Mycoplasma hyopneumoniae (M. hyopneumoniae) is the causative agent of enzootic pneumonia [1]. The disease has a worldwide impact on intensive swine production and causes substantial losses due to reduced growth of the pigs, poor feed conversion ratio, higher antimicrobial use and increased susceptibility to secondary respiratory agents [1–3]. Previous research has shown that most pigs are infected with more than one M. hyopneumoniae strain [4, 5], and recently, Michiels et al., [6] showed that a higher severity and prevalence of Mycoplasma-like lung lesions was found in batches of slaughter pigs with detection of more than one M. hyopneumoniae strain [6]. It is generally accepted that control of M. hyopneumoniae is critical for reducing economical losses in infected swine operations [7]. This is achieved by optimizing housing conditions, management practices, antimicrobial therapy and vaccination against M. hyopneumoniae [8]. It is estimated that 70% of swine herds are practising vaccination against M. hyopneumoniae worldwide [9]. Vaccination reduces clinical signs, lung lesions and improves performance, although colonization is not prevented and there is no significant reduction in transmission [3, 8, 10–12]. The benefits obtained through vaccination can vary from herd to herd [13]. Most commercially available bacterin vaccines are based on an adjuvanted whole-cell preparation of an inactivated M. hyopneumoniae strain [8]. This strain is mostly the J-strain of M. hyopneumoniae, which was isolated in 1958 [14, 15]. Recently a commercial bacterin became available, using M. hyopneumoniae strain 2940, isolated in the 1999’s from a farm in the United States facing a severe outbreak of enzootic pneumonia. Also, most vaccination studies so far used one single strain for experimental infection. As most pigs are infected with more than one M. hyopneumoniae strain [4–6], it may be more appropriate to challenge the pigs with different M. hyopneumoniae strains.

The aim of the present study was to determine the efficacy of a commercial vaccine (Hyogen®) against an experimental challenge with two genetically different (low and highly virulent) M. hyopneumoniae strains.

No animals died during or shortly after the inoculations. One out of 45 piglets in the NCG was euthanized (D22) (Release®300 mg/ml, WDT, Garbsen, Germany) due to nervous symptoms and lateral decubitus. Necropsy was performed and Haemophilus parasuis was isolated in pure culture from the meninges and pericardium. All animals in the PCG coughed, showed macroscopic lung lesions and seroconverted. Mycoplasma hyopneumoniae-DNA was detected in 19 out of 20 animals from this group.

Clinical and performance parameters

The results of the clinical parameters are summarized in Table 1. There was no coughing in the NCG throughout the study. The RDS_24–53 was 1.20 ± 0.83 and 0.55 ± 0.42 for the PCG and VG (P < 0.001), respectively. This corresponds with a reduction of 52.9% in RDS when vaccinating the piglets compared to the PCG. All groups significantly differed from each other. These results and the RDS _1–53 and RDS _1–23 results are shown in Table 1 and Fig. 1.

Parameter	Experimental Groups
	NCG (# = 5)	PCG (# = 20)	VG (# = 20)	P-value
RDS
D1–53	0 ± 0 a	0.68 ± 0.41b	0.32 ± 0.86 c	<0.001
D1–23	0 ± 0 a	0 ± 0 a	0.017 ± 0.060 a	0.983
D24–53	0 ± 0 a	1.20 ± 0.83 b	0.55 ± 0.42 c	<0.001
Weight ± SD
D0	7.04 ± 1.22a	6.97 ± 0.84 a	7.04 ± 1.02 a	0.966
D24	15.25 ± 3.96 a	16.39 ± 2.16 a	15.61 ± 2.96 a	0.581
D53	36.51 ± 8.63 a	34.63 ± 6.63 a	34.27 ± 5.35 a	0.808
ADG (g/pig/d)
D0–24	357 ± 123 a	393 ± 62.8 a	357 ± 92.0 a	0.372
D24–53	733 ± 194 a	629 ± 199 a	643 ± 102 a	0.504
D0–53	563 ± 144 a	522 ± 121 a	514 ± 87.0 a	0.714
MLL, histopathology and air analysis D53
MLL	0 ± 0 a	7.56 ± 4.73 b	0.68 ± 0.73 a	<0.001
Histopathology	1.31 ± 0.18 a	3.32 ± 0.85 b	1.92 ± 0.47 c	<0.001
Percentage of air (%)	47.72 ± 7.29 a	34.49 ± 8.22 b	45.23 ± 5.99 a	<0.001
log copies of M. hyopneumoniae DNA/ml BALF (qPCR) ± SD
D39	0.36 ± 0.64 a	1.81 ± 1.32 a	1.23 ± 1.61 a	0.062
D53	0.40 ± 0.43 a	3.99 ± 1.20 b	1.78 ± 1.36 a	<0.001
Percentage of M. hyopneumoniae qPCR positive samples in BALF (#positive samples/total number of samples)
D39	(0/4) 0a	(14/20) 70b	(7/20) 35c	<0.01
D53	(0/4) 0a	(19/20) 95b	(12/20) 60c	<0.001
M. hyopneumoniae specific AB expressed in OD-values ± SD in serum
D1	1.51 ± 0.037 a	1.48 ± 0.058 a	1.48 ± 0.16 a	0.578
D24	0.91 ± 0.044 a	0.95 ± 0.069 a	0.50 ± 0.17 b	<0.001
D53	0.90 ± 0.10 a	0.23 ± 0.11 a	0.087 ± 0.041 b	<0.001
Percentage of ELISA M. hyopneumoniae positive samples in serum
D1	(0/5) 0a	(0/20) 0a	(0/20) 0a	1.000
D24	(0/4) 0a	(0/20) 0a	(15/20) 75b	<0.001
D53	(0/4) 0a	(20/20) 100b	(20/20)100b	<0.001

