Development and use of a DNA microarray for the detection of enteritic pathogens in cattle and pigs

Enteritis is a very frequent cause of morbidity and mortality in young calves and pigs. They may be infected with 54 known pathogens, particularly in the first months of their life. Simultaneous infection with multiple pathogens occurs frequently and produces a synergistic effect in terms of the severity of clinical disease. In this study two microarray platforms (Agilent and Alere) were used to detect enteric pathogens. A total of 15993 probes were designed from viral, bacterial and parasitic sequences using four different software (UPS, Picky, eArray and GoArray). The probes for the Alere platform were assessed thermodynamically individually for secondary structure formation and hybridised to their complementary sequences in silico. Specificity and sensitivity testing was done with reference strains and porcine and bovine clinical samples and were performed with both platforms. The Alere ArrayTube platform holding 201 probes was used to identify viruses in reference and clinical samples. Among eight reference virus strains, five (PEDV, TGEV, PCV-2, BVDV and PPV) were identified correctly and of these PEDV, TGEV, PCV-2 were confirmed by PCR using designed primers Two viruses, P. rotavirus A and P. bocavirus were negative by array but were confirmed by PCR, rotavirus by designed primers and bocavirus by published primers. In two hybridisations using multiplex PCR products from two separate sets each of 5 mixed pathogens, the ArrayTube detected all viruses for one set and only one virus out of two in the other set. The specificity test using three non-enteric viruses showed high background noise for Bunyamwera virus and Schmallenberg virus, however only one probe cross-hybridised with Equine influenza virus. The sensitivity of this platform showed that it can detect an amount of 2.065 x 109 copy number (2 ng) of PCV-2 and 2.420 x 105 copy number (39 pg) of TGEV present in the sample. The results of hybridised reference viral strains and clinical samples showed that random amplification was more favourable for reference strain detection compared to specific amplification. However, specific amplification performed better for clinical samples. The Agilent microarray platform, comprising 44000 probes of enteric bacteria, viruses and parasites, was subjected to hybridisation of 12 reference strains for specificity testing (four viruses, seven bacteria and one parasite). All hybridised strains were correctly detected except P. rotavirus A which showed only 7 positive probes, however with high signal intensities. A high level of cross-hybridisation was observed with this platform due to the 16S rRNA and 18S rRNA probes as these two genes were amplified in their entirety prior to hybridisation and a high degree of similarity exists between of 16S rRNA and 18S rRNA among different strains of bacteria and parasites respectively. Hybridisation of PCR products to the Agilent platform from two sets of five multiplexed pathogens showed that all ten pathogens were correctly identified. The sensitivity results of this platform showed that it can detect 2.065 x 109 copy number of PCV-2, equivalent to a viral load of 2 ng. On the other hand the detection limit of E. coli F5 was comparable to the real-time PCR technique with a minimum of 2.089 x 108 copy number of E. coli fimbrial gene present in the sample. In bovine clinical samples, the Agilent microarray was able to identify the presence of E. coli F5 in two samples out of four tested. However, in porcine clinical samples, the array successfully detected all pathogens whose presence was confirmed by PCR. Mixed infections in porcine samples were also detected by microarray, where Clostridium difficile with its toxins (toxin A tcdA and binary toxin cdt), P. rotavirus and P. kobuvirus were detected simultaneously in one sample. It also detected the presence of the C. difficile clindamycin resistance gene (ermF) in another sample. In this study microarray technology has been shown to have the potential to detect mixtures of enteric pathogens in bovine and porcine faecal samples. It also has genotyping abilities for exploration of genetic variation. However, the sensitivity and specificity could be improved with more in silico assessments of designed probes. Eventually testing with a higher number of reference strains and clinical samples is necessary