Molecular epidemiology, antimicrobial susceptibility, and pulsed-field gel electrophoresis genotyping of Pseudomonas aeruginosa isolates from mink

Abstract

Pseudomonas aeruginosa is an important animal pathogen and contributes to hemorrhagic pneumonia in mink. Between April 2011 and December 2016, samples of lung, liver, and spleen were collected from mink with this disease on 11 mink farms in 5 Chinese provinces. From these samples, we obtained 98 isolates of P. aeruginosa that belonged to 5 serotypes: G (n = 58), I (n = 15), C (n = 8), M (n = 5), and B (n = 2); 10 isolates were not typeable (10/98). More than 90% of the isolates formed biofilms, and 85% produced slime. All 98 isolates were resistant to 10 antibiotics (oxacillin, ampicillin, penicillin G, amoxicillin, ceftriaxone, cefazolin, cefaclor, tilmicosin, tildipirosin, and sulfonamide). However, almost all were susceptible to gentamicin, polymyxin B, and amikacin. We identified 56 unique genotypes by pulsed-field gel electrophoresis. These findings have revealed genetic diversity and high antimicrobial resistance in P. aeruginosa isolated from mink with hemorrhagic pneumonia and will facilitate the prevention and control of this disease.

Introduction

Pseudomonas aeruginosa is a major cause of death and economic loss in the mink industry and is the agent of hemorrhagic pneumonia, with a mortality rate near 50% since 1983 (1–5). This disease is now prevalent on most mink farms in China and appears in the autumn, although precise epidemiologic data are lacking (3,6).

Antibiotics are commonly used to treat P. aeruginosa infections, but antibiotic resistance in isolates from mink is rapidly increasing (3). An additional threat to these animals is the ability of P. aeruginosa to form biofilms (7). Biofilm formation is a multistep process that requires the participation of bacterial structural appendages such as flagella and type IV pili. These appendages provide the bacterium with swimming, swarming, and twitching motility, which makes the infections difficult to eradicate (8,9). The combination of resistance to multiple antibiotics, biofilm formation, and motility contributes to persistence of infections caused by P. aeruginosa (10).

Knowledge of the antimicrobial susceptibility of P. aeruginosa and epidemiologic data is important for effective treatment of mink hemorrhagic pneumonia. The objective of this study was to evaluate the epidemiologic features of these infections by the following means: i) investigating genetic relatedness; ii) evaluating serotype, biofilm formation, slime production, and motility; and iii) analyzing the results of antimicrobial susceptibility testing.

Materials and methods

Isolation and identification of P. aeruginosa

Pseudomonas aeruginosa was isolated from dead or diseased mink that had been screened for hemorrhagic pneumonia between April 2011 and December 2016 in 5 Chinese provinces: Shandong, Hebei, Jiangsu, Jilin, and Liaoning. Lung, liver, and spleen samples were collected for bacterial isolation and were streaked onto nalidixic acid–cetrimide agar plates (Qingdao Haibo Biotech, Qingdao, China) and incubated at 37°C for 24 h. Single green colonies with irregular borders were picked for Gram’s staining and inoculated into sterile tryptic soy agar (Qingdao Haibo Biotech) and incubated at 37°C for 24 h. Each isolate suspension was used for biochemical and polymerase chain reaction (PCR) identification as previously described (11). Slide agglutination serotyping was done with commercial polyvalent I, II, and III group-specific antiserum (Denka Seiken Company, Tokyo, Japan) against 14 O antigens as previously described (12).

The study was approved by the animal research committee of the South China Agriculture University (protocol 2014-031). All procedures were done in a way that minimized animal suffering as defined by the guidelines issued by this committee.

