Abstract
We used virus isolation (VI) to determine tissue culture infectivity and reverse-transcription quantitative PCR (RT-qPCR) to determine the stability of porcine reproductive and respiratory syndrome virus 2 (PRRSV) strain P129 in solvent-extracted soybean meal (SBM), dried distillers grains with solubles (DDGS), complete swine feed (FEED), or medium (DMEM) at 4°C, 23°C, or 37°C for up to 3 d. Samples of each treatment were taken at regular intervals and processed. Supernatant was titrated and used to inoculate confluent MARC-145 cells to determine infectivity. RNA was extracted from each supernatant sample and tested by RT-qPCR to determine any change in detectable virus RNA across matrix type, temperature, and time. An interaction (p = 0.028) was observed for matrix × temperature × hour for live virus detected by VI. At 4°C, the concentration of infectious virus was greatest in DMEM, intermediate in SBM, and lowest in DDGS and FEED. DMEM also had the greatest concentration of infectious PRRSV at 23°C over time; a higher infectious virus concentration was maintained in SBM for longer than in DDGS or FEED. At 37°C, a greater concentration of infectious virus was sustained in DMEM than in the feedstuffs, with concentrations decreasing until 48 h post-inoculation. Only matrix type influenced the quantity of viral RNA detected by RT-qPCR (p = 0.032). More viral RNA was detected in the virus control than in DDGS; SBM and FEED were intermediate. By VI, we found that infectious virus could be harbored in SBM, DDGS, and FEED for a short time.
Porcine reproductive and respiratory syndrome virus (PRRSV; Betaarterivirus suid 2) can cause decreased productivity and significant economic impacts.9,15 PRRS is a seasonal challenge for the U.S. swine industry, with many systems filtering air for barns to prevent airborne introduction.3,18,23 PRRS can cause up to a 23% decrease in the production of weaned pigs and increases in abortions and repeat breeding in sows, leading to a need to understand infectivity and transmission to prevent the introduction of PRRSV to farms. 21
Transmission of PRRSV is believed to be via aerosol, but the virus can also survive in the environment, manure, or feed. The half-life of infectious PRRSV varies based on the matrix type, temperature, and humidity, with higher temperature and humidity decreasing the viral half-life.7,13 Although the half-life is shorter in feedstuffs than in other matrix types, there is evidence that PRRSV can survive in feedstuffs, which could be a concern for the introduction of virus into a farm. 5 Reliable detection of viral RNA is a challenge when dealing with non-animal samples such as feed or environmental samples, largely because commercial PRRSV PCR testing was originally developed for clinical samples such as tissues or serum.
Sampling feed to detect viruses of concern has been a challenge for feed and virology researchers. Increasing the PCR Ct can help detect low levels of virus in complex matrices, such as feed, but measuring the infectivity of these viruses has still been difficult. The mechanisms behind these complications have not been positively identified, but likely include the complexity of the feed matrix, the large number of products, the steps involved in processing both the feed and the ingredients, and the particle size of the matrix. Infectivity of PRRSV in feed has been documented in a large research barn, and bioassay is the most sensitive and reliable method to determine the infective potential of contaminated feed. 6 Unfortunately, bioassay is not a good option for commercial livestock producers, given the time and cost involved.
The most common method for the detection of PRRSV or any other virus in animal feed is by using PCR and non-standardized extraction methods. Detection by PCR only provides information on the presence of viral genome, not if the virus is infectious. An assay using MARC-145 cells to determine infectivity would allow for faster, more cost-effective identification of infectious PRRSV in animal feed. Therefore, we used virus isolation (VI) to determine infectivity and reverse-transcription quantitative PCR (RT-qPCR) to determine the stability of PRRSV when incubated in soybean meal (SBM), dried distillers grains with solubles (DDGS), or a corn-soy–based swine feed (FEED) at 3 temperatures over 3 d.
