Enveloped Virus Inactivation on Personal Protective Equipment by Exposure to Ozone

Assay Validation

In the literature, virus survival post-disinfection is typically measured by determining viral titer by plaque assays or TCID₅₀ (median tissue culture infectious dose) on permissive cells, or by measuring viral RNA loads via RT-qPCR.^18,25 We sought to maintain viral infectivity as the primary readout, but to do so in a simpler and faster way. Accordingly, we employed the reporter strains WSN/33-PA-2A-NLuc and RSV-luc5, replication-competent influenza A virus and respiratory syncytial virus, respectively, modified to have a cleavable NanoLuc (NLuc) reporter on the PA protein (IAV) and a luciferase between the P and M genes (RSV).^26,27 Overnight passage of any remaining virus after ozone treatment on MDCK cells (for IAV) or HEp-2 cells (for RSV) followed by cell lysis and NLuc or luciferase signal quantification provided a fast and convenient readout of viral protein load, and thereby of viral replication.

Testing with known concentrations of IAV and RSV demonstrated the dynamic range of the NLuc reporter assay (factors of 10⁶ or greater in virus dilution) to exceed those of RT-qPCR and plaque assays (10⁴-fold range) in our hands (Figure 1a-c, f-g). Asterisks in Fig. 1b indicate dilutions where large cytopathic effects were observed, such that plaques were uncountable. We also treated representative IAV samples with ozone in a standard manner (see below), and analyzed half of the samples by RT-qPCR and the other half with the NLuc reporter assay. As shown in Fig. 1d-e, the observed trends were similar; these results with a small number of samples in each group also illustrate the general observation that the NLuc assay provided somewhat smaller variance among replicates.

• Download figure
• Open in new tab

Figure 1.

Comparison of assay methods for virus viability. Variation in IAV concentration assayed by (a) RT-qPCR, (b) plaque assay, (c) NanoLuc. (d,e) Treatment of IAV on N95 material with ozone, assayed by (d) NanoLuc and (e) RT-qPCR. Variation in RSV concentration assayed by (f) RT-qPCR, (g) Luc.

Effects of exposure time, humidity, and ozone concentration on virus inactivation

We selected three types of materials in common use in healthcare environments for testing of virus deactivation by ozone: cloth face masks, Tyvek (spun high-density polyethylene) fabric used in disposable gowns and PAPR (powered air purifying respirator) hoods, and N95 respirators (Supplemental Information, Table S1). In each case, a measured amount of virus was applied to small (1 cm square) pieces of material, allowed to dry, and then subjected to disinfection. Any remaining virus was recovered by immersion of the treated material in growth media, followed by passage through susceptible cells in culture. For multi-layered materials such as the N95 respirators, the virus solution did not remain on the surface layer, but rather were absorbed into the material. For negative disinfection controls, the same types of samples were treated with virus, left to dry, and were kept at room temperature without treatment. Virus-laden samples sprayed with 70% ethanol were allowed to dry and served as positive disinfection controls. Minimum assay response was defined by the signal generated by processing of materials not treated with any virus. In each experiment, 4-6 replicates of each sample type were used. To account for experiment-to-experiment variation in absolute values, results are reported as % reduction vs. ethanol treatment, which consistently provided more than 10⁴-fold reduction in viral infectivity.

Ozone exposure was performed using three commercial instruments. Two (manufactured and provided by Global Ozone Innovations and Zono Technologies) were cabinets used primarily to disinfect sports and playground equipment in schools, daycare centers, and the like. The third was a disinfection device for home PAP (positive airway pressure) equipment (VirtuCLEAN, provided by Healthcare Logiix Systems). These devices were obtained by virtue of personal contacts with company representatives, not because of any particular technical features, and we believe they are representative of the types of equipment that can be widely obtained. Ozone and humidity were monitored with built-in or externally-introduced sensors. Hardware details are given in Tables S2 and S3; Figure S1 shows representative time courses of changes in ozone and humidity levels through a typical run. One of the instruments (made by Zono Technologies) was able to simultaneously control ozone concentration, humidity, and temperature, and so this was used for most of the experiments investigating changes in these parameters.