Tests for biofilm formation, slime production, and motility

The test for biofilm formation was based on a method described previously (13,14). Pseudomonas aeruginosa isolates were grown in Mueller–Hinton broth (MHB) at 37°C for 24 h and then diluted to 1:100 with MHB. Aliquots (200 μL) of the diluted cell suspensions were added to 96-well flat-bottom microtiter polystyrene plates, which were sealed with parafilm and incubated overnight at 37°C. The wells were washed 3 times with phosphate-buffered saline (PBS), pH 7.2, and bacterial biofilms were visualized by staining with 200 μL of 0.1% crystal violet for 15 min at room temperature. The plates were again washed 3 times with PBS, and 200 μL of 95% ethanol was added. The optical density at 450 nm (OD₄₅₀) was measured in an enzyme-linked immunosorbent assay (ELISA) plate reader with MHB used as a negative control. Each assay was done in triplicate. Relative scores for biofilm formation were assigned as follows: OD₄₅₀ = 0.050, absent; 0.050 < OD₄₅₀ ≤ 0.1, weak; 0.1 < OD₄₅₀ ≤ 0.15, moderate; and 0.15 < OD₄₅₀, strong.

Tests for slime production were carried out as previously described (15). The results were categorized as follows: −, negative; +, weak; +++, moderate; and ++++, strong.

Pseudomonas aeruginosa isolates were cultured overnight in trypticase soy broth for assays of motility (swimming, swarming, and twitching) as previously described (16). The diameter of the turbid zone was measured.

Antimicrobial susceptibility testing

Antibiotic susceptibility of the P. aeruginosa isolates to 32 antibiotics was determined by minimum inhibitory concentration (MIC) testing, including breakpoint determinations, according to the guidelines of the Clinical and Laboratory Standards Institute, Wayne, Pennsylvania, USA (17). All antibiotics were obtained from the National Institutes for Food and Drug Control, Beijing, China, with the exception of tildipirosin, which was provided by Shanghai Tongren Pharmaceutical Company, Shanghai, China. Escherichia coli ATCC (American Type Culture Collection) 25922 and P. aeruginosa ATCC 27853 served as quality-control strains.

Pulsed-field gel electrophoresis (PFGE)

The P. aeruginosa isolates were typed by means of PFGE as previously described with modifications (18). In brief, genomic DNA was embedded in 1% Seakem Gold agarose (Lonza, Allendale, New Jersey, USA) and digested in situ with SpeI (Takara, Japan) at 37°C for 4 h. Salmonella enterica serovar Braenderup H9812 standards digested with XbaI (Takara, Japan) were used as molecular size markers. Electrophoresis was carried out in a 1% (w/v) agarose gel in 0.5× Tris–borate–ethylene diamine tetraacetic acid at 14°C on a CHEF Mapper Power Module system (Bio-Rad, Hercules, California, USA) with the following settings: initial switching time, 2.16 s; final switching time, 63.8 s; voltage, 6 V/cm; angle, 120°; and run time, 21.3 h. Band profiles were visually assessed with the use of BioNumerics 6.6 software (Applied Maths, Austin, Texas, USA). Isolates with correlation coefficients greater than 85% were defined as genetically identical (19). Simpson’s index of diversity was calculated according to the method of Hunter and Gaston (20).

Results

We identified 98 clinical isolates of P. aeruginosa from mink farms in 5 Chinese provinces (Table I). The samples came from lung [58% (57/98)], liver [27% (26/98)], and spleen [15% (15/98)] (Figure 1). Some isolates produced the pigment green fluorescein, and all were Gram-negative. All the isolates metabolized glucose and xylose but not sucrose or lactose. All were oxidase-positive, hydrolyzed gelatin, and produced urea but not indole and hydrogen sulfide. Polymerase chain reaction (PCR) confirmed that the isolates were P. aeruginosa.

Table I.

Pseudomonas aeruginosa isolates from mink farms in China.

