Changes in antimicrobial susceptibility profiles of Mycoplasma bovis over time

Abstract

Mycoplasma bovis is a major cause of pneumonia, arthritis, and mastitis in cattle and can lead to significant economic losses. Antimicrobial resistance is a concern and further limits the already short list of drugs effective against mycoplasmas. The objective of this study was to examine changes in in vitro minimum inhibitory concentrations (MICs) of antimicrobials of aminoglycoside, fluoroquinolone, lincosamide, macrolide, pleuromutilin, phenicol, and tetracycline classes for 210 M. bovis isolates collected from 1978 to 2009. The MIC₅₀ values of the various antimicrobials were also compared. The MIC₅₀ levels for enrofloxacin and danofloxacin remained low (0.25 μg/mL) across all 3 decades. MIC₅₀ levels for tetracyclines, tilmicosin, and tylosin tartrate were low in the 1980s, then increased in the 1990s and remained high. In the 1980s, MIC₅₀ levels were low for clindamycin, spectinomycin, and tulathromycin, increased in the 1990s to 8 μg/mL (clindamycin) and 32 μg/mL (spectinomycin and tulathromycin), then decreased again in the 2000s. Members of the fluoroquinolone class of antimicrobials had the lowest MIC₅₀ levels across all 3 decades, which suggests in vitro susceptibility of M. bovis to this class of antimicrobials. Statistically significant associations were observed between MIC values for chlortetracycline, oxytetracycline, tylosin tartrate, and tilmicosin; between clindamycin, tulathromycin, spectinomycin, and tiamulin; and between tylosin tartrate and clindamycin. Changes in MIC levels of various antimicrobials over time show the importance of monitoring the susceptibility of mycoplasmas to antimicrobials. The number of antimicrobials that showed elevated MIC₅₀ levels, and therefore possibly reduced in vitro effectiveness against M. bovis, supports initiatives that promote prudent use of antimicrobials in agriculture.

Introduction

Mycoplasma bovis is a member of the Mollicute class, which is a group of bacteria that do not have a cell wall (1). It was first found as a causal agent of mastitis in 1961 and then later isolated from a case of respiratory disease in 1976 (2,3). In addition to causing arthritis, decubital abscesses, otitis media, keratoconjunctivitis, mastitis, abortion, and infertility in dairy and beef cattle, it has been shown that M. bovis is one of the pathogens involved in bovine respiratory disease (BRD), inducing lesions of caseonecrotic bronchopneumonia (1,4–6). A recent study found that 82% of pre-weaned dairy calves from different herds in Quebec were positive for M. bovis and that there were statistical associations between the presence of M. bovis and clinical signs of BRD, lung consolidation, and a decrease in average daily weight gain (4).

Antimicrobial resistance is a serious problem in both human and veterinary medicine. Emerging antimicrobial resistance is particularly problematic when only a few classes of antimicrobials are effective. Since mycoplasmas lack a cell wall, β-lactam antimicrobials, such as ceftiofur and penicillin, are ineffective (7). Sulfonamide antimicrobials are also ineffective because mycoplasmas do not synthesize folic acid (7). Generally, antimicrobials that act on protein or deoxyribonucleic acid (DNA) synthesis, such as tetracyclines, lincosamides, macrolides, phenicols, or fluoroquinolones, are effective (8). Recent studies have suggested, however, that point mutations cause elevated minimum inhibitory concentrations (MICs) against M. bovis for certain classes of antimicrobials (9).

While other studies have described antimicrobial resistance of M. bovis in Canada and in other parts of the world (8,10–17), to our knowledge, no published studies have compared MIC levels for M. bovis strains which were isolated over a long period of time. The objective of this study was to determine the antimicrobial susceptibility profiles of 210 M. bovis isolates from Ontario, Canada, to evaluate changes in MIC in isolates collected from 1978 to 2009, and to compare MIC₅₀ values in various antimicrobials.

