Evaluating the microbiome of two sampling locations in the nasal cavity of cattle with bovine respiratory disease complex (BRDC)

Abstract

Bovine respiratory disease complex (BRDC) is a multifactor disease, and disease incidence may be associated with an animal’s commensal bacterial populations (microbiome) in the upper nasal cavity. Identifying these commensal bacterial populations in the upper nasal cavity may help us to understand the impact of the microbiome on incidence of BRDC in cattle. Various sampling techniques have previously been utilized to evaluate the microbiome of different locations of the upper nasal cavity in cattle. Therefore, our objective was to determine whether bacterial populations of the nasal cavity vary based on these sampling locations. Two common sampling techniques were evaluated, including 6-inch nasal swabs and deep nasopharyngeal swabs. Nasal swabs from calves were collected when the animal was diagnosed with BRDC after weaning in the feedlot in addition to collection of samples from asymptomatic cohorts. Samples were pooled in groups based on year the animal was in the feedlot (2015 or 2016), when the animal was diagnosed with BRDC (1 to 5 weeks after weaning), type of sample (6-inch nasal swab or deep nasopharyngeal swab), and health status (diagnosis with BRDC or control). Variable regions 1 through 3 along the 16S rRNA gene were amplified by PCR and sequenced using next-generation sequencing (Illumina MiSeq) for identification of the bacterial taxa present. Overall, sampling site did not consistently influence diversity of the bacterial populations of the upper nasal cavity. However, the effect of disease incidence on the microbiome was depended on sampling time after weaning (P = 0.0462) for 2015, while the main effects of sampling time after weaning (P = 0.00992) and disease phenotype (P = 0.012) were significant for 2016. These data for 2016 demonstrate that in addition to bacterial profiles changing throughout weaning, calves diagnosed with BRDC have different bacterial profiles compared to their control cohorts. In addition, evaluation of the microbiome identified predominant bacteria genera in the upper nasal cavity included those previously reported to be associated with cattle diagnosed with BRDC including Mycoplasma sp., Psychrobacter sp., and Mannheimia sp. In summary, these results demonstrate that shorter, less invasive 6-inch nasal swabs produce similar results to deep nasopharyngeal swabs.

INTRODUCTION

Bovine respiratory disease complex (BRDC) is an involved multifactor disease, which is the most expensive animal disease afflicting herds in U.S. beef cattle industry, costing the industry over $1 billion annually (Griffin et al., 2010; USDA, 2013). Animals may be predisposed to suffer from BRDC by a variety of viral and bacterial agents and it is increasingly evident that disease incidence is intimately associated with an animal’s commensal bacteria (microbiome; Holman et al., 2015; Taylor et al., 2015; Timsit et al., 2016; Johnston et al., 2017; Holman et al., 2017). Therefore, evaluation of the animal’s commensal bacterial populations in the upper nasal cavity may help us to understand the impact of the microbiome on incidence of BRDC in cattle.

Deep nasopharyngeal swabs are a common sampling technique of the nasal cavity (Holman et al., 2015; Taylor et al., 2015; Timsit et al., 2016; Johnston et al., 2017), as they are proposed to provide an accurate representation of the bacterial populations present. However, the deep nasopharyngeal swabs are guarded to protect from contaminating bacteria during sampling and are therefore greater in diameter than the common unguarded nasal swab. Deep nasopharyngeal swabs are also more expensive than the 6-inch swab and traditionally require additional restraint of the calf during sampling. As a result, deep nasopharyngeal swabs may increase stress on the animal, are not ideal for multiple samplings of the same animal, and are not cost effective for sampling large numbers of calves. Previous research (Doyle et al., 2017) evaluated four common sampling techniques of the upper nasal cavity and identified similar respiratory pathogens among sampling methods including 6-inch nasal swabs and deep nasopharyngeal swabs from calves diagnosed with BRDC using culture techniques. To expand upon this research, we propose to evaluate 6-inch nasal swabs and deep nasopharyngeal swabs utilizing next-generation sequencing of the 16S ribosomal gene to evaluate all microbial pathogens present and compare the microbiomes of these sampling sites.

