Highlights
- •
-
Extracting RNA using TRIzol reagent from filter material ground with liquid nitrogen can be used to detect PRRSv and IAV from air filters.
-
•PRRSv and IAV were detected in used filters from commercial swine farms that employed air filtration. IAV was detected in more filters than PRRSv. PRRSv could be detected in used filters from farms with PRRSv negative status at the time of filter removal.
-
•Presence of PRRSv and IAV in the used filters shows likely evidence of between farm aerosol spread and the methods derived from this study open up avenues to further investigate regional airborne transmission and risk of virus introduction into farms.
Abstract
Filtering the air entering swine farms has significantly reduced the incidence of porcine reproductive and respiratory syndrome virus (PRRSv) infections in Midwestern breeding herds. Despite the significant investment made in installation of air filters, the nature and type of viruses trapped by these filters are not yet clear because of the unavailability of reliable methods to elute and detect these viruses. Here we report that eluting viral particles from the air filters by grinding filter specimens with liquid nitrogen, coupled with TRIzol reagent to extract RNA can detect both PRRSv and influenza A virus (IAV) from used minimum efficiency reporting value (MERV) 14, 15 and 16 rated air filters. PRRSv was detected in 27% (12/44) and IAV was detected in 66% (29/44) of filters that had been in installation between 08/12/13 and 07/12/17. Interestingly, PRRSv was also detected on used filters from farms with PRRSv negative status at the time of filter removal. Presence of PRRSv and IAV in the used filters show likely evidence of between herd aerosol spread for these viruses. The methods derived from this study open up avenues to further investigate airborne transmission and risk of virus introduction into farms contributing to the control of diseases in swine.
Keywords
Abbreviations
1. Introduction
Porcine reproductive and respiratory syndrome virus (PRRSv) and influenza A virus (IAV) are the two major respiratory viruses that often co-circulate in swine herds (Haden, Painter, Fangman, & Holtkamp, 2012) and can result in significant economic losses to the farmers (Holtkamp et al., 2013; Lugar, Ragland, & Stewart, 2017). PRRSv alone costs the US swine industry more than $664 million dollars/year (Holtkamp et al., 2013). Of the several routes of virus transmission, air/aerosol mediated transmission of PRRSv (Arruda et al., 2019; Otake, Dee, Corzo, Olivera & Dean, 2010; Otake, Dee, Jacobson, Torremorell, & Pijoan, 2002) and IAV (Corzo, Allerson, Gramer, Morrison, & Torremorell, 2014; Corzo, Ramagosa, Dee, Gramer, Morrison & Torremorell, 2013; Loeffen et al., 2011) is considered an important route of viral spread between farms. To reduce the risk of disease spread in herds, swine producers constantly implement several stringent biosecurity measures. One such strategy is filtering the air entering the farms (Cariolet, Marie, Moreau, & Robert, 1994; Dee, Batista, Deen, & Pijoan, 2005). Air filtration was successfully used to control the spread of Marek’s disease in commercial poultry farms in the early 1970s (Burmester & Witter, 1972; Grunder, Gavora, Spencer, & Turnbull, 1975). However, air filtration on commercial swine breeding herds in the US has gained momentum only recently (Spronk, Otake & Deen, 2010). Several field and experimental studies have shown that filtering air entering the swine facilities significantly reduce the risk of introducing new PRRS viruses (Alonso, Murtaugh, Dee, & Davies, 2013; Dee, Spronk, Reicks, Ruen, & Deen, 2010; Spronk, Otake, & Dee, 2010). Therefore, several sow farms have implemented air filtration systems in swine dense areas of the Midwestern USA (Dee, Otake, & Deen, 2010; Spronk, Otake & Deen, 2010), which has contributed to the reduction of PRRSv incidence outbreaks in large sow herds (Alonso et al., 2013). Additionally, the Morrison Swine Health Monitoring Project (MSHMP) (Perez et al., 2019) reports that the number of filtered farms has increased from 25 to 150 in the last decade, especially in the hog dense areas of Iowa and Minnesota (Vilalta, Unpublished). MSHMP also reports a statistically significant reduction in the number of PRRSv outbreaks in farms that have installed filters (Vilalta, Unpublished).
Despite significant investments made in filtering air entering the swine farms in the Midwestern USA, filtering alone does not eliminate the risk of new virus introductions through other routes. (Alonso et al., 2013). Further, there is no direct evidence on the type of viruses trapped by air filters in swine farms. Therefore, it is of great interest to identify and characterize the viruses that are captured by these air filters. Identification of such viruses from the filters will help advance the knowledge of airborne transmission and better understand the introduction of new viruses into swine farms.
