Transmission dynamics of Mannheimia haemolytica in newly-received beef bulls at fattening operations

Veterinary Microbiology 161 (2013) 295–304

Contents lists available at SciVerse ScienceDirect

Veterinary Microbiology journal homepage: www.elsevier.com/locate/vetmic

Transmission dynamics of Mannheimia haemolytica in newly-received beef bulls at fattening operations E. Timsit a,*, H. Christensen b, N. Bareille a, H. Seegers a, M. Bisgaard b, S. Assie´ a a b

LUNAM Universite´, Oniris, UMR 1300, Biology, Epidemiology and Risk Analysis, and INRA, Nantes, F-44307, France Department of Veterinary Disease Biology, Faculty of Life Sciences, University of Copenhagen, Frederiksberg, DK-1870, Denmark

A R T I C L E I N F O

A B S T R A C T

Article history: Received 1 October 2011 Received in revised form 27 July 2012 Accepted 29 July 2012

The primary objective of this study was to determine, at the lung level, whether single or multiple clones of Mannheimia haemolytica are present within a pen during a bovine respiratory disease (BRD) episode. A secondary objective was to assess whether M. haemolytica isolates obtained from nasal swabs (NS) are identical to those isolated deeper within the respiratory tract. Sixteen BRD episodes that naturally occurred in 12 pens of eight to 12 bulls (n = 112) newly-received at three fattening operations were investigated. One hundred and seventy five M. haemolytica isolates were collected from 239 pairs of trans-tracheal aspirations (TTA) and NS performed during these 16 BRD episodes. M. haemolytica isolates were characterized by pulsed-field gel electrophoresis (PFGE). PFGE types obtained from NS and TTA were then compared. M. haemolytica was isolated during 14 BRD episodes. Two to three different clones of M. haemolytica were recovered during 10 episodes whereas only one clone was recovered in four episodes. A moderate agreement (kappa = 0.50) between NS and TTA for M. haemolytica isolation was observed. Identical PFGE types were only observed in 77% of matched NS-TTA pairs. The significant withinpen diversity of M. haemolytica during BRD episodes indicates that the disease is not primarily due to the spread of a single virulent clone among cattle and highlights the importance of predisposing factors that enable the resident flora to overcome the cattle’s immune system. The results also demonstrate that isolates recovered from NS are not always representative of the isolates present deeper within the respiratory tract. ß 2012 Elsevier B.V. All rights reserved.

Keywords: Mannheimia haemolytica Cattle Bovine respiratory disease Epidemiology PFGE

1. Introduction Bovine respiratory disease (BRD) is the most prevalent disease in cattle entering fattening operations (Assie et al., 2009; Smith, 1998). Mannheimia haemolytica is the principal bacterium implicated in BRD and it is generally accepted that its control would markedly reduce the prevalence of BRD in fattening operations (Rice et al., 2007). A good understanding of the transmission dynamics of M. haemolytica is needed to adapt control measures during BRD episodes (Miles, 2009). However, to date, little

* Corresponding author. Tel.: +33 240687652. E-mail address: [email protected] (E. Timsit). 0378-1135/$ – see front matter ß 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.vetmic.2012.07.044

information on the transmission dynamics of this bacterium is available and it is not clear whether M. haemolyticaassociated BRD episodes are due to predisposing factors that enable the resident flora to overcome the cattle’s immune system or due to the contagious spread of a single virulent clone among penmates or due to both (Rice et al., 2007; Taylor et al., 2010b). Molecular typing methods provide tools to investigate the transmission dynamics of M. haemolytica. Among these methods, pulsed-field gel electrophoresis (PFGE) appears to be well adapted to monitor the transmission dynamics of M. haemolytica at the BRD episode level due to its high discriminatory power and repeatability (Klima et al., 2010; Kodjo et al., 1999). It can be applied on isolates recovered from the upper and/or lower respiratory tracts. Because

296

E. Timsit et al. / Veterinary Microbiology 161 (2013) 295–304

the sampling of the nasal flora is easier to perform than that of the lung, it is generally preferred (Klima et al., 2011; Purdy et al., 1993). However, further investigation is needed to determine whether M. haemolytica isolates obtained from the nasal cavities are representative of the isolates present deeper within the respiratory tract (Taylor et al., 2010b). The objectives of this study were therefore, firstly, to characterize by PFGE the M. haemolytica isolates collected from the lower respiratory tracts of bulls during BRD episodes to determine whether single or multiple clones of M. haemolytica are present within a pen and, secondly, to assess whether M. haemolytica isolates recovered from nasal swabs (NS) are identical to those isolated deeper within the respiratory tract. 2. Materials and methods All procedures in the present experiment were performed in accordance with the European directive and the French regulation and conform to the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1996). 2.1. Animals One hundred and twelve bulls (initial body weight  SD = 346  36 kg) were observed for 40 days following their arrival at three French fattening operations between November 2007 and March 2008 (Table 1). Bulls came from multiple sources (n = 43 farms of origin) and were purchased at auction markets. After purchase, bulls were transported by truck over a mean distance of 329  104 km (range = 70–515 km) to a central facility where they were sorted into 12 groups of eight to 12 bulls based on body weight (Table 1). The size of the group (eight to 12 bulls) was determined by the size of the bulls’ pen in the fattening operations, which ranged from 30 to 48 m2. Bulls were allowed free access to hay and water at the central facility for 36 to 48 h. During this period, they were in contact with others bulls present at the central facility. Then, bulls were transported by truck for travel distances less than 50 km to the fattening operations. None of the bulls received any vaccine or antibiotics at entry. During the study period, each group of bulls was housed in a pen separated from the other

