Letter pubs.acs.org/journal/estlcu
Avian Influenza Virus RNA in Groundwater Wells Supplying Poultry Farms Affected by the 2015 Influenza Outbreak Mark A. Borchardt,*,† Susan K. Spencer,† Laura E. Hubbard,‡ Aaron D. Firnstahl,‡ Joel P. Stokdyk,‡ and Dana W. Kolpin§ †
Agricultural Research Service, U.S. Department of Agriculture, Marshfield, Wisconsin 54449, United States Wisconsin Water Science Center, U.S. Geological Survey, Middleton, Wisconsin 53562, United States § Illinois−Iowa Water Science Center, U.S. Geological Survey, Iowa City, Iowa 52240, United States ‡
S Supporting Information *
ABSTRACT: During the 2015 outbreak of highly pathogenic avian influenza virus (HPAI) on poultry farms in the midwestern United States, concern was raised about the potential for HPAI to contaminate groundwater. Our study objective was to evaluate the occurrence of HPAI in the groundwater supply wells on 13 outbreak-affected poultry farms in Iowa and Wisconsin. We sampled 20 wells, six waste-storage lagoons, and one pond. Three wells and one lagoon were positive for the matrix gene indicative of influenza A virus. Using a semi-nested qPCR assay specific to the H5 HPAI outbreak strain, one well was H5-positive, matching the outbreak virus hemagglutinin gene. Matrix gene-positive samples analyzed for avian influenza virus (AIV) by cell culture and embryonating egg culture were negative. Seven wells were positive by PCR for a poultry-specific parvovirus, thus providing corroborating evidence of virus transport pathways between poultry fecal wastes and groundwater. Our data suggest it is possible for AIV to be transported to groundwater, and during an outbreak, the potential for poultry farm wells to become contaminated with AIV should be considered.
■
INTRODUCTION During the 2015 outbreak of highly pathogenic avian influenza virus (HPAI) in the midwestern United States, the question arose whether avian influenza virus (AIV) could be transported downward from the land surface of affected poultry farms to groundwater where it could contaminate wells supplying water to nearby farms and households. We were asked by the Wisconsin State Public Health veterinarian, J. Kazmierczak, to investigate this question. The Midwest outbreak subtype was H5N2, and human infections were not reported.1 AIV has been detected in surface waters2 but not, to our knowledge, in groundwater. Other fecal-borne viruses are well-known contaminants of groundwater3 that have resulted in notable disease outbreaks4 and sporadic illness.5 Thus, it was reasonable to suspect that AIV could be transported through the subsurface and reach groundwater. Virus contamination of groundwater requires a proximate virus source (i.e., fecal wastes from infected hosts) and is favored by infiltrating rainfall or snowmelt that drives virus transport through the vadose zone to the underlying aquifer6 and by cool groundwater temperatures that slow virus inactivation.7 These conditions were in place in the upper Midwest during the 2015 HPAI outbreak. Commercial poultry operations produce large volumes of fecal wastes (1 million layers produce 90 kilotonnes of manure per day),8 and infected poultry shed high AIV concentrations in their feces (e.g., 107 EID50/g).9 The upper Midwest outbreak happened in the spring, the season when rains © XXXX American Chemical Society
are frequent, groundwater recharge is greatest (i.e., rising groundwater levels from infiltrating water), and groundwater temperatures are relatively cool. Our primary study objective was to determine HPAI occurrence in the groundwater supply wells of outbreak-affected poultry facilities. Secondarily, we sampled wastewater lagoons of outbreak-affected facilities as these storage structures could be a source for AIV-contaminated groundwater. Knowing whether AIV can possibly reach groundwater is the first step in an exposure assessment, which then could motivate further research to assess whether AIV in groundwater presents a risk of infection to poultry or other animals.
