Sie sind auf Seite 1von 9

Antiviral Chemistry & Chemotherapy 11:4149

Prevention of murine influenza A virus pneumonitis by surfactant nano-emulsions


Brian W Donovan1, Jon D Reuter1,2, Zhengyi Cao1, Andrzej Myc1, Kent J Johnson3 and James R Baker Jr1*
Center for Biologic Nanotechnology, Department of Internal Medicine, Division of Allergy; 2Unit for Laboratory Animal Medicine and 3Department of Pathology, University of Michigan, Ann Arbor, MI 48109 0648, USA *Corresponding author: Tel: +1 734 647 2777; Fax: + 1 734 936 2990; E-mail: jbakerjr@umich.edu These authors contributed equally to this manuscript
1

Non-ionic surfactant nano-emulsions have extensive anti-microbial activity and are biocompatible with skin and mucous membranes at effective concentrations. Two nano-emulsion formulations (8N8 and 20N10) made from soybean oil, tributyl phosphate and Triton X100, were tested for their ability to prevent murine influenza virus pneumonia in vivo. In the initial study, CD-1 mice were administered various dilutions of the nano-emulsions intranasally, and safe dosages and concentrations were determined. Non-toxic concentrations of the nano-emulsions were then mixed with influenza virus and applied to the nares of mice. Animals receiving mixtures of two different emulsions (8N8 or 20N10) and a LD50 of virus survived the challenge without evidence of viral infection. To determine if the nanoemulsions could prevent influenza virus infec-

tion in vivo when used as a prophylactic treatment, the nano-emulsions (8N8 at 1.0% and 20N10 at 1.0% or 0.2%) were applied to mouse nares 90 min before exposure to 510 5 p.f.u./ml virus by nebulized aerosol. Animals pretreated with the nano-emulsions had significantly decreased clinical signs of infection. Only 26.0% (8N8 at 1.0%), 31.25% (20N10 at 1.0%) and 37.0% (20N10 at 0.2%) of animals pretreated with nano-emulsion died from pneumonitis, whereas >80.0% of mock pretreated animals succumbed to infection (P<0.005). These findings suggest that nonionic surfactant nano-emulsions have therapeutic potential for the prevention of influenza virus infection in vivo. Keywords: non-ionic surfactant; nano-emulsion; mice; enveloped virus; influenza A virus

Introduction
Influenza virus is an enveloped virus and a highly conta-

gious respiratory pathogen responsible for severe pandemic disease (Mulder & Hers, 1972). The envelope glycoproteins, haemagglutinin (HA) and neuraminidase (NA), determine the antigenic specificity of influenza viral subtypes (Lamb & Krug, 1996). These glycoproteins mutate readily, allowing the virus to evade the inhibitory effects of neutralizing antibodies and thus allowing the virus to escape vaccine-induced immunity (Lamb & Krug, 1996; Schulze, 1997). Given these issues it is important to develop effective anti-influenza therapeutics that are broadly active across viral subtypes, yet non-toxic to mammalian cells. Non-ionic surfactant nano-emulsions are a unique approach to antimicrobial therapy. In contrast to broad spectrum liposome-based antimicrobials, such as povidine-iodine-encapsulated liposomes, that have been
2000 International Medical Press 0956-3202/00/$17.00

used for topical therapy (Reimer et al., 1997), nanoemulsions are not a vehicle for the delivery of an antimicrobial agent. The physical structure of the nanoemulsion contains surfactants and solvents that have antimicrobial activity. The surfactant activity specifically disrupts pathogenic microorganisms through fusion with the membrane of the microbe, leading to the rapid lysis of the targeted organism. This mechanism of action has been documented in preliminary studies examining the in vitro virucidal, bactericidal and sporicidal effects of these compounds (Hamouda et al., 1999). In addition, emulsifying the detergent and solvent markedly reduces the toxicity of these agents to eukaryotic cells at anti-microbial concentrations. Preliminary in vitro studies demonstrated the capability of nano-emulsions to inactivate influenza virus A and B, and prevent the infection of MadinDarby
1

BW Donovan et al.