Province	City	Isolates	Number	Serotype
Shandong	Chengyang	PAM01, PAM07, PAM16, PAM19, PAM20, PAM47, PAM49, PAM65, PAM87, PAM89, PAM97	11	G, M, I, NT
	Linyi	PAM02, PAM03, PAM13, PAM72, PAM75	5	G, M
	Zhucheng	PAM04, PAM11, PAM43, PAM69	4	G
	Rizhao	PAM05, PAM32, PAM36, PAM37, PAM53, PAM55, PAM60, PAM61, PAM76, PAM80, PAM94, PAM96	12	G, NT
	Weifang	PAM06, PAM18, PAM40, PAM41, PAM68	5	G, I, C
	Jiaozhou	PAM08, PAM10, PAM27, PAM28, PAM57, PAM79	6	G, I, NT
	Weihai	PAM12, PAM24, PAM48, PAM50, PAM58, PAM64, PAM67, PAM77, PAM81, PAM85, PAM90	11	G, C, M, NT
	Jimo	PAM15, PAM22, PAM25, PAM29, PAM35, PAM51, PAM54, PAM70, PAM86, PAM88, PAM93	11	G, I, C
Hebei		PAM09, PAM31, PAM44, PAM52, PAM62, PAM71, PAM73, PAM82, PAM84, PAM92	10	G, I, B
Jiangsu		PAM14, PAM39, PAM91	3	G, NT
Jilin		PAM17, PAM21, PAM23, PAM30, PAM34, PAM38, PAM42, PAM45, PAM46, PAM78, PAM95	11	G, I, B, NT
Liaoning		PAM26, PAM33, PAM56, PAM59, PAM63, PAM66, PAM74, PAM83, PAM98	9	G, C, I, NT

NT — nontypeable.

Figure 1.

Dendrogram of banding patterns obtained by pulsed-field gel electrophoresis of digested DNA from 98 Pseudomonas aeruginosa isolates.

The isolates belonged to 5 serotypes: G (58), I (15), C (8), M (5), and B (2). Ten isolates were nontypeable. The 58 serotype G isolates were obtained from different geographic regions (Table I and Figure 1).

Of the 98 isolates 47% were strong or moderate biofilm formers, and almost as many (45%) were weak biofilm formers. Only 8% were incapable of biofilm formation (Table II).

Table II.

Biofilm formation and slime production by the 98 isolates.

Biofilm-forming ability		Slime production; number of isolates
Number of isolates	Average OD450
20	0.168	22
26	0.121	24
44	0.062	37
8	0.05	15

OD — optical density at 450 nm.

Slime production mirrored biofilm-forming ability, 47% of the isolates being strong or moderate slime producers, 38% being weak producers, and 15% producing no slime (Table II). However, 40% (6/15) of the isolates not producing slime were moderate biofilm producers (average OD₄₅₀ 0.152).

The isolates varied greatly in motility (Figure 2). The diameter of the circular turbid zone indicating swimming activity ranged from 4.0 to 30.0 mm (mean: 13.3 mm), whereas the diameter of swarming activity ranged from 6.0 to 31.0 mm (mean: 10.2 mm). Twitching was visible in all the isolates and was diverse, the zone diameter ranging from 6.0 to 27.0 mm (mean: 11.8 mm). All these behaviors concentrated within a zone with a diameter of 5.0 to 15.0 mm.

Figure 2.

Swimming, swarming, and twitching of the isolates.

All of the isolates were resistant to oxacillin, ampicillin, penicillin G, amoxicillin, ceftriaxone, cefazolin, cefaclor, sulfonamide, tilmicosin, and tildipirosin (Table III). We also found a high incidence of resistance to ceftiofur (96%), nalidixic acid (94%), trimethoprim (83%), kanamycin (76%), florfenicol (71%), cefepime (68%), tetracycline (59%), cefoperazone (56%), and enrofloxacin (51%). A lower percentage of the isolates were resistant to doxycycline (37%), lomefloxacin (35%), spectinomycin (20%), and ceftazidime (20%). Even smaller proportions (11% to 14%) were resistant to norfloxacin, ciprofloxacin, and levofloxacin. However, almost all of the isolates were susceptible to gentamicin, polymyxin B, and amikacin. There is no breakpoint for caphalonium, but the MIC were all ≥ 512 mg/L. The MIC₅₀ values of all 32 tested antibiotics were in the range of 0.5 to ≥ 512 mg/L, and the MIC₉₀ values were 1.0 to ≥ 512 mg/L.

