Clinical responses of calves following weaning
Rectal temperature was measured on experimental day 0 (D0) and every third day throughout the 28 day study period. Temperatures exceeding 40°C were considered a fever and a possible indication of a respiratory infection in the absence of other clinical signs of illness. All calves had temperatures below 40°C on D0. One calf in the Weaned + Transport group had a rectal temperature of 40.2°C on D3 post-weaning and two different calves in the Weaned + Transport group had rectal temperature of 40.2°C and 40.3°C on D12. All other calves in this group and the other two treatment groups had rectal temperatures below 40°C throughout the remainder of the study. Rectal temperatures provided evidence calves remained clinically normal throughout the study with only three calves in the Weaned + Transport group displaying a transient fever. No calves received antibiotic treatment throughout the study.
Weaning, with or without transportation, alters ADR gene expression in blood leukocytes
Analysis of ADR genes expressed in bovine blood leukocytes revealed detectable levels of transcript for all 9 genes. There were significant temporal changes in the expression of the three beta genes, ADRB1, ADRB2, and ADRB3, and one alpha gene, ADRA2A, when compared to time-matched samples from Suckling calves (Figure 1). Expression of the nine ADR genes did not change significantly when compared over time within the group of Suckling calves. However, expression of ADRB1 was significantly (p < 0.001) elevated in Weaned + Transport calves when compared to Suckling calves on D28 post-weaning (Figure 1A). Expression of the ADRB2 was significantly (p < 0.05) upregulated on D2 and D4 in Weaned calves and on D2 post-weaning in Weaned + Transport calves (p < 0.001) when compared to time-matched samples collected from the Suckling calves (Figure 1B). Expression of ADRB3 tended to upregulate (p = 0.06) in Weaned + Transport calves when compared to the Suckling calves on D28 (Figure 1C). Finally, expression of ADRA2A was significantly upregulated in Weaned calves on D4 (p = 0.03) and D8 (p = 0.05) and Weaned + Transport calves on D8 (p = 0.05) when compared to time-matched samples from the group of Suckling calves (Figure 1D).
Serum antibody responses to M. heamolytica and P. multocida
All suckling calves had detectable but low serum antibody titers to both M. haemolytica leukotoxin (Figure 2A) and bacterial lysate proteins (Figure 2B) on experimental D0. These titers remained unchanged in the Suckling group throughout the 28-day observation period. In contrast, M. haemolytica-specific antibody titers were significantly (p < 0.01) increased on D28 (Figure 2A and B) within both the Weaned and Weaned + Transport groups when compared to the Suckling group. M. haemolytica-specific antibody titers on D28 in the Weaned + Transport group were significantly (p = 0.03) greater than the Weaned group.
A similar antibody response to P. multocida was observed when comparing the three treatment groups (Figure 2C). Suckling calves in all groups were seropositive for P. multocida on experimental D0 and antibody titers did not change significantly throughout the 28-day observation period within the Suckling group. There was, however, a significant (p = 0.03) increase in antibody titers specific for P. multocida on D28 in both the Weaned and Weaned + Transport groups when compared within each group to the D0 titers. Furthermore, the increase in antibody titers observed in these two groups on D28 were significantly (p = 0.02) greater when compared to the Sucking group but there was no difference when comparing Weaned and Weaned + Transport groups.
Colonization of the URT by opportunistic pathogens
The URT of beef calves was colonized primarily by bacteria (mean relative abundance: Suckling – 95.2-97.5%; Weaned – 92.4-97.5%; Weaned + Transport – 94.9-98.5%), followed by viruses (mean relative abundance: Suckling – 0.76-3.05%; Weaned – 0.64-6.73%; Weaned + Transport – 0.43-3.33%) and archaea (mean relative abundance: Suckling – 0.03-0.05%; Weaned – 0.02-0.05%; Weaned +Transport – 0.02-0.08%) at all time points sampled (Figure S1). When the microbial composition was compared at the domain level, significant temporal variations were identified from all samples regardless of the treatment group (Table 1). Comparison of the relative abundance of opportunistic pathogens at genus level revealed that the abundance of Mannheimia, Pasteurella, Moraxella, and Histophilus varied temporally regardless of treatment group (Table 1). There were no significant temporal changes in the abundance of opportunistic pathogens identified in the Suckling group. However, the median relative abundance of Mannheimia was highest on D4 after weaning in Weaned calves (Table 1). In Weaned + Transport calves, the median relative abundance of Histophilus was highest on D4, while that of Microvirus (Enterobacteria phage phiX174 sensu lato) was lowest on D2 (Table 1).
