Search

Additional file 2. Detailed results from the statistical models used to analyse the serology (ELISA) data.

Additional file 1. Detailed results from the statistical models used to analyse the infection status (PCR) data.

Additional file 3. Optical density values for M. hyopneumoniae-specific antibodies in serum of sows and 2-day-old piglets. Farm A: endemically infected with M. hyopneumoniae; Farm B: free of M. hyopneumoniae. Optical density (OD) investigated with a commercial blocking ELISA (IDEIA™ Mycoplasma hyopneumoniae EIA kit, Oxoid Limited, Hampshire, UK). On farm A, sows were vaccinated against M. hyopneumoniae at 6 and 3 weeks before farrowing and blood samples were taken before the first vaccination (pre vacc.) and at the time of farrowing (post vacc.), on farm B blood was sampled of the sows at the time of farrowing. On farm A 47 piglets were sampled (24 cross-fostered and 23 non-cross fostered) and on farm B 24 piglets were sampled. Samples were considered positive if the OD of the sample was lower than 50% of the average OD of the buffer control.

Additional file 4: Table S2. Sequencing summary of the M. bovis MiSeq (Illumina) sequencing run with resulting genome coverages after quality-filtering and adapter-clipping.

Additional file 1.

Additional file 1: Figure S1. In-depth analyses of de novo consensus genomes between the Canu and Flye assemblers, with or without four rounds of Racon. While both Canu and Flye result in a similar number of mismatches (per 100 kbp), significantly more indels (per 10 kbp) are observed when implementing the Flye assembler in the bioinformatics workflow. Decreased Flye performance is suggested to lie within decreased insertion accuracies as compared to Canu. Implementation of four rounds of Racon allows increased Flye insertion Q-scores when custom-pg45 basecalling is used. However, Canu without four rounds of Racon still results in an overall better performance.

Additional file 3: Table S1. Sequencing summary of multiplexed M. bovis and E. coli sequencing runs (48 h) on MinION R9.4.1 flow cells and resulting coverages after qcat and NanoFilt filtering and trimming..

Additional file 2: Table S3. Sequencing summary of single M. bovis sequencing runs (24h) on Flongle R9.4.1 flow cells and resulting coverages after qcat and NanoFilt filtering and trimming.

Table S1. Broncho-alveolar lavage fluid total nucleated cell count and differential cell counts according to ultrasonographic lesion score (ULS) based on 352 group-housed calves. (DOCX 26 kb)

Table S2. Descriptives of broncho-alveolar fluid cellular characteristics from 352 group-housed calves. (DOCX 37 kb)

