College of Agriculture & Natural Resources Department of Veterinary Sciences Wyoming State Veterinary Laboratory 1174 Snowy Range Road Laramie, Wyoming 82070 (307) 766-9925 • fax (307) 721-2051 • wyovet.uwyo.edu Validation of a Novel Molecular Assay for Swine Brucellosis Meagan Soehn: Senior Honors Thesis Introduction With over 500,000 new cases annually, Brucellosis remains one of the world’s most common zoonosis.1 It is known as the world’s most common laboratory acquired infection.2 Species within the bacterial genus Brucella can all be characterized as facultative intracellular Gram- negative coccobacilli.1,3,4 Each species has its own host specificity though transmission from the original host to other species, including humans, has been documented.1 B. suis is the second most pathogenic Brucellosis species to humans in the world, and the most severe form found in the United States. This is due to the absence of B. melitensis, the most pathogenic species, in the United States.4 B. suis was first identified in the United States in 1914 within feral swine populations. Since then, sporadic cases have been seen in domestic swine when interaction has occurred between feral and domestic swine populations. Within a sounder an infection could then spread and infect 50-80% of the sounder in a few months.4 These feral populations act as a reservoir for B. suis in the same way that elk and bison populations do for B. abortus.1,5 Feral swine have also been shown to infect cattle with B. suis when their populations overlap and interact for a given time. Brucella organisms enter the host via ingestion, inhalation or through contact with skin abrasions. The hallmark of infection amongst food animals is spontaneous abortion. Pastures and water sources can become contaminated if an infected sow aborts her fetus in the area. This transmission method is then continued within the swine population, spreading it throughout the sounder. 5 There is a low infectious dose for humans. One model estimated that it would only take lung exposure of 10-100 organisms6. Concentrations of B.suis are highest in reproductive tissues and products of parturition. The placenta from an aborted fetus was shown to have 1010 organisms/g.6 Though it is rarely fatal in humans, it can be severely debilitating, and is often not diagnosed right away. This is due to its common symptoms that are often mistaken for other ailments. Symptoms include fever, joint pain, sweating, and constitutional symptoms that may affect a variety of human systems.1 Currently, there is no single test that is a direct measure, rapid, and safeguarded for the detection of swine brucellosis. Proper diagnosis of B. suis in a laboratory setting relies on tests that have a high sensitivity and are specific to B. suis. Current serological tests that look for an antibody reaction with the O-antigen polysaccharide (O/PS) in the bacteria’s lipopolysaccharides (LPS) can often cross react with other bacterial infections such as Yersinia enterocolitica giving a false positive test result.1,2 This testing method is also unable to distinguish between B. suis and B. abortus. This is dangerous when testing cattle that may have interactions with infected swine. The current “gold standard” to differentiate between the two is bacterial culture, followed by biochemical testing or sequencing. This testing must be done after an animal is euthanized and the incubation of the culture takes at least 7 days. Within this time there is risk for exposure to those in the laboratory as well as within the herd as the origin of the infection remains unknown during that time.2,3 In recent years feral swine have begun to expand into more areas of the United States. In 2014 it was reported that an increase from 17 to 38 states with feral hog populations was seen over the last 30 years.7 This increases the potential for interactions between feral swine, domesticated swine, cattle, and humans, and makes the need for high performing diagnostic tests quite apparent. Current research does not show an assay for Brucellosis that has high sensitivity, specificity, and maintains a low safety risk for laboratory personnel. Quantitative polymerase chain reaction (qPCR) is a simple yet effective enzymatic assay. It is able to detect and amplify a specific DNA sequence from a sample that may have a diverse pool of DNA. The amplified DNA is tagged with a fluorophore so that by measuring fluorescence levels one can quantify the amount of target DNA in a sample.8 This can be run from organ samples or other sources of DNA including blood or urine. To develop an assay the main component is creating primers that will target a DNA sequence unique to your pathogen.7 This is important for pathogens that