Development of Bacterial Therapeutics against the Bovine Respiratory Pathogen Mannheimia haemolytica

Bovine respiratory disease (BRD) is a significant animal health issue impacting the beef industry. Current BRD prevention strategies rely mainly on metaphylactic use of antimicrobials when cattle enter feedlots. However, a recent increase in BRD-associated bacterial pathogens that are resistant to metaphylactic antimicrobials highlights a pressing need for the development of novel mitigation strategies. Based upon previous research showing the importance of respiratory commensal bacteria in protecting against bronchopneumonia, this study aimed to develop bacterial therapeutics that could be used to mitigate the BRD pathogen Mannheimia haemolytica. Bacteria isolated from the respiratory tracts of healthy cattle were characterized for their inhibitory, adhesive, and immunomodulatory properties. In total, 6 strains were identified as having the best properties for use as intranasal therapeutics to inhibit M. haemolytica. If successful in vivo, these strains offer an alternative to metaphylactic antimicrobial use in feedlot cattle for mitigating BRD.

ABSTRACT

Bovine respiratory disease (BRD) is a major cause of morbidity and mortality in beef cattle. Recent evidence suggests that commensal bacteria of the bovine nasopharynx have an important role in maintaining respiratory health by providing colonization resistance against pathogens. The objective of this study was to screen and select bacterial therapeutic candidates from the nasopharynxes of feedlot cattle to mitigate the BRD pathogen Mannheimia haemolytica. In a stepwise approach, bacteria (n = 300) isolated from the nasopharynxes of 100 healthy feedlot cattle were identified and initially screened (n = 178 isolates from 12 different genera) for growth inhibition of M. haemolytica. Subsequently, selected isolates were evaluated for the ability to adhere to bovine turbinate (BT) cells (n = 47), compete against M. haemolytica for BT cell adherence (n = 15), and modulate gene expression in BT cells (n = 10). Lactobacillus strains had the strongest inhibition of M. haemolytica, with 88% of the isolates (n =33) having inhibition zones ranging from 17 to 23 mm. Adherence to BT cells ranged from 3.4 to 8.0 log₁₀ CFU per 10⁵ BT cells. All the isolates tested in competition assays reduced M. haemolytica adherence to BT cells (32% to 78%). Among 84 bovine genes evaluated, selected isolates upregulated expression of interleukin 8 (IL-8) and IL-6 (P < 0.05). After ranking isolates for greatest inhibition, adhesion, competition, and immunomodulation properties, 6 Lactobacillus strains from 4 different species were selected as the best candidates for further development as intranasal bacterial therapeutics to mitigate M. haemolytica infection in feedlot cattle.

IMPORTANCE Bovine respiratory disease (BRD) is a significant animal health issue impacting the beef industry. Current BRD prevention strategies rely mainly on metaphylactic use of antimicrobials when cattle enter feedlots. However, a recent increase in BRD-associated bacterial pathogens that are resistant to metaphylactic antimicrobials highlights a pressing need for the development of novel mitigation strategies. Based upon previous research showing the importance of respiratory commensal bacteria in protecting against bronchopneumonia, this study aimed to develop bacterial therapeutics that could be used to mitigate the BRD pathogen Mannheimia haemolytica. Bacteria isolated from the respiratory tracts of healthy cattle were characterized for their inhibitory, adhesive, and immunomodulatory properties. In total, 6 strains were identified as having the best properties for use as intranasal therapeutics to inhibit M. haemolytica. If successful in vivo, these strains offer an alternative to metaphylactic antimicrobial use in feedlot cattle for mitigating BRD.

INTRODUCTION

Bovine respiratory disease (BRD), also known as shipping fever, remains the costliest disease in the North American feedlot industry, despite advances in antimicrobials and vaccines against respiratory pathogens (1). Although BRD is a multifactorial disease with several viral and bacterial agents involved, Mannheimia haemolytica is considered a major pathogen in its etiology and is therefore a primary target for both BRD mitigation and treatment in cattle (2). As an opportunistic pathogen, M. haemolytica exists in the general cattle population and colonizes the nasopharynx in healthy cattle. However, when cattle experience compromised immunity due to stress and viral infection, M. haemolytica can proliferate in the nasopharynx and then translocate into the lung, where it can cause fibrinous pleuropneumonia (3).

Calves arriving at feedlots are often more susceptible to respiratory bacterial infections due to stress imposed by maternal separation and environmental and management factors (4). As a result, cattle determined to be at risk for BRD are frequently administered long-acting antimicrobials upon feedlot entry (i.e., metaphylaxis) (5). However, antibiotic resistance has been reported to be increasing in BRD-associated pathogens (6). In addition, recent feedlot studies conducted in both Canada (7, 8) and the United States (9) revealed a high prevalence of multidrug-resistant BRD bacterial pathogens displaying resistance to antimicrobials used for metaphylaxis treatment. Emergence of antimicrobial-resistant bacteria associated with BRD presents a significant risk to the beef industry, particularly if the efficacy of antimicrobials diminishes due to pathogens acquiring resistance. Novel alternatives to metaphylactic antimicrobials are therefore greatly needed.

Increasing evidence shows that bacterial communities residing within the respiratory tract are important to respiratory health and that disruption of the microbiota can reduce host resistance to colonization and proliferation of pathogenic bacteria (10, 11). The nasopharynx in cattle harbors a rich and diverse microbial community that is dynamic and has been shown to change in response to several management practices, including transportation to a feedlot (12), altering diet (13), and antimicrobial administration (14). Recent studies have also suggested an association between the nasopharyngeal microbiota and development of BRD in feedlot cattle (15, 16). This notion is further supported by studies associating a greater relative abundance of nasopharyngeal Lactobacillaceae at the time of feedlot entry with protection against BRD (17) and also specific inhibition of M. haemolytica in vitro (18). Hence, maintaining a stable microbial community in the nasopharynxes of cattle after feedlot placement may offer protection against BRD development, and bacteria colonizing the bovine respiratory tract may have potential for use as therapeutics to mitigate BRD pathogens. The objective of the present study was to develop bacterial therapeutics, with a focus on lactic acid-producing bacteria (LAB) originating from the respiratory tracts of healthy cattle for mitigation of M. haemolytica, using a stepwise approach based on pathogen inhibition, cell adherence, and immunomodulatory properties (Fig. 1).

FIG 1.

Schematic workflow diagram illustrating the process of isolation and identification and the screening criteria for commensal bacteria from the nasopharynxes of healthy feedlot cattle to identify candidate bacterial therapeutics to mitigate the bovine respiratory pathogen M. haemolytica.

RESULTS

Isolation and identification of nasopharyngeal commensal bacterial isolates.

A total of 300 isolates from De Man, Rogosa, and Sharpe (MRS) and Rogosa agar plates were isolated and identified using nearly full-length 16S rRNA gene sequences (Table 1). These isolates were from 14 different genera, with Bacillus (34%), Staphylococcus (30%), Streptococcus (12.3%), and Lactobacillus (12.0%) the more predominant genera. Although both MRS and Rogosa agar plates are semiselective for LAB, 69% of the total bacteria isolated were non-LAB species. A total of 93 isolates were taxonomically classified as LAB and comprised the genera Streptococcus (39.8% of the total LAB), Lactobacillus (38.7%), Enterococcus (10.8%), Aerococcus (9.7%), and Pediococcus (1.1%).

TABLE 1.

Bacteria identified from the nasopharynxes of healthy feedlot cattle and those selected for initial inhibition of M. haemolytica using agar slabs^a

Genus	No. of isolates	% of total isolates	No. of isolates screened for inhibition of M. haemolytica
LAB (n = 93)
Aerococcus	9	3.0	4
Enterococcus	10	3.3	5
Lactobacillus	36	12.0	33
Pediococcus	1	0.3	0
Streptococcus	37	12.3	32
Non-LAB (n = 207)
Acetobacter	5	1.7	2
Bacillus	102	34.0	53
Corynebacterium	3	1.0	1
Escherichia coli	3	1.0	2
Macrococcus	1	0.3	1
Micrococcus	1	0.3	1
Moraxella	1	0.3	0
Rummeliibacillus	1	0.3	1
Staphylococcus	90	30.0	43
Total	300		178

^aBacteria were isolated by plating swabs onto MRS or Rogosa medium and then identified using 16S rRNA gene sequencing and biochemical tests.

Growth-inhibitory effects against M. haemolytica.

Of the identified bacteria, 178 isolates representing 12 different genera were tested for the ability to inhibit M. haemolytica growth using an agar slab method (Fig. 2). A total of 74 isolates were LAB, within the genera Aerococcus (n = 4), Enterococcus (n = 5), Lactobacillus (n = 33), and Streptococcus (n = 32) (Fig. 2A). Of these, 88% inhibited the growth of M. haemolytica, with zones of inhibition (ZOI) ranging from 11 to 23 mm (Fig. 2B). Approximately 48% of the tested LAB displayed relatively strong inhibition of M. haemolytica (ZOI, 17 to 23 mm). Specifically, Lactobacillus displayed the greatest inhibition of M. haemolytica, with 88% of the tested Lactobacillus isolates showing ZOI ranging between 17 and 23 mm (Fig. 2A and B). Although 91% of the tested Streptococcus isolates inhibited M. haemolytica, 59% of them had relatively weak inhibition (ZOI, 11.1 to 14 mm), and 35% showed medium inhibition (ZOI, 14.1 to 16.9 mm). Four of 5 tested Enterococcus isolates moderately inhibited M. haemolytica. However, none of the Aerococcus isolates inhibited M. haemolytica (Fig. 2A).