Materials and methods

Mycoplasma bovis strains and culture techniques

The 210 bovine M. bovis isolates used in this study were previously isolated from various types of clinical samples submitted to the Animal Health Laboratory at the University of Guelph from 1978 to 2009. Each isolate was from a different individual. The isolates had been identified as M. bovis by indirect fluorescent antibody techniques (IFAT) (18). During this period, 20 to 60 M. bovis strains were isolated each year. Most of the isolates were from lung and bronchoalveolar lavage (BAL) (62%), milk (22%), and joints and occasionally from heart, pleural fluid, ears, and other sources. The collection of M. bovis isolates was stored either as agar blocks at −80°C or as broth cultures in liquid nitrogen.

From this collection, 5 to 7 isolates were selected from each year and from different animals and sources, instead of strictly even distribution. These isolates represented approximately 20% of the M. bovis collection. The isolates were selected to ensure that there was a variety of isolates from beef (n = 81), dairy (n = 113), veal (n = 7), and other breeds (n = 9); types of samples: respiratory (n = 105), joint (n = 35), milk (n = 30), or other (n = 40); and geographic distribution within the province (as determined by the first letter of the farm postal code). Strains of M. bovis isolated from 2 goats and 1 foal were excluded from the study.

Archived isolates were propagated in the following manner. Blocks of agar containing embedded M. bovis colonies or 50 μL of broth cultures were plated directly onto modified Hayflick’s medium containing 15% horse serum (Hyclone; Fisher Scientific, Mississauga, Ontario) and incubated at 35°C in 5% carbon dioxide (CO₂) (19). Plates were examined every 48 h for Mycoplasma spp. using a 40× stereomicroscope. Colonies were identified to the species level by IFAT to reconfirm that they were M. bovis (19). Cultures containing mixed Mycoplasma species were purified to contain only M. bovis by conducting 3 rounds of passage (individual colonies were isolated, subcultured, and identified to the species level using IFAT).

Testing for minimum inhibitory concentration (MIC)

Minimum inhibitory concentration (MIC) levels were determined as described in a previous study (17). Briefly, M. bovis cultures were inoculated into the MIC broth [modified Hayflick’s broth containing 20% fetal calf serum (Gibco through Life Technologies; Mississauga, Ontario), 5 U/mL penicillin G (Sigma Aldrich, Oakville, Ontario), 2 μg/mL thallium acetate (Fisher Scientific), and 0.05% Alamar blue (Cedarlane, Burlington, Ontario)] and incubated for 48 to 72 h at 37°C in 5% CO₂. Once grown, this culture was stored at −80°C in 1-mL aliquots until needed. Before testing, the frozen isolates were thawed at room temperature for 10 to 15 min.

A 10-fold dilution series of each M. bovis culture was prepared from 10⁻¹ to 10⁻⁶ in MIC broth. The concentration of M. bovis [colony-forming units (CFUs)/milliliter] was determined by plating each dilution onto modified Hayflick’s agar containing 20% porcine serum, incubating for 48 to 72 h at 37°C in 5% CO₂, and counting the resulting colonies.

The MIC test was carried out as described in a previous study (17). Briefly, 200 μL of 10³ to 10⁵ CFUs/mL suspension was added to each well of the Sensititre Bovine/Porcine MIC Plate (Code: BOPO6F; Trek Diagnostics, Independence, Ohio, USA). The CFUs per milliliter of the inoculum were afterward determined to be 10⁴ CFUs/mL for 103 inoculums, 10³ CFUs/mL for 102 inoculums, and 10⁵ CFUs/mL for 5 inoculums. The antimicrobials tested included ampicillin (0.25 to 16 μg/mL), ceftiofur (0.25 to 8 μg/mL), neomycin (4 to 32 μg/mL), penicillin (0.5 to 8 μg/mL), sulphadimethoxine (256 μg/mL), chlortetracycline (0.5 to 8 μg/mL), clindamycin (0.25 to 16 μg/mL), danofloxacin (0.12 to 1 μg/mL), enrofloxacin (0.12 to 2 μg/mL), florfenicol (0.25 to 8 μg/mL), gentamicin (1 to 16 μg/mL), oxytetracycline (0.5 to 8 μg/mL), spectinomycin (8 to 64 μg/mL), tiamulin (1 to 32 μg/mL), tilmicosin (4 to 64 μg/mL), trimethoprim/sulfamethoxazole (2 to 38 μg/mL), tulathromycin (1 to 64 μg/mL), and tylosin tartrate (0.5 to 32 μg/mL). As described in a previous study, 10-fold dilutions of the challenge culture were inoculated into wells without antimicrobial to confirm that the challenge was within the acceptable range of color-changing units per well (17).