MATERIALS AND METHODS

Animal Populations

All animal use was approved by the U.S. Meat Animal Research Center Animal Care and Use Committee. Data were collected in 2015 and 2016 in advanced generations of the U.S. Meat Animal Research Center GPE (Germplasma Evaluation Program; Schiermiester et al., 2015) herd, Clay Center, Nebraska. This particular GPE subset of approximately 800 animals was produced each year (Table 1) and was a product of multiple-sire matings of crossbred cows to purebred and crossbred bulls of varying breed composition. The resulting animals used within this study consisted of variable fractions of 18 breeds: Angus, Hereford, Red Angus, Brahman, Charolais, Gelbvieh, Limousin, Simmental, Brangus, Beefmaster, Shorthorn, Maine Anjou, Santa Gertrudis, Chiangus, Salers, Braunvieh, South Devon, and Tarentaise.

Table 1.

Number of animals in populations for 2015 and 2016

Year	Total animal number	Treated for BRDC	Controls sampled
2015	820	35	72
2016	794	16	78

Nasal Swabs and DNA Samples

Nasal swabs were collected from the upper nasal cavity using 6-inch nasal swabs and deep nasopharyngeal swabs of calves diagnosed with BRDC during a 1- to 5-week period after weaning (approximately 170 days of age at weaning). Animals presenting with clinical signs of BRDC (signs including lethargy, nasal discharge, elevated body temperature, and distressed breathing) were examined by the trained staff and then sampled. Twenty-four control cohorts that did not display signs of disease and had not been diagnosed with BRDC previously were also sampled each week after weaning for 3 weeks in 2015 and 2016. Additionally, in 2016 a decreased number of calves were diagnosed with BRDC in week 4 and 5 after weaning. Therefore, for week 4 and 5 an equal number of control cohorts were sampled as calves diagnosed with BRDC.

For sampling, the 6-inch nasal swab was gently inserted first into the nasal cavity at an approximate depth of 6 inches. The nasal swab was then rotated and removed. The guarded deep nasopharyngeal swab was then gently inserted into the nasal cavity at an approximate depth of 8 inches. The swab was then pushed out from the guard and advanced an additional inch, rotated, and brought back into the guard and removed. After collection of the sample, 6-inch nasal swabs were placed in buffered peptone water with 12% glycerol, drop frozen in liquid nitrogen directly after collection, and stored at −80°C for both 2015 and 2016. For the deep nasopharyngeal swab, the swab was placed in Liquid Amies, placed on ice after collection, and stored at 4°C for 2015 samples and −80°C for 2016 samples. The deep nasopharyngeal samples for 2015 were initially stored at 4°C due to the samples first being evaluated for bacterial profiles by spreading a portion of Liquid Amies media on agar plates (data not shown). This was completed within 24 hours of the samples being collected and the samples were then stored at −80°C for subsequent DNA extraction. Total DNA was extracted from each swab using a commercial kit (MoBio, Carlsbad, CA) and initial DNA quantity was evaluated with a DNA spectrophotometer. Equal amounts of DNA from each swab were then pooled within collection year (2015 or 2016), sample collection type (deep nasopharyngeal swab or 6-inch swab), time point after weaning (1, 2, 3, 4, or 5 weeks after weaning), and phenotype (diagnosed with BRDC or control cohort) for amplification of the 16S rRNA gene variable region using standard PCR (AccuPrime, Invitrogen, Carlsbad, CA) and primers that amplify variable region 1 through 3 of the 16S gene (Myer et al., 2015). Quality and quantity of the resulting 16S rRNA gene amplification was checked on the Fragment Analyzer (Advanced Analytical, Ankeny, IA) and then sequenced utilizing the MiSeq Illumina Sequencer (Illumina, San Diego, CA) with a MiSeq Reagent Kit v3 to generate 2 × 301 paired-end reads. For 2015 (Table 2), two pools were evaluated for each week (3 weeks total) after weaning for each phenotype (diagnosed with BRDC or control cohort). For 2016 (Table 3), up to two pools were evaluated each week depending on the number of calves that were diagnosed with BRDC. Number of calves represented in each year and pool is defined in Tables 2 and 3.