Materials and methods
General
Our treatment structure was a split-plot design with time nested within matrix × temperature. Three feed matrices were used (SBM, DDGS, and FEED [a swine gestation diet with 78% corn, 17% soybean meal, and 0.5% fat]), and 10 g of each were inoculated with 1 mL of type 2 PRRSV strain P129 with a concentration of 105.4 TCID50/mL for a final concentration of 104.4 TCID50/g of matrix. Each matrix was incubated at 3 temperatures (4, 23, and 37°C) for a total of 9 treatments with samples taken at 1, 24, 48, and 72 h post-inoculation (hpi). Each treatment was replicated in duplicate. At each time, a 1-g sample of each matrix type was suspended in 5 mL of Dulbecco modified eagle medium (DMEM; Gibco, Thermo Fisher) supplemented with 10% bovine calf serum (HyClone; Cytiva), ultraglutamine (2 mmol/L; Lonza), antibiotic–antimycotic (100,000 units/L of penicillin, 100 mg/L of streptomycin, and 0.25 mg/L of Fungizone; Gibco, Thermo Fisher), and Fungizone (2.5 mg/L of amphotericin B and 2.05 mg/L of sodium deoxycholate; HyClone, Cytiva) before immediate centrifugation and removal of supernatant. The supernatant was syringe-filtered and serially diluted from 10−1 to 10−6. Two aliquots of supernatant were frozen at −80°C for RT-qPCR or future analysis.
VI was conducted immediately after syringe filtration and serial dilution by applying 100 μL of the undiluted and diluted supernatant in triplicate to confluent MARC-145 cells in 96-well plates. The inoculated cells were incubated at 37°C for 48 h before being fixed and frozen at −20°C until staining. The VI plates were stained using SDOW17-A–conjugated antibodies and anti-mouse fluorescein isothiocyanate (FITC)-conjugated antibodies. The log10 TCID50/mL of each supernatant sample at each time for each treatment was calculated utilizing the Spearman–Kӓrber method and recorded.14,16
One aliquot of each sample was thawed, and 200 µL were removed for RNA extraction. Viral RNA was isolated from each sample (Quick-RNA viral kit; Zymo Research) according to the manufacturer’s instructions. One negative extraction control, consisting of all of the reagents except the sample, was included with each extraction. Extracted RNA was frozen at −80°C until assayed by RT-qPCR (EZ-PRRSV MPX 4.0 master mix and enzyme with ROX assay; Tetracore) according to the manufacturer’s instructions. Each RT-qPCR run included 4 dilutions of the stock virus to create a standard curve of PRRSV RNA and allow for the quantification of viral RNA in experimental samples. Analyzed values represent log10 of detected genomic copies.
Statistical analysis
VI data and RT-qPCR data were analyzed to determine the main effects of matrix type, temperature, and time in relation to the concentration of infectious PRRSV detected, with “sample” as the experimental unit and “tube” as a random effect and with a compound-symmetry structure to account for repeated measurement on each tube over time. Two replicates were conducted for each matrix–temperature combination.
The cell culture medium + PRRSV was aliquoted 3 times at each time to be run alongside each matrix time; “aliquot number nested within tube” was included in the model as a random intercept. One subsample for the cell culture medium + PRRSV after 24 h at 4°C was removed from analysis as a result of a studentized residual > 6.
All models were fit using PROC GLIMMIX (SAS Institute). The Kenward–Roger approach was used to approximate the degrees of freedom. Means were separated with the LSMEANS procedure, and the LINES option was used to determine means that differed significantly as determined by an F test. Results were considered significant at p ≤ 0.05.
Results
There was a matrix × temperature × hour interaction (p = 0.028) for the concentration of infectious PRRSV in the VI assay (Table 1). When inoculated into DMEM, elevated concentrations of infectious PRRSV were maintained from 1 h to 72 h (3.61 and 3.72 log10 TCID50/mL, respectively) at 4°C. The concentration of infectious PRRSV was similar (p > 0.05) from 1 to 72 hpi (2.94 and 2.22 log10 TCID50/mL, respectively) at 23°C. However, at 72 hpi, DMEM samples at 23°C had decreased (p < 0.05) concentrations of infectious virus compared to 4°C (2.22 and 3.72 log10 TCID50/mL, respectively). Finally, when incubated in DMEM at 37°C, the concentration of infectious PRRSV decreased (p < 0.05) over time from 3.5 log10 TCID50/mL at 1 hpi, to 1.94 log10 TCID50/mL at 24 hpi, and 0.78 and 0.42 log10 TCID50/mL at 48 and 72 hpi, respectively. At each time after 1 hpi, the sample incubated at 37°C contained a lower concentration of infectious PRRSV than when incubated at 4°C or 23°C.