Plates were incubated at 37°C in 5% CO₂ and examined every 24 h for up to 96 h. The lowest concentration of antimicrobial that inhibited growth, as indicated by no blue-red color change, was recorded as the MIC of that antimicrobial. The readings at 48 h were used as the endpoint result. An in-house M. bovis control strain (strain 227) had been run with the MIC test 10 times before this study in order to establish limit ranges. This same M. bovis control strain was also included in each MIC test experiment.

Statistical analyses were conducted with SAS (Version 9.4, 2013; SAS Institute, Cary, North Carolina, USA) using bacterial isolate (N = 210) as the experimental unit. The MIC level was modeled separately for each antibiotic as a general linear mixed model (GLMM) using PROC MIXED (SAS 9.4). The fixed-effects variables included year of isolation, breed, and source. All 2-way interactions among year of isolation, breed, source, and the covariates were entertained, as were the 3-way interactions among date, source, time, and the covariates. Manual backward elimination of variables with P > 0.10 was used to create the final models for MIC level for each antibiotic. Residual analyses were conducted to assess whether analysis of variance (ANOVA) assumptions were met. Residuals were tested for normality using 4 tests inherent in the SAS software (Shapiro-Wilk, Cramér-von Mises, Kolmogorov-Smirnov, and Anderson-Darling tests). Residuals were also plotted against the predicted values and explanatory variables included in the model. To generate normal distribution, data were log₂-transformed. MIC values for neomycin, enrofloxacin, florfenicol, and tiamulin were constant across time and no further statistical analyses were conducted. The effect of year of isolation was first analyzed as a continuous variable to determine significant differences in MIC values across time. Based on observations of the data, MIC values were subsequently compared among the year categories 1978 to 1990, 1991 to 2000, and 2001 to 2009 using PROC MIXED. Correlations among MIC values for the different antimicrobials were tested on an individual sample basis using the CORR procedure to determine Spearman correlation coefficients and P-values. P-values < 0.05 were considered significant. Data in all graphs are represented as a mean ± 95% confidence interval.

Results

The background information and MIC values for all of the 210 isolates used in this study are shown in Supplemental Table I available from the authors upon request. There was no correlation between MIC values of any antimicrobial and the geographic location or breed of cattle. The source of the sample, i.e., anatomic site, was not correlated to MIC values, with the exception of higher MIC values for gentamycin for isolates from joints than for those from lungs (P < 0.0031).

When the year of isolation was analyzed as a continuous variable, there were significant differences in MIC values across time for chlortetracycline, oxytetracycline, tilmicosin, tylosin, clindamycin, tulathromycin, spectinomycin, danofloxacin, and gentamycin, but not for neomycin, enrofloxacin, florfenicol, or tiamulin (Figure 1).

Figure 1.

The changes in the minimum inhibitory concentration (MIC, μg/mL) for various antimicrobials for Mycoplasma bovis isolates over time. The data show the median MIC values (and 95% confidence intervals) for each antimicrobial for the M. bovis isolates in each 2-year period from 1978 to 2009. The statistical significance of changes over time is shown in Tables I and II.