The respiratory disease score (RDS), body weight, average daily gain (ADG), M. hyopneumoniae specific antibodies, macroscopical lung lesions (MLL), histopathology and air analysis, log copies of M. hyopneumoniae DNA (qPCR), percentage of of M. hyopneumoniae qPCR positive samples, M. hyopneumoniae specific AB expressed in OD-values and Percentage of ELISA M. hyopneumoniae positive samples

^a,b,cDifferent superscripts in one row are statistically different (P < 0.05)

NCG negative control group, PCG positive control group, VG vaccination group, SD standard deviation, D Day of the study, ADG average daily gain, kg kilogram, RDS respiratory disease score, M. hyopneumoniae Mycoplasma hyopneumoniae, AB antibodies, OD optical densities, # number, MLL macroscopic lung lesions, qPCR quantitative polymerase chain reaction, DNA DeoxyriboNucleic Acid, BALF bronchoalveolar lavage fluid

The average group weight at D0, D24 and D53 and the ADG calculated during these three periods (0–24, 24–53 and 0–53 days of age) for each group are shown in Table 1.

This study showed that the vaccine was efficacious for the most part against an experimental challenge with both a low and highly virulent M. hyopneumoniae strain. The vaccinated pigs coughed significantly less, showed significantly less lung lesions, a significantly lower number of log copies in the bronchoalveolar fluid at euthanasia was shown and the IL-1 and IL-6 concentrations at euthanasia were lower compared to the non-vaccinated inoculated pigs.

The challenge infection was performed with two genetically different [27] M. hyopneumoniae isolates. The virulence of these strains was evaluated by Vicca et al. [16] and both strains were characterised at proteomic and genomic level by Calus et al. [28], Vranckx et al. [4] and Stakenborg et al. [29], respectively. The double challenge infection model was successful as all animals in the PCG showed lung lesions, coughed and seroconverted. In 19 out of 20 animals of the PCG at euthanasia, DNA of M. hyopneumoniae was detected in the bronchoalveolar lavage fluid. It is well known that there can be quite some variation between pigs in terms of responses to an experimental infection. Why this particular pig tested negative is not known. Possibly, the conditions for multiplication of the M. hyopneumoniae strains were not optimal in that pig. Although bronchoalveolar fluid is an appropriate sampling technique to detect M. hyopneumoniae during the early stages of infection, it fails to recover some positive animals in case of chronic infection [30]. Taken these factors into consideration, this might be the reason why, although the pig was challenged, it did not test positive. Although comparing to other experimental settings should be done with caution, clinical symptoms and lung lesions were comparable or more severe than the single F7.2C challenge model performed in our research group [12, 16, 24, 25, 31–38] and the standard deviation was lower [16, 31–35]. It was shown that most pigs in the field are simultaneously infected with two or sometimes three genetically different M. hyopneumoniae strains and that when batches of slaughter pigs are infected with more than one M. hyopneumoniae strain, this can result in more (severe) pneumonia lesions and fissures [6]. The current experimental setting might therefore better resemble the field situation than infection with a single strain. Finally, it was shown that anesthetizing and inoculating the piglets with seven ml of inoculum on two consecutive days is feasible, as none of the piglets showed adverse reactions or died during or shortly after the inoculations.