Effects of matrix type, hour post-inoculation, and storage temperature on infectivity of PRRSV assessed by virus isolation.
DDGS = dried distillers grain with solubles; DMEM = Dulbecco modified eagle medium (Gibco; Thermo Fisher); FEED = a swine gestation diet with 78% corn, 17% soybean meal, and 0.5% fat; SBM = soybean meal. An initial tissue culture (1 mL of diluted virus inoculum, 105.4 TCID50/mL) of type 2 PRRSV strain P129 was inoculated into 10 g of SBM, DDGS, or FEED and stored at 4, 23, or 37°C for 72 h with sampling at 1, 24, 48, and 72 h post-inoculation. Samples were suspended in cell culture medium prior to centrifugation and titration of the supernatant over MARC-145 cells for virus isolation. Means lacking a common superscript (a,b,c,d,e,f,g,h,i,j,k,l,m,n) differ (p < 0.05).
Matrix × temperature × hour interaction; n = 2.
The contaminated SBM had a high concentration of infectious virus at all times when incubated at 4°C (2.50–3.00 log10 TCID50/mL). Compared to 4°C, PRRSV had decreased (p < 0.05) concentrations of infectious virus at 1 and 24 hpi (1.67 and 1.25 log10 TCID50/mL, respectively) at 23°C. No virus was detected at 48 hpi, and low concentrations of infectious virus (0.83 log10 TCID50/mL) were detected at 72 hpi. Finally, the concentration of infectious PRRSV in SBM decreased (p < 0.05) dramatically when incubated at 37°C. Infectious virus was only detected at 1 hpi (1.00 log10 TCID50/mL). PRRSV-contaminated SBM had (p > 0.05) concentrations of infectious virus similar to the contaminated DMEM at 4°C until 72 hpi. At this time and temperature, the concentration of infectious PRRSV in SBM decreased (p < 0.05) compared to the DMEM. The contaminated SBM incubated at 23°C had lower (p < 0.05) concentrations of infectious virus at all times than the DMEM at the same temperature, and no infectious virus was detected at 48 hpi. The PRRSV-contaminated SBM also had decreased (p < 0.05) concentrations of infectious virus compared to the DMEM incubated at 37°C, with low infectivity at 1 hpi (1.00 log10 TCID50/mL), and no infectivity detected after that time.
At 4°C, the contaminated DDGS started with moderate levels of infectious PRRSV at 1 hpi (2.33 log10 TCID50/mL) and decreased (p < 0.05) until 72 hpi (0.92 log10 TCID50/mL). The PRRSV-contaminated DDGS had low levels of infectious virus after incubation at 23°C. From 1 to 48 hpi, the concentration of infectious virus ranged from 0.42 to 0.58 log10 TCID50/mL, and no infectivity was detected at 72 hpi. No infective PRRSV was detected at any time when contaminated DDGS samples were incubated at 37°C. When incubated at 4°C, PRRSV-contaminated DDGS contained decreased (p < 0.05) concentrations of infectious virus at each time compared to the DMEM at the same temperature. The difference in PRRSV infectivity between the DDGS and DMEM was even wider at 23°C, with 1 hpi having infectious concentrations of 0.42 and 2.94 log10 TCID50/mL for the DDGS and DMEM, respectively. This difference in quantity of infectious virus became even larger with decreasing-to-no infectivity (p < 0.05) detected at 72 hpi in the DDGS and 2.22 log10 TCID50/mL in the DMEM. At 37°C, no infectious virus was detected in the DDGS; the DMEM still had measurable infectious PRRSV.
Finally, the PRRSV-contaminated FEED had low concentrations of infectious virus over time at each temperature. At 4°C, low concentrations of infectious virus were detected at 1, 24, and 72 hpi (0.75, 0.42, and 0.42 log10 TCID50/mL, respectively), with no infectivity detected in either of the 2 FEED replicates at 48 hpi. When incubated at 23°C, contaminated FEED maintained similar (p > 0.05) concentrations of infectious virus over time, with 1 and 24 hpi both having log10 TCID50/mL of 1.00 before decreasing (p < 0.05) to 0.00 at 48 and 72 hpi. Finally, 1 hpi was the only time that a contaminated FEED sample had detectable infectivity (0.58 log10 TCID50/mL) when incubated at 37°C. The FEED had lower concentrations (p < 0.05) of infectious virus detected at all time and temperature combinations than were detected in the DMEM under the same conditions.