Observation of the data identified changes in MIC values in 1991 and 2001 (Supplemental Table I). These changes were confirmed by statistical analysis when the years of isolation were categorized as 1978 to 1990, 1991 to 2000, and 2001 to 2009 (Table II). The MIC₅₀ values of some antimicrobials increased progressively over time. The MIC₅₀ values of chlortetracycline, oxytetracycline, tilmicosin, and tylosin tartrate were low in isolates from 1978 to 1990 (2 μg/mL, 2 μg/mL, 2 μg/mL, and 0.5 μg/mL, respectively). Beginning in 1991, however, the MIC₅₀ values increased to 4 μg/mL for chlortetracycline and oxytetracycline, 32 μg/mL for tilmicosin, and 16 μg/mL for tylosin tartrate and remained at these high levels from 1991 to 2009 (Figure 1, Table II, and Supplemental Table I).

Table II.

Analysis of the minimum inhibitory concentrations (MICs) of various antimicrobials against isolates of Mycoplasma bovis from 1978 to 1990, 1991 to 2000, and 2001 to 2009.

Antimicrobial	Year of isolation	MIC50	Number of strains with MIC (μg/mL)									P-values
			0.125	0.25	0.5	1	2	4	8	16	32	1978 to 1990 versus 1991 to 2000	1978 to 1990 versus 2001 to 2009	1991 to 2000 versus 2001 to 2009
Neomycin	1978 to 1990	16	0	0	0	0	0	0	1	75	0	NS	NS	NS
	1991 to 2000	16	0	0	0	0	0	0	0	70	0
	2001 to 2009	16	0	0	0	0	0	0	0	63	1
Enrofloxacin	1978 to 1990	0.25	0	76	0	0	0	0	0	0	0	NS	NS	NS
	1991 to 2000	0.25	0	70	0	0	0	0	0	0	0
	2001 to 2009	0.25	0	64	0	0	0	0	0	0	0
Florfenicol	1978 to 1990	4	0	0	0	2	4	70	0	0	0	NS	NS	NS
	1991 to 2000	4	0	0	0	1	14	55	0	0	0
	2001 to 2009	4	0	0	0	0	3	61	0	0	0
Tiamulin	1978 to 1990	0.5	0	29	34	6	7	0	0	0	0	NS	NS	NS
	1991 to 2000	0.5	0	11	36	23	0	0	0	0	0
	2001 to 2009	0.5	0	7	53	3	1	0	0	0	0
Tilmicosin	1978 to 1990	2	0	0	0	0	66	1	2	3	4	***	***	**
	1991 to 2000	32	0	0	0	0	7	0	5	5	52
	2001 to 2009	32	0	0	0	0	0	0	2	2	60
Chlortetracycline	1978 to 1990	2	0	3	1	25	18	29	0	0	0	NS	NS	NS
	1991 to 2000	4	0	0	1	2	2	65	0	0	0
	2001 to 2009	4	0	0	0	1	1	62	0	0	0
Oxytetracycline	1978 to 1990	2	0	3	1	27	19	26	0	0	0	NS	NS	NS
	1991 to 2000	4	0	0	1	2	3	64	0	0	0
	2001 to 2009	4	0	0	0	1	1	62	0	0	0
Tylosin	1978 to 1990	0.5	0	4	41	15	8	3	1	4	0	***	***	*
	1991 to 2000	16	0	1	1	3	4	7	3	51	0
	2001 to 2009	16	0	0	0	0	1	1	2	60	0
Clindamycin	1978 to 1990	0.125	53	11	9	0	0	0	3	0	0	***	NS	***
	1991 to 2000	8	15	5	1	0	0	1	48	0	0
	2001 to 2009	0.25	21	36	2	0	0	0	5	0	0
Tulathromycin	1978 to 1990	4	0	0	0	2	4	46	9	8	7	***	NS	***
	1991 to 2000	32	0	0	0	0	4	8	8	2	48
	2001 to 2009	4	0	0	0	0	5	36	12	4	7
Spectinomycin	1978 to 1990	4	0	0	0	0	0	66	4	0	6	***	NS	***
	1991 to 2000	32	0	0	0	0	0	21	2	0	47
	2001 to 2009	4	0	0	0	0	0	47	11	0	6
Danofloxacin	1978 to 1990	0.25	0	76	0	0	0	0	0	0	0	***	*	*
	1991 to 2000	0.25	1	50	19	0	0	0	0	0	0
	2001 to 2009	0.25	1	55	8	0	0	0	0	0	0
Gentamycin	1978 to 1990	4	0	0	0	0	3	59	14	0	0	NS	NS	NS
	1991 to 2000	4	0	0	0	0	3	45	22	0	0
	2001 to 2009	4	0	0	0	0	2	47	15	0	0