The ability to detect several bacteria, fungi and viruses from biofilters used in the biosamplers from air, water and other environmental samples have been well documented (Andersen, 1958; Zimmerman, Reist, & Turner, 1987, Alvarez, Buttner, & Stetzenbach, 1995; Mehta, Bell-Robinson, Groves, Stetzenbach, & Pierson, 2000, Verreault, Moineau, & Duchaine, 2008; Gendron, Verreault, Veillette, Moineau, & Duchaine, 2010; Cao, Noti, Blachere, Lindsley, & Beezhold, 2011; Li et al., 2018). Such samplers equipped with a filter have made it possible to detect IAV in the air from buildings and hospitals (Blachere et al., 2009; Lindsley et al., 2010), commercial aircrafts (Yang, Elankumaran, & Marr, 2011), educational institutes (Coleman & Sigler, 2020; Xie et al., 2020) and live animal markets (Chen et al., 2009). In general, filters used in commercial biosamplers have varying but high levels of collection efficiency. These filters are specifically designed to collect and concentrate pathogen samples by filtration, which could be combined with subsequent size fractionation or ultracentrifugation as required to further concentrate the nucleic acids. Additionally, the pathogens or viral particles collected by some of the biosamplers are deposited directly in tubes with liquid media. The liquid containing the pathogens or viruses can be subjected to RNA extraction directly or after a subsequent ultracentrifugation step to further concentrate the pathogen load before RNA extraction. The scientific revolution in sequencing technologies have made available a number of RNA extraction strategies that are readily available to choose from. From this perspective, several commercial viral extraction kits and reagents such as Trizol (Chomczynski & Sacchi, 1987) are being used to extract RNA from biofilters (https://www.cdc.gov/niosh/topics/flu/transmission.html). Fabian, McDevitt, Lee, Houseman, and Milton (2009) showed that the Trizol-chloroform extraction method resulted in higher recovery of viral RNA and efficiently removed PCR inhibitors in the RT-PCR reactions, which ultimately facilitated efficient detection of influenza. The same authors further observed that the magnetic bead extraction method, resulted in poor viral RNA yield (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7197756/; Fabian et al., 2009). In contrast, (Gendron et al., 2010), reported that the viral extraction performed better in identifying phages from biofilters. Therefore, it is evident that the method/s used may affect the quality and quantity of RNA extracted which will in turn have a profound effect on precise detection of the viruses in question (Korves et al., 2011).
Commercial swine barns employ MERV (Minimum Efficiency Reporting Values) filters with ratings ranging between 14 and 16 (Spronk, Otake, & Dee, 2010). MERV rating signifies a filter’s ability to capture airborne particles between 0.3 and 10 μm (https://www.epa.gov/indoor-air-quality-iaq/what-merv-rating-1). The filter efficiency increases with the MERV ratings (https://www.ashrae.org/File%20Library/About/Position%20Documents/Filtration-and-Air-Cleaning-PD.PDF) and therefore different MERV rated filters are used for different applications. It should be noted that the MERV14 air filters generally provide 90%–95% efficiency for filtering particles between 3 and 10 μm in size, 85%–90% efficiency for filtering particles between 1 and 3 μm in size (such as legionella, lead dust, humidifier dust, coal dust, and nebulizer droplets) and 50%–75% efficiency for filtering particles between 0.30 and 1 μm in size (such as bacteria, most smoke, sneeze nuclei, pesticide and fungicide dust).
The size and complex pleated nature of the commercial HVAC filters does not provide a definite rationale to single out a specific area on the filter that can be consistently used to sample and identify the viruses until established. Though expected, there is no known evidence that corroborates an uniform distribution of IAV or PRRSv aerosols or aggregates on the surface of the HVAC filters. Unlike the biofilters, that are specifically designed to collect and concentrate the pathogens and viral particles, commercial HVAC filters are not designed to collect and readily facilitate the detection of viruses trapped or immobilized in the filters. It might be possible to concentrate the nucleic acids by extensive sample elution combined with ultrafiltration or ultracentrifugation methods especially with filters that are heavily loaded with viruses under experimental conditions. However, this might not be the case with filters removed from the field as there is no evidence on the amount of viruses or viral particles present on such filters, especially from those retrieved from swine barns. More importantly, the HVAC filters in the swine barns are not frequently replaced and are subjected to extensive fluctuations in environmental conditions (especially heat, cold and moisture) and are often covered with debris and extraneous material that could interfere with successful RNA extraction and hence in the detection of the viruses. Further, concentration of the viral particles by extensive sampling, pooling and subsequent ultrafiltration or ultracentrifugation from such battered filters may result in the degradation of nucleic acids and introduce PCR inhibitors that could result in false negatives. The difficulties underlying the detection of Influenza from electret filters from exhaled breath have been corroborated by Huynh, Oliver, Stelzer-Braid, Rawlinson, and Tovet (2008) and Anja Valen (2011). However, a simple switch from electret to teflon resulted in the successful detection of influenza from exhaled human breath (Fabian et al., 2009).