■
METHODS We sampled 20 groundwater supply wells, six wastewater lagoons, and one pond on 12 outbreak-affected poultry farms in Iowa and one in Wisconsin (Table 1). The farms produced eggs, pullets, and turkeys. The U.S. Department of Agriculture Animal and Plant Health Inspection Service coordinated farm access and trained sampling personnel in biosecurity and personal safety. Samples were collected 8 to 98 days after outbreak onset at a farm. Received: Revised: Accepted: Published: A
April 9, 2017 May 8, 2017 May 26, 2017 May 26, 2017 DOI: 10.1021/acs.estlett.7b00128 Environ. Sci. Technol. Lett. XXXX, XXX, XXX−XXX
Letter
Environmental Science & Technology Letters Table 1. Characteristics of Sampled Groundwater Supply Wells on Poultry Farms farm ID
aquifer type
1
Mississippian limestone
2
Mississippian sandstone
3
Mississippian sandstone Mississippian limestone Mississippian sandstone
4 5 6 7 8 9 10 11 12 13
Cretaceous sandstone Cretaceous sandstone NA Cretaceous sandstone sand and gravel Manson Mississippian limestone sandstone
well sample site ID
well depth (m)
well casing depth (m)
overburden thickness (m)
depth to groundwater (m)a
construction year
days between outbreak onset and sampling
1b 2c 3 1 2 1
90−136 130−149 145 64 78 142
47−136 59−148 52 44 45 NAd
16−41 46−56 47 40 43 30
17−18 15−19 18 27 26 4
2002−2003 1994−2007 2004 1994 1993 1998
45 40 38 28 28 27
1
41
40
37
21
1990
8
1 2 1
137 110 158
60 NA 59
54 NA 110
37 NA NA
2011 1990 2012
24 24 79
1 2 1 1
147 159 NA 128
147 54 NA 125
107 104 NA 55
84 86 NA NA
2001 2005 NA NA
51 51 38 49
1 1 2 1
24 77 96 151
24 49 55 49
69 55 54 13
NA 15 18 14
1950 1949 1991 1970
76 55 55 51
1 2
46 NA
45 NA
43 NA
16 NA
2014 NA
98 98
Static water level at time of well construction. bSample tap was supplied by five wells on the farm. cSample tap was supplied by four wells on the farm. dNA, no available information. a
AIV was concentrated in the field by dead-end ultrafiltration10 or glass wool filtration11 connected to farm water taps. AIV recovery by glass wool filtration is documented; AIV recoveries from Alaskan lake waters by glass wool filtration ranged from 8% to 43%.11 Dead-end ultrafiltration is easier to implement in the field and has been used to detect influenza A virus in river water samples,12 but AIV recovery efficiencies by ultrafiltration have not been reported to our knowledge. We employed both methods per well when logistically possible. Mean sample volumes were 722 L (range 426−1070 L) for well water and 242 L (range 48−636 L) for lagoon and pond water. qPCR analyses followed our established procedures for groundwater-borne viruses.5 Amplification was performed with the LightCycler 480 and LightCycler 480 Probes Master kit (Roche Diagnostics Mannheim, Germany). Reverse transcription (RT) was by random hexamers. Primers and hydrolysis probes were from Integrated DNA Technologies (Coralville, IA) and TIB Molbiol (Berlin, Germany), respectively. No-template controls for nucleic acid extraction, RT, and PCR steps were performed with every analysis batch, totaling 29, 27, and 29 controls, respectively, for the study. All were negative (i.e., no cycle of quantification value (Cq)), suggesting laboratory contamination was absent. All samples were evaluated for RT and PCR inhibition following previously established methods.13 Only two well water samples were RT inhibited, and the concentrates had to be diluted 1:5 with nuclease-free water before analysis. No samples were PCR inhibited. All samples were analyzed for the influenza A virus matrix gene using the widely adopted primers and hydrolysis probe developed by Spackman et al.14 Samples positive for the matrix gene were further analyzed to confirm the detected influenza A genomes were of the same H5 subtype as those AIV strains responsible for the outbreak by performing qPCR for North
