Eighty newly-weaned Black Angus and Black Simmental beef steer calves were transported for approximately 8 h from the ranch of origin to the Lethbridge Research and Development Centre (LeRDC), as described previously (Meléndez et al., 2020). The day after arrival, calves received a 7-way bovine clostridial vaccine (Ultrabac/Somubac, Zoetis Canada Inc., Kirkland, Quebec, Canada); a 5-way bovine viral diarrhea, rhinotracheitis, parainfluenza and bovine respiratory syncytial virus vaccine (Pyramid FP 5 + Presponse SQ, Boehringer Ingelheim, Burlington, Ontario, Canada); an antibiotic (Draxxin, Zoetis Canada Inc., Kirkland, Canada); and an anti-parasitic agent (Ivomec Pour-on for Cattle, Boehringer Ingelheim, Burlington, Ontario, Canada). The calves were vaccinated, ear tagged, and acclimatized for 26 days to eat a grain diet from the feed bunk, and drink from the water trough in the feedlot pens, prior to the start of the trial. This feedlot acclimation was done to reduce the number of stressors experienced by the calves prior to transportation. Calves were then randomly assigned to two treatments and were transported separately for an initial 36 h time period, followed by the assigned rest stop time (0H or 12H), and transported for an additional 4 h in the same trailers for delivery to the LeRDC feedlot. The treatments were (N=40 per treatment): 1) calves with an initial transport of 36 h, followed by 0 h rest and 4 h additional transport, and 2) calves with an initial transport of 36 h, followed by 12 h rest and 4 h additional transport. There were 40 calves in each treatment group, housed in 4 pens (10 animals/pen). A detailed schematic depiction of the experimental design is presented in Figure 1 .

For each treatment, 12 randomly selected calves (3 calves/pen) were assigned for repetitive sampling. Calves were sampled before loading (BL), after unloading (AU), 1 day AU, 3 days AU, and 28 days AU. Nasopharyngeal (NP) swabs were collected from the right nasal cavity while calves were restrained in a chute. Prior to sampling, the nostril was wiped clean with 70% ethanol. Sterile swabs (27 cm) within a protective outer sheath (MW 124, Medical Wire & Equipment, Corsham, England) were inserted through the nasal passage to the nasopharynx (approximately 70% the head length). The swabs were extended beyond the sheath and rotated in a circular direction for approximately 10 s to collect nasopharyngeal samples as described previously (Holman et al., 2017). Swab tips were then cut and placed into sterile 1.5mL tubes kept on ice. Samples were transported to the lab and stored at -80°C within one hour of collection.

Eighty crossbred steer calves (Black or Red Angus × Hereford/Simmental and Black or Red Angus × Charolais) were sourced from two different ranches in southern Alberta, Canada. Upon initial feedlot entry, calves were administered the same vaccines, and antimicrobial as Study 1, and were fed a silage-based diet described previously (Meléndez et al., 2022). However, in Study 2, calves were not acclimated to the feedlot. Calves were randomly assigned to two treatments and were transported separately for an initial 20 h time period, followed by the assigned rest stop time (0H or 8H), and transportation for another 15 h for delivery to the LeRDC feedlot. The treatments were (N=40 per treatment): 1) calves with an initial transport of 20 h, followed by 0 h rest and 15 h additional transport, and 2) calves with an initial transport of 20 h, followed by 8 h rest and 15 h additional transport. A detailed schematic illustrating the experimental design can be found in Figure 1 .

Each treatment group was comprised of 40 calves, which were housed in 4 pens with 10 animals per pen. Microbial parameters were measured from 12 calves/treatment (3 calves/pen). NP samples were collected before loading (BL), after unloading (AU), and then 1, 3, 6, 14, and 28 days AU, and processed as described previously (Meléndez et al., 2022).