This document contains all supplementary material associated with this paper

Salmonellose bij de mens wordt vaak veroorzaakt door Salmonella enterica subspecies enterica serovar Typhimurium (Salmonella Typhimurium). De ziekte veroorzaakt door deze bacterie is voornamelijk geassocieerd met het eten van varkensvlees. Varkens die geïnfecteerd zijn met Salmonella Typhimurium zijn meestal symptoomloze dragers. De mechanismen die Salmonella Typhimurium gebruikt om deze dragerdieren blijvend te infecteren zijn niet gekend. Het doel van deze thesis was om inzicht te verkrijgen in de mechanismen die Salmonella Typhimurium gebruikt om varkens te kolonizeren en er in te persisteren. In het eerste hoofdstuk van deze thesis werd een Salmonella Typhimurium stam geselecteerd die een persisterende infectie bij varkens kan veroorzaken. Hiervoor werd een veldstam, geïsoleerd uit een persistent geïnfecteerd varken, vergeleken met een standaard laboratorium stam die vaak wordt gebruikt voor onderzoek in muizen. De laboratorium stam was virulenter in muizen, maar de varkensstam was efficiënter in het veroorzaken van een persisterende infectie bij varkens. Deze varkensstam werd gekozen om te gebruiken in verdere experimenten. In het tweede hoofdstuk werden verschillende in vitro en in vivo modellen op punt gesteld om de pathogenese van Salmonella Typhimurium infecties bij varkens te onderzoeken. In het derde hoofdstuk werd de rol van verschillende virulentiegenen van Salmonella in de pathogenese van Salmonella Typhimurium infecties bij het varken onderzocht. In een eerste reeks experimenten werd de rol nagegaan van de genen die gelegen zijn op het Salmonella Pathogeniciteitseiland 1 (SPI-1). Deze genen waren essentieel voor Salmonella Typhimurium om varkensdarmcellen en macrofagen te kunnen binnen dringen. Alle SPI-1 mutanten waren sterk verzwakt in hun vermogen om diarree te veroorzaken. Zowel vroege als late celdood werd gezien in de macrofagen, maar enkel de vroege celdood bleek SPI-1 afhankelijk. Wanneer varkens peroraal geïnfecteerd werden met een combinatie van de Salmonella Typhimurium veldstam enerzijds en een SPI-1 deletiemutant anderzijds, bleek dat de mutant stam sterk verzwakt was in het invaderen en kolonizeren van de darmen, maar niet in het kolonizeren van de tonsillen. In een tweede reeks experimenten werd de rol nagegaan van de genen die gelegen zijn op het Salmonella Pathogeniciteitseiland 2 (SPI-2). Hierbij werd gebruik gemaakt van een mutant in het ssrA gen. Deze mutant was minder goed in staat om in vitro in macrofagen te vermeerderen en om de organen te kolonizeren van varkens die intraveneus geïnfecteerd werden. Bij biggen die oraal geïnoculeerd werden met de ssrA mutant stam verliep de infectie evenwel gelijkaardig in vergelijking met biggen die geïnoculeerd waren met de wild type stam. In een derde reeks experimenten, werd het belang van het fibronectine bindend eiwit ShdA in het ontstaan van een persisterende infectie onderzocht. Alhoewel dit gen bij muizen recent geïdentificeerd is als een belangrijke factor in de intestinale kolonizatie en persistentie in muizen, kon dit in deze experimenten niet worden bevestigd voor Salmonella Typhimurium infecties bij varkens. De resultaten van deze thesis tonen aan dat zowel de gastheermodellen als de Salmonella stammen met zorg gekozen moeten worden om relevant onderzoek te kunnen doen naar de pathogenese van Salmonella infecties. Tevens kan geconcludeerd worden dat SPI-1 afhankelijke invasie cruciaal is voor de kolonizatie van de darmen, maar niet van de tonsillen. De bijdrage van SPI-2 en shdA tot de persistentie van Salmonella Typhimurium in varkens is veel kleiner dan beschreven voor muizen.

Additional file 1. Results of the M. hyopneumoniae-specific antibodies measured at different time points in serum and in BAL fluid. Pigs were prime-boost vaccinated on D0 and D14 with three different experimental M. hyopneumoniae bacterins (Lipo_DDA:TDB, PLGA_TLR, SWE_TLR), challenge infected on D28–29 and euthanized on D56. M. hyopneumoniae-specific antibodies were determined by the IDEIA™ Mycoplasma hyopneumoniae EIA kit (Oxoid Limited, Hampshire, UK) and by indirect in-house ELISAs. For the in-house ELISAs, NetOD-values were calculated by subtracting the OD-value of the blank from the OD-value of the sample. BAL, bronchoalveolar lavage; NCG, non-challenge control group (PBS-injected, non-challenge infected); PCG, PBS-injected control group (PBS-injected, challenge infected); OD, optical density.

Additional file 1. Results of the M. hyopneumoniae-specific antibodies measured at different time points in serum and in BAL fluid. Pigs were prime-boost vaccinated on D0 and D14 with three different experimental M. hyopneumoniae bacterins (Lipo_DDA:TDB, PLGA_TLR, SWE_TLR), challenge infected on D28–29 and euthanized on D56. M. hyopneumoniae-specific antibodies were determined by the IDEIA™ Mycoplasma hyopneumoniae EIA kit (Oxoid Limited, Hampshire, UK) and by indirect in-house ELISAs. For the in-house ELISAs, NetOD-values were calculated by subtracting the OD-value of the blank from the OD-value of the sample. BAL, bronchoalveolar lavage; NCG, non-challenge control group (PBS-injected, non-challenge infected); PCG, PBS-injected control group (PBS-injected, challenge infected); OD, optical density.