canine kidney (MDCK) cells at concentrations as low as 1.0% (Myc et al., 1999). Kinetic studies using a plaque reduction assay have also demonstrated that nano-emulsions completely inactivate influenza virus (4.23 log10 reduction) at a concentration of 1.0% within 30 min. On the basis of this information, the present studies were designed to evaluate the toxicity of two non-phospholipid nano-emulsion formulations on murine nasal epithelium and pulmonary tissue. Once non-toxic concentrations were identified, the ability of these compounds to prevent influenza A virus infection when applied as prophylaxis was assessed in a murine influenza pneumonitis model previously developed for the testing of antiviral agents (Huang et al., 1991; Karaivanova & Spiro, 1998; Sidwell, 1999; Wyde et al. 1977).

accordingly with free access to food and water. The animals were anaesthetized with metofane in a bell jar and inoculated with 50 l total (25 l/nare) test material (PBS, 8N8 at concentrations of 10.0, 1.0 or 0.1% or 20N10 at a concentration of 1.0%). The mice were observed for abnormalities during recovery, and checked at least daily after inoculation over a period of 5 days. On day 5, all surviving mice were humanely euthanized with metofane and tissues were harvested. The lungs, trachea and nasal turbinates were inflated and fixed in 10.0% buffered formalin for histopathology. Nasal turbinates were demineralized for 48 h in Cal-Ex solution (Fischer Scientific, N.J., USA) before paraffin embedding. In vivo virus inhibition study Two models of predictable and consistent murine viral pneumonia were established. Once the models were verified, CD-1 mice were divided into groups of at least five animals to perform the studies on. In the first model, anaesthetized animals were administered 50 l (25 l/nare) intranasally of either PBS, 105 p.f.u./ml influenza A virus in PBS or 105 p.f.u./ml virus mixed with either 1.0, 0.2 or 0.1% 8N8 or 1.0, 0.2, 0.1 or 0.02% 20N10. The development of viral pneumonia was monitored for a 5 day period. In the second model, mice were anaesthetized and pretreated intranasally with 25 l (12.5 l/nare) of test material (8N8 at 1.0% or 20N10 at 1.0 or 0.2%). Control mice received PBS only. After 90 minutes, fully recovered animals were secured within a nebulization chamber and exposed to nebulized influenza virus at a concentration of 5105 p.f.u./ml for 5 min. The mouthpiece of a PARI LC Jet+ nebulizer (PARI Respiratory, Richmond, Va., USA) (delivering 0.32 ml/min with a particle size from 0.55 m) modified with a short section of clear tubing was inserted into one end of the nebulization chamber. A Proneb compressor was used to deliver the aerosol into the sealed chamber and an exhaust tube was run through a bleach bath. Owing to the confined space and the restricted movement of the animals within the tubes, the 5 minute exposure time was chosen to significantly reduce stress during the experimental procedure. Upon completion of the exposure period, mice were removed from the chamber and monitored daily for signs of illness over a period of 14 days. Animals in the first model developed appreciable clinical signs of infection within 3 days. In the second model, clinical signs of morbidity emerged 67 days after a 5 minute exposure to aerosolized influenza virus [5105 p.f.u./ml; mortality ratio (n=4/5) 80.0%]. In both disease models, rectal core body temperatures were recorded with a Model BAT-12 digital thermometer fitted with a RET-3 type T mouse rectal probe

Materials and Methods: Virology


H Maassab (School of Public Health, University of Michigan, Ann Arbor, Mich., USA) graciously provided influenza virus strain A/AA/6/60 (H2N2). This virus was passaged in mice and documented to cause significant mortality. The virus was propagated in chicken embryos using standard methods (Barrett & Inglis, 1985). Infectious allantoic fluid was aliquoted into samples containing ~107 p.f.u./ml, and stored at 80C until use.