When supernatant samples were analyzed by RT-qPCR, no significant interaction (p = 0.479) was observed for matrix × temperature × hour genomic copies (Table 2). However, a main effect of matrix type was significant (p = 0.032) for genomic copy number from RT-qPCR (Table 3). The recovery of genomic copies was highest from the contaminated DMEM (105.26 genomic copies/mL); DDGS had the lowest (102.48 genomic copies/mL); and SBM and FEED were intermediate (104.18 and 103.86 genomic copies/mL, respectively).
Interactive effects of matrix × temperature × hour on log10 genomic copies of PRRSV/mL of sample supernatant as detected by RT-qPCR.
DDGS = dried distillers grain with solubles; DMEM = Dulbecco modified eagle medium (Gibco; Thermo Fisher); FEED = a swine gestation diet with 78% corn, 17% soybean meal, and 0.5% fat; SBM = soybean meal. An initial tissue culture (1 mL of diluted virus inoculum, 105.4 TCID50/mL) of type 2 PRRSV strain P129 was inoculated into 10 g of SBM, DDGS, or FEED and stored at 4, 23, or 37°C for 72 h with sampling points at 1, 24, 48, and 72 h post-inoculation. Samples were suspended in cell culture media prior to centrifugation and RT-qPCR analysis of supernatant.
Matrix × temperature × hour interaction; n = 2.
Main effect of matrix type on log10 genomic copies of PRRSV/mL of sample supernatant as detected by RT-qPCR.
DDGS = dried distillers grain with solubles; DMEM = Dulbecco modified eagle medium (Gibco, Thermo Fisher); FEED = a swine gestation diet with 78% corn, 17% soybean meal, and 0.5% fat; SBM = soybean meal. An initial tissue culture (1 mL of diluted virus inoculum, 105.4 TCID50/mL) of type 2 PRRSV strain P129 was inoculated into 10 g of SBM, DDGS, or FEED and stored at 4, 23, or 37°C for 72 h with sampling at 1, 24, 48, and 72 h post-inoculation. Samples were suspended in cell culture medium prior to centrifugation and RT-qPCR analysis of supernatant with 6 replicates per matrix type and 4 repeated measures per replicate. Means lacking a common superscript (a,b) differ (p < 0.05).
Discussion
We found common trends among several matrix types over certain temperatures. When incubated at 4°C, the concentration of infectious PRRSV was maintained in DMEM, SBM, and FEED at each time. The highest concentration of infectious PRRSV was maintained in DMEM, followed closely by the SBM; FEED had a very low concentration of infectious virus across the whole time period while refrigerated. At both 23°C and 37°C, PRRSV concentrations in DDGS and FEED were similar and had very low or no infectious virus detected across all sampling times. SBM and DDGS both had decreased concentrations of infectious PRRSV with increasing temperature. A similar decrease in infectious virus concentration was not shown in DMEM until 37°C.
Several researchers have investigated the stability of PRRSV in cell culture medium over time and temperatures. Previous research demonstrated a decrease in the half-life of the virus from 140 h at 4°C to 20 h at 21°C, and just 3 h at 37°C. 1 Although we did not calculate half-life, the loss of infectious virus in DMEM at each temperature was consistent with these reported decreases in virus stability. The half-life of PRRSV also has been demonstrated to undergo an exponential decrease in cell culture medium with increasing temperature. 11
Some viruses, including PRRSV, were also found to be infectious in some feedstuffs 30 d after inoculation, whereas no infectivity was detected in other products4,5; SBM, DDGS, and a complete diet were included as matrices, and neither SBM nor DDGS had evidence of infectious virus when evaluated using VI, but had infectivity when evaluated using bioassay. This is particularly interesting for our study given that several of our sampling points for DDGS did not contain detectable PRRSV via VI, but may still have had infectivity potential when fed. The half-life estimates for individual viruses also changed based on the matrix type in the same, aforementioned studies.4,5
PRRSV in manure had a longer infectious half-life than controls incubated in culture medium at the same temperature. 13 Because of the substantial difference in moisture between feces and feed and the increased presence of fats or other molecules that could impact the viral envelope, it would be reasonable to expect a greater decrease of viral infectivity in feed and feedstuffs compared to other environmental samples given the presence of other reactive compounds, fat, and the complexity of the matrix. In particular, DDGS may have had a higher fat concentration than some of the other matrices that we investigated. Although we chose methods to reduce cytopathic effect, it is possible that the cells from the DDGS VI experienced cytotoxicity and this had an adverse effect on VI. We demonstrated that the decrease in infectivity can change based on which ingredient is being tested as well as the difference between single ingredients and complete feeds.