^*P < 0.05;

^**P ≤ 0.001;

^***P ≤ 0.00001.

NS — not statistically significant.

The MIC₅₀ values of some antimicrobials showed similar changes in 1991, with further changes in 2001. From 1978 to 1990, the MIC₅₀ values of clindamycin, spectinomycin, and tulathromycin were low (0.125, 4, and 4 μg/mL, respectively). Beginning in 1991 and continuing to 2000, the MIC₅₀ values were subsequently increased to 8 μg/mL for clindamycin and 32 μg/mL for spectinomycin and tulathromycin (Table I). Beginning in 2001 and continuing to 2009, however, the MIC₅₀ values of clindamycin, spectinomycin, and tulathromycin declined to 0.25, 4, and 4 μg/mL, respectively (Figure 1, Table II, and Supplemental Table I).

Table I.

Multivariate analysis of the relationship between date of sampling (analyzed as a continuous variable), breed, or source and minimum inhibitory concentrations of various antibiotics against isolates of Mycoplasma bovis.

Antibiotic	P-values			Pearson r
	Breeda	Sourceb	Date	Date
Neomycin	NS	NS	NS	NA
Enrofloxacin	NS	NS	NS	NA
Florfenicol	NS	NS	NS	NA
Tiamulin	NS	NS	NS	NA
Tilmicosin	NS	NS	< 0.0001	0.80
Chlortetracycline	NS	NS	< 0.0001	0.58
Oxytetracycline	NS	NS	< 0.0001	0.59
Tylosin	0.021	NS	< 0.0001	0.80
Clindamycin	NS	NS	< 0.0001	0.46
Tulathromycin	NS	0.090	< 0.0001	0.40
Spectinomycin	0.680	0.028	< 0.0001	0.48
Danofloxacin	NS	NS	0.003	0.22
Gentamycin	NS	0.006	0.005	0.31

^aDairy, beef, veal, other.

^bMilk, lung, joint, other.

NS — not statistically significant.

NA — not applicable.

The MIC₅₀ values of some antimicrobials did not change over time (Figure 1, Table II). Isolates from all 3 decades had low MIC₅₀ values for enrofloxacin (0.25 μg/mL), danofloxacin (0.25 μg/mL), and tiamulin (0.5 μg/mL). Isolates from all 3 decades had medium MIC₅₀ values for florfenicol and gentamicin (4 μg/mL) and high MIC₅₀ values for neomycin (16 μg/mL). The MIC values for neomycin, enrofloxacin, florfenicol, and tiamulin were constant across time and thus could not be fitted to the statistical model. As expected, all M. bovis isolates had high MIC₅₀ values for cell-wall active antimicrobials, including ampicillin, ceftiofur, and penicillin (data not shown).

Correlations were observed among MIC values for many antimicrobials. For example, the MIC values for tilmicosin and tylosin tartrate were correlated [Spearman correlation coefficient (ρ) = 0.948], i.e., isolates that had high MIC values for tilmicosin also had high MIC values for tylosin tartrate. Similarly, the MIC values of chlortetracycline and oxytetracycline were correlated (ρ = 0.967), as were tulathromycin with spectinomycin (ρ = 0.684) and clindamycin (ρ = 0.803). Other associations of note were between antimicrobials from the aminoglycoside, tetracycline, pleuromutilin, macrolide, and lincosamide classes. Specifically, there were statistically significant correlations between clindamycin and antimicrobials from the macrolide class, including tulathromycin (ρ = 0.803), tilmicosin (ρ = 0.561), and tylosin tartrate (ρ = 0.647) (Table III).