Mycoplasma bovis is a leading, primary cause of pneumonia, arthritis, otitis and mastitis in cattle, resulting in impaired animal welfare and huge economic losses in all cattle sectors worldwide. This small bacterium lost its cell wall and several physiological mechanisms through evolution, whereupon it acquired inherent resistance against many conventional antimicrobials (e.g. penicillines, cephalosporines, sulfonamides, ..). Next to this natural resistance, M. bovis may acquire resistance against other antimicrobials as well. Currently, isolation and identification of M. bovis by culture takes 1-2 weeks, and subsequent antimicrobial susceptibility testing is currently not performed in routine diagnostics. No standard protocol is available and the lack of clinical breakpoints limits the translation of in vitro results to clinical outcome predictions of antimicrobial treatment. At a higher price, faster identification is possible with PCR (2 days). Although diagnostic accuracy of PCR is expected to be higher than culture, scientific information on this matter is limited. To be able to control M. bovis and start appropriate antimicrobial treatment immediately, there is a great need for rapid and reliable diagnostic tools for this pathogen. However, next to control, prevention of M. bovis spreading into/within the herd is also very important. How M. bovis is exactly transmitted, and whether there are specific M. bovis strains associated with antimicrobial resistance or sectors, has not been elucidated yet. Key factors for successful control and prevention are the formulation of specific biosecurity protocols and guidelines targeted to M. bovis. To achieve this, a rapid diagnosis of infected or carrier animals and better insights into the spread of M. bovis between herds, sectors, and countries are needed. Therefore, the general aim of this thesis was to develop new Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry (MALDI-TOF MS) and nanopore sequencing based diagnostic methods for rapid identification, strain typing, and antimicrobial susceptibility testing of M. bovis, and to apply those methods on Belgian field samples, gaining better insight into the epidemiology of M. bovis. In the general introduction (Chapter 1) a literature overview is provided, presenting current state-of-the-art on disease course, risk factors and treatment of M. bovis. Subsequently, the many different diagnostic techniques available for identification, strain typing, and susceptibility testing are described. Next to the existing techniques, more innovative techniques, such as MALDI-TOF MS and nanopore sequencing and their potential as rapid diagnostic methods are explained. In the first experimental study different methods were explored to identify M. bovis cultures grown on solid medium with MALDI-TOF MS (Chapter 3.1). The most straight-forward method, being the direct transfer method, is broadly applied for most bacteria, but faced several problems for M. bovis identification. In this study, these problems were better identified and it was shown that medium-related peaks (mostly obtained from horse serum and colistin) can result in false positive Mycoplasma alkalescens and Mycoplasma arginini identification. Unfortunately, it was not possible to obtain a more reliable direct transfer protocol. Therefore, in Chapter 3.2 the identification of M. bovis with MALDI-TOF MS from liquid medium was further explored and optimized. Here it was shown that identification was possible within 24 hours after inoculation of one colony from a solid medium into liquid medium. Supplementing pleuropneumonia-like organism broth (PPLO-broth) with pyruvate prolonged the possibility of M. bovis identification to at least 120 hours after inoculation. Also, supplementation with antimicrobials prevented overgrowth with other bacteria, and did not influence the identification score. Although with the previous two methods, a step towards more rapid identification of M. bovis was set, prior isolation of M. bovis from any sample is still necessary and could easily take 5-10 days. Therefore, methods to identify M. bovis directly from bronchoalveolar lavage fluid (BALf) with MALDI-TOF MS (Chapter 3.3) and nanopore sequencing (Chapter 3.4) were developed and validated in a Bayesian latent class model on 104 and 100 BALf from calves, respectively. It was possible to identify M. bovis with MALDI-TOF MS within 2-3 days with a sensitivity and specificity of 86.6% (CI95%: 69.4-97.6%) and 86.4% (76.1-93.8%), respectively. While sensitivity and specificity of nanopore sequencing were 77.4% (58.6-92.3%) and 97.3% (91.1-99.7%), respectively. Also when 5 BALf were pooled, both methods were still reliable, and therefore very cost-effective possibilities. In addition, the in Belgium widely used selective-indicative agar method based on lipase-activity, which was never validated before using a large number of field samples, showed a sensitivity of 70.5% (52.1-87.1%) and specificity of 93.9% (85.9-98.4). All three methods are useful in routine laboratories, depending on the diagnostic needs