bulls fed in the fattening operation by fences that allowed nose-to-nose contact. Bulls were fed with a total mixed ration formulated to meet the French National Institute for Agriculture Research recommendations (Garcia et al., 2007). Feed was mixed and delivered once daily at 9 a.m. Throughout the study period, animals had unlimited access to water. 2.2. Study design During the 40-day-study period, owners observed twice daily all bulls for the detection of the following signs: depression, decreased rumen fill compared with penmates, nasal or ocular discharge, cough and increased respiratory rate. As soon as the owner detected a bull displaying at least one of the above signs in a pen under study, a veterinarian with experience in cattle disease diagnostics restrained, one by one, each bull housed in the pen including the in-contact apparently healthy bulls in a conventional cattle handling chute, to perform a close physical examination, a NS sample and a transtracheal aspiration (TTA). Bulls with a rectal temperature  39.7 8C and, at least, one other sign of respiratory tract disease were diagnosed as clinically BRD affected and received a single subcutaneous injection (2 mL per 15 kg of body weight) of a product containing 16.5 mg/mL of flunixin meglumin and 300 mg/mL of florfenicol (Resflor, MSD, Angers, France). Physical examinations and clinical samples (NS and TTA) were then repeated every three days on nonpreviously treated bulls until the end of the BRD episode i.e. until no new BRD affected animal was detected. 2.3. Sampling procedures and bacteriology Prior to NS sampling, the nostril was disinfected using 90% alcohol. A guarded swab (Dryswab Veterinary Laryngeal, Medical Wire and Equipment, Corsham, England) was used to sample the nasal cavities. This swab was enclosed in a sterile plastic sleeve in order to reduce contamination while it was introduced through the nostril to a depth of approximately 20 cm (dorsal conchae). At this point, the swab was exposed by withdrawing the sleeve and it was moved back and forth several times against the nasal mucosae as previously described (Allen et al., 1991).

Table 1 Characteristics of 12 pens of newly-received bulls at three fattening operations. Fattening operation no.

Date of arrival

Pen no.

No. of bulls per pen

No. of different origins per pen

Mean body weight at arrival  SD, kg

1

16/11/2007

1 2 3 4 5 6 7 8 9 10 11 12

9 9 9 9 8 8 8 8 8 12 12 12

1 2 7 6 6 5 5 5 2 3 4 9

314  28 326  26 378  14 355  43 330  19 334  20 350  25 359  22 364  27 379  34 350  31 303  19

17/01/2008 2

21/11/2007

06/12/2007 3

27/11/2007 11/01/2008

E. Timsit et al. / Veterinary Microbiology 161 (2013) 295–304

The swab was then pulled back into the sleeve to protect it during the withdrawal. Afterwards, each swab was immediately inserted into a transport medium (Venturi Transsystem, Copan, Bovezza, Italy) for subsequent bacterial culture. Transtracheal aspirations were performed as previously described (Espinasse et al., 1991). An area of skin covering the ventral aspect of the middle third of the trachea was shaved and sterilised. An intravenous catheter (Centracath, Vygon, Ecouen, France) 75 cm length with an external diameter of 2 mm was used for the aspiration. The trochar part of the catheter was passed between two tracheal rings and the flexible part of the catheter was introduced for about 40 cm in the direction of the lung. 50 mL of physiologically balanced saline (i.e. NaCl 0.9%) was pushed in the catheter using a 50 mL syringe. The fluid was then immediately retrieved by gentle suction using the syringe. On average, 5–10 mL of fluid was recovered and immediately placed into sterile plain tubes. After sampling, NS and TTA samples were transported in a container at a temperature of 4 8C to the laboratory of the Nantes-Atlantic National College of Veterinary Medicine, Food Science and Engineering (Oniris) and processed within 6 h. The primary isolation of M. haemolytica from clinical samples was done by directly streaking each NS or by inoculating 0.1 mL of each TTA aspirate onto 5% sheep blood agar (Columbia, Oxoid, Hampshire, UK) containing 15 mg/mL of bacitracin in order to limit the growth of bacteria other than Pasteurellaceae (Catry et al., 2007; Klima et al., 2011). Sheep blood agar plates were then incubated for 24 h at 37 8C in aerobic atmosphere. Afterwards, one colony resembling M. haemolytica was selected and subcultured for another 24 h and identified by standard biochemical procedures (Quinn et al., 1994) and API (Biomerieux, Marcy l’Etoile, France). M. haemolytica isolates were stored at 80 8C until further analysis. 2.4. Pulsed field gel electrophoresis Plugs preparation was performed as previously described (Hedegaard et al., 2009). Briefly, M. haemolytica isolates kept at 80 8C were plated onto 5% sheep blood agar (Columbia, Oxoid, Hampshire, UK) and incubated overnight at 37 8C. A single typical colony was subcultured overnight in 10 ml brain heart infusion (BHI) broth (Oxoid, Hampshire, UK) at 37 8C with shaking. To measure and standardize cell density, the optical density of a solution containing 800 mL SE buffer (75 mM NaCl, 25 mM ethylene-diamine-tetraacetic acid [EDTA], pH 7.4) and 200 mL culture in BHI broth was determined at 578 nm. Cells contained in a 3 mL sample of the overnight culture in BHI broth were then washed by centrifugation and resuspended in sufficient SE buffer to obtain a standardized cell density. Afterwards, 500 mL of the cells suspension was mixed with 500 mL agarose solution (0.02 g agarose, 1 mL SE buffer) and dispensed into disposable plug moulds (BioRad, Marnes-la-Coquette, France). To release DNA from the cells, plugs were incubated overnight in a water bath at 56 8C in 950 mL proteinase K solution (0.5 mg proteinase K [Roche Diagnostics,