The pigs in the vaccinated group coughed significantly less (−53%) from inoculation until euthanasia compared to the positive control group, demonstrating the efficacy of the vaccine. The degree of coughing is an important efficacy parameter [39, 40] not only under experimental but also under field conditions. A coughing index can be used to estimate if a higher prevalence of enzootic pneumonia in a herd might be present [41] and to determine the most optimal timing of serological sampling of the pigs or collecting bronchoalveolar fluid to be tested with PCR [41, 42]. In this study, no statistical significant differences were observed in the parameters weight and ADG, although the pigs in the VG grew slightly faster compared to the PCG from day of inoculation until day of euthanasia onwards. Although the parameter weight and ADG are of great use to evaluate in field circumstances, in experimental trials they are merely of descriptive value as the low number of animals used, the duration of the trial and the sometimes high standard deviations are the reason why only numerical differences are found [10, 16, 31].

The MLL score in the vaccinated pigs was very low (0.68) and statistically different from the PCG (7.56), implying a reduction with 91%, and demonstrating the efficacy of the vaccine. A recent field study showed that when pigs are infected with more strains, more lung lesions are detected [6]. When comparing the MLL scores from the double inoculated challenge in PCG with those from challenged unvaccinated pigs in the single inoculation model from previous studies (highest MLL was 6.69 for F7.2C challenge in Villarreal et al., [25]), higher MLL in this study are obtained. However, comparing different experimental settings, should be done with caution. To make a true comparison, the resulting MLL from the double infection model should be compared with the single inoculation model from both strains F7.2C and F1.12A in one experimental setting.

In the vaccinated animals significantly less lymphohistiocytic infiltration (1.92) was observed compared to the PCG (3.32). These results are in accordance with the results of previous studies [25, 31, 32]. Although many aspects of the effect of vaccination on the immunological response remain unclear [43], the present results confirm that vaccination has a regulating function on the immune system, causing less infiltration of macrophages in the lung tissue [34]. A significantly lower area occupied by air or air percentage was measured in the non-vaccinated animals of PCG (34.49%) compared to VG (45.23%). Infection with M. hyopneumoniae results in intrabronchiolar- and bronchial exudate and infiltration of lymphohistiocytic cells in the lung tissue. These responses may result in compressing the airways and alveoli, leading to less air volume in the lungs [15]. Vaccination also significantly reduced the log copies of M. hyopneumoniae DNA in the bronchoalveolar lavage fluid compared to the non-vaccinated animals (1.78 versus 3.99), which may lead to less shedding of M. hyopneumoniae in vaccinated pigs. Similar reductions were found in previous studies [34, 44]. These results, although very promising, confirm once more that vaccination is not able to prevent pigs from being colonized as commonly known [8].

Mycoplasma hyopneumoniae-specific antibody (Ig G and Ig A) levels in tracheo-bronchial washings were not significantly different between vaccinated and non-vaccinated pigs. Our findings are in accordance with Djordevic et al. (1997) [24, 45], but Thacker et al. (2000) suggested that secretion of M. hyopneumoniae- antibodies induced by vaccination into the bronchoalveolar washings might be implied in resolving mycoplasmal pneumonia [43, 46, 47]. Further research is needed to clarify a possible protective role of M. hyopneumoniae-specific antibodies in trachea-bronchial washings.

There were no significant differences between the vaccinated and non-vaccinated animals for the TNF-α concentration in the bronchoalveolar fluids. In the study of Marchioro et al. [24], vaccination significantly decreased TNF-α concentrations, but the difference was small and only borderline significant. This highlights the need to further elucidate the pathogenesis of M. hyopneumoniae and the exact mechanisms in which way M. hyopneumoniae-bacterins provide protection to the animal.