of the applicant. Currently the prevalence of M. bovis in the Belgian dairy and beef sector is estimated at 30%, whereas 100% of the veal calf herds tested positive. Together with the high antimicrobial use in the veal sector, the question has been raised whether there is a possible reservoir of multi-resistant and sector-specific M. bovis strains in this sector, as previously shown for other respiratory bacteria. To better understand the molecular epidemiology and genetic relatedness of different M. bovis isolates, the full genome of 100 Belgian M. bovis isolates collected from dairy, beef and veal herds was obtained using nanopore sequencing (Chapter 4). A single nucleotide polymorphism (SNP) analysis was performed and the phylogenetic tree showed five separate genomic clusters of M. bovis isolates and one outlier circulating in Belgium between 2014 and 2019. No sector-specific isolates and no association with spatial location in Belgium were identified. At world-scale, the Belgian M. bovis isolates clustered together with European, American and Israeli strains. These results contribute to emphasizing the importance of purchase protocols and biosecurity to prevent M. bovis from entering the country or herd. In Chapter 5.1, antimicrobial susceptibility testing of 141 M. bovis isolates retrieved from Belgian dairy, beef and veal calf herds was performed with broth microdilution. Minimum inhibitory concentration values were used to establish the epidemiological cut-off (ECOFF) with visual and statistical methods to distinguish the population in wild type M. bovis and those with acquired antimicrobial resistance (non-wild type). The results showed high percentages of acquired resistance for macrolides (tilmicosin, tylosin, and gamithromycin), but no acquired resistance for tetracyclines (oxytetracycline, doxycycline). Only little acquired resistance was observed for florfenicol, gentamicin, and tiamulin, while there was limited acquired resistance to enrofloxacin. Only M. bovis isolates from beef cattle or the third genomic cluster had a significantly higher change to have acquired resistance against gamithromycin than those collected from other sectors or genomic clusters. These results support the current national formulary for respiratory disease associated with M. bovis, recommending florfenicol as first choice, and oxytetracycline and macrolides as second choice. Possibly, a small remark for gamithromycin is needed, as higher risk for acquired resistance for this antimicrobial was seen in beef cattle. In vitro susceptibility testing results should be interpreted carefully, as the association with in vivo efficacy has not confirmed yet, due to the lack of clinical breakpoints. Finally, in Chapter 5.2, upgraded genomes derived from Chapter 4 and the susceptibility data from Chapter 5.1 were combined to compare genotype and phenotype antimicrobial susceptibility of M. bovis isolates. A genome wide association study to reveal genetic markers for antimicrobial resistance in M. bovis and verifying the ECOFF values obtained by the previously used different methods was executed. Many point mutations were associated with antimicrobial resistance against the critically important antibiotics of the macrolide (A2058G in the 23S rRNA gene, Gln83His in the L22 protein) and fluoroquinolone classes. For enrofloxacin the combination of different mutations in the GyrA and ParC gene showed the step-wise acquired resistance. Also previously described mutations for tilmicosin (G478A mutation in 23S rRNA alleles), and new markers for gentamicin (A1408G and G1488A in 16S rRNA) were identified. The visual estimation of de ECOFF showed to be a reliable method, although statistical methods can help when step-wise resistance results in difficult to interpret “tailing”. Even when phenotypical resistance is not yet obtained, in case of first-step mutations it should be discouraged to use fluoroquinolones as antimicrobial therapy, as selection pressure will eventually result in phenotypical resistance as well. In the general discussion (Chapter 6), the innovations in M. bovis diagnostics achieved with this thesis are discussed. In the second part, practical recommendations for diagnostics in M. bovis outbreak management, purchase policy, and eradication or herd status certificates are proposed. In this thesis new methods to identify, strain type, and access the antimicrobial susceptibility of M. bovis were developed. When rapid identification of M. bovis with MALDI-TOF MS (Chapter 3.3) is followed by the determination of antimicrobial resistance with nanopore sequencing (Chapter 5.2) it is now possible to obtain identification, strain typing and an antibiogram for critically important antibiotics within 3-5 days. This is a major step towards better control of M. bovis in clinical outbreaks and prevent herd infection when purchasing animals. Together with these new methods, also substantial epidemiological information came to light, showing the importance of a more national approach for the prevention of introducing M. bovis into the herd and country.