FIG 2.

Growth-inhibitory effects of bovine respiratory bacteria against M. haemolytica. (A to D) Summaries of LAB (A) and non-LAB (C) and their respective zones of inhibition (B and D). (E) Lactobacillus species displayed the greatest inhibition. (F) Example of the agar slab method to measure inhibitory properties of screened bacteria. The results are presented as mean zones of inhibition (plus standard deviations [SD]) from three replicates.

The non-LAB isolates (n =104) tested for inhibition of M. haemolytica taxonomically belonged to 8 different genera and consisted mainly of Bacillus (51% of the total non-LAB isolates tested) and Staphylococcus (41%) isolates (Fig. 2C). Of these non-LAB isolates, 46% displayed growth-inhibitory effects against M. haemolytica (Fig. 2C). Among the inhibition-positive isolates, 48% had relatively weak inhibition (ZOI, 11.1 to 14 mm) and 31% showed medium inhibition (ZOI, 14.1 to 16.9 mm) (Fig. 2D). Of the Bacillus isolates tested, 55% did not inhibit growth of M. haemolytica. Only 6% of the tested isolates showed relatively strong inhibition (ZOI, >17 mm). The growth of M. haemolytica was inhibited by 49% of the tested Staphylococcus isolates, with 76% of the inhibition-positive isolates showing weak to medium inhibition (ZOI, 11.1 to 16.9 mm). Corynebacterium and Macrococcus isolates displayed relatively strong inhibition (ZOI, 17 to 19.9 mm). Moderate inhibition of M. haemolytica was also observed with one Escherichia coli isolate. However, isolates within the genera Acetobacter, Micrococcus, and Rummeliibacillus did not inhibit M. haemolytica.

Adherence of selected isolates to BT cells.

A total of 47 isolates were selected for adherence to bovine turbinate (BT) cells, based on their ability to inhibit M. haemolytica (ZOI, ≥15 mm). These isolates were from 6 different genera (Bacillus, Lactobacillus, Macrococcus, Enterococcus, Staphylococcus, and Streptococcus). All the tested isolates were able to colonize BT cell monolayers, with mean adherences ranging between 3.4 and 8.0 log₁₀ CFU per 10⁵ BT cells (Fig. 3). Of these, 32 isolates were Lactobacillus spp. and were from 7 different species: Lactobacillus amylovorus (n = 1), Lactobacillus brevis (n =2), Lactobacillus buchneri (n = 23), Lactobacillus paracasei (n = 3), Lactobacillus hilgardii (n = 1), Lactobacillus curvatus (n = 1), and Lactobacillus sunkii (n = 1) (Fig. 3A). Among these Lactobacillus isolates, 47% displayed mean adherences greater than 5.0 log₁₀ CFU per 10⁵ BT cells. Interestingly, the adherences to BT cells differed for strains within the same species, which was more obvious within the species L. buchneri (Fig. 3A). Of the 15 non-Lactobacillus isolates, the Enterococcus faecium strains, two Staphylococcus strains (28C and 98C), and one Streptococcus strain showed mean adherences greater than 5.0 log₁₀ CFU per 10⁵ BT cells (Fig. 3B).

FIG 3.

(A and B) Adherence of bovine respiratory bacterial isolates to BT cell monolayers. In total, 32 strains of Lactobacillus (A) and 15 strains of Bacillus, Enterococcus, Macrococcus, Staphylococcus, and Streptococcus (B) were evaluated. The results are presented as means and standard errors (SE) of bacterial adherence (log₁₀ CFU) to BT cell monolayers (10⁵ cells) obtained from three independent experiments performed on different days. (C) Confocal image of L. curvatus 103C (stained green with Alexa Fluor 594) adhering to bovine turbinate cells (stained blue with DAPI) shown as an example of bacterial adherence.

Antagonistic competition activities of selected isolates against M. haemolytica.

Fifteen isolates were tested for competitive exclusion ability. The isolates were selected based on high inhibition of M. haemolytica (ZOI, >15 mm) and strong adherence to BT cells (≥5.0 log₁₀ CFU per 10⁵ BT cells). Reduction of M. haemolytica adherence to BT cells was observed with all the tested isolates (Fig. 4). The mean reduction of M. haemolytica adherence to BT cell monolayers ranged from 32% to 78%. Lactobacillus amylovorus (72B) displayed the strongest inhibition of M. haemolytica adherence to BT cells, with adherence greater than that of 9 other tested strains (P < 0.05). In contrast, L. buchneri (65A) showed the weakest antagonistic competition against M. haemolytica. The remaining strains had essentially comparable levels of M. haemolytica adherence inhibition (P > 0.05).

FIG 4.

(A) Antagonistic competition of bovine respiratory bacteria (n = 15) against M. haemolytica. The results are presented as mean (+SE) reduction in M. haemolytica adherence to BT cell monolayers by commensal bacteria obtained from 6 replicates. Different letters indicate that the mean values differ (P < 0.05). (B) Representative confocal image showing adherence of M. haemolytica (stained red with Alexa Fluor 488) and L. paracasei 3E (stained green with Alexa Fluor 594) to bovine turbinate cells (stained blue with DAPI).

Antimicrobial susceptibilities of selected isolates.

The MICs of 20 antibiotics against 15 isolates were determined. According to the interpretive criteria provided by CLSI (M45) (19), all tested bacteria were susceptible to clindamycin, erythromycin, linezolid, meropenem, and penicillin (Table 2). The majority of Lactobacillus isolates were susceptible to daptomycin. Only L. amylovorus 72B grew at the maximum antibiotic plate concentration tested (2 μg/ml). Given that the resistance breakpoint for daptomycin against Lactobacillus is defined as greater than 4 μg/ml, it was not possible to define whether L. amylovorus 72B was daptomycin resistant. The L. amylovorus strains were susceptible to vancomycin, while the MIC values for all the other Lactobacillus strains were greater than the maximum concentration tested (>4 μg/ml) and could not be defined. Only two of the tested Lactobacillus strains (L. buchneri 38C and L. amylovorus 72B) were not inhibited by levofloxacin at the maximum concentration tested (4 μg/ml). All the other tested strains of L. buchneri were not inhibited by tetracycline at the maximum concentration of 8 μg/ml, which is 16-fold lower than the resistant cutoff value for L. buchneri (128 μg/ml), thus limiting evaluation of their tetracycline resistance. The L. paracasei strains were susceptible to chloramphenicol.

TABLE 2.

MICs of antibiotics against 15 bacterial strains isolated from the nasopharynxes of feedlot cattle^a

Antibiotic	Tested range (μg/ml)	MIC (μg/ml) for:
		L. plantarum NCDO1193	L. amylovorus (72B)	L. buchneri (65G)	L. buchneri (67A)	L. buchneri (63B)	L. buchneri (65E)	L. buchneri (63A)	L. buchneri (38C)	L. buchneri (65A)	L. buchneri (86D)	L. curvatus (103C)	L. paracasei (3E)	L. paracasei (57A)	E. faecium (64C)	S. epidermidis (6E)	S. chromogenes (28C)
Amoxicillin/clavulanic acid, 2:1 ratio	2/1–16/8	≤2/1	≤2/1	≤2/	≤2/1	≤2/1	≤2/1	≤2/1	≤2/1	≤2/1	≤2/1	≤2/1	≤2/1	≤2/1	≤2/1	≤2/1	≤2/1
Azithromycin	0.25–2	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	>2
Cefepime	0.5–8	≤0.5	4	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	>8	8	8	≤0.5	1
Cefotaxime	0.12–4	≤0.12	1	≤0.12	≤0.12	≤0.12	≤0.12	0.25	≤0.12	≤0.12	≤0.12	0.25	>4	>4	4	0.25	0.5
Ceftriaxone	0.12–2	≤0.12	>2	0.25	≤0.25	0.25	0.25	0.25	≤0.12	≤0.12	≤0.12	0.5	>2	>2	>2	1	2
Cefuroxime axetil	0.5–4	≤0.5	4	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	≤0.5	4	4	>4	≤0.5	≤0.5
Chloramphenicol	1–8	≤4	2	≤1	2	≤1	≤1	4	≤1	≤1	4	≤1	4	2	4	4	8
Clindamycin	0.12–1	≤0.12	≤0.12	≤0.12	≤0.12	≤0.12	≤0.12	≤0.12	≤0.12	≤0.12	≤0.12	≤0.12	≤0.12	≤0.12	>1	≤0.12	0.25
Daptomycin	0.06–2	0.12	>2	≤0.06	0.12	≤0.06	≤0.06	0.12	2	1	1	0.12	0.25	0.25	2	0.5	>2
Ertapenem	0.5–4	1	2	2	4	2	2	2	≤0.5	≤0.5	≤0.5	2	>4	4	4	≤0.5	≤0.5
Erythromycin	0.25–2	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	1
Levofloxacin	0.5–4	4	>4	4	4	4	4	4	>4	4	4	1	1	1	≤0.5	≤0.5	≤0.5
Linezolid	0.25–4	2	2	1	2	1	0.5	2	4	4	2	0.5	1	1	2	0.5	2
Meropenem	0.25–2	≤0.25	0.5	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25	0.5	2	1	>2	≤0.25	≤0.25
Moxifloxacin	1–8	≤1	>8	≤1	≤1	≤1	≤1	≤1	2	≤1	≤1	≤1	≤1	≤1	≤1	≤1	≤1
Penicillin	0.03–4	>4	0.12	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.06	0.5	0.5	0.25	0.5	2
Tetracycline	1–8	>8	>8	>8	>8	>8	8	>8	>8	>8	>8	>8	≤1	≤1	≤1	≤1	≤1
Tigecycline	0.015–0.12	0.12	0.03	0.12	>0.12	0.12	0.06	>0.12	>0.12	>0.12	>0.12	0.06	≤0.015	≤0.015	0.03	0.06	0.06
Trimethoprim/ sulfamethoxazole	0.5/9.5–4/76	≤0.5/9.5	>4/76	≤0.5/9.5	≤0.5/9.5	≤0.5/9.5	≤0.5/9.5	≤0.5/9.5	2/38	2/38	2/38	>4/76	≤0.5/9.5	≤0.5/9.5	≤0.5/9.5	≤0.5/9.5	>4/76
Vancomycin	0.5–4	>4	≤0.5	>4	>4	>4	>4	>4	>4	>4	>4	>4	>4	>4	1	2	4