Although FEED had the lowest concentration of infectious virus over time and temperature, DDGS had the lowest recovery of viral genetic material. This could indicate that components of DDGS are able to bind the virus and prevent virions from suspending in the supernatant. One explanation is that residual ethanol in the DDGS could interact with the viral envelope and disrupt the virion, leading to degradation of viral RNA. 22 Of the 3 matrices, DDGS are the most acidic, and the lower pH might have created an issue with viral detection. Another explanation would be that fatty acids in the DDGS disrupted the viral envelope and aided in denaturing the RNA. The antiviral effects of fatty acids, particularly medium-chain fatty acids, and monoglycerides have been demonstrated when supplied via milk or feed additives; hence, DDGS may be able to disrupt the viral envelope.2,10,19 When palmitate (C16H32O2) was added to Sindbis virus, an enveloped alphavirus, a lipid-glycoprotein complex was formed 17 ; the antiviral effect of this fatty acid on Sindbis virus was to prevent the virus from budding from the host cell. A decrease in viral infectivity was demonstrated when free linoleic acid was added to a cell culture at the same time as Sindbis virus, and electron microscopy indicated that the viral envelope was destroyed by the fatty acid. 12 In the same study, the same reduction in infectivity did not occur for nonenveloped viruses. Also, lowering the pH of the environment increased the antiviral effect of fatty acids. 8 This is particularly relevant because the U.S. Grains Council suggests that the pH of DDGS is 3.6–5, 20 which may explain the decreased detection of PRRSV genetic material in our study.
The relation between PCR results and VI results is not easy to describe. Each of these tests measures a different aspect of the virus in question. Although PCR is sufficiently sensitive to measure even one virion in a mL of sample, a positive result indicates only that the RNA is present and not that the virion is viable. Infectivity assays, such as VI, are required to determine if the virus is able to bind to its target, replicate in the cell, and proliferate. A limitation of using a VI assay is the limit of detection of the assay. Although more expensive, bioassay could be useful to confirm negative infectivity results from a VI model because bioassay is a more sensitive test.
The fact that PRRSV infectivity was higher in DDGS than in FEED adds complexity to our results. This would indicate that the DDGS had more active, replicating PRRSV per mL compared to the FEED samples; however, the DDGS has a higher Ct value than the FEED, indicating less PRRSV RNA per mL in the DDGS than in the FEED. There was no difference in RT-qPCR detection of PRRSV RNA over time or temperature, whereas an infectivity assay demonstrated a decrease in infectivity over time and as temperature increased. 11 We saw a similar effect, with VI demonstrating a matrix × temperature × time effect on the infectivity of PRRSV, whereas the RT-qPCR data indicated only an impact of time on the detection of the virus. More investigation is warranted to understand the different responses between VI and RT-qPCR of feed matrices inoculated with PRRSV.
SBM and DDGS are both capable of harboring infectious PRRSV for up to and potentially beyond 3 d at cold temperatures, but this is less likely in FEED that contains ingredients such as the swine gestation diet that we tested (78% corn, 17% soybean meal, 0.5% fat). Reduced infectivity over time was more pronounced as temperatures increased. When comparing the RT-qPCR to VI analysis to detect and confirm live and infectious virus in feed material, we saw VI as being of intermediate outcome compared to the expensive animal inoculation assay. The low recovery of PRRSV RNA from contaminated DDGS warrants more investigation into sample collection and processing procedures to maximize recovery and detection of viral RNA.
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the Kansas State global food system seed grant program.