The biggest challenge in identifying the viruses from the HVAC filters from the swine barn is to efficiently sample and elute the viral particles for RNA extraction. Though Korves et al. (2011) and Goyal et al. (2011) identified influenza from HEPA filters retrieved from aircrafts and HVAC filters installed in buildings, they did not identify the model or the manufacturer of the filters used in their studies and the logic behind their sampling strategy was not clearly detailed. Furthermore, performance of the filters vary with the type and efficiency of the filter and the period they are in use, which may have a profound effect on the sampling strategy and hence the outcome of RNA extraction and viral detection. Therefore, we reasoned that it might be important to nail down the sampling and extraction strategies, to identify respiratory viruses specifically for the MERV rated HVAC commercial filters from the swine barns. To the best of our knowledge, there is no well recognized method currently available to readily sample and detect viruses from air filters retrieved from swine or poultry barns. Therefore, the main goal of this study is to evaluate the feasibility of identifying PRRSv and IAV on used air filters retrieved from swine farms The objectives of this study were to 1) identify the best combination of sample elution and RNA extraction methods to detect PRRSv from MERV14 filters aerosolized with PRRSv under experimental conditions and 2) exploit the best method identified to detect PRRSv and IAV from used filters from farms located in pig dense areas of the Midwestern USA.
2. Materials and methods
2.1. Pilot studies to identify PRRSv in MERV14 filters under laboratory conditions by spiking MERV14 filters with VR-2332 PRRSv reference strain
To evaluate our ability to identify PRRSv from MERV14 filters and understand the detectable limits, we spiked one square inch area of brand new MERV14 filters with 1000 μL of varying dilutions (10^-1 to 10^-8) of VR-2332 PRRSv reference strain. We used MERV14 grade commercial HVAC filters (Model E736, MERV A13 V-Bank with Gasket with a dimension of 24″ X 24″ X 12”, 3M, St. Paul, MN) because they are commonly used filter types in the swine farms. The spiked filters were air dried for 60–75 min at room temperature in a biological safety cabinet in a BSL2 facility. The dried filter material was then cut into small pieces and transferred into 50 ml conical tubes, soaked in 5 ml of Minimum Essential Media (MEM) (Tiwari, Patnayak, Chander, Parsad, & Goyal, 2006), vortexed briefly and incubated at 37°C for 15 min. Such processed filters were subjected to four different methods to elute the viral particles as follows: a) elbow shake, b) vortex and c) freeze thaw and d) liquid nitrogen grinding. The later method was reasoned by considering that cryogenic grinding of the filter materials with liquid nitrogen will be more efficient in releasing the viral particles from the MERV rated HVAC commercial filters and therefore prevent the subsequent degradation of the viral RNA. The details of the materials were as follow:
Elbow shake: Tubes containing the soaked material were inverted upside down 20 times and 250 μL of the MEM media was retrieved for RNA extraction.
Vortex: Tubes containing the soaked filter material were vortexed at 3000 rpm for 3 min and 250 μL of the MEM media was retrieved for RNA extraction.
Freeze thaw: Tubes containing the soaked filter material were frozen at −70°C overnight and thawed on ice the next day, from which 250 μL of the media was retrieved and used for RNA extraction.
Liquid nitrogen grind: The soaked filter material was transferred to a clean pestle and mortar and finely ground by pouring liquid nitrogen as required. The ground filter material was pressed against the walls of the mortar using the pestle and 250 μL of the sample eluate was retrieved and used for RNA extraction.
2.2. RNA extraction using TRIzol™ LS reagent (TRIzol)
RNA was extracted from all samples exploiting the methods described above using TRIzol reagent (Applied Biosystems, Foster City, CA). Two hundred and fifty μL of the sample eluate from all the above four methods was added to 750 μL of TRIzol reagent. The contents were mixed and then incubated at room temperature for 5 min. 200 μL of chloroform was added to the above mix and mixed by inversion, followed by incubation at room temperature for another 3 min. The sample was centrifuged at 13,000 rpm for 15 min at 4°C. The upper aqueous phase was collected into a new tube and 0.5 ml of 100% isopropanol were added to the tube. The mixture was then incubated at room temperature for 10 min and centrifuged at 13,000 rpm for 15 min at 4°C. The supernatant was discarded and the RNA pellet was washed by applying one ml of 75% ethanol, and centrifuging the tube at 10,000 rpm for 5 min at 4°C. The wash was discarded and the RNA pellet was air-dried for 5–10 min. The RNA pellet was then resuspended in 50 μl of RNase-free water by flicking the tubes several times, followed by incubation at 55–60°C for 5–15 min in a water bath. 5 μL of RNA was used for detecting PRRSv by real time PCR (RT-PCR) from the spiked filter.