^aBreakpoint interpretations are provided in Materials and Methods.

E. faecium 64C was susceptible to tetracycline, chloramphenicol, vancomycin, and erythromycin. According to the CLSI VET01S guidelines, Staphylococcus chromogenes 28C was resistant to penicillin while Staphylococcus epidermidis 6E was susceptible to penicillin. Both of these Staphylococcus strains were also resistant to amoxicillin-clavulanate and tetracycline but were susceptible to vancomycin, clindamycin, erythromycin, and chloramphenicol.

Stimulation of innate and adaptive immune responses in BT cell monolayers.

Among the 84 genes tested, 33 genes showed significant differences in transcriptional gene expression between the BT cells cocultured with at least one tested bacterium and the BT cells incubated under the same conditions without bacteria (P < 0.05) (Table 3). Of the Lactobacillus isolates, L. curvatus (strain 103C), L. amylovorus (72B), and L. buchneri (67A, 63A, 65G, and 63B) upregulated the CXCL8 gene, with 3.98-fold to 8.43-fold difference in expression from the control (P ≤ 0.02). The transcription of IL-6 was also upregulated by 5 of the 6 Lactobacillus isolates (changes ranged from 8.63- to 22.75-fold; P ≤ 0.04). The expression of NFKB1 was upregulated (P ≤ 0.03) in BT cells cocultured with L. curvatus (103C) and L. buchneri (63A, 63B, 65G, and 67A) isolates relative to the control (changes ranged from 1.5- to 2.13-fold). L. amylovorus (72B), L. buchneri (63B and 65G), and L. curvatus (103C) induced overexpression of interferon alpha/beta receptor alpha chain, with 1.6- to 2.0-fold changes in gene expression compared to the control (P ≤ 0.02). Of the 84 genes, very few were downregulated in BT cells in response to Lactobacillus inoculation. Transcription of chemokine receptor 6 was downregulated by L. paracasei (57A) and L. buchneri (63B and 38C) isolates (P ≤ 0.04), and L. paracasei (57A) downregulated the gene encoding solute carrier family 11 member 1 by 1.94-fold (P = 0.03).

TABLE 3.

Selected genes that had altered expression in BT cells after incubation with bacteria isolated from the nasopharynxes of cattle^a

Gene	Description	M. haemolytica		L. buchneri (63A)		L. amylovorus (72B)		L. paracasei (57A)		L. buchneri (63B)		L. buchneri (65G)		L. curvatus (103C)		L. paracasei (3E)		L. buchneri (38C)		L. buchneri (86D)		L. buchneri (67A)
		Fold changeb	P value	Fold change	P value	Fold change	P value	Fold change	P value	Fold change	P value	Fold change	P value	Fold changeb	P value	Fold change	P value	Fold change	P value	Fold change	P value	Fold change	P value
BOLA	Major histocompatibility complex class I, heavy chain	3.68	0.00	1.06	0.73	1.09	0.60	1.04	0.79	1.17	0.36	1.09	0.60	1.06	0.70	1.01	0.87	1.01	0.90	−1.05	0.70	1.08	0.66
CCL2	Chemokine (C-C motif) ligand 2	6.36	0.00	−1.47	0.10	1.08	0.73	1.30	0.39	−1.43	0.11	−1.39	0.15	1.17	0.51	1.05	0.82	−1.10	0.68	1.04	0.86	−1.47	0.15
CCL5	Chemokine (C-C motif) ligand 5	782	0.01	1.35	0.59	1.40	0.61	−1.01	0.76	1.29	0.67	1.08	0.94	1.17	0.82	1.00	0.96	−1.13	0.97	1.04	0.99	1.15	0.88
CCR6	Chemokine (C-C motif) receptor 6	−1.76	0.13	−1.82	0.07	−1.09	0.99	−2.10	0.02	−1.69	0.04	−1.46	0.28	−1.56	0.09	−1.54	0.15	−1.79	0.02	−1.25	0.31	−1.06	0.76
CCR8	Chemokine (C-C motif) receptor 8	−1.07	0.77	2.54	0.05	−1.08	0.88	−1.00	0.74	1.87	0.13	1.94	0.05	1.37	0.33	−1.46	0.44	1.19	0.62	1.29	0.47	2.76	0.06
CD40	CD40 molecule, TNFc receptor superfamily member 5	26.88	0.00	1.21	0.56	1.23	0.54	1.22	0.56	1.40	0.27	1.17	0.70	1.28	0.46	1.20	0.62	−1.05	0.67	1.08	0.95	1.18	0.71
CD80	CD80 molecule	3.86	0.01	1.05	0.78	1.67	0.20	−1.02	0.75	−1.32	0.22	−1.05	0.97	1.23	0.39	1.20	0.44	−1.08	0.79	−1.01	0.75	1.06	0.88
CSF2	Colony-stimulating factor 2 (granulocyte-macrophage)	5.90	0.00	−1.43	0.35	1.11	0.94	−1.29	0.40	−1.68	0.27	−1.62	0.31	−1.55	0.31	−1.41	0.52	−1.65	0.29	−1.16	0.69	1.43	0.70
CXCL10	Chemokine (C-X-C motif) ligand 10	1321	0.00	−1.67	0.35	−1.08	0.78	−1.92	0.14	−1.53	0.30	−1.48	0.34	−1.41	0.34	−1.28	0.56	−1.50	0.33	−1.15	0.71	1.58	0.68
CXCL8	Interleukin 8	35.40	0.00	5.32	0.00	8.11	0.00	2.15	0.36	3.98	0.02	4.62	0.00	8.43	0.02	−1.10	0.61	−1.31	0.12	1.01	0.98	6.99	0.01
DDX58	DEAD (Asp-Glu-Ala-Asp) box polypeptide 58	54.62	0.00	−1.10	0.82	−1.04	0.91	−1.26	0.42	−1.02	0.99	−1.28	0.20	−1.16	0.39	−1.01	0.94	1.08	0.61	−1.07	0.85	−1.08	0.99
FAS	Fas (TNF receptor superfamily, member 6)	2.02	0.01	−1.11	0.55	−1.24	0.15	−1.14	0.28	1.02	0.95	−1.05	0.66	−1.03	0.73	−1.04	0.80	−1.00	0.95	−1.18	0.25	−1.02	0.88
ICAM1	Intercellular adhesion molecule 1	10.05	0.02	1.00	0.87	1.01	0.95	−1.04	0.73	−1.37	0.17	−1.05	0.73	−1.10	0.65	1.06	0.83	1.15	0.53	1.09	0.71	−1.20	0.41
IFNAR1	Interferon (alpha, beta, and omega) receptor 1	1.43	0.02	1.45	0.05	1.65	0.00	−1.08	0.74	1.72	0.02	1.60	0.01	2.00	0.01	1.06	0.70	1.00	0.96	1.05	0.76	1.50	0.07
IFNGR1	Interferon gamma receptor 1	−1.79	0.03	−1.11	0.78	1.01	0.78	−1.36	0.47	−1.14	0.58	−1.02	0.98	−1.16	0.50	1.03	0.73	−1.00	0.78	1.10	0.56	1.06	0.66
IL15	Interleukin 15	19.84	0.01	−1.53	0.24	−1.60	0.47	−1.85	0.09	−2.27	0.10	−2.18	0.13	−1.78	0.13	−1.37	0.32	−1.61	0.12	−1.27	0.45	1.07	1.00
IL18	Interleukin 18 (interferon gamma-inducing factor)	3.45	0.04	−1.23	0.69	1.14	0.75	−1.07	0.72	−1.12	0.66	−1.12	0.75	−1.25	0.54	1.09	0.79	−1.00	0.98	1.20	0.61	1.31	0.56
IL1R1	Interleukin 1 receptor, type I	1.11	0.42	1.13	0.33	1.11	0.35	−1.16	0.28	1.34	0.01	1.29	0.03	1.31	0.01	−1.04	0.68	1.04	0.76	1.08	0.53	1.19	0.20
IL4	Interleukin 4	−1.83	0.33	−1.56	0.33	−1.19	0.75	−2.12	0.12	−1.69	0.27	−1.63	0.31	−1.55	0.31	−1.41	0.52	−1.39	0.31	−1.23	0.75	1.43	0.70
IL6	Interleukin 6 (interferon beta 2)	289.5	0.01	11.68	0.01	8.04	0.02	3.73	0.11	8.63	0.04	9.13	0.04	22.75	0.04	1.36	0.34	1.30	0.35	1.75	0.27	11.44	0.12
IRF3	Interferon regulatory factor 3	4.48	0.00	−1.43	0.32	−1.10	0.96	−1.37	0.21	1.12	0.63	−1.14	0.64	−1.18	0.32	−1.30	0.42	−1.19	0.47	−1.05	0.92	−1.37	0.52
IRF7	Interferon regulatory factor 7	67.01	0.00	−1.40	0.55	−1.08	0.85	−1.53	0.28	−1.08	0.53	−1.43	0.38	−1.12	0.63	−1.14	0.75	−1.37	0.39	−1.11	0.75	1.41	0.70
JAK2	Janus kinase 2	5.74	0.00	−1.10	0.71	1.06	0.71	1.04	0.82	−1.00	1.00	−1.04	0.88	1.03	0.79	1.23	0.44	1.18	0.48	1.26	0.35	1.18	0.41
LOC512672	Major histocompatibility complex, class I	13.20	0.00	−1.28	0.50	1.09	0.81	−1.03	0.88	1.33	0.29	1.24	0.36	1.34	0.20	1.01	0.94	1.13	0.65	1.26	0.31	1.01	0.85
LOC616942	Major histocompatibility complex, class I, A-like	7.21	0.00	2.18	0.23	1.18	0.51	−1.19	0.66	−1.14	0.71	1.21	0.46	1.47	0.15	−1.23	0.31	−1.11	0.45	−1.00	1.00	1.21	0.50
MX1	Myxovirus (influenza virus) resistance 1, interferon-inducible protein p78 (mouse)	329.9	0.00	1.55	0.49	1.24	0.91	−1.02	0.62	1.08	0.87	1.28	0.81	1.42	0.66	1.07	0.81	1.47	0.52	1.03	0.82	1.13	0.84
NFKB1	Nuclear factor of kappa light polypeptide gene enhancer in B cells 1	8.80	0.00	1.67	0.00	1.27	0.15	1.27	0.30	1.81	0.00	1.75	0.00	2.13	0.00	1.12	0.30	1.15	0.34	1.14	0.25	1.50	0.03
NFKBIA	Nuclear factor of kappa light polypeptide gene enhancer in B cells inhibitor, alpha	19.05	0.00	−1.14	0.58	1.06	0.69	1.39	0.39	−1.14	0.28	−1.25	0.10	1.42	0.16	−1.07	0.54	−1.17	0.34	1.02	0.86	−1.32	0.11
SLC11A1	Solute carrier family 11 (proton-coupled divalent metal ion transporters), member 1	2.18	0.05	−1.64	0.27	−1.64	0.47	−1.94	0.03	−1.69	0.08	−1.41	0.21	−1.81	0.10	−1.32	0.32	−1.30	0.25	−1.24	0.46	1.04	0.95
STAT1	Signal transducer and activator of transcription 1, 91 kDa	8.17	0.00	−1.16	0.37	−1.05	0.71	−1.12	0.40	−1.07	0.57	−1.04	0.74	−1.09	0.57	−1.04	0.71	−1.06	0.74	−1.14	0.50	−1.12	0.45
STAT3	Signal transducer and activator of transcription 3 (acute-phase response factor)	2.34	0.00	1.19	0.35	1.32	0.20	−1.03	0.76	1.28	0.19	1.23	0.29	1.40	0.12	−1.02	0.81	−1.06	0.70	1.01	0.99	1.24	0.31
TLR2	Toll-like receptor 2	2.39	0.03	1.00	0.82	1.76	0.04	−1.03	0.78	−1.13	0.84	−1.20	0.62	1.54	0.20	1.20	0.47	−1.00	0.73	1.10	0.63	1.18	0.57
TLR3	Toll-like receptor 3	4.04	0.00	1.06	0.66	1.08	0.65	−1.10	0.83	−1.12	0.80	−1.17	0.85	1.07	0.65	1.09	0.62	1.01	0.73	1.22	0.36	−1.11	0.54
TLR4	Toll-like receptor 4	9.28	0.00	1.20	0.75	1.28	0.62	−1.12	0.95	1.24	0.86	1.61	0.35	1.32	0.69	1.28	0.63	2.07	0.14	1.06	0.81	1.35	0.56