Time-dependent variation in the URT microbial community composition
PCA plot visualization of taxonomic profiles generated through metagenomics sequencing of samples from all calves at all time points revealed no effect of weaning, with or without transport, on the URT microbiome (Figure 3A, ANOSIM-R = 0.0091; p = 0.18). All microbial profiles clustered closely except a few individual animals were outliers from the population at individual time points. When the same analysis approach was used within treatment groups, taxonomic profiles clustered closely in both Suckling calves (Figure 3B, ANOSIM-R = 0.2662; p < 0.01) and Weaned calves (Figure 3C, ANOSIM-R = 0.1337; p < 0.01), regardless of sampling time point. In Weaned + Transport calves, however, the PCA plot revealed the URT community no longer clustered closely together (Figure 3D, ANOSIM-R = 0.3418; p < 0.01) and sampling time had a significant effect on microbial community composition.
Comparison of the three treatment groups at individual sampling time points revealed no significant differences in the URT microbial taxonomic composition (Figure S2). In addition, when the same analysis approach was used to understand the effect of stressors and sampling time on microbial functions, all functional profiles were tightly clustered for all groups at all time points (Figure S3).
Colonization of the URT by opportunistic pathogens changes with time but not with stressors
Logistic regression analysis was used to understand the relationship between colonization of the URT by opportunistic pathogens and the possible impact of stressors relative to Suckling calves. This analysis revealed the abundance of most opportunistic pathogenic bacterial genera was not linked to weaning and associated stress, except for Haemophilus (Table 2). Colonization by Haemophilus was significantly decreased in Weaned calves compared to Suckling calves, reflecting a 70% decreased likelihood of high abundance (adjusted OR 0.30, 95%CI 0.11-0.81, p = 0.02; Table 2) when adjusted for sampling time. In addition, Microvirus (Enterobacteria phage phiX174 sensu lato) colonization tended to be increased in Weaned + Transport calves compared to suckling calves, reflecting more than a two-fold increased likelihood (adjusted OR 2.32, 95%CI 0.88-6.43, p = 0.09; Table 3) when adjusted for sampling time.
Colonization by opportunistic bacterial genera, Mannheimia, Pasteurella, Histophilus, Haemophilus, and Moraxella displayed significant temporal variations independent of stressors (Table 2). Colonization by Mannheimia increased significantly on days 2, 4, 8 and 14 when compared to D0, while colonization by Pasteurella increased significantly on days 2, 4, and 8 compared to D0 (Table 2). High abundances of Histophilus and Haemophilus were observed on D2 and D4 compared to D0 (Table 2). In addition, Moraxella colonization increased on D2 and tended to be increased on D4 when comparing to D0 (Table 2). The same analysis was performed using microbial colonization data after weaning (days 2, 4, 8, 14, and 28), which showed a reduction in the colonization of Haemophilus in Weaned calves compared to Suckling calves (adjusted OR 0.33, 95%CI 0.11-0.91, p = 0.04). A low abundance of Haemophilus was observed on D28 (adjusted OR 0.28, 95%CI 0.07-0.98, p = 0.05), whereas a lower abundance of Histophilus was observed on D8 and D14 than D2 (D8 – adjusted OR 0.21, 95%CI 0.50-0.76, p = 0.02, D14 – adjusted OR 0.18, 95%CI 0.05-0.66, p = 0.04).
When logistic regression analysis was performed within each treatment group to further understand the temporal variations in opportunistic pathogen colonization with stressors, use of neither whole microbial data (before and after treatment assignment) or after treatment assignment (post-weaning) data displayed temporal variation in Suckling calves. In contrast, Weaned calves had a higher abundance of Moraxella on D8 (adjusted OR 25, 95%CI 1.7-1058, p = 0.04) and D28 (adjusted OR 15, 95%CI 1.4-408, p = 0.05) as well as a higher abundance of Histophilus on D2 (adjusted OR 14, 95%CI 1.2-384, p = 0.05) compared to D0. In Weaned + Transport calves, Pasteurella and Mannheimia tended to be higher after weaning than D0 (Pasteurella D8 – adjusted OR 15, 95%CI 0.95-665, p = 0.08; Mannheimia D4 – adjusted OR 15, 95%CI 0.90-0.665, p = 0.08; Mannheimia D8 – adjusted OR 15, 95%CI 0.90-665, p = 0.08). Analysis of only post-weaning data revealed a lower abundance of Mannheimia (adjusted OR 0.11, 95%CI 0.01-0.91, p = 0.05) and Histophilus (adjusted OR 0.02, 95%CI 0.005-0.26, p = 0.01) on D14 compared to D2 only in Weaned calves but there was no temporal variation in Weaned + Transport calves.