American and Eurasian H5 subtypes following National Veterinary Services Laboratory (NVSL) protocol AVPRO1510.04. Matrix and hemagglutinin (HA) positive controls were gBlocks gene fragments (Integrated DNA Technologies). Standard curve parameters for the matrix gene: efficiency = 0.936 (PCR efficiency = 10−1/slope − 1), mean squared error = 0.0065. The lowest standard detected had a Cq value of 40, which is equivalent to 1.4 genomic copies per reaction. No standard curve was developed for the HA gene, and results are reported as positive or negative. Additionally, matrix gene positive samples were cultured for viable AIV using MDCK cells and embryonating chicken eggs, the latter performed at the U.S. Geological Survey National Wildlife Health Center. The cell culture AIV positive control was a low pathogenicity H5N2 subtype (NVSL A/TY/MN/36891551/81, MN Turkey). To provide corroborating evidence for AIV transport potential to groundwater, we analyzed all samples for a poultry-specific parvovirus.15 Three parvovirus assays were performed per sample following previously described procedures:15 (1) nested PCR for nonstructural protein 1, (2) nested PCR for virion proteins 1 and 2 (VP1/VP2), and (3) qPCR for VP1/VP2. PCR amplicons were viewed by gel electrophoresis; the identity of all PCR amplification products were confirmed by DNA sequencing. Standard curve parameters for the VP1/VP2 qPCR: efficiency = 1.001, mean squared error = 0.0225, Cq of the lowest standard detected = 40 (equivalent to 1.4 genomic copies per reaction). All no-template controls for nucleic acid extraction (n = 10), PCR (n = 10), and qPCR (n = 10) were negative.
■
RESULTS AND DISCUSSION Three wells were positive for the influenza A virus matrix gene (Table 2). One lagoon sample was matrix gene positive. The B
DOI: 10.1021/acs.estlett.7b00128 Environ. Sci. Technol. Lett. XXXX, XXX, XXX−XXX
Letter
Environmental Science & Technology Letters
Table 2. Groundwater Supply Wells, Lagoons, and Pond on Poultry Farms That Were Positive for Avian Influenza RNA or Poultry Parvovirus DNA avian influenza qPCR
poultry parvovirus
farm ID
well sample site ID or lagoon/pond
concentration methoda
matrix gene (genomic copies/L)
H5 modified primers
H5 semi-nested primers
nonstructural protein PCR
VP1/VP2 PCR
VP1/VP2 qPCR (genomic copies/L)
1 1 2 4 7 7 7 7 9 12 1 12 12 8
1 3 2 1 1 1 2 2 1 1 lagoon 1 lagoon 1 lagoon 2 pond
UF GW GW GW GW UF GW UF GW UF GW PEG PEG GW
NDb ND ND 1.8 ND ND 1.8 0.9 3.8 ND 334 ND ND ND
ND ND ND ND ND ND ND ND ND ND positive ND ND ND
ND ND ND positive ND ND ND ND ND ND ND ND ND ND
positive positive positive ND positive positive positive positive positive positive positive positive positive positive
positive positive positive ND positive positive positive positive positive positive positive positive positive positive
ND 0.8 ND ND 0.8 0.5 ND ND ND ND 45 962 2166 ND
GW, glass wool filtration; UF, dead end ultrafiltration; PEG, concentration of 3 L sample with polyethylene glycol (PEG) following previously established methods.5 bND, not detected. a
for the ultrafiltration sample is considered suspect because the ultrafiltration equipment blank performed at this site was also positive. Glass wool filtration uses a separate apparatus from ultrafiltration but no glass wool blanks were performed. All matrix gene positive samples analyzed by following the NVSL protocol were negative for the H5 gene, the subtype responsible for the outbreak. However, at the same time as we were performing these analyses, other research teams were depositing in GenBank the H5 gene sequences of HPAI isolates from outbreak-affected farms, including the same farms we had sampled. Examination revealed there were four base pair mismatches with the NVSL primers (Figure 2), which likely reduced qPCR amplification efficiency and increased the likelihood of false-negative results as virus concentrations in environmental samples are typically low. We modified the NVLS H5 primers to match the consensus H5 sequence from 12 HPAI outbreak-isolates from Iowa, Minnesota, and Arkansas (Figure 2) and reanalyzed