^aAll values are presented as the mean fold change in gene expression for 4 replications. The difference in gene expression between BT cells (control) that were not cocultured with bacterial cells and the BT cells cocultured with bacterial cells was assessed by Student’s t test for each gene using the Rt² Profiler PCR array analysis software, version 3.5 (Qiagen). The level of statistical significance was set at a P value of <0.05. Genes that showed significant differences in expression between the control and at least one bacterium cocultured with BT cells are listed. Significant changes (P < 0.05) in expression are shown in boldface.

^bFold change (2^−ΔΔCT) is the normalized gene expression (2^−ΔΔCT) in the test sample divided by the normalized gene expression (2^−ΔΔCT) in the control sample. Fold change values greater than 1 indicate positive or upregulation, and fold change values less than 1 indicate negative or downregulation.

^cTNF, tumor necrosis factor.

In contrast to Lactobacillus spp., significantly high immune stimulation in BT cells was observed with M. haemolytica, which upregulated 28 genes with changes ranging from 1.43- to 1,321-fold (P < 0.05). The greatest responses to M. haemolytica were in the transcription of C-X-C motif chemokine 10 (1,321-fold change), CXCL5 (782-fold change), myxovirus resistance 1 (330-fold change), and IL-6 (290-fold change). The IRF7, DDX58, CXCL8, CD40, IL-15, NFKB1A, ICAM1, and LOC512672 genes exhibited relatively high overexpression in BT cells incubated with M. haemolytica (10- to 67-fold changes) (P < 0.05). Moderate upregulation by M. haemolytica was observed for the genes encoding chemokine ligand 2, colony-stimulating factor 2, Janus kinase 2, major histocompatibility complex class 1 (A like), NFKB1, STAT1, and Toll-like receptor 4 (TLR4) (5- to 10-fold changes).

M. haemolytica-inhibitory mechanisms of 6 Lactobacillus isolates as candidate bacterial therapeutics.

A total of 6 Lactobacillus strains (listed in Table 4) from four different species were selected as the best candidates for the development of intranasal bacterial therapeutics to mitigate M. haemolytica infection based on the selection criteria of inhibition (strongest inhibition), adhesion (strongest adhesion), competition exclusion (strongest exclusion), antimicrobial susceptibility (limited resistance), and immunomodulation (moderate stimulation of immune genes). To understand the potential mechanisms through which these selected Lactobacillus strains confer direct inhibition of M. haemolytica, we determined their lactic acid, H₂O₂, and putative bacteriocin production capacities and their effects on the cell morphology of M. haemolytica.

TABLE 4.

Antimicrobial properties of selected bacterial therapeutic strains (n = 6) evaluated by the measurement of their lactic acid and H₂O₂ production and bacteriocin-encoding genes

Strain	Lactic acid concn (mM)a	H2O2 concn (nmol/ml)a	Bacteriocin-encoding genes detected from the whole-genome sequences of the selected strains by BAGEL 4
			No. of genes	Class of bacteriocin
L. amylovorus (72B)	102.6 ± 0.61	29.0 ± 4.63	5	64.3 (enterolysin A); 70.3 (helveticin J); 6.3 (bacteriocin helveticin J); 64.3 (enterolysin A); 70.3 (helveticin J)
L. buchneri (63A)	79.9 ± 0.87	4.7 ± 4.21	0	NDd
L. buchneri (86D)	93.6 ± 7.34	5.2 ± 1.85	0	ND
L. curvatus (103C)	110.7 ± 4.74	9.4 ± 3.16	0	ND
L. paracasei (3E)	133.6 ± 1.35	0	4	142.2 (LSEI_2163); 62.3 (enterolysin A); 51.2 (carnocin CP52); 97.2 (enterocin X chain beta)
L. paracasei (57A)	131.6 ± 3.61	0	NSc	ND
Cocktail of 6 strains	142.1 ± 8.72	NAb	NA	ND
Control (MRS broth)	0	0	NA	ND

^aThe results are reported as the mean concentration (±SE) of lactic acid and H₂O₂ produced by the selected strains over the 24-h incubation period. The mean was obtained from triplicate samples.

^bNA, not applicable.

^cNS, not sequenced.

^dND, bacteriocin not detected in genomes.

Lactic acid production and growth inhibition by lactic acid against M. haemolytica.