A standard ozone concentration of approximately 20 ppm was produced by all of the units. Under humid conditions (80% relative humidity, RH, at room temperature), effective inactivation of influenza virus (>70% of the ethanol control) was achieved even at the shortest exposure time tested (18 min), with a tendency to increase at longer times (up to 90 min), as shown in Figure 2a. In contrast, when the humidity was not controlled, disinfection was much more variable and required up to 4 h of ozone treatment to achieve results comparable to 70% ethanol for all three materials tested (Fig. 2b).

• Download figure
• Open in new tab

Figure 2.

(a-f) Inactivation of influenza A virus by ozone under various conditions. Vertical axis = decrease in IAV infectivity relative to the results obtained from saturation of virus-deposited materials with 70% ethyl alcohol relative to samples receiving no disinfection treatment. (e) 80% RH at 24 °C = approx. 17.4 g/m³ water vapor; 53% RH at 48°C = approx. 40 g/m³. (g) Inactivation of RSV by ozone. RH = relative humidity. All experiments were performed at room temperature (24 ± 1°C) unless otherwise indicated. 2way ANOVAs with Tukey’s multiple comparison tests were performed for the data in panels a, c-e, and g. A 2way ANOVA with a Tukey’s multiple comparison test was run for panel b, except for N95 samples, for which a t-test was performed. A Brown-Forsythe and Welch ANOVA with a Dunnet’s T3 multiple comparisons test was run for panel f. ^* p < 0.02, ^** p < 0.01, ^*** p < 0.001, and ^**** p < 0.0001.

The results of a series of experiments controlling relative humidity (Fig. 2c) demonstrated the critical nature of this parameter. A clear threshold was observed for all three materials, with 40% RH being ineffective and 50-70% RH giving maximum performance. These results are quite similar to those reported recently by Duchaine and colleagues on the deactivation of aerosolized nonenveloped viruses (phage and norovirus examples).²¹ Increasing to 80% RH provided no significant additional benefit. Increasing ozone concentration from 20 ppm to 50 ppm gave a predictably better response, but the lower concentration was nevertheless quite effective (Fig. 2d). Higher temperature (48°C vs. room temperature) under humid conditions allowed for significant reduction of IAV infectivity on both Tyvek and N95 material in only 5 minutes, but this of course is less convenient to perform in practice (Fig. 2e). In this experiment, a higher water vapor concentration was maintained at the higher temperature even though the relative humidity value was lower. In contrast, increasing the temperature to 45°C while maintaining humidity at a relatively low level did not significantly enhance the rate of virus inactivation by ozone (Figure S2).

Extended (90 min) ozone treatment under humid conditions also successfully inactivated IAV on samples of Tyvek from a standard disposable coverall (“bunny suit”) and a hospital PAPR hood, as well as the polycarbonate PAPR face shield (Fig. 2f). Lastly, RSV was completely deactivated by ozone under the same conditions, after either 40 or 90 minutes exposure (Fig. 2g). Therefore, if IAV and RSV are representative, these results suggest that ozone treatment at 20 ppm or greater, 70% or greater relative humidity at room temperature, and for at least 40 minutes should reliably inactivate enveloped viruses on a variety of materials of use in medical PPE.

Effects of ozone on materials properties

We exposed larger swaths of Tyvek fabric from various sources, as well as cloth masks, N95 respirator material, and polycarbonate face shields to multiple cycles of ozone exposure (20 ppm) adding up to two hours or more. With one important exception, no perceptible changes occurred in any material, including goggle transparency, the integrity of seams and glued components, and the appearance and function valve assemblies and other plastic components (such as zippers). While we did not perform quantitative fit testing of the respirators post-treatment, no deformations were observed following any of the ozone treatment variations. A summary of mechanical assessments is provided in Table S4. Quantitative rheological measurements were not performed.