2.3. Studies to identify aerosolized PRRSv in MERV14 filters
The distribution of particles within a filter depends on the way a filter is loaded (Montgomery, Green, Rogak, & Bartlett, 2012; Thomas, Penicot, Contal, Leclerc, & Vendel, 2001; Sun & Woodman, 2009). In this objective we loaded filters by aerosolizing a VR-2332 derived PRRSv modified live vaccine virus (Ingelvac, Boehringer Ingelheim, St. Joseph, MO) into a MERV14 commercial grade HVAC filter. To identify the best sample elution and RNA extraction strategy, three different concentrations (undiluted, 10^-1 and 10^-2 dilutions) of a modified live VR-2332 PRRSv vaccine strain was aerosolized onto three different MERV14 filters (Camfil, L6, P/N 85508361D44V-PB-2424-L6-30/SP, Riverdale, NJ). PRRSv vaccine virus was aerosolized using two aerosol generators (Large Particle Tank and Air Techniques International (ATI) tank at the same time) in the ASHRAE Standard 52.2 wind tunnel available in the Particle Calibration Laboratory at the University of Minnesota Department of Mechanical Engineering, University of Minnesota, Minneapolis, MN. Approximately between 2310–2320 ml of the PRRSv vaccine solution were aerosolized for 120 min and air samples collected upstream and downstream from the filter to indicate virus filter loading. After aerosolization, the filters were removed from the wind tunnel and sampled as shown in Fig. 1. One square inch piece of the filter(s) was excised from five different spots on each of the five different panels of the MERV14 rated HVAC filter. The spot was replicated thrice (total of 15 replicates per panel/side) and all the filter samples were cut into small pieces. There was a total of 75 samples evaluated per filter per dilution. The finely cut filter material was soaked in 5 ml of MEM media and vortexed at 3000 rpm/min or until all the filter material was soaked up and incubated at 37°C for 15 min. The samples were then subjected to three different elution methods namely: vortex, freeze thaw and grinding with liquid nitrogen as described above. Two different RNA extraction strategies (MagMAX viral RNA extraction system (Applied Biosystems, Foster City, CA) or TRIzol were also evaluated for their ability to detect PRRSv from the MERV14 filters that had been aerosolized. 250 μL of the eluate retrieved from the different methods were subjected to RNA extraction. RNA was extracted by the TRIzol method as described above and from MagMAX viral extraction kit following the instructions provided by the manufacturer. RNA was eluted in a final volume of 15 μL of RNase free water in the case of TRIzol, and elution buffer in the case of the MagMAX method.
Fig. 1. Sampling of the MERV14 filters aerosolized with the porcine reproductive and respiratory syndrome (PRRSv) vaccine virus.
The elution and RNA extraction combinations tested were: 1) MagMAX + Freeze thaw, 2) MagMAX + Liquid nitrogen grinding, 3) MagMAX + Vortexing, 4) TRIzol + Freeze thaw, 5) TRIzol + Liquid nitrogen grinding and 6) TRIzol + Vortexing. The best method was identified based on statistical significance for both, the cycle threshold (Ct) RT-PCR values and RNA copies/ml.
2.4. Quantitative RT-PCR
Five μL of RNA isolated by the different methods were subjected to quantitative real time PCR to detect PRRSv by targeting the open reading frame 6 (ORF-6) of the virus. Both, Ct values and RNA copies/ml were derived. The best method to detect PRRSv was identified based on statistical significance for both the Ct values and RNA copies/ml.
The primer sequences with their concentration and volume of the various components used in the RT-PCR reactions for detecting PRRSv are shown in Table 1.
Table 1. Porcine reproductive and respiratory syndrome virus primer sequences with their concentration and volumes of the various components used in the RT-PCR reactions.
| PCR component and concentration used | Volume of individual PCR component/PCR reaction (μL) | Primer sequence 5′-3′ |
|---|---|---|
| 00 NA For 1 (ORF6) (40 μM) | 0.47 | GTA GTY GCR CTC CTT TGG GGR GTG T |
| 09 NA For 2 (ORF6) (40 μM) | 0.47 | AGR TGC CGT YTG TGC YTG CTA |
| 2010 NA Rev 1(ORF 6) (40 μM) | 0.47 | GAC GCC GRA CGA SAA AYG CGT GGT TA |
| 2010 NA Rev 2 (ORF6) (40 μM) | 0.47 | CCY GCR GCA CTT TCV ACG |
| PRRS NA/EU FAM labeled (ORF6)TAMRA probe (22 mer) | 0.63 | FAM- TACATTCTGGCCCCTGCCCAYC-(TAMRA) |
| 2X RT Buffer | 12.50 | |
| 25X Enzyme | 1.00 | |
| Detector Enhancer | 1.67 | |
| RNA template | 5.0 | |
| Water | 1.29 | |
| Total PCR volume | 25 |
Water was used as a negative control in the RT-PCR reactions in place of the RNA template. RNA extracted from MEM minimal media was also used as an extraction negative control in the RT-PCR reactions.
The samples were subjected to RT-PCR, using the Ag Path-ID One Step RTPCR Kit (Applied Biosystems, Foster City, CA) in an ABI 7500 Fast Thermal cycler (Applied Biosystems, Foster City, CA) using the following conditions: 45°C: 10 min, 95°C for 10 min followed by 45 cycles of 95°C for 15 s and 60°C for 45 s. Water was used as a negative control in place of the RNA template in the RT-PCR reactions. RNA extracted from MEM minimal media was also used as an extraction negative control in the RT-PCR reactions. Real-time RT-PCR samples with Ct values ≤ 35 were considered positive, between >35 and < 40 were suspect, and ≥40 were negative for PRRSv.