We determined the lactic acid production of 6 Lactobacillus strains individually, as well as when combined (Table 4). The concentrations of lactic acid detected from the supernatants ranged between 80 and 142 mM. The supernatants of the 6 strains combined and L. paracasei strains 3E and 57A contained the most lactic acid (132 to 142 mM), followed by L. curvatus 103C (111 mM) and L. amylovorus 72B (103 mM). The supernatants of L. buchneri strains contained the least lactic acid, with 80 and 94 mM, respectively. No lactic acid was detected from the negative control (MRS broth).

The inhibitory effects of lactic acid covering the range of concentrations produced by the 6 strains were tested against growth of M. haemolytica (Fig. 5). At 8 h postinoculation, M. haemolytica growth was inhibited when the lactic acid concentrations ranged from 18.75 to 150 mM compared to 0 and 9.38 mM concentrations. Although there was cell growth in media supplemented with 18.75 mM lactic acid within the first 8 h of incubation, this concentration of lactic acid resulted in lower cell counts than at 0 and 9.38 mM concentrations. After 24 h, no viable cells were detected in media containing 100 and 150 mM lactic acid.

FIG 5.

Growth inhibition effects of lactic acid on M. haemolytica. The values are the means of two replicates. The error bars indicate SD.

H₂O₂ production.

The concentrations of H₂O₂ in cell-free culture supernatants obtained from Lactobacillus strains after 24 h of incubation varied among different Lactobacillus species, with L. amylovorus the most predominant H₂O₂ producer (Table 4). L. buchneri (63A and 86D) and L. curvatus (103C) strains produced similar amounts of H₂O₂; however, no H₂O₂ was detected from the supernatants of L. paracasei strains (3E and 57A).

Encoded bacteriocins.

Sequences obtained for L. paracasei (57A) were too poor to evaluate and thus were eliminated from bacteriocin evaluation. The genomes of 5 selected strains revealed that L. amylovorus 72B and L. paracasei 3E contained bacteriocin-encoding genes while L. buchneri (63A and 86D) and L. curvatus (103C) strains did not have genes encoding bacteriocins (Table 4). Lactobacillus amylovorus (72B) had 5 genes encoding enterolysin A and helveticin J, both of which belong to the bacteriolysin class. The genome of L. paracasei 3E contained four different bacteriocin genes encoding LSEI_2163, enterolysin A, carnocin-CP52 immunity protein, and enterocin Xβ.

Effect of selected strains on cell morphology of M. haemolytica.

The morphological effects of supernatants from candidate strains on M. haemolytica were examined using scanning electron microscopy (SEM). Noticeable changes in the cell structure, including shrinkage of the cell surface, irregular rod shape, and holes in the cell envelope, were observed when M. haemolytica was incubated in culture supernatants from Lactobacillus strains compared to untreated cells (Fig. 6). Supernatant from L. amylovorus (72B) reduced the cell density of M. haemolytica to the greatest extent (data not shown) and caused the most apparent destructive changes in M. haemolytica cell structure (Fig. 6B). L. buchneri (63A) demonstrated minor cell damage compared to other strains tested (Fig. 6C), followed by L. paracasei (3E) (Fig. 6D). The bacteria L. paracasei (57A), L. curvatus (103C), and L. buchneri (86D) exhibited similar degrees of cell damage (Fig. 6E to G).

FIG 6.

Scanning electron microscopy images of M. haemolytica after incubation with cell-free culture supernatants of selected bacterial therapeutic strains. Bacteria were incubated with cell-free culture supernatants before fixation and microscopy. Shown are untreated cells (A) and cells incubated with cell-free culture supernatants of L. amylovorus 72B (B), L. buchneri 63A (C), L. paracasei 3E (D), L. paracasei 57A (E), L. curvatus 103C (F), and L. buchneri 86D (G).

DISCUSSION

Bacteria originating from the host target site are more likely adapted for successful recolonization (20). To develop bacterial therapeutics to mitigate the BRD pathogen M. haemolytica, we therefore performed our screening on commensal bacteria isolated from the upper respiratory tracts of healthy feedlot cattle. This is also the site where BRD bacterial pathogens colonize and proliferate before translocating into the lung to induce lung infection. In previous studies comparing healthy and BRD-affected cattle phenotypes, an increased abundance of LAB was observed in healthy animals (15, 21). Thus, bacterial species taxonomically belonging to the LAB (order Lactobacillales) were originally targeted using selective media. A wide range of commensal bacterial species were isolated and identified from two different sources of healthy feedlot cattle. Of these isolates, 31% were LAB strains. A subset of isolates comprised mainly of LAB strains and non-LAB species from genera such as Bacillus and Staphylococcus, which have a known history of probiotic use (22, 23), were selected for evaluation.

In summary, a number of diverse bacteria were able to inhibit the growth of M. haemolytica to varying extents. This suggests that the environment within the bovine respiratory tract is highly competitive, with multiple bacteria capable of producing factors that inhibit the opportunistic pathogen M. haemolytica. In support of this, Corbeil et al. (24) also described several genera from the nasal cavities of cattle that could inhibit bovine respiratory pathogens, including M. haemolytica, Pasteurella multocida, and Histophilus somni. In the present study, the bacteria displaying the strongest inhibition of M. haemolytica were within the genus Lactobacillus. Previously, Lactobacillus was identified as being more abundant in the lungs of healthy feedlot calves than in those diagnosed with BRD (21). Lactobacillus spp. therefore appear to be involved in maintaining the respiratory health of cattle and may achieve this by inhibiting respiratory pathogens through the production of antimicrobial factors. Direct inhibition of pathogens is an important attribute of bacterial therapeutics and probiotics and occurs through production of lactic acid, bacteriocins, and H₂O₂ or through the mechanical property of autoaggregation (25, 26). While few studies have investigated bacterial therapeutics to mitigate BRD bacterial pathogens, the inhibitory effects of commensal bacteria against the growth of a wide range of bacterial pathogens involved in human intestinal and respiratory tract infections have been documented (27, 28). In order to develop bacterial therapeutics with the greatest potential to mitigate M. haemolytica, only a subset of screened commensals with the strongest inhibition were further evaluated, the majority of which were Lactobacillus spp.

We used BT cells for adhesion assays because the cell type is found within the upper respiratory tract, which M. haemolytica colonizes (29) and where inhibition of its growth would be desired in order to limit proliferation and subsequent translocation and infection of the lungs. All the strains tested were capable of adhering to BT cells to varying extents, which was expected given that they were originally isolated from the nasopharynxes of cattle. Surprisingly, E. faecium isolates had the strongest adherence. Enterococcus colonizes the lower gastrointestinal tract, though the bacteria are capable of adhering to extraintestinal epithelial cells (30). While it is likely a transient inhabitant of the respiratory tract, this probably explains why Enterococcus has been consistently observed in the upper respiratory tract sof cattle (12, 17, 31). Several Lactobacillus strains also displayed a high level of adherence. Although no studies have investigated the adhesion of commensal respiratory tract bacteria to bovine upper respiratory tract cells, we previously showed that commercial Lactobacillus spp. displayed greater adhesion to bovine bronchial epithelial cells than Streptococcus and Paenibacillus strains (18). Lactobacillus spp. have the ability to colonize both mucus (32) and underlying epithelial cells of the respiratory tract (33). The high level of adherence to BT cells observed with some of the Lactobacillus strains might be attributed to both specific (surface-dependent proteins and surface layer proteins) and nonspecific (cell surface hydrophobicity and lipoteichoic acid) adhesion mechanisms (34). Different adherence capacities among Lactobacillus species (L. brevis versus L. amylovorus), and even strains of the same species (L. buchneri), were observed in the present study. It is likely that variation in strain adherence resulted from differences in cell surface structure of the isolates, as has been described previously (35). Compared to Lactobacillus spp., the strains within the genera Bacillus, Macrococcus, and Streptococcus showed weaker adhesion to BT cells and were therefore excluded from further screening.

A total of 15 isolates with the strongest inhibition of M. haemolytica and adhesion to BT cells were further evaluated for the ability to competitively inhibit M. haemolytica colonization of BT cells. All the tested strains were able to inhibit the adherence of M. haemolytica to BT cell monolayers. This likely occurred through a combination of direct inhibition and also competition for binding sites on the epithelial cells (36, 37). Of note, when evaluating inhibition, adhesion, and competition results, no single strain was ranked as the top candidate across all of the criteria. For example, L. amylovorus 72B showed the strongest inhibition and competition against M. haemolytica, but the adherence of the strain to BT cells was less than that of L. buchneri and L. paracasei strains. In contrast, L. buchneri (63A) and E. faecium (64C) showed moderate growth inhibition of M. haemolytica but strong adherence to BT cells and antagonistic competition against M. haemolytica in comparison to other strains. Thus, utilizing several criteria is important in defining potentially effective bacterial therapeutics. In addition, designing therapeutic strains with varying antimicrobial properties may promote broader efficacy of bacterial therapeutics against pathogens. In support of this, multistrain cocktails containing probiotics with different mechanisms of action have shown better antipathogenic activity and modulation of the mouse gut microbiome and short chain fatty acid production (38), as well as greater attenuation of pathogen-induced inflammatory response in the intestinal epithelium (39), than single-strain probiotics.