Materials and Methods: Chemistry


Based on previous in vitro studies of antiviral activity (Myc et al., 1999), two formulations of nano-emulsion were prepared. An oil phase of tri-n-butyl phosphate, soybean oil and Triton X-100 was blended and heated at 86C for 1 h, then emulsified in water using a reciprocating syringe pump or commercial emulsifier (Silverson Machines, East Longmeadow, Mass., USA). One nanoemulsion formulation (8N8) contained 8.0% Triton X100 and 8.0% tri-n-butyl phosphate, whereas the second (20N10) was 20.0% Triton X-100 and 10.0% tri-n-butyl phosphate. Nano-emulsion particle sizes were measured using a Coulter LS 130 laser-sizing instrument. The size of nano-emulsions ranged from 223228 nm and 95% of particles had a diameter less than 226 nm. The 8N8 emulsion had a hydration ratio of 1:4 and pH 7.68, whereas the 20N10 emulsion had a hydration ratio of 41:9 and pH 7.69. All subsequent dilutions of the emulsions were made in PBS. Murine respiratory toxicity assay Four-to 5-week-old specific-pathogen-free CD-1 mice (Charles River, Portage, Mich., USA) of either sex were randomly assigned into groups of three and housed

2000 International Medical Press

In vivo antiviral activity of nano-emulsions

Table 1. Murine dose-dependent survival response of nano-emulsion/influenza virus mixtures


Body temperature (P value) 0.0036*
NS NS

Treatment groups Influenza A virus Influenza A virus and: 8N8 1.0% 8N8 0.2% 8N8 0.1% 20N10 1.0% 20N10 0.2% 20N10 0.1% 20N10 0.02%
NS,

Mortality Ratio 5/5 0/4 3/4 4/4 0/3 0/4 2/4 4/4

Mean survival time 4.2 5.0 (P<0.005) 4.75 (NS) 4.25 (NS) 5.0 (P<0.02) 5.0 (P<0.005) 5.0 (NS) 4.75 (NS)

Mean clinical sign score 3.0 0.25 (P<0.00001) 2.25 (P<0.02) 3.0 (NS) 0.0 (NA) 0.25 (P<0.00001) 1.75 (NS) 3.0 (NS)

Mean gross pathology score 3.2 0.5 (P<0.002) 2.25 (NS) 4.0 (NS) 0.0 (NA) 0.25 (P<0.002) 1.25 (NS) 3.0 (NS)

0.0027* 0.0021* 0.0312*


NS

not significant. not aplicable. *Significant change in body temperature is defined as a difference between the virus control group and nano-emulsion treated groups based on the mean temperatures found in Figure 2. An acute death during nano-emulsionvirus mixture instillation owing to obstruction of airway.
NA,

(Physitemp, Clifton, N.J., USA) (Wong et al., 1997). Temperature measurement time-points were carefully adhered to because of the natural diurnal fluctuation of body temperature in mice (Clement et al., 1989). Clinical signs of illness were graded on a scale of 03, where 0 indicated no significant clinical abnormality; 1 indicated mild symptoms including piloerection, hunched back and loss of movement; 2 indicated cyanosis, dyspnea, circulatory compromise, tachypnea and a rectal temperature <33C; and 3 indicated death of the animal. Mice with core body temperatures falling below 33C were judged to be terminally moribund and humanely euthanized. Gross pathology of the lungs was quantified using a standard technique (Scott & Sydiskis, 1976) with minor modification. Focal lung lesions without consolidation were given a score of 0, whereas a score of 1 represented 25% consolidation, a score of 2 50% consolidation and a score of 4 signified 100% pulmonary consolidation coinciding with death of the animal. The right lung lobes were aseptically removed, weighed and frozen at 80C for determination of virus presence via plaque assay. Identification of virus within the lung tissue The presence of influenza A virus in lung tissue was detected by plaque assay, as described elsewhere (Myc et al., 1999). The right lung was thawed at room temperature and homogenized in a 7 ml sterile tissue grinder (Wheaton Science Products, Millville, N.J., USA) in serum-free MEM. This contained 0.1 mM MEM nonessential amino acid solution, 2 mM L-glutamine, 1.0 mM MEM sodium pyruvate, 100 U/ml penicillin 100 g/ml streptomycin (GibcoBRL) with 3.0 g/ml TPCK-treated trypsin (TPCK) (Worthington Biochemical Corporation, Lakewood, N.J., USA) to

make a 10.0% (w/v) solution. Following homogenization, the sample was centrifuged at 20003000 r.p.m. for 10 min. The supernatant was serially diluted and plated on MDCK cells for plaque assay titration and viral presence was recorded. Statistics and data interpretation The means, standard deviation, standard error, and 2 analysis with Yates correction were calculated. To compare the control group to the study groups, Cox regression (Cox, 1972) was used. Difference between the study groups and the control group was tested using the log-likelihood ratio test.