the matrix gene positive samples. With these new primers, one sample, the lagoon sample, was H5 positive (Figure 1C). To increase the analytical sensitivity of the H5 assay, we designed a new semi-nested qPCR based on the 12 HPAI outbreak-isolates (Figure 2). Primer concentrations and thermal cycling parameters are presented in the Supporting Information. Using the semi-nested qPCR, the matrix gene positive well at Farm 4 was positive for the H5 gene (Table 2, Figure 1C). This result suggests that the outbreak HPAI strain had been transported to the underlying groundwater on this farm. Two samples, the lagoon and Farm 9 well, exhibited cytopathic effect in the first and second passages of the cell cultures, but subsequent qPCR of the culture lysates were AIV-negative. The cytopathic effect must have resulted from the growth of virus types in the groundwater other than AIV. The egg cultures were also negative for AIV. Thus, the AIV RNA detected in the wells by qPCR could have been from inactivated virions or the culture methods may not be as analytically sensitive as qPCR. The poultry-specific parvovirus was detected in seven wells out of the 20 wells sampled (Table 2). This parvovirus has been found in nearly 80% of chicken and turkey samples from USA farms.16 The pond and three of the six lagoons were also
amplification curves were unambiguous, indicating true positives (Figure 1A and B). Both glass wool filtration and ultrafiltration samples from Farm 7, Well 2 were positive. However, the result
Figure 1. qPCR amplification curves for samples and positive controls analyzed for the influenza A virus matrix gene (A and B) and HA gene (C). Amplification curves are for well samples unless indicated. Samples were concentrated by glass wool filtration unless indicated by “UF” (ultrafiltration). C
DOI: 10.1021/acs.estlett.7b00128 Environ. Sci. Technol. Lett. XXXX, XXX, XXX−XXX
Letter
Environmental Science & Technology Letters
Our data likely underestimate AIV occurrence in groundwater because the study wells were sampled only once and usually many days after an outbreak began (Table 1) and the poultry depopulation completed. Viruses in groundwater are transient,6 and after a contamination event, concentrations decline from inactivation, dilution by well pumping, reduction of infected hosts, changes in the contamination source, and changes in virus transport factors. Perhaps not coincidently, the farm with the H5positive well had the least time between outbreak onset and groundwater sampling, only 8 days. The scope of our study was limited to evaluating the occurrence of AIV in groundwater. The study was not designed to evaluate whether AIV is transmitted to poultry via groundwater. Our data suggests AIV can contaminate groundwater wells on poultry farms, and this observation may motivate additional research to evaluate the importance of groundwater as an AIV transmission route. Waterborne AIV transmission to poultry has been demonstrated,18 and groundwater was suspected as a route for AIV exposure in turkeys.19 Although peak concentrations of AIV in groundwater may be brief, well water usage at that time could result in exposures. For example, rinsing with well water as part of biosecurity or decontamination procedures could inadvertently transport AIV to another location or reinfect a facility. If AIV contamination of a poultry farm’s groundwater wells is suspected, taking precautions such as point-of-use disinfection could mitigate potential exposures.
Figure 2. Primers and hydrolysis probe for the standard National Veterinary Services Laboratory H5 assay (Eurasian strains) and the modified primers and semi-nested primer developed for this study. The consensus sequence shown is from 12 2015 HPAI isolates from Iowa, Minnesota, and Arkansas, GenBank accession numbers KT002473.1, KT002481.1, KT002489.1, KT002497.1, KT002505.1, KT002513.1, KT002528.1, KT762924.1, KT762948.1, KR234022.1, KR234006.1, and KR233982.1.