Safety concerns exist over probiotic or therapeutic bacteria that are resistant to antibiotics, especially if resistance elements are encoded on mobile elements (40, 41). The isolates used in our study were from cattle that were not administered antibiotics before or during the time of isolation, and thus, pressure for resistance to develop or for resistant bacteria to be selected was low. Indeed, when evaluated, the selected Lactobacillus strains were generally susceptible to antibiotics with defined break points, though some antibiotics could not be fully tested due to their breakpoint concentrations not being reached in the antibiotic panel. Although the majority of L. buchneri strains were not inhibited by the highest concentration of tetracycline tested (8 μg/ml), tetracycline resistance (128 μg/ml) in L. buchneri has been reported to be intrinsic (42). An L. buchneri strain (NRRL B-50733) that showed a tetracycline MIC value of 32 μg/ml has been considered safe for use as a silage additive for livestock (43). Thus, higher MICs for tetracycline in L. buchneri strains would not likely limit the use of these bacteria as therapeutics, although further evaluation of the L. buchneri isolates in our study with an antibiotic panel containing higher concentrations of tetracycline would be beneficial. L. amylovorus strain 72B had higher levofloxacin and moxifloxacin MICs. Currently, there are no breakpoints available for these two fluoroquinolone antibiotics against L. amylovorus strains. However, some species within the genus Lactobacillus are known to be intrinsically resistant to fluoroquinolones, including levofloxacin ( 44, 45).

Staphylococcus strains were resistant to either penicillin (28C) or amoxicillin-clavulanate and tetracycline (28C and 6E). Therefore, these strains were excluded from further evaluation. Despite being susceptible to the antibiotics tested and showing strong inhibition of M. haemolytica, the E. faecium strain 64C was also not considered for the immunostimulation assay due to E. faecium being a potential opportunistic pathogen (46) and defined as a level 2 pathogen in some countries. Considering the data from the inhibition, competition, and antimicrobial susceptibility assays, in addition to a history of safe use in both food and the feed industry (47), the strains of Lactobacillus, therefore, were selected for immunomodulation properties.

Modulation of host innate and adaptive immunity by therapeutic bacteria can potentially increase resistance to pathogen infection (48). Bovine respiratory epithelial cells are an initial point of contact between the host and respiratory microbiota (49) and were therefore used to evaluate immunomodulation. In the present study, the tested Lactobacillus strains induced moderate gene expression of the chemokine CXCL8 in BT cells, showing the ability of these nasopharyngeal strains to stimulate an immune response. In contrast, M. haemolytica induced strong overexpression of CXCL8 in BT cells. The protein CXCL8 is a potent chemoattractant and plays an important role in inflammation and wound healing through activation of neutrophils and other immune cells (50). Whether CXCL8 has a beneficial role in protecting the host through the inflammation healing process or is detrimental by promoting pathogenesis likely depends on its level of expression (50). In support of this, Gärtner et al. (51) reported that commensal Lactobacillus spp. present in the bovine uterus upregulated CXCL8 gene expression by 2- to 6-fold in endothelial epithelial cells after 6 to 8 h of coculture. The authors suggested that excessive expression of CXCL8 might contribute to the development of uterine disease, while moderate stimulation of CXCL8 may be necessary for bacterial clearance.

Similar to CXCL8, expression of NFKB1 and the proinflammatory cytokine IL-6 in BT cells was stimulated by most of the Lactobacillus strains tested. However, upregulation of these genes in BT cells by Lactobacillus spp. was considered moderate compared to M. haemolytica. Lipopolysaccharide from M. haemolytica has been shown to induce secretion of IL-6 in pulmonary epithelial cells (52), and activation of the NF-κB pathway causes excessive inflammation in the lower respiratory tract, thus enhancing infection by the pathogen (53). While we did not coinoculate Lactobacillus strains with M. haemolytica for the immunomodulation assays using BT cells, expression of IL-6 in bovine endothelial epithelial cells was moderately upregulated in response to commensal Lactobacillus spp. present in the bovine uterus (46). Future studies evaluating coinoculation would provide insight into whether the virulence of M. haemolytica could be reduced by bacterial therapeutics.

Overall, most of the 84 genes encoding the TLR pathway, cytokine and chemokine receptors, inflammation response, NF-κB signaling, apoptosis, and innate immune and defense responses to bacteria were not influenced by the Lactobacillus strains tested. However, M. haemolytica induced overexpression of 28 genes, including cytokine and chemokine receptors, inflammatory responses, NF-κB signaling, and defense response to bacteria. None of the tested Lactobacillus strains caused an excessive immune response, as observed with M. haemolytica. The effects of Lactobacillus on immune modulation in BT cells was species and strain specific. Similarly, immunomodulation has been shown to vary at the strain level in commensal bacteria (54, 55). Combined, the results regarding immune stimulation in BT cells suggest that Lactobacillus may have a role in modulating immunity in cattle; however, future studies are warranted to elucidate the mechanisms by which modulation occurs and its impact on respiratory pathogens.

All 6 of the candidate therapeutic strains produced lactic acid to varying extents. This acid has been shown to alter the membrane structure of pathogens and is a mechanism by which probiotic Lactobacillus spp. inhibit pathogen growth (56). In agreement with our study, Neal-McKinney et al. (56) reported a similar range of lactic acid production by Lactobacillus spp. (Lactobacillus acidophilus, Lactobacillus crispatus, Lactobacillus gallinarum, and Lactobacillus helveticus), as well as concentrations varying according to the strains tested. For our study, given that each therapeutic candidate, or a combination of all 6, produced concentrations of lactic acid greater than the MIC (37.5 mM), lactic acid is likely a common metabolite by which these candidates directly inhibited M. haemolytica.

In contrast to lactic acid, production of H₂O₂ and encoded bacteriocins varied among the candidate strains. Hydrogen peroxide produced from Lactobacillus has previously been shown to have bactericidal activity against gut pathogens (25) and may be a mechanism of inhibition of M. haemolytica for four of the candidate strains in our study. Only two of the candidate therapeutics encoded bacteriocins, which can have bactericidal or bacteriostatic activity (57, 58). Similar to our study, it has been observed that bacteriocins encoded by Lactobacillus are species, strain, and origin dependent (59). The different classes of bacteriocins detected from L. paracasei 3E were similar to those observed with the probiotic strain L. paracasei SD1 (60). Expression of these bacteriocins may have led to L. paracasei 3E being one of the strongest inhibitors of M. haemolytica (ZOI, >22 mm). The strain L. amylovorus 72B, was predicted to produce bacteriolysins, including enterolysin A and helveticin J. Although L. amylovorus has been reported to produce the bacteriocin amylovorin L471 (59, 61), genes encoding this bacteriocin were not detected in strain 72B. It is interesting that while L. amylovorus (72B) did not have the greatest adhesion to BT cells, it did have one of the greatest inhibition and antagonistic values. The 72B strain also caused the greatest morphological damage to M. haemolytica. Thus, a combination of lactic acid, H₂O₂, and bacteriocin production may have resulted in the strong inhibition of M. haemolytica observed for 72B.

In summary, we isolated and identified commensal bacteria from 14 different genera residing in the nasopharynxes of healthy feedlot cattle as part of the normal flora. Of these commensal isolates, using a stepwise approach, we screened isolates comprising 12 different genera for their abilities to inhibit the growth of the respiratory pathogen M. haemolytica, to adhere to BT cells, and to compete against M. haemolytica adherence to the BT cells. We then evaluated the best candidates for antimicrobial resistance and their immunomodulation effects in BT cells. Based on the data generated, 6 Lactobacillus strains from four different species (L. amylovorus strain 72B, L. buchneri strains 63A and 86D, L. curvatus strain 103C, and L. paracasei strains 3E and 57A) were selected as the best candidates for the development of intranasal bacterial therapeutics to mitigate M. haemolytica infection in cattle. Their selection was based on high inhibition of M. haemolytica directly and through competition, high adherence to BT cells, lack of antibiotic resistance, and moderate immunomodulation in BT cells. The potential mechanisms by which these 6 selected strains inhibited M. haemolytica were also investigated. Lactic acid production was common among the strains, but production of H₂O₂ and encoded bacteriocins varied. Currently, in vivo studies are being conducted to evaluate the effects of intranasal administration of these strains on the microbiota of feedlot cattle.

MATERIALS AND METHODS

Isolation of commensal bacteria from the nasopharynxes of feedlot cattle.

A schematic of the general study design, including bacterial isolation, is shown in Fig. 1. Nasopharyngeal bacteria were isolated as part of previous studies (17, 31). Briefly, two groups of animals were used to increase the diversity of the bacterial isolates. In the first group (group 1), 70 crossbred recently weaned steers purchased from a local auction market were sampled by deep nasopharyngeal swab (DNS) on day 0 (feedlot entry) and 60 days after placement at the Lethbridge Research and Development Centre (LRDC) feedlot (Alberta, Canada). The steers were fed diets typical of feedlots in western Canada and remained healthy from days 0 to 60 (17). In the second group (group 2), 30 Angus beef steers were sourced from a cow-calf ranch. These steers were sampled by DNS at weaning while still on the cow-calf ranch and then upon arrival at the LDRC feedlot and 40 days after arrival (31). The DNS samples were processed for bacterial isolation using semiselective medium (MRS or Rogosa plates) for LAB, as described by Holman et al. (17). Isolates were subcultured and stored in cryopreservative.

Identification of nasopharyngeal commensal bacterial isolates.