Microbial functions are related to weaning associated stressors
Use of logistic regression analysis to investigate the relationship between the presence/absence of a microbial function (KEGG Orthology at level 2) with weaning-associated stressors revealed that microbial functions related to “membrane transport”, “replication and repair”, and “metabolism of cofactors and vitamins” were linked to weaning (Table S1). Presence of microbial functions related to “membrane transport” (adjusted OR 0.38, 95%CI 0.14-1, p = 0.05) and “metabolism of cofactors and vitamins” (adjusted OR 0.23, 95%CI 0.08-0.63, p = 0.01) displayed a decreased likelihood in Weaned calves compared to Suckling calves. In contrast, the presence of microbial functions related to “replication and repair” (adjusted OR 3.7, 95%CI 1.3-10.8, p = 0.01) displayed an increased likelihood in Weaned calves compared to Suckling calves. Microbial functions related to “cell motility”, “transport and catabolism”, “signal transduction”, folding, sorting and degradation”, and “transcription” varied only with time when compared to D0 (before assigning treatments) (Table S1). A decreased likelihood of “membrane transport” function was evident post-weaning in both Weaned (adjusted OR 0.27, 95%CI 0.09-0.79, p = 0.02) and Weaned + Transport (adjusted OR 0.27, 95%CI 0.08-0.83, p = 0.03) calves compared to suckling calves. Microbial functions related to “replication and repair” were higher in Weaned calves (adjusted OR 3.4, 95%CI 1.1-10.6, p = 0.03) compared to suckling calves after weaning. “Cell communication”, “cell growth and death”, “cell mortality”, “carbohydrate metabolism”, and “xenobiotics biodegradation and metabolism” were linked only to sampling time points post-weaning (Table S1).
Weaning, with or without transportation, decreases total bacterial density
Estimation of bacterial densities collected via deep nasopharyngeal swabs revealed that calves assigned to the three treatments groups had similar bacterial densities (D0) prior to allocating treatments (Figure 4A). Bacterial densities were, however, significantly lower (p < 0.01) in weaned calves, with (7.18±0.07 log10 16S rRNA gene copy/swab) or without (7.10±0.07 log10 16S rRNA gene copy/swab) transportation, on day 4 after weaning when compared to Suckling (7.76±0.07 log10 16S rRNA gene copy/swab) calves (Figure 4B). Bacterial densities were not statistically different on D8, D14, and D28 post-weaning when comparing among treatment groups (Figure 4B).
It is important to note that total bacterial density in URT samples also displayed temporal variation throughout the experimental period. In Suckling calves, bacterial density on D4 was higher than D2 and D28, whereas in Weaned and Weaned + Transport calves bacterial density on D8 was higher than D2, D4, and D28 (Figure 4B).
Density of opportunistic bacterial pathogens varies with time but not stressor
Estimation of the density of opportunistic pathogenic bacteria in the URT revealed M. haemolytica (MH) and P. multocida (PM) had colonized all suckling beef calves (Figure 4A). Densities of M. haemolytica and P. multocida were not statistically different among the three treatment groups either before (Figure 4A) or after allocation to treatments (Figure 4C and 4D). Similar to total bacterial density, the density of opportunistic bacterial pathogens also displayed temporal variations during the experimental period. The highest density of M. haemolytica in all treatment groups was observed on D7 and D14 (Figure 4C), while the highest density of P. multocida in all treatment groups was observed on D14 (Figure 4D).
Serum antibody responses to M. heamolytica and P. multocida are not related to bacterial abundance in the URT
A negative binomial (NB) regression analysis was used to analyze possible relationships between URT bacteria and serum antibody IgG responses to M. haemolytica and P. multocida. No significant association was identified between serum antibody responses and the relative abundance of M. heamolytica and P. multocida (Table S2). Consistent with previous analyses, the NB model also revealed an increase in antibody responses with time only in Weaned and Weaned + Transport calves but not in Suckling calves (Table S2).
Use of mediation analysis further suggested that the URT microbiota did not mediate the systemic immune responses to opportunistic pathogenic bacteria in any of the calf groups (Figure 5). Once again mediation analysis revealed that antibody responses varied in a time dependent manner in Weaned and Weaned + Transport calves but not in Suckling calves (Figure 5).