Only the elastic headbands of some N95 respirators were found to be degraded by ozone treatment, and only when the material was strained during exposure, either by function of a stapled attachment to the respirator or by tying off in a stretched state in the ozone cabinet (Figure S3). The degradation of strained rubber upon exposure to ozone has been previously described,²⁸ presumably due to scission of C=C double bonds in the polymers which results in cracking orthogonal to the strain. We found that even a short exposure of strained elastics to 20 ppm O₃ (15 minutes, less than generally required for viral inactivation) often resulted in compromised headband integrity or complete failure.

A subset of NIOSH-approved N95 respirators cleared by the Food and Drug Administration (FDA) are referred to as Surgical N95 Respirators.²⁹ While we were unable to obtain samples of these FDA-approved respirators, many are manufactured with stapled headband attachments. This presents a significant practical barrier to the use of ozone to disinfect N95 respirators. However, headbands made of thermoplastic elastomer or polyisoprene with polypropylene overbraid can be attached by ultrasonic welding (such as the 3M 8210 and ZKG9501 KN95 respirators that we were able to obtain for this study). These headbands showed no degradation upon several hours of repeated ozone treatments (Figure S4), and so we surmise that welded attachments do not induce significant elastomer strain at the point of attachment.

N95 particle filtration efficiency

Full-thickness sections of N95 respirator material before and after extensive ozone exposure were tested for filtration ability using sodium chloride aerosols in a manner closely analogous to the NIOSH N95 respirator penetration testing guidelines.³⁰ No substantial loss of filtration performance was found for these ozone-treated materials (Figure S5, Table S5).

Description of fabric sample swatches

Fabric samples for viral inoculation were prepared by cutting 1 cm × 1 cm sample swatches from four different materials: surgical face masks, Tyvek (disposable gown), N95 respirators, bunny suits, and PAPR hoods (both the fabric and plastic face shield).

Cell lines and Viruses

MDCK (ATCC) and HEp-2 (ATCC) cells were maintained in DMEM + 10% fetal bovine serum and 1% pen-strep at 5% CO₂ and 37°C. H1N1 Influenza A WSN/33-PA-2A-NLuc (BEI) and RSV-luc5 (Viratree) reporter strains were obtained from the indicated suppliers, and designated here as “IAV” and “RSV,” respectively.

Viral stocks were prepared as follows. IAV was propagated in MDCK cells, where virus was added to cells at a multiplicity of infection (MOI) of 0.01 for 45 minutes before being removed and replaced with virus growth medium (DMEM with 1% pen-strep, 2 mM L-glutamine, 0.2% bovine serum albumin, and 25 mM HEPES buffer). Virus was harvested when 90% of the cytopathic effects were visualized. Media was collected, spun down at 300xg for 15 minutes to remove cell debris and supernatant was aliquoted. RSV was propagated in HEp-2 cells, where virus was added to cells at a MOI of 0.1 for 1 h before adding virus growth medium (DMEM with 10% fetal bovine serum and 1% pen-strep) to the inoculum. Cell-associated virus was harvested by scraping the cells when 90% of cytopathic effects were visualized. Virus-containing media was then vortexed and aliquoted for use.

Sample preparation and disinfection

Virus solutions (50 μL) in growth media at ~10⁵ PFU/ml were deposited by pipette onto 1 cm × 1 cm pieces of each candidate material, as evenly as possible over at least half of the surface area. The samples were allowed to dry in air at room temperature for 30 minutes and were transported to the disinfection unit in plastic cases. For initial treatment experiments, samples were left inside open cases during ozone exposure. To better approximate the treatment of actual intact PPE garments and materials, most subsequent samples were pinned to cotton shirts and hung in the ozone cabinets so as to allow air/ozone access to both sides of the sample swatch. For the small zippered pouch of the VirtuCLEAN PAP cleaning device, the samples were pinned to an N95 respirator placed inside the pouch.