have various species or strains as causative agents to allow for a high level of accuracy in the diagnosis process. PCR has been shown as an excellent diagnostic assay. Multiple studies have shown PCR to have higher sensitivity and specificity than culture methods for diagnosis of bacteria and viruses.9,10,11 The objective of this study is to develop a diagnostic test for distinguishing B. abortus and B.suis infections that has high sensitivity, specificity, and is low risk for laboratory personnel. Our hypothesis is that a qPCR is an assay that will accomplish the desired goals. Materials and Methods Bioinformatics/Bench Validation Utilizing Genious software we analyzed the Brucella suis genome to design primer and probe sets for a qPCR assay. 13 candidate sets were identified to be used. We selected our preferred set to be used and began to optimize our qPCR protocol for that set. We optimized our assay to find the prober primer concentration, probe concentration, annealing temperature, and amount of added magnesium chloride (MgCl2) that yielded the best results. G blocks (synthetic amplicon sequences) were ordered and a dilution curve was performed in order to find the limit of detection and quantification of the assay. A specificity experiment was conducted to asses if any serological cross-reactors would amplify on our preferred primer/probe set. This included additional Brucella spp. Sample Collection Samples were collected from feral hogs from southeastern Texas. Methods of capture included trapping, stand hunting, and receiving of already euthanized animals from locals. Tissues, aqueous solutions (including blood and or urine), and swabs were collected for each animal. While working proper decontamination of utensils was performed in between tissues and animals to reduce the risk of cross contamination. The tissue samples were immediately frozen at -80°C post-collection. Tissues remained frozen for up to one month. They were then shipped to the I.D.E.A Lab over dry ice. Samples were stored at -20°C until they were thawed for processing. Control animals were taken from Albany County, WY from a production facility that was undergoing depopulation for other disease issues. Samples were taken and processed from head tissues in an identical manner to those from the feral hogs. Tissue Processing All tissues once thawed were trimmed of excess fat. Working utensils were decontaminated in between tissues and animals to minimize cross-contamination risk. Post trimming, half of the tissues were cross-hatched and placed in either 50mL or 15mL high-speed conical tubes. The other half was placed in whirl-packs and sent to TVMDL for culture. The conical tubes contain garnet zirconia/silica beads and 1x PBS solution (5-20mL depending on conical size) to aid in tissue shearing. Conical tubes containing tissues were homogenized post-trimming using an Omni Bead Ruptor. The homogenized tissues were frozen at -20°C for 12-24 hours before extraction. DNA Extraction and Concentration All tissues, blood samples, and swabs collected from each animal were extracted using kits that were predetermined to work well for eluting Brucella spp. Swabs and tissues were extracted using an OMEGA kit and blood was extracted on an IBI kit. 400 µL of sample supernatant solution was obtained for extraction. Each set included a control to assess if environmental factors and/or kit reagents were introducing contamination into sample tubes. These controls would be concentrated and run on our qPCR before continuing with concentration of the extracted samples. Concentration of the eluted DNA was performed using a ZYMO DNA kit which had previously been shown to increase concentration and purity of Brucella spp. Using 250uL of eluted DNA from the extraction step a 10x concentration was performed leaving 25uL of DNA. Another control was used for these sessions to assess reagent or environmental contamination that may have been introduced to sample tubes. qPCR Reagents utilized for our qPCr included: nuclease free water, forward annealing primer, reverse annealing primmer, probe, MgCl2, and universal probe supermix. These reagents were prepared in a dedicated hood to eliminate environmental contamination. The sample template (1 µL) was added to the qPCR plate in another separate hood to eliminate environmental and cross- contamination. All qPCr assays were run in a separate room from that where extraction, concentration and qPCR preparation took place to further eliminate contamination possibilities. A positive control containing B. suis DNA from the National Veterinary Services Laboratory (NVSL) was included in every run. A non-template control of nuclease free water and controls from the concentration sessions were