■
parvovirus positive. DNA sequencing confirmed that PCR amplicons matched the poultry parvovirus genetic sequence (genomic regions for nonstructural protein 1, 100% identity; and virion protein 1 and virion protein 2, 99% identity; BLAST search). These data provide strong evidence for virus transport pathways between poultry fecal wastes and the groundwater supply wells of the outbreak-affected farms. The poultry-specific parvovirus may have been detected more frequently than AIV because of differences in analytical sensitivity of the assays (the two PCR assays for parvovirus were nested), transport potentials through the vadose zone, or survival times in groundwater. In surface water at 10 °C, the time needed for a 90% reduction in the initial concentration of three subtypes of AIV ranged from 10 to 14 days.17 Survival times for AIV and parvovirus in groundwater are not available to our knowledge. Two equipment blanks were performed during the study, one at Farm 1 and the other at Farm 7, both for ultrafiltration. The Farm 1 blank was negative for all AIV assays. The Farm 7 blank was positive for the matrix gene but negative for the H5 modified and semi-nested qPCR assays. Both blanks were negative for all three poultry parvovirus assays. During sampling, the tap is flame-sterilized, the ultrafiltration and glass wool filtration equipment is sterile and preassembled, and from the tap to the filtrate outlet the system is closed. We believe the Farm 7 blank was positive because it was conducted in a recently depopulated hen house and the filtered sterile water was held in an open 25 L carboy. The Farm 1 blank was conducted in the farm’s well pump house. In all, three wells showed evidence of AIV RNA, suggesting that avian influenza virus can be transported to groundwater. The detected AIV was likely HPAI; one well was positive for the H5 subtype matching the HA gene sequence of the outbreakisolates. Moreover, on these outbreak-affected farms with 10,000 to 1,000,000 of HPAI infected poultry, just by its high abundance in the farm environment, HPAI H5 was the most likely subtype to be found in the underlying groundwater.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.estlett.7b00128. Primer concentrations and thermal cycling conditions for the semi-nested qPCR for the H5 gene. (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Phone: (715) 387-4943. Fax: (715) 384-9157. ORCID
Mark A. Borchardt: 0000-0002-6471-2627 Dana W. Kolpin: 0000-0002-3529-6505 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Dr. Hon Ip, U.S. Geological Survey National Wildlife Health Center, for conducting embryonating egg cultures and providing the AIV cell culture positive control, the USDA Animal and Plant Health Inspection Service for their assistance with identifying potential poultry operations for sampling, the U.S. Geological Survey Toxic Substances Hydrology Program for providing a portion of funding, and Kimberlee Barnes of the U.S. Geological Survey, Iowa Water Science Center, for her assistance with sample collection. We also thank poultry producers for permitting access to their farms. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.
■
REFERENCES
(1) Hvistendahl, M. Enigmatic bird flu strain races across the U.S. Midwest. Science 2015, 348, 741−742.