Results
Toxicity When a 10.0% concentration of either the 8N8 or 20N10 nano-emulsion formulations was applied to mice nares, acute dyspnea, serosanguinous nasal discharge and cyanosis occurred in all animals and 4 of 5 mice died before anaesthetic recovery. Histopathology demonstrated marked nasal epithelial erosion and patchy necrosis of the bronchial mucosa with a focal pneumonitis. No residual lipid was observed in the lung, but the pulmonary macrophages appeared foamy and contained several clear vacuolated areas of presumed lipid phagocytosis. At a concentration of 1.0%, both the 8N8 and 20N10 formulations were well tolerated after direct intranasal instillation in mice. Mice recovering from anesthesia appeared healthy the following day and survived the 5 day observation period without problems. Upon euthanization at day 5, no clinical, gross or histopathologic changes were evident in mice receiving the 1.0% nano-emulsion pretreatment. There was also

Antiviral Chemistry & Chemotherapy 11:1

BW Donovan et al.

Figure 1. Photomicrographs of mouse lung tissue haemotoxylin and eosin stain

(a) Lung tissue of mouse given influenza virus mixed with 1.0% 8N8 showing intact alveoli and reduction in pneumonitis. (b) Lung tissue of mouse given influenza virus mixed with 0.2% 8N8 and (c) influenza virus mixed with 0.1% 8N8, each showing marked bronchointerstitial pneumonia and consolidation with a mixed mononuclear and neutrophilic inflammatory cell infiltrate. (d) Lung tissue of mouse given influenza virus mixed with 1.0% 20N10 and (e) virus mixed with 0.2% 20N10 showing intact alveoli and reduction in pneumonitis. (f) Lung tissue of mouse given influenza virus mixed with 0.1% 20N10 showing intact alveolar tissue with some inflammation. Magnification 20.

no histopathologic evidence of injury to the nasal cavity or lungs in these animals. Concurrent mixture of influenza A and nanoemulsion applied in vivo Based on toxicity data, nano-emulsion concentrations of 1.0% or lower were studied further. Animals were administered influenza virus mixed with either 8N8 (at a concentration of 1.0, 0.2 or 0.1%) or 20N10 (at a concentration of 1.0, 0.2, 0.1 or 0.02%). As shown in Table 1, significant mortality resulting from viral pneumonitis was seen in mice that received 8N8 at concentrations of 0.1% and 0.2% and mice that received 20N10 at concentrations of 0.1% and 0.02%. The lungs of infected mice were red, swollen and failed to collapse when the thoracic cavity was entered. Histopathology demonstrated a marked bronchointerstitial pneumonia and consolidation with a mixed mononuclear and neutrophilic inflammatory cell infiltrate (Figure 1b, c and f ). There were no deaths related to the development of pneumonitis in groups of mice receiving influenza virus mixed with 1.0% 8N8, 1.0% 20N10 or 0.2% 20N10 within the 5 day observation period (Figure 1a, d and e). This survival was dose-dependent as declining concentrations of nano-emulsion mixed with virus did not prevent pneumonitis or death (Figure 2).

Nano-emulsion prophylaxis of murine pneumonitis caused by inhalation of influenza virions Pretreatment of mouse nares with nano-emulsion followed by exposure to aerosolized influenza virus was performed to determine the ability of these compounds to protect mice against infection from inhaled virus particles. As the survival curve in Figure 3 shows, both the 8N8 (1.0%) and 20N10 (1.0 and 0.2%) nano-emulsions administered intranasally 90 min before exposure were significant prophylactic barriers to influenza infection. Cox regression model analysis revealed that there was no significant difference in survival between the three emulsion pretreatment groups. The difference between study groups compared to the control group was significant (P<0.005). Control mice, mock pretreated with an intranasal application of PBS, remained unprotected from a 5 min exposure to nebulized influenza A virus at a concentration of 5105 p.f.u./ml. This concentration killed 81.25% (n=13/16) of control mice. In contrast, only 31.25% (n=5/16) of mice pretreated with 1.0% concentration of 20N10 died, whereas 37.5% (n= 6/16) of mice treated with a 0.2% concentration died (Table 2). 8N8 also showed significant ability to protect mice from influenza infection. Only 26% (n=4/15, 1 mouse did not recover from anesthesia) of mice died when pretreated