297

Meylan, France], 1 mL ES-buffer [1% N-Lauroylsarcosin, 0.5 M EDTA, pH 9.5]). Then, plugs were washed 3 times with TE buffer (10 mM Tris, 10 mM EDTA, pH 7.4) and stored at 4 8C until further use. Plugs were digested overnight at 37 8C with 20 U of SalI and 152 mL of appropriate buffer provided by the manufacturer (New England BioLabs, Hitchin, UK). The digested DNA was separated by PFGE using a CHEF DRIII device (Bio-Rad, Marnes-la-Coquette, France) according to the program previously described by Villard et al. (2006). A low range PFGE Marker (N0350S, New England BioLabs, Hitchin, UK) was placed in the first, middle and last wells of each gel. After electrophoresis, gels were stained in 1 mg/ mL ethidium bromide aqueous solution for 30 min, then washed in distilled water and eventually photographed using ultraviolet light transillumination. To study the reproducibility of the PFGE procedure (cell lysis, washing, endonuclease digestion steps, gel and electrophoretic conditions), a M. haemolytica reference strain (CCUG 12392) was processed 11 times, along with the M. haemolytica field isolates. To assess the genetic diversity of M. haemolytica within a TTA sample, eight TTA samples from a subset of positive bulls were cultured and eight to ten M. haemolytica colonies per sample were characterized by PFGE. 2.5. Serotyping One isolate per PFGE type (n = 23) was serotyped at the Agriculture and Agri-food Canada Research Centre (Lethbridge, Canada) using the published protocol described by Klima et al. (2011). 2.6. Data analysis All PFGE-related analyses were performed using BioNumerics V5.1 (Applied Maths Inc., Austin, TX, USA). Dendrograms were created by UPGMA cluster analysis using Dice coefficients of similarity with 1.0% optimization and 3.0% position tolerance settings. Only DNA fragments in the range of the marker (i.e. between 23.1 and 339.5 kb) were used for the cluster analysis. Isolate relatedness was assessed using the modified Tenover criteria described by van Belkum et al. (2007) i.e. isolates with PFGE patterns differing by one to four bands were assigned to subtypes of the same main types (or clones). Each main type was arbitrarily identified by a letter. Subtypes were identified by a suffix to the assigned main type. For example, subtypes of the main type E were E1 and E2. Typeability of the PFGE procedures was reported as the proportion of M. haemolytica isolates that were assigned a PFGE pattern. Reproducibility of the PFGE procedures was reported as the proportion of M. haemolytica reference strain (CCUG 12392) patterns that had 100% similarity based on UPGMA cluster analysis using Dice coefficients of similarity. Associations between PFGE types recovered from the lower respiratory tract during the BRD episodes and (1) the sources of bulls (n = 43 farms of origin) and (2) the arrival groups of bulls (i.e. bulls arriving the same day at the fattening operation; n = 6) were explored using Fisher’s

E. Timsit et al. / Veterinary Microbiology 161 (2013) 295–304

298

exact test in SAS 9.2 (SAS Institute Inc., Cary, NC, USA). When more than one PFGE type were recovered from the same bull during a BRD episode, only the PFGE type which corresponded to the day of treatment or to the last sample obtained from animals that remained apparently healthy during the BRD episode was taken into account for these analyses. Agreement between NS and TTA for the isolation of M. haemolytica was measured using the Kappa statistic (Cohen, 1960). The strength of agreement for the Kappa coefficient was interpreted using the scale of Landis and Koch (1977): 0 = poor, 0.1–0.20 = slight, 0.21–0.40 = fair, 0.41–0.60 = moderate, 0.61–0.80 = substantial and 0.81– 1 = almost perfect. Differences between the isolation rates of M. haemolytica from clinical samples (NS or TTA) in clinically BRD affected animals and apparently healthy animals were evaluated by contingency table chi-square analysis. Isolation rates of M. haemolytica from NS and TTA samples were compared using McNemar’s chi-square test for two correlated proportions. 3. Results

88 (79%) of the 112 bulls entering the study were treated for BRD. Descriptive statistics of BRD episodes are shown in Table 2. 3.2. Clinical samples Because, at least, one BRD episode occurred in every pen, NS and TTA samples were collected from all the bulls entering the study. Depending on the duration of the BRD episodes and the number of bulls diagnosed and treated for BRD, from eight to 30 paired samples were collected during each BRD episode (Table 2). In total, 239 pairs of NS and TTA samples were performed (Table 3). Out of the 239 pairs of NS and TTA samples, 103 (43%) were performed on clinically BRD affected animals (i.e. sampling after the onset of BRD signs) and 136 (57%) were performed on apparently healthy animals that were subsequently diagnosed and treated for BRD (i.e. sampling prior to the onset of BRD signs) or that remained apparently healthy during the BRD episode (Table 3 and Supplementary Table S1). 3.3. M. haemolytica isolation

3.1. Health data Sixteen BRD episodes were observed in the 12 pens under study during the 40 days following the arrival of bulls at the fattening operations (Table 2). One BRD episode occurred in eight pens whereas two BRD episodes occurred more than 10 days apart from each other (mean  SD = 15 days  4) in the four remaining pens. Depending on the pen, the duration of the BRD episodes ranged from one to ten days (median = 4 days) with an intrapen morbidity ranging from 44 to 100% (mean = 69%). In total,

The results of bacterial cultures for the isolation of M. haemolytica from NS and TTA samples are shown in Table 3. M. haemolytica was isolated more frequently from clinically BRD affected animals than apparently healthy ones (P < 0.001). Indeed, of the 103 pairs of NS and TTA samples collected from clinically affected animals, 67 (65%) were positive for M. haemolytica on either NS or TTA or both, whereas only 48 (35%) of the 136 pairs collected from apparently healthy animals were positive for this bacterium on either NS or TTA or both. In total, 175 M.