No-disinfection control samples remained at room temperature dry and untreated, while ethanol controls were sprayed with 70% ethanol and then allowed to dry and remain at room temperature for the duration of the disinfection procedure. After treatment (or control non-treatment), fabric samples were immersed in 300 μL of virus growth media (for IAV, DMEM with 1% pen-strep, 2 mM L-glutamine, 0.2% bovine serum albumin, and 25 mM HEPES buffer; for RSV, DMEM with 10% fetal bovine serum and 1% pen-strep) for 30 minutes at room temperature in the top of a 0.65 μm filter. Each sample was then spun at 12000xg for 2 minutes to extract virus/media, designated as “virus inoculum.” For virus applied to hard plastic surfaces, the surface was swabbed with a sterile cotton swab dipped in virus growth media. The cotton swab was then incubated in 300 μL of virus growth media for 30 minutes at room temperature, followed by removal of the swab and centrifugation at 12000×g for 2 minutes. MDCK cells (IAV) or HEp-2 cells (RSV) plated in a 24 well plate at confluency were washed once with 1×PBS and then incubated with 100 μL of virus inoculum for 1 hour at 37°C, being rocked every 15 minutes. After 1 hour, the inoculum was removed and the cells were washed once with 1×PBS. 500 μL of virus growth media was added to cells, which were then incubated at 37°C overnight before analysis.

Plaque Assay

MDCK cells were plated in 6-well plates until confluent. Cells were washed twice and incubated with 200 μL of virus inoculum, diluted in DMEM + 0.1% trypsin-TPCK, for 1 h at 37°C, being rocked every 15 minutes. The inoculum solution was then removed, the cells were washed once, and then were incubated at 37°C with an Avicel-DMEM overlay media containing a 1:1 ratio of 2.4% Avicel (FMC Biopolymer RC-481) and DMEM, supplemented with 0.2% BSA, 1% pen-strep, 2 mM L-glutamine, 0.1% trypsin-TPCK, and 25 mM HEPES buffer. After 3 days, the overlay was removed and cells were fixed in 4% paraformaldehyde for 30 minutes, before being stained with crystal violet (0.1% in 20% ethanol) for 30 minutes. After the crystal violet was removed, the monolayer was air dried and plaques were counted.

RT-qPCR

The supernatant was removed and the cells were washed and then processed with an RNAeasy PLUS RNA extraction kit (Qiagen) to extract RNA. cDNA was then generated from isolated RNA with a High-Capacity cDNA Reverse Transcription Kit (ThermoFisher). RT-qPCR was then performed for IAV PB1, with Taqman probe (ACACGAGTGGACAAGCTGACACAA) and primers (ATCTTTGAGACCTCGTGTCTTG and CAGCAGGCTGGTTCCTATTTA) or for RSV F, with Taqman probe (TGCCATAGCATGACACAATGGCTCCT) and primers (AACAGATGTAAGCAGCTCCGTTATC and CGATTTTTATTGGATGCTGTACATTT) (ThermoFisher).

NLuc Reporter Assay

The cell media was removed and the cells were washed once with 1×PBS. 250 uL of Nano-Glo Luciferase Assay Reagent (Promega) (1:100 ratio of Nano-Glo Luciferase Assay Substrate:Nano-Glo Luciferase Assay Buffer) was added per well and incubated for 3 minutes at room temperature. 100 uL of Nano-Glo Luciferase Assay Reagent was pipetted from each well as technical duplicates into a white 96 well plate and luminescence was quantified.

Luciferase Reporter Assay

Supernatants were removed and cells were washed once with 1×PBS. 250 ul of Bright-Glo Luciferase Assay Reagent (Promega) was added per well and incubated for 3 minutes at room temperature. 100 ul of Bright-Glo Luciferase Assay Reagent was pipetted from each well as technical duplicates into a white 96 well plate and luminescence was quantified.