Total nucleic acids were extracted from NP swabs as described previously (Holman et al., 2017), including samples collected at all time points from both studies. Briefly, total DNA was extracted from NP swabs using the DNeasy Tissue Kit (Qiagen Inc., Mississauga, ON, Canada), with modifications to optimize bacterial lysis. The cotton tip were removed from the swab and placed in a sterile tube containing 180 µL of enzymatic lysis buffer, which included mutanolysin (300 U/mL) and lysozyme (20 mg/mL). The mixture was vortexed and incubated at 37°C for 1 hour. After enzymatic digestion, 25 µL of proteinase K and 200 µL of Buffer AL (without ethanol) were added to the tube, and the mixture was thoroughly vortexed before being incubated at 56°C for 30 minutes. Approximately 300 mg of sterile 0.1 mm zircon/silica beads were added, and the samples were disrupted using a TissueLyser II (Qiagen) at maximum amplitude for 3 minutes. The samples were then centrifuged at 13,000 × g for 5 minutes, and 200 μL of ethanol was added to the supernatants, followed by vortexing. The remaining steps were conducted according to the manufacturer’s protocol for the DNeasy Tissue Kit. In DNA from all swabs, 16S rRNA gene sequence libraries were constructed using a two-step PCR approach. In the first step, the V4 region of the 16S rRNA gene was amplified utilizing primers 515-F (5′-GTGYCAGCMGCCGCGGTAA-3′) and 806-R (5′-GGACTACNVGGGTWTCTAAT-3′). The second PCR step introduced a unique 10-base pair barcode at the 5′ end of each amplicon and incorporated Illumina adapter sequences (Illumina, San Diego, CA, USA). All PCR amplification and sequencing procedures were carried out at Genome Quebec (Montreal, Quebec, Canada). The amplicon was sequenced on a MiSeq instrument (2 x 250 bp with an expected library size of 25,000 reads, Genome Quebec, Montreal, Quebec, Canada) with the MiSeq Reagent Kit v2.

Sequences were analyzed as described previously (Uddin et al., 2023). Briefly, quality verification and summarization of raw reads for V4 amplicons were conducted using FastQC (v0.11.9) and MultiQC (v1.12) (Ewels et al., 2016). Primer removal and trimming of low-quality read regions were performed using Trimmomatic (v0.39) (Bolger et al., 2014) with specific parameters: HEADCROP:18, SLIDINGWINDOW:5:18, LEADING:3, and TRAILING:3. Subsequent statistical analyses were carried out in R (v4.1.0) (R Core Team, 2022). The trimmed paired reads were filtered and merged using DADA2 (v1.22.0) (Callahan et al., 2016). Bimeric sequences were removed with DADA2’s removeBimeraDenovo function, and taxonomic classification was assigned using SILVA 138 database (Quast et al., 2012). Updates to the genus Mycoplasma taxonomy were applied manually due to database limitations. Statistical analysis of the data were performed using phyloseq 1.38.0 (McMurdie and Holmes, 2013), vegan 2.5–7 (Oksanen et al., 2019), and DESeq2 1.34.0 (Love et al., 2014). Alpha diversity metrics, including Shannon diversity index and observed richness, were calculated using vegan (v2.5-7) (Oksanen et al., 2019) and visualized with ggplot2 (v3.3.5) (Wickham, 2011). Noise was minimized by filtering the ASV table to retain ASVs present in at least 1% of samples with counts ≥ 2. Two-way ANOVA was used to analyze alpha diversity across treatments and time points. Beta diversity was assessed by normalizing filtered ASV counts using DESeq2’s GMPR size factors, with Bray-Curtis distances calculated and visualized using detrended correspondence analysis (DCA) through phyloseq. Microbial community structure differences were evaluated using PERMANOVA (Adonis function) with 9999 permutations in vegan to identify treatment effects at each time point. Taxonomic abundances at the phylum and genus levels were determined using phyloseq and visualized accordingly. Differentially abundant genera were identified using DESeq2 with a negative binomial model with the following predictors: Treatment + Time + Time: Treatment. Genera with significant differences (P < 0.05, log2(FC) > 2 or <-2) were visualized using heatmaps, detailing both genera-level and ASV-specific changes.

In Study 1, during the feeding period, 5 cases of illness were reported. All 5 cases of illness occurred in the 12H rest group, with 4 cases of fever and 1 case of foot rot. In Study 2, two cases of illness were reported throughout the study, both of which were from the 8H rest group. There were no reported cases of mortality in either study.

The raw ASV table contained 6,623 ASVs with a total of 2,825,036 merged paired reads assigned to 118 samples (data not shown). The median number of sequences per sample was 26,027.5 with a minimum of 0 and maximum of 44,819. After filtering, the ASV table contained 1,244 SVs with a total of 2,409,893 merged paired reads assigned to 116 samples. The median number of sequences per sample was 21,386 with a minimum of 1,178 and maximum of 37,204.