2.5. Statistical methods
For both quantity of virus and Ct value, we fitted a linear mixed model for each dilution, with treatment (elution strategies: liquid nitrogen grinding, vortex or freeze thaw), method (RNA extraction methods: TRIzol or MagMAX), and the interaction as fixed effects, and replicate and treatment within replicate as random effects. Quantity of virus was modeled on the log scale to meet the assumptions of normality and equal variance, and back transformed for reporting. For each model, we report estimated marginal means and 95% confidence intervals for the six treatment/method combinations, with pairwise comparisons computed using the Tukey’s Honestly Significant Difference (HSD) correction for multiple comparisons. Differences were considered statistically significant at the 0.05 level. Additionally, the values at the detection limit were apparent outliers, so to evaluate the effect they had on the model, data was truncated at the nearest value below the detection limit, and models were refit. Models did not differ meaningfully, so the results using the truncated data are reported. All calculations were performed in R version 3.6.0 (2019-04-26).
2.6. Evaluating used filters from swine farms
Exploiting the procedure standardized for extracting RNA from PRRSv treated MERV14 HVAC commercial filters, we set out to identify both PRRSv and IAV from used filters obtained from different swine barns. Though air filtration is primarily used to prevent new PRRSv airborne infections into the herds, we targeted both IAV and PRRSv because both these viruses are airborne and co-circulate in swine farms. Further, these two viruses are the major respiratory pathogens of swine that result in significant yield losses to the swine farmers (Haden et al., 2012).
A set of 44 filters, originating from 13 different farms and 4 different agricultural companies were tested. These farms were located in high pig dense areas and had a history of PRRSv infection prior to filtration. The filters investigated included various brands and MERV ratings 14, 15 and 16, and were installed between 08/12/13 and 07/12/17. PRRSv status at the time of obtaining the filter was known for farms from one company. One pool of three samples per side was evaluated for each of the five sides separately i.e. five pooled samples/filter. Each of the three samples/side was processed separately starting from one square inch of the filter material. 250 μL of elute from each of the three samples were then pooled and 250 μL of the pooled sample was used for RNA extraction using TRIzol as described earlier. RNA was suspended in 15 μL of RNase free water. Eight microliters of RNA was used for detecting PRRSv by targeting the ORF-6 gene as described above and IAV by targeting the Matrix gene using a modified method of Slomka et al. (2010) separately.
The primer sequences with their concentration and volume of the various components used in the RT-PCR reactions for detecting IAV are shown in Table 2. The samples were subjected to RT-PCR, using the Ag Path-ID One Step RTPCR Kit (Applied Biosystems, Foster City, CA) in an ABI 7500 Fast Thermal cycler (Applied Biosystems, Foster City, CA) using the following conditions: 45°C for 10 min, 95°C for 10 min followed by 45 cycles of 94°C for 1 s and 60°C for 30 s. Water was used as negative control in place of the RNA template in the RT-PCR reactions. RNA extracted from MEM minimal media was also used as a negative extraction control. Real-time RT-PCR samples with Ct values ≤ 35 were considered positive, between >35 and < 40 were suspect, and ≥40 were negative for both PRRSv and IAV.
Table 2. Influenza A virus primer sequences with their concentration and volumes of the various components used in the RT-PCR reactions.
| PCR component and concentration used | Volume of individual PCR component/PCR reaction (μL) | Primer sequence 5′-3′ |
|---|---|---|
| M+25 F (20 μM) | 0. 25 | AGA TGA GTC TTC TAA CCG AGG TCG |
| M-124 R (20 μM) | 0.25 | TGC AAA AAC ATC TTC AAG TCT CTG |
| M-124 SIV R (20 μM) | 0.25 | TGC AAA GAC ACT TTC CAG TCT CTG |
| M+64 Probe (6 μM) | 0.25 | FAM/TCA GGC CCC/ZEN/CTC AAA GCC GA/IABk®FQ |
| 2X RT Buffer | 12.50 | |
| 25 X Enzyme | 1.00 | |
| Detector Enhancer | 1.67 | |
| RNA template | 5.00 | |
| Water | 3.83 | |
| Total PCR volume | 25.0 |
Water was used as a negative control in the RT-PCR reactions in place of the RNA template. RNA extracted from MEM minimal media was also used as an extraction negative control in the RT-PCR reactions.
3. Results and discussion
We first assessed the feasibility to identify PRRSv on brand new MERV14 HVAC commercial filters under laboratory conditions by spiking them with PRRSv reference strain VR-2332. We evaluated four different elution strategies and were able to successfully detect PRRSv by all the four elution methods namely elbow shake, vortex, filter thawing and grinding with liquid nitrogen (Fig. 2). This is not surprising since the filters were infiltrated with a high concentration of the virus inoculum. Elbow shake and freeze thaw methods were able to positively detect PRRSv only to 10^-5 and 10^-6 dilution, respectively, while vortex and grinding the filters with liquid nitrogen positively identified PRRSv up to 10^-8 dilution. However, grinding with liquid nitrogen yielded numerically lower Ct values than vortex. It is interesting to note that the Ct values were lower by1.5 log values approximately, when the infiltrated filters were subjected to liquid nitrogen grinding than the virus by itself (virus alone). This is probably because liquid nitrogen breaks down the viruses more efficiently and release the nucleic acids better. Grinding tissues with liquid nitrogen has been successfully employed to extract viral RNA from starchy and lignocellulosic materials (Xiao, Kim, & Meng, 2015). To the knowledge of the authors, this is the first time that grinding filters with liquid nitrogen has been reported to detect viruses from air filters.