2000 International Medical Press

In vivo antiviral activity of nano-emulsions

Figure 2. Relationship of core body temperature and survival time after intranasal application of either influenza virus or concurrent mixture of influenza virus and nano-emulsion

Figure 3. Survival curve of mice pretreated with nano-emulsions applied as prophylactic barriers to influenza A infection and mice mock pretreated with PBS

Individual animals with core body temperatures <33.0C were humanely euthanized.

with 8N8 at a concentration of 1.0% (Table 2). The core body temperature in mice pretreated with either nanoemulsion did not drop below 33C, however, this was commonly observed in the control group mice (Figure 4). Surviving mice pretreated with nano-emulsion were

Cox regression model analysis revealed that there was no significant difference in survival between the three study groups. The overall difference between study groups compared to control group was significant (P<0.005). Individual animals with core body temperatures <33.0C were humanly euthanized.

free of influenza virions on day 14. In contrast, in either control mice or pretreated mice that died from pneumonitis, influenza virus was present in the lungs. The

Table 2. Individual mouse survival times and clinical response to aerosolized influenza virus following prophylactic application of nano-emulsion
Treatment groups Mock pretreatment Ear tag No 1676 14 1677 10 1678 9 1679 9 1680 7 1681 7 1682 7 1683 7 1684 14 1685 8 1686 14 1687 8 1688 10 1689 10 1690 8 1691 6 Mean survival time 9.3 Mortality ratio 13/16 Mortality (%) 81.2 Scoring Mean clinical sign 2.6 Mean gross pathology 3.5
ND,

1.0% 20N10 1612 1613 1614 1615 1616 1637 1638 1639 1652 1653 1654 1655 1656 1657 1658 1659 14 9 14 14 14 14 14 14 14 9 14 10 14 9 9 14 12.5 (P<0.001) 5/16 (P<0.02) 31.2 1.2 (P<0.02) 1.3 (P<0.005) 1617 1618 1619 1620 1621 1642 1643 1644 1660 1661 1662 1663 1664 1665 1666 1667

0.2% 20N10 Ear tag No Survival (days) 14 14 14 14 10 10 14 14 14 9 8 8 9 14 14 14 12.125 (P<0.001) 6/16 (P<0.05) 37.5 1.2 (P<0.02) 1.5 (P<0.01)

1.0% 8N8 Ear tag No 1622 1623 1624 1625 1626 1647 1648 1649 1668 1669 1670 1671 1672 1673 1674 1675 Survival (days) 14 14 14
ND

Survival (days) Ear tag No Survival (days)

14 10 10 14 9 8 14 14 14 14 14 14 12.7 (P<0.001) 4/15 (P<0.01) 26.0 0.9 (P<0.005) 1.3 (P<0.005)

no data, this mouse died while under anaesthesia.

Antiviral Chemistry & Chemotherapy 11:1

BW Donovan et al.

Figure 4. Measurement of hypothermic response in mice

Mice were anesthetized and pretreated intranasally with 25 l (12.5 l/nare) of test material (8N8 at 1.0% and 20N10 at 1.0 and 0.2%; control mice received PBS only) then exposed to nebulized influenza virus. All mice were monitored daily for signs of illness over a period of 14 days.

lungs of control mice pretreated with PBS also show degradation and necrosis of alveolar epithelia and neutrophilic inflammatory cell infiltrate indicative of pneumonitis (Figure 5a). In contrast, the lungs of emulsion pretreated mice surviving the 14 day observation period show only minimal inflammation with intact alveolar epithelia and an absence of obvious degradation and necrosis (Figure 5b, c, d).