A total of 300 banked isolates were randomly selected for inclusion in the present study. For identification, the nearly full-length 16S rRNA gene (>1,400 bp) was sequenced for each isolate and used for taxonomic identification, as described previously (17). In instances where taxonomic identification at the species level was not possible by 16S rRNA gene sequence analysis, biochemical tests were also employed. For this, the isolates were subcultured on MRS agar (Lactobacillus and Enterococcus) or tryptic soy agar (TSA) (Staphylococcus) at 39°C. Colony morphologies were observed after 24 to 48 h at both 27°C and 39°C. Anaerobic growth was tested on MRS agar or TSA at 39°C in an anaerobic chamber with an atmosphere of 85% nitrogen, 10% hydrogen, and 5% CO₂. Acid production from carbohydrates was determined with the API 50CHL gallery (bioMérieux, Saint-Laurent, QC, Canada) (Lactobacillus and Enterococcus) according to the manufacturer’s instructions or using Difco purple agar base medium (BD Canada, Mississauga, ON, Canada) containing 1% carbohydrate (Staphylococcus). Confirmatory identifications were obtained through comparison with published results.

Growth-inhibitory effects of nasopharyngeal commensal bacteria against M. haemolytica.

For all assays, M. haemolytica L024A was used. The isolate originated from a feedlot steer that succumbed to BRD in Alberta, Canada, and was confirmed as serotype 1 (62). Except for Moraxella and Pediococcus, isolates within all the genera identified were included for inhibition of M. haemolytica. A total of 178 commensal isolates were studied for inhibition of M. haemolytica. The isolates were selected to include a diverse group of LAB and non-LAB strains from the respiratory tracts of cattle. Lactic acid-producing bacteria are defined as the order Lactobacillales, which encompasses the following six families: Aerococcaceae, Carnobacteriaceae, Enterococcaceae, Lactobacillaceae, Leuconostocaceae, and Streptococcaceae (63). Growth-inhibitory effects against M. haemolytica were determined using an agar slab method according to the instructions of Dec et al. (64) with some modifications. Briefly, 100 μl of an 18-h culture from each isolate grown in Difco Lactobacilli MRS broth (BD, Mississauga, ON, Canada) was spread as a lawn onto MRS plates and incubated at 37°C with 5% CO₂ for 24 h. Agar slabs (10 mm in diameter) were cut from the 24-h-incubated MRS plates using a sterile hollow punch (Tekton: 12-piece hollow punch set; catalog no. 6588) and placed with the culture side down onto a lawn of M. haemolytica on TSA plates containing 5% sheep blood. The lawn of M. haemolytica was prepared by spread plating a 100-μl aliquot of M. haemolytica culture suspended in Dulbecco’s phosphate-buffered saline (DPBS) (pH 7.4) to obtain the target bacterial concentration of 1 × 10⁸ CFU per ml. Up to four agar plugs were placed onto a single lawn of M. haemolytica, including a control plug containing no bacteria. The agar plug-M. haemolytica lawns were then incubated at 37°C with 5% CO₂ for 24 h. After incubation, the plates were checked for ZOI. The ZOI were measured with a ruler. The results were presented as the mean diameter of the inhibition zone for three independent experiments.

Adherence of commensal bacteria to BT cell monolayers.

A subset of isolates (n = 47) displaying the greatest inhibition of M. haemolytica (ZOI ≥ 15 mm) were evaluated for adhesion to BT cells (ATCC 1390; American Type Culture Collection, Manassas, VA, USA) using an assay described previously (18, 19) with some modifications. The isolates comprised the genera Lactobacillus (n = 32), Bacillus (n = 2), Enterococcus (n = 3), Macrococcus (n = 1), Staphylococcus (n = 6), and Streptococcus (n = 3). The BT cells were seeded onto 6-well flat-bottom tissue culture plates at 1 × 10⁵ cells per well and incubated in Dulbecco’s modified Eagle’s medium (DMEM) (Thermo Fisher Scientific, Oakville, ON, Canada) supplemented with 10% horse serum (American Type Culture Collection) and 50 μg/ml gentamicin (Sigma-Aldrich, Oakville, ON, Canada) at 37°C with 5% CO₂ until a complete monolayer was obtained. The BT cell monolayer was washed twice with antibiotic-free DMEM. Then, 2 ml of antibiotic-free DMEM was added to each well, and the plates were incubated at 37°C with 5% CO₂ for 1 h before inoculation with bacteria. Eighteen-hour cultures of bacterial isolates were diluted in DMEM to give bacterial concentrations of approximately 1 × 10⁹ CFU/ml. Then, 200 μl of the bacterial suspension was pipetted into each well of cell monolayers to achieve 1 × 10⁸ CFU bacterial cells per 10⁵ BT cells per well. The plates were incubated for 3 h at 37°C with 5% CO₂. After 3 h of incubation, unbound bacterial cells were removed by washing four times with DMEM. The monolayers were then lysed with 0.1% Triton X-100 in DPBS for 30 min at room temperature on an orbital shaker. The detached bacterial cells were aspirated and serially diluted with DPBS and then plated onto Lactobacillus MRS agar medium. The plates were incubated for 24 to 48 h at 37°C with 5% CO₂, and colonies were counted (CFU per milliliter). The assay was performed three times in independent experiments on different days.

Antagonistic competition activity of commensals against M. haemolytica on BT cells.

A total of 15 isolates from 3 different genera (Lactobacillus, n = 12; Enterococcus, n = 1; and Staphylococcus, n = 2) that had adherence values of ≥5 log₁₀ CFU/10⁵ BT cells were evaluated in competition assays against M. haemolytica. Antagonistic competition of M. haemolytica was performed using a method described previously (18) with some modifications. Monolayers of BT cells (1 × 10⁵ cells per well) on 6-well plates were washed twice with DMEM and then incubated with 2 ml of antibiotic-free DMEM at 37°C with 5% CO₂ for 1 h before inoculation of bacteria. M. haemolytica and probiotic bacteria suspended in DMEM to achieve individual concentrations of 1 × 10⁸ CFU/ml/well were simultaneously added to the well, and the plates were incubated for 1 h at 37°C with 5% CO₂. For control wells, only M. haemolytica was added. At the end of the experiment, the cell monolayers were washed and lysed as described above, and the adherent M. haemolytica cells were enumerated by plating onto blood agar plates supplemented with 15 μg/ml bacitracin to inhibit Gram-positive bacteria. The reduction in M. haemolytica cells adhering to BT cells was calculated as follows: enumerated M. haemolytica cells in control wells (B₀ [CFU]) − enumerated M. haemolytica cells from wells coinoculated with commensal bacteria (B₁ [CFU])/B₀ × 100%. The competition assay was replicated six times with a minimum of four different culture days for the cell line.

Fluorescence microscopy of bacteria adhering to BT cells.

Bovine turbinate cells were seeded onto two-well chamber slides (Nunc Lab-Tek II chamber slide system; Sigma) and incubated under the growth conditions described above until confluent. The BT cell monolayers were washed three times with DMEM (1 ml each time) and then stained by adding 1 ml of antibiotic-free DMEM (prewarmed) with 5 μl DAPI (4',6-diamidino-2-phenylindole dihydrochloride) (Molecular Probes, Inc., Eugene, OR). The slides were gently rotated to mix the DAPI stain in each chamber, followed by incubation at 37°C, 5% CO₂ for 90 min. After incubation, the cell monolayers were washed three times with DMEM to remove unbound DAPI stain. One milliliter of DMEM was then added to each well and incubated for 30 min before the addition of labeled bacteria. For bacteria, 500-μl aliquots of cultures grown for 18 h were centrifuged at 8,000 × g for 5 min, and the pellets were resuspended with phosphate-buffered saline (PBS) prior to fluorescence-labeling reactions. Mannheimia haemolytica was labeled with Alexa Fluor 488 Microscale Protein, and Lactobacillus strains were labeled with Alexa Fluor 594 microscale protein (Molecular Probes, Inc., Eugene, OR) according to the manufacturer’s instructions. Lactobacillus strains were either added to the stained BT cells alone to visualize adherence or in combination with M. haemolytica to visualize competition. Labeled bacterial cells were suspended in DMEM to achieve concentrations of 1 × 10⁸ CFU/ml/well per bacterium prior to addition to wells of DAPI-stained BT cell monolayers and incubated for 1 h. Subsequently, the cell monolayers were washed four times with 1 ml DMEM to remove all unbound bacterial cells. The cell monolayers were fixed with 2% paraformaldehyde solution (diluted in PBS) and examined using an Olympus Fluoview FV1000 laser confocal scanning microscope.

Evaluation of antibiotic susceptibilities of selected isolates.

A total of 15 isolates that were evaluated in the competition assays were analyzed for their antimicrobial susceptibilities. MICs of 20 antibiotics were determined by the microdilution method (Sensitre; Thermo Fisher Scientific, Nepean, ON, Canada) using a commercially available panel (YSTP6F; Trek Diagnostic Systems, Cleveland, OH, USA). The antimicrobial susceptibility testing was performed according to the procedures recommended for the YSTP6F panel with the exception that Lactobacillus MRS broth was used for Lactobacillus strains that did not grow well in cation-adjusted Mueller-Hinton broth with lysed horse blood. The antimicrobials and the range of concentrations tested are listed in Table 2. Isolates were inoculated into plates using a Sensititre AIM delivery system (Sensitre; Thermo Fisher Scientific), and after incubation, the plates were evaluated with a Vision imager (Sensitre; Thermo Fisher Scientific). The reference strain Lactobacillus plantarum NCDO1193 served as the quality control.