Table 2 Descriptive statistics of 16 bovine respiratory disease (BRD) episodes observed in 12 pens of bulls during the first 40 days post-arrival at three fattening operations. The number of paired samples (nasal swabs and transtracheal aspirations) collected from the bulls during each BRD episode is also indicated. Fattening operation no.

Date of arrival

1

16/11/2007

Pen no.

1 2

2

17/01/2008

3

21/11/2007

4 5 6

06/12/2007 3

27/11/2007 11/01/2008

a

7 8 9 10 11 12

BRD episodea no.

Days from arrival to first BRD treatmentb

Days from first to last BRD treatment

No. of paired samples collectedc

No. (%) of bulls treated for BRD

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

6 20 7 28 1 21 1 2 1 31 2 4 2 3 8 3

1 10 4 4 7 1 7 4 10 1 7 1 4 10 4 4

9 17 13 13 16 9 16 13 20 8 18 8 12 30 20 17

4 9 5 5 8 5 9 6 8 4 6 4 5 9 6 10

(44) (100) (56) (56) (89) (46) (100) (75) (100) (50) (75) (50) (63) (75) (50) (83)

In four pens, a second BRD episode was observed more than 10 days after the first one. b Bulls with a rectal temperature  39.7 8C and, at least, one other sign of respiratory tract disease such as abnormal pulmonary sounds, coughing, polypnea/dyspnea, or nasal or ocular discharge were treated for BRD. c Nasal swabs and transtracheal aspirations were collected from all the bulls housed in the pen on the day of the first BRD case detection and then every three days on non-previously treated bulls until the end of the BRD episode.

E. Timsit et al. / Veterinary Microbiology 161 (2013) 295–304

299

Table 3 Agreement between nasal swabs (NS) and trans-tracheal aspiration (TTA) samples for the isolation of M. haemolytica. TTA

NS

Total

+ (a) All paired samples (n = 239 pairs)a +

60b 22 Total 82 (b) Paired samples performed on clinically BRD affected animals (n = 103 pairs)c + 38d 11 Total 49 (c) Paired samples performed on apparently healthy animals (n = 136 pairs)a + 22e 11 Total 33 a b c d e

33 124 157

93 146 239

18 36 54

56 47 103

15 88 103

37 99 136

Kappa = 0.50. Identical PFGE types in 77% of matched pairs. Kappa = 0.44. Identical PFGE types in 71% of matched pairs. Identical PFGE types in 86% of matched pairs.

haemolytica isolates were recovered during the study period, with 82 and 93 isolates recovered, respectively, from NS and TTA. At the BRD episode level, M. haemolytica were isolated from the lower respiratory tract of bulls in 14 of the 16 BRD

episodes (Table 4). The prevalence of bulls positive for M. haemolytica on TTA samples during these BRD episodes varied from 11 to 89% (median = 56%). In total, 75 of the 112 bulls were positive for M. haemolytica on TTA samples during the study period: 58 bulls were positive once, 16

Fig. 1. Dendrogram and schematic representation of PFGE fragments with SalI digestion of 175 field M. haemolytica isolates and M. haemolytica reference strain (CCUG 12392). The dendrogram was created by UPGMA cluster analysis using Dice coefficients of similarity with optimization and tolerance settings of 1.0% and 3.0%, respectively. Isolates with PFGE pattern differing by one to four bands were assigned to subtypes of the same main types (van Belkum et al., 2007). Each main type was arbitrarily identified by a letter (A–L). One isolate per PFGE type (n = 23) was serotyped using the published protocol described by Klima et al. (2011).

E. Timsit et al. / Veterinary Microbiology 161 (2013) 295–304

300

Table 4 Number of bulls culture-positive for M. haemolytica on trans-tracheal aspiration (TTA) samples and number of isolates (n) assigned to specified types by pulsed-field gel electrophoresis (PFGE) observed during 16 bovine respiratory disease (BRD) episodes which occurred in 12 pens of bulls during the first 40 days post-arrival at three fattening operations. Fattening operation no.

Date of arrival

1

16/11/2007

Pen no.

1 2

2

17/01/2008

3

21/11/2007

4 5 6

06/12/2007 3

27/11/2007 11/01/2008

7 8 9 10 11 12

BRD episodea no.

No. (%) of bulls culture positive on TTA

PFGE typesb (n)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

0 6 5 6 5 1 8 4 0 4 3 6 5 6 9 10

–c L1 (3), L3 (2), B (1) B (5) L3 (6) G2 (2), J1 (2), J2 (1) L1(1) G2 (5), G4 (1), J1 (1), E1 (1) G3 (3), G5 (1) – G3 (2), D (1), I3 (1) C (1), G3 (1), I1 (1) G5 (4), I2 (1), E2 (1) G5 (3), I1 (2) H (3), F2 (3) F2 (6), B (1), H (2) K (5), J1 (4), G1 (1)

(0) (66) (56) (66) (56) (11) (89) (50) (0) (50) (38) (75) (63) (50) (75) (83)

a

In four pens, a second BRD episode was observed more than 10 days after the first one. b When more than one PFGE type were recovered from the same bull during a BRD episode, only the PFGE type which correspond to the day of treatment or to the last sample performed in animals that remained apparently healthy during the BRD episode is shown in this table. c –: not applicable.

bulls were positive twice and one bull was positive three times (Supplementary Table S1). It is interesting to note that the isolation of M. haemolytica from TTA samples was not always associated with BRD clinical illness. Indeed, 23 bulls that tested positive for M. haemolytica in one or more TTA samples during a BRD episode remained apparently healthy (Supplementary Table S1).