The cytokines IL-1 and IL-6, next to TNF-α are pro-inflammatory cytokines [48]. Previous research stated that they mediate lymphocyte infiltration and activation in the pneumonic lung [49, 50], and thus are associated with the induction of pneumonia lesions [24, 51, 52]. The pro-inflammatory cytokines TNF-α, IL-1 and IL-6 were correlated with MLL at D39 and D53. The parameter TNF-α was weakly and not significantly correlated with MLL on both sampling days (D39 and D53). Interleukin 1 was strongly and significantly correlated with MLL at euthanasia, and IL-6 was moderately and significantly correlated with MLL on D39 and D53. This suggests that the vaccine shows a protective effect mainly by modulation of the pro-inflammatory cytokines IL-1 and IL-6, but not TNF-α (data not shown). The IL-1 concentration and the standard deviation in the non- vaccinated pigs was very high. This result is in accordance with the result of Meyns et al. [38] at 28 days post inoculation with the F7.2C isolate. One can conclude that a part of the piglets in the non- vaccinated group showed very high IL-1 levels. The reason why this occurred is not clear, however the individual susceptibility of the pig to an M. hyopneumoniae infection must be kept into account causing the challenge infection to be more successful. The IL-1 and IL-6 concentrations at necropsy in the vaccinated pigs were significantly lower than in the non-vaccinated pigs and not significantly different from the levels in negative control group. These results substantiate the statement that was made in Marchioro et al. [24] based on the IL-1 result of the vaccinated pigs at D36 that IL-1 and IL-6 may have a regulatory role upon M. hyopneumoniae vaccination [24].

In this study, the pigs were vaccinated approximately 4 weeks before challenge, as vaccination is most likely to be effective if active immunity is established before the piglets are exposed to the pathogen. Under field conditions, however, the time between vaccination and infection is not known, and very likely highly variable between pigs. Hence, it is possible that the vaccine used in the field is less efficacious when the timing of vaccination is not optimal. As this is an experimental study, the results cannot be extrapolated as such to field conditions. Under field circumstances, more factors challenging the efficacy of a vaccine are present, compared to the controlled environment of an experimental facility. First, multi-factorial and multi-pathogen enzootic pneumonia outbreaks are typically found in the field, rather than isolated M. hyopneumoniae outbreaks [2]. The piglets originated from a herd free of important diseases which can influence the outcome of an M. hyopneumoniae infection, such as PRRSv Pasteurella multocida, and Actinobacillus pleuropneumoniae. [1, 53, 54] and according to some authors infection with certain pathogens can determine the effectiveness of M. hyopneumoniae vaccination as well [54–56]. Second, the impact of weaning stress should be taken into account when vaccinating piglets at weaning age in the field. Although the impact of weaning stress on vaccine efficacy might not be clear [31], it could not have influenced the efficacy of the vaccine, as the pigs were transported and vaccinated some days after weaning. Thirdly, passively acquired maternal derived immunity could not have influenced the efficacy of the vaccine as these piglets were originating from a herd free of M. hyopneumoniae and no vaccinations against M. hyopneumoniae in the sows, nor the piglets were performed. This is a double sided given. On the one hand, Hodgins et al. [57] stated that maternal antibodies in the piglets were associated with reduced antibody responses to vaccination. On the other hand, the piglets in the study could not benefit from the passively transferred cell-mediated immunity. Bandrick et al. [30] showed that vaccination of piglets against M. hyopneumoniae in the face of antigen-specific maternal-derived immunity results in cell-mediated immunity priming and anamnestic cell-mediated immunity responses following the exposure to M. hyopneumoniae antigen. On the other hand, some factors in the experimental study were comparable to field circumstances or challenged the vaccine more. First, although the piglets were free from certain pathogens, other than that, the piglets were originating from a conventional farm. Second: the piglets were transported from the herd of origin to the experimental facilities and the piglets needed to establish a new hierarchy after comingling. These events are associated with stress, which might influence the efficacy of the vaccine. This can resemble the stress that piglets experience in the field, when sorted and moved from one facility to another in a multi-site production system. Thirdly: the vaccine was highly challenged in this experimental study, because as stated by Villarreal et al., [35] and Meyns et al. [37], it can be assumed that the M. hyopneumoniae challenge dose used to infect the pigs was higher than might be reached under natural conditions resulting in a faster and higher colonization level and presumably challenging the vaccine more than under field circumstances [35, 37]. Although extrapolation of the efficacy testing of the vaccine results obtained in experimentally M. hyopneumoniae infected pigs should be done with caution, this infection model enabled to study the effect of the vaccine in M. hyopneumoniae infected pigs in a reproducible and standardized way.

Using a double challenge infection model with a low and highly virulent M. hyopneumoniae strain, one-dose vaccination of piglets was efficacious for the most part, as coughing was reduced by more than 50% and macroscopical lung lesions by 91%. In addition, the lymphohistiocytic infiltration in lung tissue was lower, the number of log copies of M. hyopneumoniae DNA detected in bronchoalveolar lavage fluid and the IL-1 and IL-6 concentrations at euthanasia were lower in the vaccinated animals compared to non-vaccinated animals.