Table III.

Results of antibiotic susceptibility testing of the 98 isolates.

Antibiotic	Minimum inhibitory concentration (MIC), mg/L; number of isolates												(S)	(I)	(R)	MIC50	MIC90	Resistance %
	≥ 512	256	128	64	32	16	8	4	2	1	0.5	0.25
Nalidixic acid	74	2	10		6	2	4						≤ 16	—	≥ 32	≥ 512	≥ 512	93.80
Lomefloxacin		1		1	1	10	21	17	4	5	10	28	≤ 2	4	≥ 8	4	16	34.69
Ciprofloxacin		1				3	5	4	8	22	17	38	≤ 1	2	≥ 4	0.5	8	13.27
Norfloxacin	1	1			4	5	2	15	18	8	8	36	≤ 4	8	≥ 16	1	16	11.22
Levofloxacin	1						13	12	19	10	6	37	≤ 2	4	≥ 8	1	8	14.29
Enrofloxacin			3	5	9	16	17	11	11	11	3	12	≤ 2	4	≥ 8	8	32	51.02
Oxacillin	98												≤ 1	2	≥ 4	≥ 512	≥ 512	100
Ampicillin	98												≤ 8	16	≥ 32	≥ 512	≥ 512	100
Penicillin G	56	30	12										≤ 1	2	≥ 4	≥ 512	≥ 512	100
Amoxicillin	98												≤ 8	16	≥ 32	≥ 512	≥ 512	100
Ceftiofur	45	6	8	20	6	4	5	4					≤ 2	4	≥ 8	256	≥ 512	95.92
Ceftriaxone	41	11	15	10	7	6	8						≤ 1	2	≥ 4	256	≥ 512	100
Cefepime	27	18	14	4	4	3	5	9	6	6	2		≤ 8	16	≥ 32	128	≥ 512	68.37
Ceftazidime	10	7		3		15	14	12	15	16	5	1	≤ 8	16	≥ 32	4	256	20.41
Cefoperazone	43	4		8	6	21	10	6					≤ 16	32	≥ 64	64	≥ 512	56.12
Cefazolin	98												≤ 2	4	≥ 8	≥ 512	≥ 512	100
Cefaclor	98												≤ 8	16	≥ 32	≥ 512	≥ 512	100
Cephalonium	98												—	—	—	≥ 512	≥ 512	—
Spectinomycin	1	3	16	38	27	11	2						≤ 32	64	≥ 128	64	128	20.41
Gentamicin		1							5	35	37	20	≤ 4	8	≥ 16	0.5	1	1.02
Kanamycin	1		3	70	24								≤ 16	32	≥ 64	64	64	75.51
Amikacin							2	28	56	12			≤ 16	32	≥ 64	2	4	0
Florfenicol	17	16	21	16	16	9	2	1					≤ 4	8	≥ 16	128	≥ 512	71.43
Tetracycline		1		6	13	38	27	10	3				≤ 4	8	≥ 16	16	32	59.18
Doxycycline		1			1	34	36	18	4	3	1		≤ 4	8	≥ 16	8	16	36.73
Sulfonamide	98												≤ 256	—	≥ 512	≥ 512	≥ 512	100
Trimethoprim				3	65	13	10	7					≤ 8	—	≥ 16	32	32	82.65
Lincomycin	98												—	—	—	≥ 512	≥ 512	—
Tylosin	70	21	7										—	—	—	≥ 512	≥ 512	—
Tilmicosin	10	30	45	13									≤ 8	16	≥ 32	128	256	100
Tildipirosin	60	38											≤ 16	32	≥ 64	≥ 512	≥ 512	100
Polymyxin B		1					2	17	42	34	2		≤ 2	4	≥ 8	2	4	3.06

R — resistance.