Fig. 2. Effect of different sample elution methods on the qRT-PCR cycle threshold (Ct) values of porcine reproductive and respiratory syndrome virus (PRRSv), identified from MERV14 filters infiltrated with VR-2332 PRRSv reference virus strain under laboratory conditions.
We next evaluated the three elution methods, freeze thaw, vortex, and liquid nitrogen grinding, by aerosolizing PRRSv into MERV14 HVAC commercial filters under experimental conditions to mimic how air filters become loaded under normal operating conditions. Elbow shake was not included in this test because of lower level of sensitivity (higher Ct values reached at lower concentration of the virus) to detect the virus in the pilot study reported above and the physical labor involved in processing the samples. Further, we reasoned that elbow shake might not be the best method to elute viral particles from filters that might have very low levels of the virus. In addition to the different elution methods, we also evaluated two different methods of RNA extraction strategies (MagMAX viral and TRIzol), in different combinations with the different elution methods and 3 different concentrations of the PRRSv vaccine. Interestingly, we were able to detect PRRSv by all the elution and extraction methods (Fig. 3, Fig. 4). The liquid nitrogen/TRIzol combination had the lowest Ct for all three dilutions (Table 3). For direct (undiluted vaccine virus), Ct value had a mean of 29.04, and was statistically significantly smaller than all other combinations; for 10^-1 dilution, Ct had a mean of 32.5, and only vortex/TRIzol was not statistically different, with a mean of 33.0; for 10^-2 dilution, Ct had a mean 35.4, and only vortex/TRIzol was statistically different, with a mean of 38.6. The liquid nitrogen/TRIzol combination also had the highest response in terms of quantity of the virus for all three dilutions. For the undiluted vaccine, Ct had a back-transformed mean of 7.37 × 105, and was statistically significantly higher than all other combinations; for 10^-1, dilution, Ct had a back-transformed mean of 8.97 × 104, and only vortex/TRIzol was not statistically different, with a back-transformed mean of 6.42 × 104; for 10^-2, dilution it had a back-transformed mean of 1.08 × 105, and only liquid nitrogen/MagMAX was not statistically different, with a back-transformed mean of 4.85 × 103. Therefore, grinding the filter material in liquid nitrogen combined with TRIzol extraction was reasoned to be the best method to detect PRRSv from experimentally loaded filters based on lower Ct values (Fig. 3 and Table 3) and higher quantity of the virus (Fig. 4 and Table 4). RNA extraction with TRIzol is a common method of total RNA extraction using guanidinium isothiocyanate and it is used to extract RNA from a wide variety of tissues (Chomczynski & Sacchi, 1987; 2006). Though, grinding the filters with liquid nitrogen and extracting the RNA by TRIzol is a laborious and lengthy process, it is a feasible method to identify PRRSv and IAV, as shown in this study. Other viruses and microbes from filters have been identified (Fabian et al., 2008; Fabian et al., 2009; Goyal et al., 2011; Korves et al., 2011) from electret, HVAC and aircraft filters by other strategies.
Fig. 3. Estimated marginal means and 95% confidence intervals for the six treatment/method combinations, with pairwise comparisons computed using the Tukey’s honestly significant difference test correction for multiple comparisons on the cycle threshold (Ct) values detected by qRT-PCR from MERV14 filters aerosolized with porcine reproductive and respiratory virus (PRRSv). Groups that do not share a letter are statistically significant at the 0.05 level. Ct value was modeled on the log scale and back transformed for display.
Fig. 4. Estimated marginal means and 95% confidence intervals for the six treatment/method combinations, with pairwise comparisons computed using the Tukey’s Honestly significant difference test correction for multiple comparisons on the quantity of virus detected by qRT-PCR from MERV14 filters aerosolized with porcine reproductive and respiratory syndrome virus (PRRSv). Groups that do not share a letter are statistically significant at the 0.05 level. Quantity or RNA copies was modeled on the log scale and back transformed for display.