Discussion
Nano-emulsions offer a unique approach to the treatment of microbial pathogens. In contrast to liposomeencapsulated antimicrobials, which have been used for topical therapy (Reimer et al., 1997), nano-emulsions do not serve as a vehicle for the delivery of an antimicrobial compound. It is the physical structure of the nanoemulsion itself that is the active agent. These compounds contain surfactants that specifically disrupt pathogenic microorganisms through membrane fusion and lysis of the targeted organism. In addition, emulsifying the detergent in oil significantly reduces toxicity to human and animal skin, mucous membranes and gastrointestinal tissues at antimicrobial concentrations. Nano-emulsions have the added advantages of being chemically stable, resistant to heat and adaptable for varying clinical and environmental applications. Because the compounds are predominately non-phospholipid-based, rapid oxidization of the lipid does not

occur. The formulations are also relatively inexpensive to produce on an industrial scale (Azmin et al., 1985; Baille et al., 1986; Rogerson et al., 1988). Previous preliminary studies documented that a 30 min incubation period of 8N8 at a concentration of 1.0% with influenza virus and several other enveloped viruses yielded greater than a 4 log10 reduction in viral particles in vitro. This suggested that these nano-emulsions could be used to inhibit infection caused by pathogenic enveloped viruses and provided the impetus for testing as prophylactic barriers to enveloped virus infection in vivo. Intranasal application of these compounds and subsequent exposure to airborne influenza virions confirmed and expanded our in vitro findings. A significant proportion of mice pretreated with either of the nano-emulsion formulations survived exposure to a lethal dose of aerosolized influenza virus. The protection provided by the compounds lasted at least 90 min post-treatment, indicating that nano-emulsions sustained the ability to inhibit viral infection in vivo. These findings suggest the nano-emulsion may form a protective layer on the upper respiratory tract epithelia that prevents viral adhesion and simultaneously destroys free virus. The concept of a protective layer formed by the nano-emulsion is supported by both the lack of observable clinical signs indicative of influenza infection and the absence of inflammation in histopathological analysis of airway tissue from animals pretreated with nano-emulsion. Thus, the nanoemulsion appears to either prevent or attenuate the viral

2000 International Medical Press

In vivo antiviral activity of nano-emulsions

Figure 5. Lung tissue of mice pretreated intranasally with nano-emulsions applied as prophylaxis and control mice pretreated with PBS

(a) Control mouse showing degradation and necrosis of alveolar epithelia and neutrophilic inflammatory cell infiltrate. This mouse survived 6 days after exposure to nebulized influenza (number 1691). Mice pretreated with nano-emulsion show mild inflammation with intact alveolar epithelia with an absence of obvious degradation and necrosis [(b) 1.0% 8N8 (number 1649); (c) 1.0% 20N10 (number 1615); (d) 0.2% 20N10 (number 1620)]. These mice survived the entire 14 day observation period. Haemotoxylin and eosin stain. Magnification 40.

infection, resulting in survival of the animals. The issue of prevention versus attenuation is clarified in the body temperature curves of mice exposed to virus in the model two (nebulization chamber) studies. A consistent pattern seen in the temperature curves of these mice shows a decrease in core body temperature at 2 days post-exposure. This pattern was observed in all mice placed inside the nebulization chamber regardless of virus exposure. This early loss of temperature appeared to occur because of stress connected with experimental procedure and mice regained normal core temperatures by day 4 or 5. However, a distinct pattern was observed in infected mice at 6 to 7 days post-exposure. This latter temperature loss appeared to be the result of viral infection because the loss of temperature coincided with the emergence of clinical signs indicative of influenza infection. Most control mice were unable to

recover and exhibited a continued decline in core body temperatures until death or euthanization occurred. Pretreated mice exhibited a similar decrease in core body temperature, but the mean temperature never reached 33C and clinical signs were minimal. Temperature curves of pretreated mice uniformly returned to normal levels by the completion of the observation period at day 14. Overall, the decrease in body temperatures correlated well with other clinical signs of progressive viral infection in mice. Thus, it appeared that the pretreated mice had attenuated viral infections rather than absolute protection against infection. This is not surprising given the animals were only pretreated on one occasion and the emulsion probably dissipated before the virus was completely destroyed. Ongoing administration of the emulsion both before and after exposure may totally prevent infection by continuing to disrupt virus, and this

Antiviral Chemistry & Chemotherapy 11:1

BW Donovan et al.