Mechanical test methods

We treated various pieces of equipment for multiple cycles to assess mechanical tolerance of PPE to ozone treatment compared to control items. Items tested are listed in Table S4.

Particle filtration efficiency testing methods

Seven different respirators were assessed at 20 ppm ozone, with minimum treatment durations ranging from 160 to 320 min. Sample material from each of the respirators that underwent mechanical tests were used to assess post-treatment particle filtration efficiency. Circular samples of 2.5 cm diameter were punched from each respirator and placed into a filter holder. Aerosol size distribution and number concentration with and without the sample were measured by a Scanning Mobility Particle Sizer (SMPS) system⁵⁰ to determine particle filtration efficiency. The SMPS system consists of a Differential Mobility Analyzer (TSI 3080) and a Condensation Particle Counter (TSI 3775). Polydispersed aerosols were generated via atomization of 0.05 M sodium chloride solution. Experiments were conducted at 1.92 LPM. Two sets of testing were performed per each sample and averaged. We note that these tests were not completely equivalent to the NIOSH standard, and differences include but not limited to the following:

• Tested respirators were not pre-conditioned at 38°C and 85% relative humidity

• Aerosol mass concentration was substantially lower in this test

• Aerosol size distribution was skewed toward smaller size compared to NIOSH standard.

Ozone treatment methods

From April 2 to April 24, 2020, we assessed four different ozone treatment devices from three manufactures, Global Ozone Innovations (GOI), Zono Technologies (ZT), and VirtuCLEAN (VC) (Table S2). These devices are designed to disinfect recreational or medical equipment, and have been largely evaluated against bacterial pathogens,⁵¹ but few studies have assessed the efficacy of these technologies against encapsulated viruses. The GOI and VC units rely on ozone alone, whereas the ZT cabinet also treats with elevated (80%) humidity via a water reservoir and ultrasonic humidifier. The GOI and VC devices generate ozone via corona discharge and the ZT cabinet, which additionally uses elevated humidity, uses 184 nm photolytic UV ozone generators because humidity reduces corona plate generation efficiency. When not otherwise noted, all experimental runs were performed at ambient temperature (ca. 24-25°C) and pressure.

All the devices tested use atmospheric oxygen to generate a nominal ozone concentration of approximately 20 ppm. Whereas the GOI and ZT cabinets control this by virtue of internal sensing and real-time regulation, the less expensive VC cleaner turns ozone generation on and off to give rise to cycles of ozone concentrations ranging from 10-30 ppm, averaging 17-20 ppm (Figure S1). To extend runs beyond the time limits imposed by the manufacturer’s hard-coded time limits, the VC device was simply restarted immediately at the end of each treatment cycle. Both GOI and ZT provided extensive remote or on-site engineering support to permit modification of time and concentration set points in support of these experiments.

The Zono Technologies cabinet and VirtuCLEAN device do not offer data logging and export, so a SPEC Sensors Digital O₃ Sensor was placed inside the devices and used to log ozone concentration, temperature, and relative humidity levels during runs (see Figure S1). While these sensors are rated from 0-5 ppm O₃, we observed they matched the readings from an EcoSensor SM-7 0-50 ppm range sensor in the ZT cabinet up to at least 45 ppm. These sensors, as well as the EcoSensor SM-7, required a warm up period for reliable operation.

Each instrument device uses a ventilation or ozone destruction cycle (reversing airflow through a catalyst) following treatment to return ozone concentrations back to safe levels. The ZT instrument passes the internal atmosphere continuously over a catalyst until ozone levels are reduced, and the VC unit completes a 5 minute vent through a catalyst (which worked well in our hands). All devices tolerated back-to-back runs over multiple hours with no change in ozone generation, and require no consumables chemicals or reagents, other than water.