D
DOI: 10.1021/acs.estlett.7b00128 Environ. Sci. Technol. Lett. XXXX, XXX, XXX−XXX
Letter
Environmental Science & Technology Letters (2) Stallknecht, D. E.; Goekjian, V. H.; Wilcox, B. R.; Poulson, R. L.; Brown, J. D. Avian influenza virus in aquatic habitats: what do we need to learn? Avian Dis. 2010, 54, 461−465. (3) Hynds, P. D.; Thomas, M. K.; Pintar, K. D. M. Contamination of groundwater systems in the US and Canada by enteric pathogens, 1990−2013: A review and pooled-analysis. PLoS One 2014, 9, e93301. (4) Wallender, E. K.; Ailes, E. C.; Yoder, J. S.; Roberts, V. A.; Brunkard, J. M. Contributing factors to disease outbreaks associated with untreated groundwater. Groundwater 2014, 52, 886−897. (5) Borchardt, M. A.; Spencer, S. K.; Kieke, B. A., Jr.; Lambertini, E.; Loge, F. J. Viruses in non-disinfected drinking water from municipal wells and community incidence of acute gastrointestinal illness. Environ. Health Persp, 2012, 120, 1272−1279. (6) Bradbury, K. R.; Borchardt, M. A.; Gotkowitz, M.; Spencer, S. K.; Zhu, J.; Hunt, R. J. Source and transport of human enteric viruses in deep municipal water supply wells. Environ. Sci. Technol. 2013, 47, 4096− 4103. (7) John, D. E.; Rose, J. B. Review of factors affecting microbial survival in groundwater. Environ. Sci. Technol. 2005, 39, 7345−7356. (8) Schmitt, M.; Rehm, G. Fertilizing Cropland with Poultry Manure; AG-FO-5881-C; Minnesota Extension Service, University of Minnesota, 1992 . (9) Lu, H.; Castro, A. E.; Pennick, K.; Liu, J.; Yang, Q.; Dunn, P.; Weinstock, D.; Henzler, D. Survival of avian influenza virus H7N2 in SPF chickens and their environments. Avian Dis. 2003, 47, 1015−1021. (10) Smith, C. M.; Hill, V. R. Dead-end hollow-fiber ultrafiltration for recovery of diverse microbes from water. Appl. Environ. Microbiol. 2009, 75, 5284−5289. (11) Millen, H. T.; Gonnering, J. C.; Berg, R. K.; Spencer, S. K.; Jokela, W. E.; Pearce, J. M.; Borchardt, J. S.; Borchardt, M. A. Glass wool filters for concentrating waterborne viruses and agricultural zoonotic pathogens. J. Visualized Exp. 2012, 61, e3930. (12) Heijnen, L.; Medema, G. Surveillance of influenza A and the pandemic influenza A (H1N1) 2009 in sewage and surface water in the Netherlands. J. Water Health 2011, 9, 434−442. (13) Gibson, K. E.; Schwab, K. J.; Spencer, S. K.; Borchardt, M. A. Measuring and mitigating inhibition during quantitative real time PCR analysis of viral nucleic acid extracts from large-volume environmental water samples. Water Res. 2012, 46, 4281−4291. (14) Spackman, E.; Senne, D. A.; Myers, T. J.; Bulaga, L. L.; Garber, L. P.; Perdue, M. L.; Lohman, K.; Daum, L. T.; Suarez, D. L. Development of a real-time reverse transcriptase PCR assay for type A influenza virus and avian H5 and H7 hemagglutinin subtypes. J. Clin. Microbiol. 2002, 40, 3256−3260. (15) Carratalà, A.; Rusinol, M.; Hundesa, A.; Biarnes, M.; RodriguezManzano, J.; Vantarakis, A.; Kern, A.; Suñen, E.; Girones, R.; Bofill-Mas, S. A novel tool for specific detection and quantification of chicken/ turkey parvoviruses to trace poultry fecal contamination in the environment. Appl. Environ. Microbiol. 2012, 78, 7496−7499. (16) Zsak, L.; Strother, K. O.; Day, J. M. Development of a polymerase chain reaction procedure for detection of chicken and turkey parvoviruses. Avian Dis. 2009, 53, 83−88. (17) Nazir, J.; Haumacher, R.; Ike, A.; Stumpf, P.; Böhm, R.; Marschang, R. E. Long-term study on tenacity of avian influenza viruses in water (distilled water, normal saline, and surface water) at different temperatures. Avian Dis. 2010, 54, 720−724. (18) Jones, J. C.; Sonnberg, S.; Webby, R. J.; Webster, R. G. Influenza A(H7N9) virus transmission between finches and poultry. Emerging Infect. Dis. 2015, 21, 619−628. (19) Halvorson, D. A.; Kelleher, C. J.; Senne, D. A. Epizootiology of avian influenza: effect of season on incidence in sentinel ducks and domestic turkeys in Minnesota. Appl. Environ. Microbiol. 1985, 49, 914− 919.
E
DOI: 10.1021/acs.estlett.7b00128 Environ. Sci. Technol. Lett. XXXX, XXX, XXX−XXX