The raw ASV table contained 13,773 ASVs with a total of 4,906,901 reads assigned to 168 samples (data not shown). The median number of sequences per sample was 24,743.5 with a minimum of 4,123 and maximum of 82,673. After filtering, the ASV table contained 324 SVs with a total of 4,016,702 reads. The median number of sequences per sample was 19,528.5 with a minimum of 2,124 and maximum of 79,463.

The characteristics of a bacterial community within a specific niche include the number of species present, their numerical composition, and bacterial diversity (Kim et al., 2017). The NP microbial diversity did not change in Study 1, while in Study 2, it increased across feedlot sampling, mainly due to increased presence of less abundant taxa. Similar to Study 2, previous publications have demonstrated an increase in the NP bacterial and archaeal diversity after cattle are placed in a feedlot (Holman et al., 2017; Amat et al., 2019). This was likely due to the feedlot acting as a source of bacteria that colonized the respiratory tract of calves in Study 2 via aspiration. In addition, all calves received tulathromycin upon arrival at the feedlot as part of a standard management practice aimed at mitigating BRD. It has been reported that while tulathromycin can reduce the total bacterial load in the nasopharynx of cattle, its administration increases NP bacterial diversity (Amat et al., 2023). Thus exposure to the feedlot and antimicrobial administration likely selected for an increase in microbial diversity in Study 2. In contrast, these selection pressures were already experienced 26 days prior by calves in Study 1, due to feedlot acclimating, limiting their effect after transportation and unloading in Study 1 animals.

Across both studies, the predominant bacterial phyla were consistent with previous publications investigating NP bacteria of feedlot calves (Timsit et al., 2016; Holman et al., 2017; Stroebel et al., 2018) and included Proteobacteria, Mycoplasmatota, Actinobacteriota, Bacteroidota, and Firmicutes. However, variations have been noted previously and were present in Studies 1 and 2. For instance, Firmicutes accounted for 14.8% of relative abundance in Study 1 but only 2.7% in Study 2. Stroebel et al. (2018) reported that Firmicutes made up only 3% of the nasopharyngeal microbiota in newly weaned feedlot heifers derived from auction or cow-calf ranches, while Holman et al. (2017) found a relative abundance exceeding 20% in calves transported to a feedlot. Similarly, while Proteobacteria accounted for 31.4% and 35.3% of the microbiota in Studies 1 and 2, respectively, this phylum has ranged from 16 - 43% relative abundance in NP microbiota form other studies (McMullen et al., 2018; Uddin et al., 2023). These discrepancies show that while certain bacterial phyla consistently dominate the nasopharyngeal microbiota of calves, their relative abundances can vary due to factors such as the feedlot environment, sampling time, diet, and host-related differences (Alexander et al., 2020) including stress.

In both Studies 1 and 2, numerous genera changed in relative abundance after unloading, compared to the baseline before loading, with some genera sustaining increases or decreases across time. Thus, the microbiota of the bovine NP is dynamic and experiences both transient and stable shifts in populations. Notably, several genera associated with the gastrointestinal tract (e.g. Bififobacterium, Methanobrevibacter, Ruminococcus, and UCG-005) were present in NP samples, which has been observed previously (Holman et al., 2019). Except for Fusobacterium, which typically lines the wall of the rumen (Tadepalli et al., 2009) and sustained increased abundance after unloading in Study 2 calves, most genera associated with the rumen were variable in abundance. This likely highlights their periodic introduction to the upper respiratory tract from the rumen during rumination.

Mycoplasma, Histophilus, and Acinetobacter were three of the most relatively abundant taxa and mainly increased after unloading in both Studies 1 and 2. These genera are often reported to be part of the dominant NP genera and increase after feedlot placement (Holman et al., 2017; Zeineldin et al., 2017; Holman et al., 2018; Stroebel et al., 2018). In both studies, bacteria associated with BRD (Mannheimia, Pasteurella, Histophilus, and Mycoplasma; Booker et al., 2008; Confer, 2009), were part of the 9 most abundant genera, as was Moraxella, which has been shown to increase in cattle that develop bronchopneumonia (McMullen et al., 2020). In Study 1, Mannheima and Moraxella were variable between rest treatments but were observed to increase, while Pasteurella did not change after unloading. In contrast, from one day after unloading, and onwards, each of these genera were persistently reduced in both rest treatment groups in Study 2. The reduction was likely due to the administration of tulathromycin shortly before collection of NP swabs in Study 2 animals. Tulathromcyin administration has been observed to reduce prevalence or abundance of Mannheimia, Pasteurella, and Moraxella in the nasopharynx of cattle (Holman et al., 2019; Uddin et al., 2023; Amat et al., 2023).