run with each qPCR to assess environmental or reagent- based contamination. ***All methods were approved by the Institutional Animal Care and Use Committee (IACUC) and the Institutional Biosafety Committee (IBC). Results Optimized Preferred Set The 104bp-82var- variant of the pan-B. abortus 82bP set was chosen as our preferred primer/probe set. Various ranges for primer concentration, probe concentration, MgCl2 concentration and annealing temperature were tested. The resulting protocol calls for 30 µM primer 30 µM probe and .5mM MgCl2 concentrations as well as an annealing temperature of 58°C. Specificity Experiment Our preferred primer/ probe set was ran with additional Brucella spp. including: B. meltensis, B. abortus field strain, vaccine strain 19, and vaccine strain RB51 as wells as known cross-reacting species including; O. anthropi, S. enterica serovar Urbana, V. cholerae, Y. enterocolitica O:9, and E. coli O157:H7. The experiment showed that only B. suis was amplified (Figure 1). Limit of Detection gBlocks® Gene Fragments (Integrated DNA Technologies) were designed based off the primer sequences of the preferred primer/probe set and were thus used as synthetic Brucella suis DNA to make a dilution series of 100 -108 target copies per microliter to generate a standard curve for this preferred primer/probe set (Figure 2). Each dilution was run in triplicate using one microliter of template (gBlocks® Gene Fragments) per well, combined with the previously mentioned qPCR reagents in the same concentrations and ran under the same conditions. Non-template DNA and a positive control (known Brucella suis DNA-biovar 1, strain 1330) were incorporated into the assay to determine if the qPCR reagents were contaminated and/or if environmental contamination was present, as well as to ensure the gBlocks® Gene Fragments were being recognized and amplified and that the assay was functioning appropriately. The purpose of generating standard curve was to calculate limit of detection (LoD- number of bacterium able to de detected as defined by, the lowest concentration that can be detected with a 95% confidence interval), limit of quantification (LoQ-number of bacterium able to be detected as defined by, the lowest concentration that can be accurately quantified), and efficiency (increase in number of target DNA sequences per cycle, with 100% being absolutely efficient) to assess the performance of this optimized assay protocol. It was found that this novel assay using this optimized primer/probe set was able to detect target DNA as low as 102 target copies per microliter where the LoD is 137 copies, the LoQ is 321 copies, and the efficiency is 78.89%. Hog demographics 213 feral hogs (Sus scrofa) were collected from Southeast Texas. This region was chosen as it had previously reported prevalence of 12% for B. suis.12 34 control hogs were collected from Southeast Wyoming, a non-endemic region. In total 3374 samples were collected. This included 634 swabs and urine, 510 blood samples, and 2230 tissue samples. The number of samples collected per animal varied based on sex and pregnancy status. (Figure 3) Both blood samples of serum and EDTA tubes were stored on ice post-collection for 24-48 hours. The serum tubes were sent to Texas Veterinary Medical Diagnostic Laboratory (TVMDL) for serologic testing using the Rose Bengal Test. The EDTA tubes were centrifuged and the blood fractions of plasma, buffy coat, and red blood cells were separated off to be extracted. Of the 34 control hogs 32 were female and 2 were male. 30 of the female hogs were adult (over 1 year of age) and the remaining 2 were juvenile. Both male control hogs were juvenile. There were 109 female feral hogs of the 213 collected. 57 were adult and 52 were juvenile. None of our control hogs tested as positive for Brucellosis on any of the diagnostic tests (serology, culture, or qPCR) Measures of Agreement, Sensitivity and Specificity To compare the agreement of the three diagnostic tests, Cohen’s Kappa was calculated on an animal to animal level. Comparison were made of qPCR vs Culture, qPCR vs Serology and Culture vs Serology. Culture and qPCR agreement were also calculated on a tissue to tissue matching basis. qPCR had a moderate agreement (K=0.41-0.60) to both serology and culture. Serology was highest at K=0.58, then animal to animal culture K=0.49 and finally tissue to tissue culture K=0.43. When the current validated tests of serology and culture, they had good agreement (K=0.61-0.80) of K=0.63 (Figure 4). Sensitivity and specificity were also calculated with a 95% confidence interval (Figure 5). In comparing the current diagnostic tests of serology and culture a lower specificity,73.3%, is seen. This lowering means that serology