Table III.

Correlation of minimum inhibitory concentrations (MICs) among various antimicrobials for 210 isolates of Mycoplasma bovis accounting for effects of source and time.

	CLRT	OXYT	TILM	TYLO	CLIN	TULA	TIAM	SPEC	DANO	GENT	NEOM	FLOR
CLRT		0.97	0.66	0.72	0.48	0.39	0.40	0.38	0.14	0.16	0.12	0.12
OXYT	0.97		0.69	0.76	0.51	0.41	0.42	0.40	0.15	0.13	0.12	0.11
TILM	0.66	0.69		0.95	0.56	0.44	0.36	0.47	0.23	0.10	0.10	0.02
TYLO	0.72	0.76	0.95		0.65	0.50	0.44	0.52	0.23	0.15	0.11	0.08
CLIN	0.48	0.51	0.56	0.65		0.80	0.60	0.77	0.33	0.15	0.06	−0.05
TULA	0.39	0.41	0.44	0.50	0.80		0.63	0.68	0.29	0.15	0.04	0.01
TIAM	0.40	0.42	0.36	0.44	0.60	0.63		0.39	0.14	0.11	0.08	0.26
SPEC	0.38	0.40	0.47	0.52	0.77	0.68	0.39		0.29	0.11	0.00	−0.13
DANO	0.14	0.15	0.23	0.23	0.33	0.29	0.14	0.29		0.15	0.00	−0.09
GENT	0.16	0.13	0.10	0.15	0.15	0.15	0.11	0.11	0.15		0.08	0.23
NEOM	0.12	0.12	0.10	0.11	0.06	0.04	0.08	0.00	0.00	0.08		0.15
FLOR	0.12	0.11	0.02	0.08	−0.05	0.01	0.26	−0.13	−0.09	0.23	0.15

The data show Spearman correlation coefficients (ρ). Values in bold are significant at P < 0.05.

CLRT — chlortetracycline; OXYT — oxytetracycline; TILM — tilmicosin; TYLO — tylosin; CLIN — clindamycin; TULA — tulathromycin; TIAM — tiamulin; SPEC — spectinomycin; DANO — danomycin; GENT — gentamycin; NEOM — neomycin; FLOR — florfenicol. MIC values for enrofloxacin were constant and are not included.

Discussion

Since there are no MIC breakpoints approved by the Clinical and Laboratory Standards Institute (CLSI) for mycoplasmas of animals (17), we report results as MIC values. When MIC values increase, it would suggest that the antimicrobial may become less effective. Without using standardized methods and established CLSI-approved MIC breakpoints, however, it is difficult to predict the clinical efficacy of the antimicrobial (17). In addition to in vitro susceptibility, pharmacokinetics are expected to affect the in vivo interpretation of the in vitro susceptibility data.