Table 3. Estimated marginal means and 95% confidence intervals for the six treatment/method combinations, with pairwise comparisons computed using the Tukey’s Honestly Significant Difference correction for multiple comparisons on the cycle threshold values detected by qRT-PCR from MERV14 filters aerosolized with porcine reproductive and respiratory syndrome virus (PRRSv). Groups that do not share a letter are statistically significant at the 0.05 level. Ct value was modeled on the log scale and back transformed for display.
| PRRSv filter loading dilution | Treatment | Method | Ct values (95% CI) | Lower CL | Upper CL | Group |
|---|---|---|---|---|---|---|
| Direct | Liquid nitrogen | MagMAX | 30.00 | 29.84 | 30.16 | b |
| Trizol | 29.04 | 28.84 | 29.20 | a | ||
| Vortex | MagMAX | 31.32 | 31.16 | 31.48 | d | |
| TRIzol | 30.13 | 29.96 | 30.29 | bc | ||
| Freeze thaw | MagMAX | 31.32 | 31.16 | 31.48 | d | |
| TRIzol | 30.38 | 30.22 | 30.54 | c | ||
| 10^-1 | Liquid nitrogen | MagMAX | 33.91 | 33.46 | 34.36 | bc |
| TRIzol | 32.51 | 32.10 | 32.96 | a | ||
| Vortex | MagMAX | 34.59 | 34.14 | 35.05 | cd | |
| TRIzol | 32.98 | 32.53 | 33.43 | cd | ||
| Freeze thaw | MagMAX | 35.03 | 34.58 | 35.49 | d | |
| TRIzol | 33.79 | 33.34 | 34.25 | bc | ||
| 10^-2 | Liquid nitrogen | MagMAX | 35.69 | 34.38 | 37.00 | a |
| TRIzol | 35.42 | 34.14 | 36.73 | a | ||
| Vortex | MagMAX | 37.83 | 36.53 | 39.15 | ab | |
| TRIzol | 38.61 | 37.3 | 39.92 | b | ||
| Freeze thaw | MagMAX | 36.22 | 34.91 | 37.53 | ab | |
| TRIzol | 37.21 | 35.90 | 38.52 | ab |
Upper CL: upper confidence level; Lower CL: lower confidence level.
Table 4. Estimated marginal means and 95% confidence intervals for the six treatment/method combinations, with pairwise comparisons computed using the Tukey’s Honestly Significant Difference correction for multiple comparisons on the quantity of virus detected by qRT-PCR from MERV14 filters aerosolized with porcine reproductive and respiratory syndrome virus (PRRSv). Groups that do not share a letter are statistically significant at the 0.05 level. Quantity was modeled on the log scale and back transformed for display.
| PRRSv filter loading dilution | Treatment | Method | Quantity of viral RNA copies/ml (95% CI) | Lower CL | Upper CL | Group |
|---|---|---|---|---|---|---|
| Direct | Freeze Thaw | MagMAX | 1.77E+05 | 1.53E+05 | 2.05E+05 | a |
| TRIzol | 2.75E+05 | 2.37E+05 | 3.19E+05 | b | ||
| Vortex | MagMAX | 2.05E+05 | 1.77E+05 | 2.38E+05 | a | |
| TRIzol | 3.17E+05 | 2.74E+05 | 3.68E+05 | bc | ||
| Liquid Nitrogen | MagMAX | 3.94E+05 | 3.39E+05 | 4.57E+05 | c | |
| TRIzol | 7.38E+05 | 6.36E+05 | 8.55E+04 | d | ||
| 10^-1 | Freeze Thaw | MagMAX | 1.60E+04 | 1.17E+04 | 2.19E+04 | a |
| TRIzol | 2.19E+04 | 2.88E+04 | 5.41E+04 | bc | ||
| Vortex | MagMAX | 2.22E+04 | 1.62E+04 | 3.04E+04 | ab | |
| TRIzol | 6.42E+04 | 4.69E+04 | 8.80E+04 | cd | ||
| Liquid Nitrogen | MagMAX | 2.49E+04 | 1.82E+04 | 1.82E+04 | ab | |
| TRIzol | 8.97E+04 | 6.54E+04 | 1.23E+05 | d | ||
| 10^-2 | Freeze Thaw | MagMAX | 9.79E+02 | 4.86E+04 | 1.97E+03 | a |
| TRIzol | 1.24E+02 | 6.15E+02 | 2.50E+03 | ab | ||
| Vortex | MagMAX | 1.64E+03 | 8.14E+02 | 3.30E+03 | ab | |
| TRIzol | 1.30E+03 | 6.43E+02 | 2.61E+03 | ab | ||
| Liquid Nitrogen | MagMAX | 2.61E+03 | 2.41E+03 | 9.77E+03 | bc | |
| TRIzol | 1.08E+04 | 5.35E+03 | 2.17E+04 | c |
Upper. CL: Upper confidence level and Lower CL: Lower confidence level.