3.4. PFGE typeability and reproducibility Twelve PFGE main types were observed including two singletons and 10 clusters with between two (E) and 55 (G) isolates (Figs. 1 and 2). A PFGE pattern was obtained for all the M. haemolytica isolates analysed, meaning a typeability of 100%. The PFGE procedures were found to be 100% reproducible. Indeed, all the M. haemolytica reference

Fig. 2. PFGE pattern of M. haemolytica isolates with SalI digestion. First and last lanes: molecular weight marker (mixture of lambda DNA-HindIII fragments and lambda concatemers [N0350S, New England BioLabs, Hitchin, UK]). Lanes 2–14 represent PFGE types A, B, C, D, E1, F1, G1, H, I1, J1, K and L1, respectively.

E. Timsit et al. / Veterinary Microbiology 161 (2013) 295–304

strain (CCUG 12392) patterns (n = 11) had 100% similarity based on UPGMA cluster analysis using Dice coefficients of similarity with 1.0% optimization and 3.0% position tolerance settings. 3.5. Genetic diversity of M. haemolytica within a TTA sample The study of the genetic diversity of M. haemolytica within a TTA sample did not reveal any heterogeneity. Indeed, identical PFGE profiles were obtained from the eight to ten colonies typed per TTA sample (n = 8).

301

types were observed in only 77% of the matched pairs. It is interesting to note that the agreement between NS and TTA samples was not improved when only paired samples collected from clinically affected animal were considered (n = 103; Kappa = 0.44) (Table 3b). There was no difference between the isolation rates of M. haemolytica from NS and TTA samples in all paired samples (n = 239, P = 0.148) or when considering only paired samples collected from clinically BRD affected animals (n = 103; P = 0.19) or apparently healthy animals (n = 136; P = 0.438). 4. Discussion

3.6. Serotyping results Serotyping revealed that most of the M. haemolytica isolates were either serotype 6 (PFGE types E1, E2, F1, F2, G1, G2, G3, G4, G5, H, I1, I2 and I3) or serotype 1 (PFGE types J1, J2, K, L1, L2 and L3) and that only M. haemolytica isolates from PFGE types A, B, C and D were serotype 2 (Fig. 1). 3.7. PFGE types recovered from the lower respiratory tract of bulls during BRD episodes Two to three different PFGE main types (clones) defined according to the modified Tenover criteria described by van Belkum et al. (2007) were recovered from the lower respiratory tract of bulls during ten BRD episodes whereas only one main type was recovered in four BRD episodes (Table 4). In the four pens where two BRD episodes occurred, the main types recovered from the second BRD episodes were not recovered during the first episodes (Table 4). It is interesting to note that, at the bull level, identical PFGE types were recovered from 10 of the 14 bulls positive more than once during the same BRD episode (Supplementary Table S1). The PFGE types isolated from the lower respiratory tract during the BRD episodes were strongly associated with the arrival groups of bulls (P < 0.001). Indeed, only three PFGE types (L1, G5, I1) were shared between bulls that arrived at different dates in the same fattening operation, only two PFGE types (B, J1) were shared between two fattening operations and no PFGE type was shared between the three fattening operations (Table 4). The association between PFGE types and the sources of bulls could not be computed (too high number of farms of origin; n = 43) but identical PFGE types were recovered from bulls of different sources which were part of the same arrival group even if they were not housed in the same pen (Supplementary Table S1). Full details of BRD episodes, PFGE types recovered from NS and TTA samples and farm of origins can be found in the Supplementary Table S1. 3.8. Agreement between NS and TTA samples Agreement between NS and TTA samples (n = 239 pairs) for the isolation of M. haemolytica was moderate (Kappa = 0.50) (Table 3a). Furthermore, when M. haemolytica isolates were recovered at the same time from both NS and TTA samples (60 matched pairs), identical PFGE

This study provides unique information concerning the transmission dynamics of M. haemolytica in bulls newlyreceived at fattening operations. To our knowledge, this is the first time that the genetic diversity of M. haemolytica isolates recovered from the lower respiratory tract during a BRD episode is reported. The PFGE analysis revealed significant within-pen diversity of M. haemolytica during BRD episodes with up to three different clones identified. This study also demonstrated a moderate agreement between nasal swabs (NS) and trans-tracheal aspiration (TTA) samples for the isolation of M. haemolytica. Furthermore, when both NS and TTA samples were positive for M. haemolytica, the same PFGE type was only recovered in 77% of the matched pairs. The significant within-pen diversity of M. haemolytica during BRD episodes indicates that the disease is not primarily due to the contagious spread of a single virulent clone among cattle and highlights the importance of predisposing factors such as viral infections and stressors (shipping, commingling, etc.) that enable the resident flora to overcome the cattle’s immune system. This finding is consistent with a previous study conducted by Purdy et al. (1993). In their study, they observed that M. haemolytica enzyme profiles were more significantly associated with the calves’ farms of origin than with feedlot commingling and, thus, concluded that M. haemolytica tended to recrudesce from individual calves, rather than being transferred from feedlot penmates. Recently, Taylor et al. (2010a), who studied the genetic diversity of Pasteurella multocida isolates recovered from fatal cases of BRD in cattle in a commercial feedlot, also concluded that P. multocida tended to recrudesce from individuals as a strictly opportunistic pathogen, rather than being horizontally transferred. Even if M. haemolytica-associated BRD did not seem to be due to contagious dissemination of a single virulent clone in the present study, horizontal transfer of M. haemolytica between bulls also occurred. Indeed, despite the high genetic diversity of M. haemolytica isolates, identical PFGE types were recovered from bulls of different sources which were part of the same arrival group and further housed in different pens. This finding could indicate that a horizontal transfer of M. haemolytica occurred early at the central facility or during shipment. Furthermore, in the four pens where two BRD episodes occurred, the PFGE types recovered from the second BRD episodes were not recovered from the first episodes. This finding could indicate that a transfer of M. haemolytica also