Table 2.

Number of animals in pools for 2015

Pool number	Weeks after weaninga	Phenotypeb	Number of animals in pool
1	1	Control	5
2	1	Control	5
3	2	Control	5
4	2	Control	5
5	3	Control	5
6	3	Control	5
7	1	BRDC	5
8	1	BRDC	5
9	2	BRDC	5
10	2	BRDC	5
11	3	BRDC	5

^aCalves were sampled each week for a total of 3 weeks after weaning.

^bPhenotypes included control cohorts (Control) that were sampled each week and animals diagnosed with BRDC (BRDC).

Table 3.

Number of animals in pools for 2016

Pool number	Weeks after weaninga	Phenotypeb	Number of animals in pool
1	1	Control	4
2	1	Control	4
3	2	Control	6
4	2	Control	6
5	3	Control	6
6	3	Control	6
7	4	Control	2
8	4	Control	2
9	5	Control	1
10	5	Control	1
11	1	BRDC	1
12	2	BRDC	2
13	2	BRDC	2
14	3	BRDC	4
15	3	BRDC	4
16	4	BRDC	1
17	5	BRDC	1
18	5	BRDC	1

^aCalves were sampled each week for a total of 5 weeks after weaning.

^bPhenotypes included control cohorts (Control) that were sampled each week and animals diagnosed with BRDC (BRDC).

Data Analysis

The paired-end data files (fatsq.gz format) for each DNA pool were downloaded from the MiSeq Illumina Sequencer and initially processed through Geneious (10.2.3; www.geneious.com). Briefly, the paired-end files were transferred to Geneious where paired-end reads were identified, merged, and low quality (<Q20) sequence reads were removed. Reads that did not merge were removed from the data set. Resulting sequence reads were then submitted to GreenGenes (13.5, May 2013; BaseSpace, Illumina, San Diego, CA) for identification of 16S genus classification.

Statistical Methods

Comparison of samples for evaluation of bacterial populations requires a multivariate approach because there are many microbial genera to be evaluated as response variables in the characterization of the microbiome. Effects tested for this study included sampling site (S, 6-inch nasal swab versus deep nasopharyngeal swab), time (T, weeks after weaning, 1 to 5 weeks), disease phenotype (D, diagnosed with BRDC versus control cohort), and all three 2-way interactions and the 3-way interaction using permutational multivariate analysis of variance (PERMANOVA; Tang et al., 2016). Euclidean distance with the PERMANOVA test was used, which would be heavily influenced by the most abundant taxa at the genus level. Analyses were conducted and are reported separately for birth year (2015 and 2016). Cattle were sampled for 3 weeks after weaning in 2015 and 5 weeks after weaning in 2016. All statistical tests were based on 500,000 permutations. In addition, a nested model with sampling site nested within animal pools was tested. Similarity between “sampling site within animal” was compared to “between animal” distances using the average distance among sampling sites within pool. A small average distance among sampling sites within pool would indicate that the difference between sampling sites was small compared to other differences among pooled DNA samples. Microbiome diversity (α, β, and γ) was also estimated using the R package entropart (Marcon and Hérault, 2015). The function DivPart() was used with q = 1 and Biased = FALSE.