type of protocol needs to be examined. Unfortunately, because of the need to anaesthetize the animals this is impossible to accomplish in mice. The therapeutic index observed in these studies for the nano-emulsions is relatively narrow. Initial toxicity testing of the nano-emulsion formulations by installation of the compound to the nasal cavities of mice indicated that a 1.0% concentration of each nano-emulsion was well tolerated in experimental animals. However, to prevent viral pneumonitis, 8N8 requires a concentration of 1.0%, which is only 10 times lower than the concentration that causes mammalian toxicity. In contrast, the 20N10 formulation, which demonstrated a similar toxicity to 8N8, effectively inhibited influenza virus at concentrations down to 0.1%, yielding a broader therapeutic index (50100 fold). This suggests that altering the mix of the components can develop improved formulations of the nano-emulsions with increased safety. It is also significant to note that current studies may overstate the toxicity of nano-emulsions. This is mainly because much of the nano-emulsion directly entered the lungs of anaesthetized animals. The immediate effect of aspirated oil on the lung is local haemorrhage and oedema from the obstruction of alveoli (Ikeda, 1935). This phenomenon may explain most of the acute respiratory compromise and death observed in mice after receiving the nano-emulsion at concentrations of 10.0%. Lipid introduced into the lung in humans causes pneumonia and severe respiratory compromise, requiring aggressive and prolonged treatment. The type of oil and the amount of free fatty acid both determine the specific reaction to aspirated lipid (Annobil et al, 1997). The unique composition of the nano-emulsion may explain the absence of lipid pneumonia in these mice, despite spillage of the material into the airways. The nano-emulsions employ soybean oil, which has a lower potential for causing lipoid pneumonia and pulmonary fibrosis than other oils which are known to cause severe pulmonary inflammation (mineral oil, animal-derived oils and those vegetable oils high in free fatty acids) (Ikeda, 1935; Pinkerton, 1927, 1928). In these studies, no residual compound was observed in the lung 5 days post-inoculation, although macrophages appeared to be vacuolated and may have contained the phagocytized lipid. This information suggests that an intranasal application of nano-emulsion at concentrations lower than 1.0% should have safety comparable to other respiratory therapeutics (Annobil et al., 1997; Ikeda, 1935; Pinkerton, 1927). The application of these results to the prevention of natural influenza virus infections in humans is not clear. The murine model of pneumonitis is well accepted for testing antiviral agents, however there may be differ-

ences between these models and natural infection in humans. In mice, influenza infection has been reported to initially develop in the alveoli and spread to the upper respiratory tract (Hers et al., 1962). In contrast, the infection in humans starts in the lower respiratory tract and descends to the lung parenchyma (Webster, 1994). Application of nano-emulsions to mice under general anaesthesia allowed the compounds to reach the lungs because of a loss of mucociliary clearance and epiglottal function. It is presumed that nano-emulsion pretreatment resulted in local inactivation of influenza virus in these mice and with administration under anaesthesia this could have occurred in either the upper or lower airway. Thus, the capability of these nano-emulsions to prevent a natural influenza infection when applied only to the upper airway mucosa in humans cannot be determined using this model. It is interesting to postulate that only the upper airway needs to be treated in humans because it might prevent contamination at the initial site of influenza virus replication in the lower airway. However, confirmation of this will require clinical studies of naturally acquired respiratory virus infections in humans or primates. In conclusion, initial in vivo studies with surfactant nano-emulsions suggest that these compounds offer a novel approach to preventing respiratory influenza infections. Further formulation evolution and the testing of efficacy in a clinical model of exposure that mimics human disease are warranted.

Acknowledgements
This work was supported by DARPA (Defense Advanced Research Project Agency) contract #MDA 972-1-007 of the Unconventional Pathogen Countermeasures Program and National Institutes of Health (NIH) Training grant #RR0 7008-21. The authors wish to express thanks to H Maassab for his generous donation of influenza virus and AT Galecki for assistance with the statistical analysis of our data.

References
Annobil SH, Tahir ME, Kameswaran M & Morad N (1997) Olive oil aspiration pneumonia (lipoid) in children. Tropical Medicine and International Health 2:383388. Azmin MN, Florence AT, Handjani-vila RM, Stuart JFB, Vanlerberghe G & Whittaker JS (1985) The effect of non-ionic surfactant vesicle (niosome) entrapment on the absorption and distribution of methotrexate in mice. Journal of Pharmacy and Pharmacology 37:237242. Baillie AJ, Coombs GH, Dolan TF & Laurie J (1986) Non-ionic surfactant vesicles, niosomes, as a delivery system for the antileishmanial drug, sodium stibogluconate. Journal of Pharmacy and Pharmacology 38:502505.