302

E. Timsit et al. / Veterinary Microbiology 161 (2013) 295–304

occurred in bulls from adjacent pens or from the environment. A rapid horizontal transfer of M. haemolytica between feedlot cattle within days of commingling has been previously reported by Briggs et al. (1998). In that study, a strain of M. haemolytica containing a unique plasmid profile was used to inoculate the palatine tonsils of 12 calves. This specific strain of M. haemolytica was recovered from the nasopharynx of all the inoculated calves and also from 10 of the 89 in-contact calves within few days after challenge. This study therefore demonstrated the existence of a horizontal transfer of M. haemolytica. Therefore, based on the current results and this previous study, it seems that horizontal transfer of M. haemolytica occurs in feedlot cattle, albeit at an inconsistent rate. The presence of M. haemolytica serotypes A1 and A6 in the lower respiratory tract of bulls, although defined as major BRD pathogens, was not always associated with BRD clinical illness in the present study. Indeed, numerous bulls tested positive for M. haemolytica on one or more TTA samples and were, however, not detected as clinically BRD affected during the study period. This finding is in agreement with previous studies (Allen et al., 1991; Angen et al., 2009; Espinasse et al., 1991) and indicates that the changes caused by these pathogens, if any, are often subclinical and that it seems extremely difficult to predict the clinical outcome based on the microbial flora alone. It is also noteworthy that, during two BRD episodes, not a bull tested positive for M. haemolytica on NS or TTA samples. This finding highlights that, even if M. haemolytica is considered the most common bacteria isolated from BRD in feedlot cattle, other pathogens such as P. multocida, Mycoplasma bovis, etc. can also be responsible for the BRD episodes observed during the first weeks following the entry of cattle into a fattening operation (Griffin et al., 2010). The moderate agreement observed in the present study between the isolation of M. haemolytica from NS and TTA samples indicates that a NS culture does not accurately predict the presence of M. haemolytica deeper within the respiratory tract. Furthermore, in this study, when both samples were M. haemolytica positive, the same PFGE types were only obtained in 77% of the matched pairs. These two findings are in agreement with previous studies conducted by Allen et al. (1991) and DeRosa et al. (2000). Indeed, in the study of Allen et al. (1991) conducted among 119 feedlot calves, a moderate agreement (Kappa = 0.47) between nasopharyngeal and bronchoalveolar lavage for the isolation of M. haemolytica was observed. Unfortunately, no typing method was applied to the isolates obtained in that study. DeRosa et al. (2000) compared nasal and trans-tracheal swabs performed in 40 feeder calves suffering from BRD. Their results show that the same bacterial species (M. haemolytica, P. multocida and/or Histophilus somni) were recovered from nasal and transtracheal swabs in only 68.4% of the paired samples. Furthermore, they performed ribotyping of M. haemolytica isolates and observed the same profiles in both nasal and transtracheal swabs in only 70% of the matched pairs (n = 24). Knowing that the discriminatory power of ribotyping is

much lower than PFGE (Kodjo et al., 1999), it can be hypothesized that less than 70% of their paired samples would have turned out similar if PFGE had been used. The moderate agreement between NS and TTA samples could firstly be explained by a difference in the M. haemolytica populations between the nasal cavities and the lung. It is generally accepted that M. haemolytica initially proliferates in the upper respiratory tract before gaining access to the lungs via aerosolized droplets (Rice et al., 2007). However, the source of the M. haemolytica that eventually colonize the lung is still unknown: it can originate from the nasal cavities and/or from the tonsils (Rice et al., 2007). Indeed, both nasal cavities and tonsils have been identified as a reservoir for M. haemolytica (Frank and Briggs, 1992; Briggs et al., 1998) and cattle may be negative for M. haemolytica on a culture of NS and positive on a culture of the tonsils (Frank et al., 1994). Furthermore, Briggs et al. (1998) have shown that the M. haemolytica population in the tonsils is not always identical to the population of the nasal cavities within the same animal. Therefore, the difference between NS and TTA samples observed in this study could be explained by the fact that the bacterium colonizing the lung came from the tonsils and not from the nasal cavities. Another explanation could be a bias due to the sampling methods. Indeed, only one colony of M. haemolytica per clinical sample was subcultured and typed by PFGE in the present study. It can therefore be hypothesized that the same clone of M. haemolytica would have been more frequently recovered from both samples, if more than one colony per sample were subcultured and typed by PFGE. However, previous studies have shown that the genetic diversity of Pasteurellaceae within a NS is low. Indeed, Briggs et al. (1998), who subcultured on average eight M. haemolytica colonies from each NS performed on healthy and sick feedlot calves (n = 55 NS), observed colonies with different plasmid profiles in only four of the 55 samples. Klima et al. (2011), who recently examined the genetic diversity by PFGE of M. haemolytica collected from the nasopharynx of healthy feedlot cattle, typed at least three colonies per NS and reported that, in the majority of cases, the isolates were identical or displayed only one band shift (C. Klima, personal communication). Finally, Hotchkiss et al. (2009), who studied the genetic heterogeneity of P. multocida in 15 NS by typing up to 10 colonies per plate by RAPD, also observed undistinguishable RAPD profiles in 12 of 15 NS. Based on these previous studies, it seems that one colony may be sufficient to represent the bacterial population within a NS. Because no data was available in the literature concerning the genetic diversity of M. haemolytica in TTA samples, we assessed its genetic diversity by typing eight to ten M. haemolytica colonies from eight randomly selected TTA samples. Isolates from each TTA sample showed undistinguishable PFGE profiles for all eight samples. This finding indicates that the genetic diversity of M. haemolytica within a TTA sample is low and that one colony may be sufficient to represent the M. haemolytica population within this sample. Therefore, based on these