RESULTS AND DISCUSSION

A comparison of two common sampling techniques (6-inch nasal swab and deep nasopharyngeal swab) was completed in cattle diagnosed with BRDC and control cohorts weaned in 2015 and 2016 to evaluate sampling techniques and the bacterial populations present. Mean distance between sampling sites in taxa composition within animal pool was small compared to the null-permuted distribution (Figure 1; Table 4). The mean distance in taxa composition between sampling sites within pool was 0.325 for 2015 and 0.286 for 2016, which were in the extreme lower tails of the permuted distribution of distances between sampling site and between animal pool; P = 0.0182 for 2015 and P = 2 × 10⁻⁶ for 2016 (Figure 1a and b). Together, these data indicate that microbial communities are more similar between sampling sites within animals (or animal pools) than between animals (or animal pools) within or between sampling site and that these differences and similarities among sampling sites and animals are stable from year to year.

Figure 1.

The mean distance between sampling sites within pool. Mean distance between sampling sites (deep nasopharyngeal swab and 6-inch nasal swab) was determined for 2015 (a) and 2016 (b) samples. The mean distance between sampling site within pool was 0.325 for 2015 and 0.286 for 2016, as identified by the red dashed line (a and b). Mean distance in taxa composition (microbiome) between sampling sites was lower than expected by chance; P = 0.018 for 2015 and P = 2 × 10^–6 for 2016.

Table 4.

P values from analysis of the effect of sampling site (S), time (T), and disease phenotype (D) on bacterial profiles in calves after weaning

Effect	2015	2016
Sampling site (S)	0.470	0.822
Time (T)	0.895	0.00992*
Disease phenotype (D)	0.607	0.0120*
S × T	0.873	0.995
S × D	0.235	0.473
T × D	0.0462*	0.169
S × T × D	0.929	0.910

*P < 0.05.

In addition to comparing the two sampling sites, changes in microbial profiles during the weeks after weaning were also evaluated for 2015 and 2016. Overall, in 2015, the effect of disease on bacterial profiles was depended on time after weaning as evidenced by the time by disease interaction (T × D; P = 0.0462) and nonsignificant main effects for time and phenotype (T, D, S × T, S × D, and S × T × D; Table 4). In contrast, in 2016, the main effects of time after weaning (P = 0.00992) and disease phenotype (P = 0.012) were significant (Table 4) and there was no evidence of differences due to S × T, S × D, T × D, and S × T × D. These data for 2016 demonstrate that in addition to bacterial profiles changing throughout the 5 weeks after weaning, calves diagnosed with BRDC had different bacterial profiles compared to their cohorts that did not display symptoms of BRDC. Furthermore, predominant bacteria genera for the 2 years evaluated for this study included those previously reported to be associated with cattle diagnosed with BRDC including Mycoplasma sp., Psychrobacter sp., and Mannheimia sp. (Figures 2 and 3; Holman et al., 2015; Taylor et al., 2015; Timsit et al., 2016; Johnston et al., 2017; Holman et al., 2017). Throughout the 3 weeks after weaning for 2015 (Figure 2), Mycoplasma sp. was the predominant bacterial genera in both calves diagnosed with BRDC and their control cohorts (average 24.94%). However, for 2016, the predominant bacterial genera present changed throughout the 5 weeks after weaning and between calves diagnosed with BRDC and their control cohorts (Figure 3). A greater percentage of sequence reads was classified as Mycoplasma sp. at 2 weeks after weaning (7.66–44.75%), while Psychrobacter sp. was the predominant genus at most other time points after weaning (0.91–48.38%). When comparing bacterial genera present in calves diagnosed with BRDC to their control cohorts, Psychrobacter sp. was greater in calves diagnosed with BRDC (28.27–48.38%) compared to their control cohorts (1.88–14.83%) in week 4 and 5, while Morexella sp. was greater in calves diagnosed with BRDC (6.40–26.21%) compared to their control cohorts (0.42–11.02%) throughout all 5 weeks after weaning. Together these data suggest that bacterial genera change throughout weaning and may be playing a role in BRDC development in calves after weaning. Further research needs to be completed to fully understand their role in development of BRDC in feedlot calves. Furthermore, while previous research has documented the change in bacterial profiles after weaning in beef cattle (Holman et al., 2015; Taylor et al., 2015; Timsit et al., 2016; Johnston et al., 2017; Holman et al., 2017), the data presented herein are the first to document this change in commensal bacterial profiles across multiple weeks after weaning in addition to the comparison between calves diagnosed with BRDC and control cohorts.