The pilot aerosolization experiments with IAV onto MERV filters could not be performed because the wind tunnels in our experiments are not supported for evaluating potential human pathogens like IAV. Though the liquid nitrogen grinding + TRIzol combination method was initially standardized to detect PRRSv from MERV14 filters, we strongly believed that it might be possible to detect IAV from the RNA isolated from the barn filters. In our experience, the RNA extraction method remains the same for both IAV and PRRSv and it has been possible to identify both PRRSv and IAV from the same RNA source. Exploiting the liquid nitrogen grinding + TRIzol combination, a set of 44 used filters from swine farms were evaluated for PRRSv and IAV. We were able to identify both PRRSv and IAV from MERV14, 15 and 16 rated filters. PRRSv was identified in 27% (12/44) of the used filters from 31% (4/13) of the farms, while IAV was identified in 66% (29/44) of the filters from 77% (10/13) of the farms (Table 5). Ct values for samples positive for PRRSv ranged from 32.21 to 34.57, while it varied from 26.96 to 34.93 for IAV. Results of the number of PRRSv and IAV positive and suspect filters by farm are shown in Table 6. It is interesting to note that two farms, farms 5 and 6, which were negative for PRRSv at the time of filter removal, had filters that were positive for PRRSv suggesting that the filters were able to prevent the introduction of viable PRRSv into the farm (Table 6). This further asserts the airborne transmission of PRRSv and IAV and the ability of air filters to trap such viruses entering the farms. More number of filters tested positive for IAV than for PRRSv. It is possible that the techniques described in this study might work well on other filter types and enable us to identify other viruses of economic importance. The primers and PCR conditions used in this study are well established and routinely used in several institutions and diagnostic laboratories in the US and are proven to identify only PRRSv and IAV. It would be ideal to have the sequences of the PRRSv and IAV segments targeted by the RT-PCR from the filters evaluated from the farms. It is possible that the viral RNA may not be intact due to RNA degradation in the filter material given the long installation time of the filters. For instance, in this study the filters evaluated had been installed between 1 and 3 years. Further, the contaminants and the inhibitors in the samples or the extracted RNA could interfere with the amplification of the targeted genes (Fabian et al., 2008; Fabian et al., 2009; Korves et al., 2011). Despite these, further studies are necessary to amplify and sequence the viral segments found in the filters. Unfortunately, sequencing the viral RNA was outside the scope of this study.
Table 5. Summary of number of farms, filters and samples that were positive for porcine reproductive and respiratory syndrome virus (PRRSv) and influenza A virus (IAV) by qRT-PCR.
| No. PRRSv positive/total (%) | No. IAV positive/total (%) | |
|---|---|---|
| Filters | 12/44 (27) | 29/44 (66) |
| Samples | 25/246 (10) | 74/246 (30) |
| Farms | 4/13 (31) | 10/13 (77) |
Table 6. Number of filters qRT-PCR positive and suspect for porcine reproductive and respiratory syndrome virus (PRRSv) and influenza A virus (IAV) by farm. The status of the farm for PRRS virus if known is shown and was provided by the producer or was obtained from the MSHMP database.
| Farm code | No. filters/farm | No. of PRRSv positive filtersa | No. of filters suspected for PRRSv | No. filters positive for IAV | No. filters suspected for IAV | PRRSv farm status at time of filter removal |
|---|---|---|---|---|---|---|
| 1 | 5 | 1 | 0 | 5 | 5 | Positive |
| 2 | 4 | 1 | 0 | 1 | 4 | Positive |
| 3 | 1 | 0 | 0 | 1 | 1 | Positive |
| 4 | 1 | 0 | 0 | 0 | 1 | Positive |
| 5 | 7 | 1 | 3 | 4 | 7 | Negative |
| 6 | 12 | 9 | 2 | 11 | 7 | Negative |
| 7 | 2 | 0 | 0 | 2 | 2 | N/A |
| 8 | 2 | 0 | 0 | 1 | 2 | N/A |
| 9 | 2 | 0 | 0 | 0 | 2 | N/A |
| 10 | 2 | 0 | 0 | 0 | 2 | N/A |
| 11 | 2 | 0 | 0 | 2 | 2 | N/A |
| 12 | 2 | 0 | 0 | 1 | 2 | N/A |
| 13 | 2 | 0 | 0 | 1 | 2 | N/A |
| Total | 44 | 12 | 5 | 29 | 39 |
- a
-
Filters were considered positive if real-time RT-PCR cycle threshold (Ct) values were ≤35, suspect >35 and < 40, and negative if Ct ≥ 40.
4. Conclusions
Extracting RNA using TRIzol from filter material ground with liquid nitrogen can be used to detect PRRSv and IAV virus from air filters. Furthermore, PRRSv and IAV were detected in used filters installed in swine farms. IAV was detected in more filters than PRRSv, and PRRSv could be detected in used filters from farms with PRRSv negative status at the time of filter removal. Presence of PRRSv and IAV in the used filters indicates evidence of aerosol spread and the methods derived from this study open up avenues to further investigate airborne transmission and risk of virus introduction into farms or other facilities.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This study was funded by the Swine Disease Eradication Center, University of Minnesota. The authors would like to acknowledge the contributions of Dr. Bob Morrison (deceased May 2, 2017) for instigating critical thinking and raising questions related to this research; the University of Minnesota Veterinary Diagnostic Laboratory for providing the primer sequences and PCR conditions for detecting ORF 6 of PRRSv; Ian Marabella and Austin Andrews, Department of Mechanical Engineering at the University of Minnesota for their help with the wind tunnel experiments.
References
- Alonso et al., 2013
-
C. Alonso, M.P. Murtaugh, S.A. Dee, P.R. DaviesEpidemiological study of air filtration systems for preventing PRRSV infection in large sow herdsPreventive Veterinary Medicine, 112 (1–2) (2013), pp. 109-117