Specifically, our study did not account for the levels of the different antimicrobial drugs present in lung and other tissue after administration to cattle according to label directions, although drug levels in tissue are relevant to interpretation of these MIC values. Furthermore, caseonecrotic bronchopneumonia is a characteristic feature of M. bovis pneumonia and drug delivery to areas of tissue necrosis may be reduced, which further affects drug levels in tissue at the site of infection. Finally, of the antimicrobials tested in this panel, only tulathromycin (Draxxin; Zoetis, Parsippany, New Jersey, USA, introduced in Canada in 2006) has a label claim for preventing or treating M. bovis pneumonia. Gamithromycin (Zactran; Merial, Ingelheim, Germany) also has a label claim for preventing or treating M. bovis pneumonia, but this antimicrobial, which was introduced in 2010, was not included in this panel. Many of the other antimicrobials tested have a label claim for use in BRD caused by other bacterial pathogens, including ampicillin (Polyflex; Boehringer-Ingelheim, Burlington, Ontario, introduced in 2010), ceftiofur (Excenel; Zoetis, introduced in 1989), danofloxacin (A180; Zoetis, introduced in 2007), enrofloxacin (Baytril; Bayer, Mississauga, Ontario, introduced in 2004), florfenicol (Nuflor; Merck, Kirkland, Quebec, introduced in 1996), oxytetracycline (various trade names and manufacturers), penicillin (various trade names and manufacturers), tilmicosin (Micotil; Elanco Canada, Guelph, Ontario, introduced in 1990), trimethoprim-sulfadoxine (Borgal; Merck), and tylosin tartrate (Tylan 200; Elanco Canada). Several of the remaining antimicrobials tested do not have a label claim for use in preventing or treating BRD (clindamycin, gentamycin, neomycin, spectinomycin, and tiamulin).

Throughout each of the 3 decades, the MIC₅₀ remained low, with values of 0.25 μg/mL for both enrofloxacin and danofloxacin. When administered at the recommended daily dose, it has been reported that mean maximum concentrations (Cmax) ± standard deviations of enrofloxacin (0.24 ± 0.08 μg/mL) and danofloxacin (0.23 ± 0.05 μg/mL) were detected in the plasma of calves (20). Results from our study suggest that M. bovis has remained sensitive to the fluoroquinolones enrofloxacin and danofloxacin over the last 3 decades. Other studies also concluded that the fluoroquinolone class of antimicrobials was the most active compound in vitro against M. bovis in Europe, Japan, and the United States (12,14,16,17).

Conversely, another European study has shown that older isolates of M. bovis isolated from 1978 to 1979 in France had lower MIC levels (0.25 μg/mL) for enrofloxacin, whereas those isolated from 2010 to 2012 had increased MIC levels (0.5 μg/mL), which suggests that M. bovis isolates in France are becoming more resistant to enrofloxacin (13). Interestingly, another Canadian study that we were part of examined the antimicrobial susceptibility of M. bovis isolated from bison from 2012 to 2014 and found that fluoroquinolones, specifically enrofloxacin and danofloxacin, were not an effective class of antimicrobials in vitro. The M. bovis isolates used in that study were cultured from bison in Alberta, Saskatchewan, and Manitoba from 2012 to 2014 and MIC testing was done at the Animal Health Laboratory at the University of Guelph (8). Further studies may be needed to determine whether M. bovis isolated from cattle since 2010 have developed reduced susceptibility to enrofloxacin and danofloxacin or if this phenomenon is only observed in M. bovis isolates from bison. Enrofloxacin and danofloxacin were licensed for treatment of BRD in Canada in 2004 and 2007, respectively (21). This could potentially explain the increase in MIC levels reported in M. bovis isolated from bison by Suleman et al (8).

Our study also showed low MIC values for M. bovis to spectinomycin, tulathromycin, tiamulin, and clindamycin. Similar to our study, spectinomycin was effective in vitro towards M. bovis isolates from Japan (16). In contrast, European studies have reported decreased in vitro effectiveness of spectinomycin towards M. bovis in Europe (12,13,15). Although spectinomycin is licensed for use in cattle in Canada, it does not have a label claim for bovine respiratory disease (BRD). In the current study, M. bovis isolates from Ontario, Canada also showed an increase in MIC₅₀ values of spectinomycin during 1990/1991, which suggests that M. bovis was becoming less sensitive to these antimicrobials. Beginning in 2001, this phenomenon seems to have reversed as M. bovis isolated from 2001 to 2009 returned to low MIC₅₀ levels.