had more positives than culture and as culture is the gold standard these were deemed, “false positives.” Comparing qPCR to serology showed a lowered sensitivity of 62.2% in contrast to the 93.8% specificity. This indicates more animals were called positive by serological testing then by qPCR. Finally, in comparing culture to qPCR both on the animal and tissue level contrasting results were seen. The animal to animal comparison had a high sensitivity, 94.12%, but a lowered specificity,60.7%. On tissue to tissue matching there was a low sensitivity of 43.1% while specificity was much higher at 96%. These two comparisons indicate that PCR called more animals positive than culture and that even on animals that both tests called positive those results often came from differing tissues B.suis in a Household Pet In collaboration with TVMDL we found what we know to be the first reported case of B. suis in a household pet. A canine presented with similar clinical signs to B. canis including swelling of external genitalia. Figures Figure 1: qPCR from specificity experiment showing amplification of only B. suis Figure 2: Standard curve for the preferred primer/probe set (104 base pairs in length). The standard curve was generated using gBlocks® Gene Fragments (synthetic Brucella suis DNA) to make a dilution series of 100 – 108 target copies per microliter Male Female (Not Pregnant) Pregnant Female Mandibular LN Mandibular LN Mandibular LN Gastro Hepatic LN Gastro Hepatic LN Gastro Hepatic LN Medial Retropharyngeal LN Medial Retropharyngeal LN Medial Retropharyngeal LN Internal Iliac Internal Iliac Internal Iliac Tonsil Tonsil Tonsil Superficial Inguinal Superficial Inguinal Superficial Inguinal Nasal Swab Nasal Swab Nasal Swab Oral Swab Oral Swab Oral Swab Prostate Vaginal Swab Vaginal Swab Testes Uterus Uterus Epididymis Ovaries Ovaries Bulbourethral Kidney Kidney Seminal Vesicles Urine Placenta Kidney Serum Blood Tube Amniotic Fluid Urine EDTA Blood Tube Fetal Lung Serum Blood Tube Urine EDTA Blood Tube Serum Blood Tube EDTA Blood Tube Figure 3: Tissues collected for all hogs based on sex and pregnancy status, LN =Lymph node Test 1 Test 2 Agreement Serology qPCR K= 0.58 Moderate Agreement Culture Serology K=.63 Good Agreement Culture (Animal to Animal) qPCR K=.49 Moderate Agreement Culture (Tissue to Tissue) qPCR K=.43 Moderate Agreement Figure 4:Cohen’s Kappa calculations showing levels of agreements amongst diagnostic assays Test 1 (Standard) Test 2 (Clinical Test) Sensitivity and Initial Specificity Interpretation Serology qPCR Se= 62.2% False negatives in Sp=93.8% qPCR Culture Serology Se=91.7% False positives in Sp=73.3% serology Culture (Animal to qPCR Se= 94.12% False Positives in Animal) Sp=60.7% qPCR Culture (Tissue to qPCR Se=43.1% False Negatives in Tissue) Sp=96% qPCR Figure 5: Sensitivity and Specificity calculations (95% confidence interval) of diagnostic assays. Assay that is standard is assumed to be corrent for calculations. Discussion Our measures of sensitivity and specificity are very telling of each of our diagnostic tests. Serology tends to detect more positives than both culture and qPCR. However, as previously mentioned serological cross reactors are the likely cause for such false positives. Our test has high sensitivity when compared with culture on an animal to animal basis. This indicates that our test is able to accurately detect all disease positive animals that culture is. In addition, our test has found more positives than culture. This is what brings down our specificity, but it does not definitively indicate an ineffectiveness of our assay. This may be due to low bacterial burdens within the tissues of infected animals. Infected animals may only have 1-100 bacterium present.11 This makes it difficult for any diagnostic test to detect infection. However, culture is also hindered in that Brucella is a difficult bacterium to grow and is a slow growing organism. This means no growth may be seen even after a prolonged time. It is thought that only 30-50% of seropositive animals are culturable.2 These two variables mean that a sample may contain Brucella, but the organism is too few and grows to slow to out compete other microorganisms that are present. Based on what we know about Brucella, it would suggest that the discrepancies between culture and qPCR are true positives missed by culture. When looking at culture and qPCR on a tissue to tissue basis, our values were inversed from the animal to animal matching. This indicates that even on animals that both tests called positive, the two tests are not detecting the organism on the same samples. In previous studies we have seen that the bacteria localize within tissue which could explain some discordance between