NP-v3′ and PB1-AUG PPMO here is intended to further their development as possible anti-FLUAV drugs. The H3N8 virus utilized in this study is of significance to the human and veterinary community. As mentioned previously, the H3N8 subtype was responsible for the influenza pandemic of 1890 [16]. Despite vaccination, there are frequent outbreaks of H3N8 virus each year in horses [28], costing the horse racing industry hundreds of millions of dollars [4]. In August 2007, an outbreak of H3N8 occurred in horses in Australia, a continent previously unaffected by equine influenza. The recent transmission of H3N8 from horses to domestic dogs [10, 47] is of notable concern as well, as both species are in frequent and close contact with humans. In light of the inexorable evolution of FLUAV, efficacy of a potential anti-FLUAV drug against a range of FLUAV subtypes is clearly desirable. This study suggests that further testing of PPMOs against a variety of strains in vivo is warranted.