2000 International Medical Press

In vivo antiviral activity of nano-emulsions

Barrett T & Inglis SC (1985) Growth purification and titration of influenza viruses. In Virology: a Practical Approach, pp. 119151. Edited by WJ Mahy. Washington: IRL Press. Clement JG, Mills P & Brockway B (1989) Use of telemetry to record body temperature and activity in mice. Journal of Pharmacological Methods 21:129140. Cox DR (1972) Regression models and life-tables (with Discussion). Journal of the Royal Statistical Society. Series B, 34:187220. Hamouda T, Hayes M, Cao Z, Tonda R, Johnson K, Wright D, Brisker J & Baker JR Jr (1999) A novel surfactant nanoemulsion with broad-spectrum sporicidal activity against Bacillus species. Journal of Infectious Diseases 180:19391949. Hers JFP, Mulder J, Masurel N, Kuip LV & Tyrrelln DAJ (1962) Studies on the pathogenesis of influenza virus pneumonia in mice. Journal of Pathology and Bacteriology 83:207217. Huang R, Dietsch E, Lockoff O, Schuller M & Reutter W (1991) Antiviral activity of some natural and synthetic sugar analogues. FEBS Letters 291:199202. Ikeda K (1935) Oil aspiration pneumonia (lipoid pneumonia): clinical, pathological and experimental considerations. American Journal of Diseases of Children 49:9851006. Karaivanova VK & Spiro RG (1998) Sulphation of N-linked oligosaccharides of vesicular stomatitis and influenza virus envelope glycoprotiens: host cell specificity, subcellular localization and identification of substituted saccharides. Biochemical Journal 329:511518. Lamb RA & Krug RM (1996) Orthomyxoviridae: the viruses and their replication. In Fields Virology, third edition, pp. 13531395. Edited by N Fields, DM Knipe & PM Howley. Philadelphia: Lippincott-Raven. Mulder J & Ph Hers JF (1972) Influenza. Groningen: WolterNoordhoff Publishing. Myc A, Anderson MJ & Baker JR Jr (1999) Optimization of in situ cellular ELISA performed on influenza A virus-infected monolay-

ers for screening of anti-viral agents. Journal of Virological Methods 77:165177. Pinkerton H (1927) Oils and fats: their entrance onto and fate in the lungs of infants and children. American Journal of Diseases of Children 33:259285. Pinkerton H (1928) The reaction of oils and fats in the lung. Archives of Pathology 5:380401. Reimer K, Fleischer W, Brgmann B, Schreier H, Burkhard P, Lanzendrfer A, Gmbel H, Hoekstra H & Behrens-Baumann W (1997) Povidine-iodine liposomes an overview. Dermatology 195 (Suppl. 2):9399. Rogerson A, Cummings J, Willmott N & Florence AT (1988) The distribution of doxorubicin in mice following administration in niosomes. Journal of Pharmacy and Pharmacology 40:337342. Schulze IT (1997) Effects of glycosylation on the properites and functions on influenza virus hemagglutinin. Journal of Infectious Diseases 176 (Suppl. 1):S24S28. Scott GH & Sydiskis RJ (1976) Responses of mice immunized with influenza virus by aerosol and parenteral routes. Infection and Immunity 13:696703. Sidwell RW (1999) The mouse model of influenza virus infection. In Handbook of Animals Models of Infection: Experimental Models in Antimicrobial Chemotherapy, pp. 981987. San Diego: Academic Press. Webster RG (1994) Influenza viruses, general features. In Encyclopedia of Virology, pp. 709715. Edited by RG Webster & A Granoff. London, San Diego: Academic Press. Wong JP, Saravolac EG, Clement JG & Nagata LP (1997) Development of a murine hypothermia model for study of respiratory tract influenza virus infection. Laboratory Animal Science 47:143147. Wyde PR, Couch RB, Mackler BF, Cate TR & Levy BM (1977) Effects of low- and high-passage influenza virus infection in normal and nude mice. Infection and Immunity 15:221229.

Received 19 April 1999; accepted 1 November 1999

Antiviral Chemistry & Chemotherapy 11:1