E. Timsit et al. / Veterinary Microbiology 161 (2013) 295–304

results and previous studies, it seems that the 77% identity of M. haemolytica isolates recovered from paired NS and TTA samples observed in our study was more likely due to a difference in the bacterial populations between NS and TTA samples than to an artifact resulting from sampling methods. However, further research is needed to confirm this statement. Indeed, the absence of a strain (or a clone) from a sample may be inferred with a certain degree of confidence only after a minimum required number of colonies from the sample are tested (Do¨pfer et al., 2008). Therefore, in the future, at least a minimum required number of colonies (and not only one colony) should be characterized per sample to ensure that all the strains (or clones) that are present in a sample are detected with sufficient confidence. This minimum number of colonies can be derived from Bayesian inference as previously described by Do¨pfer et al. (2008). 5. Conclusion In conclusion, the PFGE analysis of M. haemolytica isolates recovered from the lower respiratory tract of bulls during BRD episodes revealed a significant within-pen genetic diversity. This finding suggests that BRD episodes associated with M. haemolytica are not primarily due to the contagious spread of a single virulent clone among cattle and highlights the importance of predisposing factors such as viral infections and/or stressors that enable the resident flora to overcome the cattle’s immune system. The moderate agreement between the isolation of M. haemolytica from NS and TTA samples observed in the present study indicates that a NS culture does not accurately predict the presence of M. haemolytica deeper within the respiratory tract. Furthermore, when both samples are M. haemolytica positive, isolates recovered from NS are not always representative of the isolates recovered from TTA. Therefore, to study the genetic diversity of M. haemolytica during a BRD episode, it seems more relevant to sample the lower respiratory tract than the nasal cavities. Conflict of interest statement None of the authors has financial or personal relationships with other peoples or organisations that could inappropriately influence the present work.

Acknowledgements The authors want to thank F. Leray, E. Blandin, T. Guyot, P.R. Mortensen and T.P. Bønnelycke for skillful assistance with PFGE and A. Lehebel for help with the statistical analyses. T.A. McAllister, T.W. Alexander and C.L. Klima are also kindly acknowledged for the serotyping of M. haemolytica isolates and their helpful comments. The study was part-funded by MSD Animal Health and partfunded by the French ministry of Agriculture under project MOZAE.