Figure 2.

Taxonomy classification for 2015. Percentage of 16S rRNA gene sequence reads classified to genus level for 2015 nasal swabs. Top eight bacterial genera in abundance are reported and remaining genera of low abundance are grouped and identified as other. Sequence reads that could not be classified to the genus level and are identified as unclassified.

Figure 3.

Taxonomy classification for 2016. Percentage of 16S rRNA gene sequence reads classified to genus level for 2016 nasal swabs. Top eight bacterial genera in abundance are reported and remaining genera of low abundance are grouped and identified as other. Sequence reads that could not be classified to the genus level and are identified as unclassified.

When evaluating taxa sequence coverage depth for the 6-inch and deep nasopharyngeal swabs, sequence coverage depth for deep nasopharyngeal samples for 2015 was much lower when compared to 2016. Mean taxa coverage for deep nasopharyngeal samples was 117,659 (range 49,983 to 249,644) sequence reads compared to 1,164,190 (range 596,938 to 1,522,511) sequence reads for 6-inch nasal swab samples in 2015; mean deep nasopharyngeal depth was 0.1X of 6-inch nasal swabs. In comparison, mean taxa coverage for deep nasopharyngeal samples was 1,315,460 (range 68,456 to 2,579,348) sequence reads compared to 638,748 (range 253,328 to 1,786,088) sequence reads for 6-inch nasal swabs in 2016; deep nasopharyngeal depth was 2.06X of 6-inch nasal swabs. This decrease in number of high-quality sequence reads for the deep nasopharyngeal samples in 2015 may be influenced by the initial storage of the sample at 4°C and may have contributed to the observed lack of significant differences due to T, D, S × T, S × D, and S × T × D.

Because of this difference in sequence coverage between the deep nasopharyngeal swabs from 2015 and all other nasal swabs collected in 2015 and 2016, we evaluated α, β, and γ diversity to see if these measures were reduced by the low sequence coverage of deep nasopharyngeal swabs in 2015. In 2015, α, β, and γ diversity were 33.0, 2.2, and 71.6, respectively, for deep nasopharyngeal swabs and 31.7, 1.8, and 54.4, respectively, for 6-inch nasal swabs collected in 2015 giving no indication of less bacterial diversity correlated with the reduced sequencing depth of deep nasopharyngeal swabs collected in 2015. Similarly in 2016, α, β, and γ diversity were 26.6, 2.04, and 54.4, respectively, for deep nasopharyngeal swabs and 27.9, 2.0, and 50.5, respectively, for 6-inch swabs. Overall, α, β, and γ diversity indicates that bacterial diversity was not consistently influenced by sampling site or year of sample collection.

Conclusions

Overall, we were able to demonstrate through next-generation sequencing that bacterial populations are more similar between sampling sites than between animals (individually and in pools) for both common sampling techniques of the nasal cavity. These sequence data support the previous findings by Doyle et al. (2017), which utilized culturing techniques to identify similarities in bacterial profiles from different sampling sites and techniques. The data presented herein also identified bacteria genera previously reported to be associated with cattle diagnosed with BRDC and demonstrates that bacterial profiles change throughout weaning in calves diagnosed with BRDC. Together these data suggest that the less invasive 6-inch nasal swab can be used to study the upper respiratory microbial populations in larger numbers of animals in relation to calves diagnosed with BRDC for a lower cost and inflicting less stress on individual animals. Furthermore, more comprehensive sampling of larger numbers of animals by using the less invasive swab would be expected to yield greater experimental power and greater scientific progress.