For the Lactobacillus sp. strains, the interpretation of the MIC values of clindamycin (resistant; ≥2 μg/ml), daptomycin (susceptible; ≤4 μg/ml), erythromycin (resistant; ≥8 μg/ml), linezolid (susceptible; ≤4 μg/ml), meropenem (resistant; ≥4 μg/ml), penicillin (susceptible; ≤8 μg/ml), and vancomycin (resistant; ≥16 μg/ml) were based on the interpretive criteria provided by CLSI document M45 (65). The L. paracasei (>4 μg/ml) and L. plantarum (>32 μg/ml) strains were defined as resistant to tetracycline according to the breakpoints provided by the European Food Safety Authority (EFSA) (66). The breakpoints provided by the EFSA (65) were also used to define L. paracasei (>4 μg/ml) and L. plantarum (>8 μg/ml) strains as resistant to chloramphenicol. For L. buchneri strains, resistance to chloramphenicol (>4 μg/ml) and erythromycin (>1 μg/ml) were determined according to the guidance of the FEEDAP Panel (67), while the breakpoint for resistance to tetracycline (128 μg/ml) was determined according to the methods of Feichtinger et al. (42).

For E. faecium, resistance breakpoints for tetracycline (>4 μg/ml), chloramphenicol (≥16 μg/ml), vancomycin (>4 μg/ml), and erythromycin (>4 μg/ml) were determined according to EFSA guidelines (66). The breakpoints provided by CLSI supplement VET01S (68) were used to define Staphylococcus strains (6E and 28C) as resistant to penicillin (≥2 μg/ml; horse), amoxicillin-clavulanate (≥1/0.5; cat), vancomycin (≥16 μg/ml; human), clindamycin (≥4 μg/ml; dog), erythromycin (≥8 μg/ml; human), chloramphenicol (≥32 μg/ml; human), and tetracycline (1 μg/ml; dog). At the time of the experiment, there were no breakpoints or interpretive criteria provided by CLSI or in other literature for the antibiotics azithromycin, cefepime, cefotaxime, ceftriaxone, cefuroxime, ertapenem, levofloxacin, moxifloxacin, tigecycline, and trimethoprim-sulfamethoxazole.

Effects of Lactobacillus sp. isolates on the expression of genes associated with adaptive and innate immune responses in BT cells.

A total of 10 selected commensal isolates from Lactobacillus spp. were evaluated for their effects on the expression of genes associated with adaptive and innate immune responses in BT cell monolayers. These isolates were selected based on their ability to compete against M. haemolytica for adherence to BT cells and their antimicrobial susceptibility phenotypes. The BT cells were seeded onto 12-well flat-bottom tissue culture plates at 1 × 10⁴ cells per well and incubated using the standard culture conditions described above. BT cell monolayers were washed twice with antibiotic-free DMEM and then incubated with 1 ml antibiotic-free DMEM at 37°C with 5% CO₂ for 1 h before inoculation with bacteria. M. haemolytica or commensal bacteria were suspended individually in DMEM and added to BT cells to achieve a concentration of 1 × 10⁷ CFU per well; then, the plates were incubated for 6 h at 37°C with 5% CO₂. Controls included BT cells without addition of bacteria. At the end of the experiment, the cell monolayers were washed four times with DMEM and then lysed with 350 μl of RLT buffer (RNeasy Mini kit; Qiagen, Valencia, CA) for 10 min at room temperature on an orbital shaker. The cell lysates were aspirated and immediately stored at −80°C for further analysis. The immune stimulation assay was performed four times in independent experiments on different days.

Total RNA from the BT cell lysates was extracted using an RNeasy Mini kit (Qiagen) according to the manufacturer’s instructions. The integrity of the extracted total RNA was evaluated with an Agilent Bioanalyzer using an RNA 6000 Nano LabChip (Agilent Technologies, Waldbronn, Germany). Total RNA from each sample (0.5 μg each; RNA integrity number [RIN] > 7) was reverse transcribed into cDNA using an RT² first-strand kit (Qiagen). The cow innate and adaptive immunity response RT² Profiler PCR array (Qiagen) with 84 test genes related to host response to bacterial infection and sepsis was used to evaluate the effects of bacteria on gene expression. Real-time PCR was performed using RT² SYBR green Mastermix (Qiagen) and a CFX Connect real-time system (Bio-Rad, Hercules, CA). Real-time PCR conditions were according to the PCR array manufacturer’s manual (Qiagen). Data normalization was performed with the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and actin beta (ACTB) housekeeping genes, and the fold change in gene expression was calculated using the 2^−ΔΔCT method (69).

M. haemolytica-inhibitory mechanisms of 6 Lactobacillus isolates as candidate bacterial therapeutics.

Based on the rankings of selection criteria (Fig. 1), 6 Lactobacillus strains from four different species (L. amylovorus, L. buchneri [n = 2], L. curvatus, and L. paracasei [n = 2]) were selected as the best therapeutic candidate strains. The potential inhibitory mechanisms by which the selected strains inhibit M. haemolytica were investigated by testing their abilities to produce lactic acid and H₂O₂ and by screening their genomes for bacteriocin-encoding genes. In addition, their effects on the cell morphology of M. haemolytica were evaluated.

Determination of lactic acid production and antimicrobial activity of lactic acid against M. haemolytica.

The selected Lactobacillus strains were inoculated individually into 5 ml MRS broth at an optical density at 610 nm (OD₆₁₀) of 0.05 and incubated aerobically at 37°C with shaking at 200 rpm for 24 h. In addition, all the strains were combined in equal concentrations (OD₆₁₀ = 0.05) and grown similarly to individual strains. MRS broth without any Lactobacillus cells was used as a control. After incubation, supernatants were collected from 1.5 ml of culture by centrifugation (10,000 × g; 5 min), followed by filtering through a 0.22-μm syringe filter to remove bacterial cells. Metaphosphoric acid (25% [vol/wt]) was added to the cell culture supernatant at a ratio of 1:5 and was mixed and immediately stored at −20°C until analysis. The concentrations of lactic acid (dl-lactic acid) in the cell-free supernatants were measured using a 5890A gas-liquid chromatograph (Phenomenex, Torrance, CA, USA) as described by Wang et al. (70).

To determine whether the observed lactic acid range produced by these selected Lactobacillus strains could inhibit the growth of M. haemolytica, M. haemolytica was grown in brain heart infusion (BHI) medium supplemented with different concentrations of lactic acid. A 100-μl overnight M. haemolytica culture was inoculated into 2.5 ml BHI medium containing 0, 9.38, 18.75, 37.5, 75, 100, or 150 mM lactic acid (dl-lactic acid lithium salt; Sigma-Aldrich Canada, Oakville, ON, Canada), and incubated at 37°C, 200 rpm, for 24 h. After 0, 8, and 24 h of incubation, the cultures were serially diluted with DPBS (pH 7.4) and plated onto TSA blood agar plates. The plates were incubated at 37°C for 24 h to determine the number of CFU.

Determination of H₂O₂ production.

Aliquots of cell-free culture supernatants of the 6 selected strains prepared for lactic acid production were subjected to H₂O₂ measurement. A hydrogen peroxide assay kit (colorimetric/fluorometric; ab102500; Abcam Inc., Toronto, ON, Canada) was used according to the manufacturer’s instructions.

Whole-genome sequencing and screening of bacteriocin-encoding genes.

Whole-genome sequencing was performed on the selected strains. The details of genomic DNA extraction, whole-genome sequencing, and analysis were described previously (71). A Web-based tool, BAGEL 4 (72), was used to search for the genes encoding bacteriocin from the genomes of the Lactobacillus strains as described by Flórez and Mayo (73).

Scanning electron microscopy.

The effects of selected Lactobacillus strains on the cell morphology of M. haemolytica were evaluated using SEM. A single colony of each Lactobacillus strain was inoculated into 5 ml BHI and incubated for 24 h in BHI (37°C; 200 rpm). For a negative control, BHI containing no cells was used. After incubation, the cell culture supernatants were obtained as described above. A 100-μl overnight M. haemolytica culture (1 × 10⁸ to 2 × 10⁸ CFU per ml) was centrifuged at 10,000 × g for 5 min, and the cell pellets were suspended with 1 ml cell culture supernatants obtained from the Lactobacillus strains and incubated for 10 h (37°C; 200 rpm) before harvesting cells. The treated M. haemolytica cells were centrifuged at 10,000 × g for 5 min, and the cell pellets were fixed with 4% glutaraldehyde. Further sample processing and SEM imaging procedures were described previously (74).

Statistical analysis.

Competition assay data were analyzed as a one-way analysis of variance (ANOVA) using Proc Glimmix in SAS (version 9.4; SAS Institute Inc., Cary, NC). The LSMEANS statement was used to compare the group means. The difference in gene expression between BT cells (control) that were not cocultured with bacterial cells and the BT cells cocultured with bacterial cells was assessed by Student's t test for each gene using the Rt² Profiler PCR array analysis software, version 3.5 (Qiagen). The level of statistical significance was set at a P value of <0.05.