303

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.vetmic.2012.07.044. References Allen, J.W., Viel, L., Bateman, K.G., Rosendal, S., Shewen, P.E., PhysickSheard, P., 1991. The microbial flora of the respiratory tract in feedlot calves: associations between nasopharyngeal and bronchoalveolar lavage cultures. Can. Vet. J. 55, 341–346. Angen, O., Thomsen, J., Larsen, L.E., Larsen, J., Kokotovic, B., Heegaard, P.M., Enemark, J.M., 2009. Respiratory disease in calves: microbiological investigations on trans-tracheally aspirated bronchoalveolar fluid and acute phase protein response. Vet. Microbiol. 137, 165–171. Assie, S., Seegers, H., Makoschey, B., Desire-Bousquie, L., Bareille, N., 2009. Exposure to pathogens and incidence of respiratory disease in young bulls on their arrival at fattening operations in France. Vet. Rec. 165, 195–199. Briggs, R.E., Frank, G.H., Purdy, C.W., Zehr, E.S., Loan, R.W., 1998. Rapid spread of a unique strain of Pasteurella haemolytica serotype 1 among transported calves. Am. J. Vet. Res. 59, 401–405. Catry, B., Dewulf, J., de Kruif, A., Vanrobaeys, M., Haesebrouck, F., Decostere, A., 2007. Accuracy of susceptibility testing of Pasteurella multocida and Mannheimia haemolytica. Microb. Drug Resist. 13, 204–211. Cohen, J., 1960. A coefficient of agreement for nominal scales. Educ. Psychol. Meas. 20, 37–46. Do¨pfer, D., Buist, W., Soyer, Y., Munoz, M.A., Zadoks, R.N., Geue, L., Engel, B., 2008. Assessing genetic heterogeneity within bacterial species isolated from gastrointestinal and environmental samples: how many isolates does it take? Appl. Environ. Microbiol. 74, 3490–3496. DeRosa, D.C., Mechor, G.D., Staats, J.J., Chengappa, M.M., Shryock, T.R., 2000. Comparison of Pasteurella spp. simultaneously isolated from nasal and transtracheal swabs from cattle with clinical signs of bovine respiratory disease. J. Clin. Microbiol. 38, 327–332. Espinasse, J., Alzieu, J.P., Papageorgiou, C., Beguin, J.C., Van Gool, F., 1991. Use of transtracheal aspiration to identify pathogens in pneumonic calves. Vet. Rec. 129, 339. Frank, G.H., Briggs, R.E., 1992. Colonization of the tonsils of calves with Pasteurella haemolytica. Am. J. Vet. Res. 53, 481–484. Frank, G.H., Briggs, R.E., Loan, R.W., Purdy, C.W., Zehr, E.S., 1994. Serotypespecific inhibition of colonization of the tonsils and nasopharynx of calves after Pasteurella haemolytica serotype A1 after vaccination with the organism. Am. J. Vet. Res. 55, 1107–1110. Garcia, F., Agabriel, J., Micol, D., 2007. Alimentation des bovins en croissance et a` l’engrais. In: INRA (Eds.), Feeding of Cattle, Sheep and Goats: Animals’ Requirements and Feed Value. Quae Editions. Versailles, France, pp. 89–120. Griffin, D., Chengappa, M.M., Kuszak, J., McVey, D.S., 2010. Bacterial pathogens of the bovine respiratory disease complex. Vet. Clin. North Am. Food Anim. Pract. 26, 381–394. Hedegaard, L., Christensen, H., Chadfield, M.S., Christensen, J.P., Bisgaard, M., 2009. Association of Streptococcus pluranimalium with valvular endocarditis and septicaemia in adult broiler parents. Avian. Pathol. 38, 155–160. Hotchkiss, E.J., Hodgson, J.C., Zadoks, R.N., 2009. Heterogeneity of Pasteurella multocida in Scottish calves at animal, farm and regional levels. In: Proceedings of the 12th Symposium of the International Society for Veterinary Epidemiology and Economics, Durban, South Africa, p. 103. Klima, C.L., Alexander, T.W., Read, R.R., Gow, S.P., Booker, C.W., Hannon, S., Sheedy, C., McAllister, T.A., Selinger, L.B., 2011. Genetic characterization and antimicrobial susceptibility of Mannheimia haemolytica isolated from the nasopharynx of feedlot cattle. Vet. Microbiol. 149, 390–398. Klima, C.L., Alexander, T.W., Selinger, L.B., Read, R.R., Shewan, P.E., Gow, S.P., Booker, C.W., McAllister, T.A., 2010. Comparison of repetitive PCR and pulsed-field gel electrophoresis for the genotyping of Mannheimia haemolytica. J. Microbiol. Methods 81, 39–47. Kodjo, A., Villard, L., Bizet, C., Martel, J.L., Sanchis, R., Borges, E., Gauthier, D., Maurin, F., Richard, Y., 1999. Pulsed-field gel electrophoresis is more efficient than ribotyping and random amplified polymorphic DNA analysis in discrimination of Pasteurella haemolytica strains. J. Clin. Microbiol. 37, 380–385.

304

E. Timsit et al. / Veterinary Microbiology 161 (2013) 295–304

Landis, J.R., Koch, G.G., 1977. The measurement of observer agreement for categorical data. Biometrics 33, 159–174. Miles, D.G., 2009. Overview of the North American beef cattle industry and the incidence of bovine respiratory disease (BRD). Anim. Health Res. Rev. 10, 101–103. Purdy, C.W., Scanlan, C.M., Loan, R.W., Foster, G.S., 1993. Identification of Pasteurella haemolytica A1 isolates from market-stressed feeder calves by use of enzyme and antimicrobial susceptibility profiles. Am. J. Vet. Res. 54, 92–98. Quinn, P.J., Carter, M.E., Markey, B.K., Carter, G.R., 1994. Pasteurella species. In: Quinn, P.J., Carter, M.E., Markey, B.K., Carter, G.R. (Eds.), Clinical Veterinary Microbiology. Mosby, Edinburgh, UK, pp. 254–259. Rice, J.A., Carrasco-Medina, L., Hodgins, D.C., Shewen, P.E., 2007. Mannheimia haemolytica and bovine respiratory disease. Anim. Health Res. Rev. 8, 117–128. Smith, R.A., 1998. Impact of disease on feedlot performance: a review. J. Anim. Sci. 76, 272–274.

Taylor, J.D., Fulton, R.W., Dabo, S.M., Lehenbauer, T.W., Confer, A.W., 2010a. Comparison of genotypic and phenotypic characterization methods for Pasteurella multocida isolates from fatal cases of bovine respiratory disease. J. Vet. Diagn. Invest. 22, 366–375. Taylor, J.D., Fulton, R.W., Lehenbauer, T.W., Step, D.L., Confer, A.W., 2010b. The epidemiology of bovine respiratory disease: what is the evidence for preventive measures? Can. Vet. J. 51, 1351–1359. van Belkum, A., Tassios, P.T., Dijkshoorn, L., Haeggman, S., Cookson, B., Fry, N.K., Fussing, V., Green, J., Feil, E., Gerner-Smidt, P., Brisse, S., Struelens, M., 2007. Guidelines for the validation and application of typing methods for use in bacterial epidemiology. Clin. Microbiol. Infect. 13 (Suppl. 3), 1–46. Villard, L., Gauthier, D., Lacheretz, A., Abadie, G., Game, Y., Maurin, F., Richard, Y., Borges, E., Kodjo, A., 2006. Serological and molecular comparison of Mannheimia haemolytica and Pasteurella trehalosi strains isolated from wild and domestic ruminants in the French Alps. Vet. J. 171, 545–550.