qPCR results and culture. Our protocol required that culture and qPCR were testing different sections of the same organs. IF the bacterium were localized to only one side of the tissue, this would explain why only one test was able to detect it. The main strength of our research is that we have developed a novel molecular assay that accurately detects B. suis DNA. We successfully collected many samples that harbored B. suis DNA. This was done with no evidence of cross contamination during field collection, trimming, extracting, concentrating, or qPCR plating. When compared with the current gold standard of culture our assay had near perfect sensitivity. Additionally, our assay had an increased detection ability than culture. Finally, we were also able to show for what we believe to be the first time of B. suis transmission to a Texas ranch dog. This has implications in minimizing all risks of B. suis infections spilling over into human populations. Our research was limited due to localization and low bacterial burden of Brucella bacteria. Our current limit of detection (LoD) is 137 copies of bacteria. As mentioned, infected animals may have a bacterial burden that is lower than this. This warrants development of a protocol for whole tissue homogenization in order to increase our DNA capture potential and overall assay sensitivity. Further optimization of our primer/probe set is also warranted to attempt to lower our LoD. Another limitation is that, as of now we have 17 culture positive and 28 culture negative hogs. The ideal amount for validation would be 30 of each. Not all hogs have been processed on culture so our total amount could change, but for now this limits our ability to validate our diagnostic test. However, based on our current results our qPCR has near perfect sensitivity, increased detection, is faster, and safer than current gold-standard testing with culture. Future directions for this project include the continuation of sending animals to TVMDL for culture. From there we may move to the development of a multiplex qPCR for the detection of B. suis and B. abortus at the same time for faster turnaround on serologic reactor livestock. Amplicon sequencing may be done to determine if there is B. suis DNA in our PCR positive but culture negative samples. This will help to strengthen the results of our qPCR despite discordance with culture. Finally, we plan to publish all of our data in a peer-reviewed journal. Work Cited [1-12] 1. Godfroid, J., et al., A “One Health” surveillance and control of brucellosis in developing countries: Moving away from improvisation. Comparative Immunology, Microbiology and Infectious Diseases, 2013. 36(3): p. 241-248. 2. Hull, N.C. and B.A. Schumaker, Comparisons of brucellosis between human and veterinary medicine. Infection ecology & epidemiology, 2018. 8(1): p. 1500846. 3. Weinstein, R.A. and K. Singh, Laboratory-acquired infections. Clinical Infectious Diseases, 2009. 49(1): p. 142-147. 4. Algers, B., et al., Porcine brucellosis (Brucella suis). EFSA Journal, 2009. 7(6). 5. Wyckoff, A.C., et al., Feral swine contact with domestic swine: a serologic survey and assessment of potential for disease transmission. Journal of wildlife diseases, 2009. 45(2): p. 422-429. 6. Bevins, S.N., et al., Consequences Associated with the Recent Range Expansion of Nonnative Feral Swine. BioScience, 2014. 64(4): p. 291-299. 7. Garibyan, L. and N. Avashia, Polymerase Chain Reacition. The Journal of Investigative Dermatology, 2013. 133(3): p. 4. 8. Lau, L., et al., Quantitative real‐time polymerase chain reaction versus culture: A comparison between two methods for the detection and quantification of Actinobacillus actinomycetemcomitans, Porphyromonas gingivalis and Tannerella forsythensis in subgingival plaque samples. Journal of clinical periodontology, 2004. 31(12): p. 1061- 1069. 9. Weinberg, G.A., et al., Superiority of Reverse-Transcription Polymerase Chain Reaction to Conventional Viral Culture in the Diagnosis of Acute Respiratory Tract Infections in Children. The Journal of Infectious Diseases, 2004. 189(4): p. 706-710. 10. Aly, B.H., et al., Polymerase Chain Reaction (PCR) Versus Bacterial Culture in Detection of Organisms in Otitis Media with Effusion (OME) in Children. Indian journal of otolaryngology and head and neck surgery : official publication of the Association of Otolaryngologists of India, 2012. 64(1): p. 51-55. 11. Gamazo, C., et al., Factors affecting detection of Brucella melitensis by BACTEC NR730, a nonradiometric system for hemocultures. Journal of clinical microbiology, 1993. 31(12): p. 3200-3203. 12. Olsen, S.C., Brucellosis in the United States: Role and significance of wildlife reservoirs. Vaccine, 2010. 28 Suppl 5: p. F73-6.