Similarly, MIC₅₀ values of tulathromycin and clindamycin increased during 1990/1991, then declined in 2001. These findings suggest that changes in selective pressure for antimicrobial resistance may affect the susceptibility of pathogens for a particular antimicrobial (22,23). It is notable, however, that of these antimicrobials, clindamycin is not licensed for use in cattle, spectinomycin was used in cattle during all periods of the study but was never licensed for use in BRD in Canada, and tulathromycin was not released for use in Canada until 2006. The reason for these changing patterns of MIC₅₀ values does not seem to be directly related to when these particular antimicrobials were first available for use in Canada, although these patterns could reflect cross-resistance to other antimicrobial drugs and perhaps the use of these antimicrobials for indications other than BRD.

Antimicrobials, such as oxytetracycline, chlortetracycline, tilmicosin, and tylosin tartrate, became less effective over the 3 decades. The MIC₅₀ levels for each of these antimicrobials increased in about 1990/1991. Of these, tilmicosin was released for use in Canada in 1990, whereas oxytetracycline, chlortetracycline, and tylosin tartrate were licensed for use in all years of the study.

It is well-documented that, in many bacterial species, antimicrobial resistance can be conferred and disseminated by resistance genes located on mobile genetic elements, such as transposons, integrons, gene cassettes, and plasmids, through horizontal gene transfer (24–26). It is not known if antimicrobial resistance can be transferred by mobile genetic elements as no plasmids have been identified in M. bovis (27). Several studies have shown, however, that macrolide, tetracycline, pleuromutilin, fluoroquinolone, and phenicol antimicrobial resistance in M. bovis can occur by point mutations that alter the antimicrobial targets (9,27–31). In vitro experiments showed that elevated MIC levels to some antimicrobials occurred in as few as 2 to 19 passages (9).

In our current study, statistical associations were observed between MIC values of tylosin tartrate, chlortetracycline, oxytetracycline, and tilmicosin. Sulyok et al (9) found that tetracycline resistance, both naturally and during in vitro experiments, was conferred in M. bovis by mutations in the rrs2 gene. These same mutations were also found in some tylosin-resistant strains. This same study showed that all macrolide-resistant in vitro mutants were resistant to both tylosin and tilmicosin (9). While a correlation between resistance to spectinomycin and tetracycline has also been previously observed, this correlation was weak, although statistically significant, in our study.

Our study also found correlations between MICs of the lincosamide and macrolide families. Specifically, the MIC of clindamycin was statistically correlated with tilmicosin, tulathromycin, and tylosin tartrate resistance. Similarly, other studies have also shown cross-resistance to macrolides and lincomycin, which is another lincosamide related to clindamycin (9,31). Antimicrobials from both of these families target the 50S subunit of the 70S ribosome and inhibit protein synthesis. It has been shown that point mutations at the macrolide-binding site cause decreased susceptibility to both macrolides and lincomycin (9). These findings therefore suggest a relationship between the 1990 release of tilmicosin for use in Canada and the changes observed in 1990/1991 in the form of increased MIC₅₀ values to tilmicosin, oxytetracycline, chlortetracycline, and tylosin tartrate. Because these data are based on diagnostic laboratory samples and many of these animals may have been treated with antibiotics, including tilmicosin, the findings do not necessarily indicate emergence of resistant strains in the general population.

It has been reported that tetracycline, spectinomycin, and macrolides have been used to treat diseases associated with M. bovis, but resistance and decreased effectiveness have been reported worldwide (9). Similar to the results of our study, fluoroquinolones are still considered effective agents against bovine mycoplasmal disease. In vitro experiments showed, however, that mutations resulting in fluoroquinolone resistance were selected for in as little as 5 to 7 passages (9). Our study was on the isolates up to 2009 and may not reveal more recent shifts in MIC values.

In conclusion, our study shows changes in MIC levels of various antimicrobials over time and demonstrates the importance of monitoring the susceptibility of M. bovis to antimicrobials. The number of antimicrobials that showed elevated MIC₅₀ levels, and therefore reduced in vitro effectiveness against M. bovis, also supports initiatives that promote prudent use of antimicrobials in agriculture.