BRD is polymicrobial in nature but genera Histophilus, Mannheimia, Mycoplasma, and Pasteurella are most often associated with infection (Booker et al., 2008; Confer, 2009; Timsit et al., 2017). Mannheimia and Pasteurella can be the primary cause of fatal secondary infections after feedlot arrival (Malmuthuge et al., 2021) but Histophilus is increasingly being implicated and presents later in feedlot placement (Timsit et al., 2017). While not typically associated with bronchopneumonia, Moraxella has been observed as a potential risk factor for BRD in received cattle (McMullen et al., 2020) thus was included in our summary of BRD-associated bacteria.

Overall, calves provided rest during transportation were more likely to have increases in abundances of these genera, at more time points after unloading, compared to those provided no rest. This was evident for both feedlot-acclimated (Study 1) and unacclimated (Study 2) calves. Interestingly, in Study 2, there was a sustained increase in Streptococcus abundance in calves provided rest which has been observed previously (Uddin et al., 2023). Although the impact on respiratory health is unknown, this genus encodes sialidases which degrade mucus glycans (Xu et al., 2008, 2011), and therefore could potentially alter protection conferred by mucus in the respiratory tract. Despite colonization by these genera in animals across both rest stop treatments, and healthy and BRD-afflicted cattle in other studies (Timsit et al, 2018), their abundance is important to BRD progression. Enrichment of Mannheimia and Pasteurella in the nasopharynx and lung of BRD morbidities has been observed (Timsit et al., 2018; Centeno-Martinez et al., 2022), and M. haemolytica prevalence has been associated with increased risk of BRD in cattle sampled at feedlot entry (Noyes et al., 2015). The same observations reported here, were also found in our previous study showing rest stop to be associated with increases in Histophilus, Mannheimia, Mycoplasma, Pasteurella, and Streptococcus (Uddin et al., 2023). However, unlike Uddin et al. (2023) we did not observe a reduction in commensal Lactobacillus due to a rest stop. While Lactobacillus can be negatively associated with Pasteurellaceae (Amat et al., 2019), their abundance can vary in cattle. Overall, based solely on increases in abundance of BRD-associated genera in the nasopharynx of calves, providing rest during transportation may be a risk factor for BRD.

Stress during transportation is typically evaluated using physiological and behavioral metrics, with circulating cortisol levels serving as a primary indicator of hypothalamic-pituitary-adrenal axis activation (Buckham-Sporer et al., 2023). Elevated cortisol concentrations have been consistently observed in cattle transport studies (Gupta et al., 2007; Odore et al., 2004; Buckham-Sporer et al., 2007), peaking at the start of the journey (Agnes et al., 1990; Fell and Shutt, 1986; Sartorelli et al., 1992), suggesting that handling during loading is the most stressful phase (Agnes et al., 1990; Grigor et al., 2001; María et al., 2004; Pettiford et al., 2008; Buckham-Sporer et al., 2023). This may explain why cattle provided rest had greater increases in BRD-associated genera, with the additional loading and unloading events adding to host stress. Although it has not been defined how stress affects BRD pathogen colonization, we have previously shown blood stress parameters, such as cortisol and haptoglobin, to be associated with bovine NP respiratory bacteria (Uddin et al., 2023) implying a host-bacterial relationship. In addition, while not evaluated in this study, it is possible that rest stations may increase chance of pathogen exposure, as these facilities are routinely utilized by multiple cattle groups over time. Overall, off-loading and reloading for a rest may elevate stress in calves, potentially leading to increased pathogen abundance. While noteworthy that only animals in rested groups were treated for illness, there were too few animals in our studies to support the association between BRD incidence and rest during transportation, and therefore larger scale studies would help define any relationship.