Typing by PFGE differentiated the 98 P. aeruginosa isolates into 56 distinct types, indicating wide genetic diversity. The discriminatory ability of PFGE was calculated with the Simpson’s index of diversity and resulted in a value of 0.983. Some isolates had 100% similarity, with exact PFGE bands: examples are PAM22 and 26; PAM11 and 12, 90, and 91; and PAM95, 85 to 87, 07, and 09. Moreover, these 56 PFGE patterns included the 5 serotypes, as shown in Table IV.

Table IV.

Correlation analysis of serotype and other variables.

Serotype	Total number of isolates	Biofilm formation		Slime production		Number of PFGE patterns
		Number of biofilm formers	Number not producing biofilms	Number of slime producers	Number not producing slime
G	58	52	6	46	12	40
I	15	15	0	14	1	13
C	8	8	0	8	0	7
M	5	4	1	5	0	4
B	2	2	0	1	1	2
NT	10	9	1	9	1	8
Total	98	90	8	83	15	56

PFGE — pulsed-field gel electrophoresis.

Analyzing the data and comparing serotype with biofilm formation, slime production, motility, antibiotic susceptibility, and PFGE pattern, we found that none of these variables were significantly correlated (Table IV). More than 80% of the 98 isolates, however, were resistant to more than 10 antibiotics, and 92% of the isolates produced biofilms. Therefore, there were no obvious differences between biofilm producers and isolates unable to form biofilms in terms of antibiotic resistance. However, resistance to 8 antibiotics was more prevalent among the biofilm producers than among the nonproducers (Table V).

Table V.

Correlation analysis of antibiotic resistance and biofilm formation.

Antibiotic	Number (%) of isolates	Forming biofilms (n = 90)	Not forming biofilms (n = 8)
	Resistant (n = 98)
Nalidixic acid	92 (94)	90 (100)	8 (100)
Lomefloxacin	34 (35)	27 (30)	7 (88)
Ciprofloxacin	13 (13)	10 (11)	3 (38)
Norfloxacin	11 (11)	9 (10)	2 (25)
Levofloxacin	14 (14)	10 (11)	4 (50)
Enrofloxacin	50 (51)	43 (48)	7 (86)
Oxacillin	98 (100)	90 (100)	8 (100)
Ampicillin	98 (100)	90 (100)	8 (100)
Penicillin G	98 (100)	90 (100)	8 (100)
Amoxicillin	98 (100)	90 (100)	8 (100)
Ceftiofur	94 (96)	89 (99)	5 (62)
Ceftriaxone	98 (100)	90 (100)	8 (100)
Cefepime	67 (68)	60 (67)	7 (88)
Ceftazidime	20 (20)	17 (19)	3 (37)
Cefoperazone	55 (56)	49 (54)	6 (75)
Cefazolin	98 (100)	90 (100)	8 (100)
Cefaclor	98 (100)	90 (100)	8 (100)
Cephalonium	—	—	—
Spectinomycin	20 (20)	14 (16)	6 (75)
Gentamicin	1 (1)	1 (1)	0 (0)
Kanamycin	74 (76)	70 (78)	4 (50)
Amikacin	0 (0)	0 (0)	0 (0)
Florfenicol	70 (71)	65 (72)	5 (62)
Tetracycline	58 (59)	56 (62)	2 (25)
Doxycycline	36 (37)	34 (38)	2 (25)
Sulfonamide	98 (100)	90 (100)	8 (100)
Trimethoprim	81 (83)	77 (86)	4 (50)
Lincomycin	—	—	—
Tylosin	—	—	—
Tilmicosin	98 (100)	90 (100)	8 (100)
Tildipirosin	98 (100)	90 (100)	8 (100)
Polymyxin B	3 (3)	3 (3)	0 (0)

Discussion

The P. aeruginosa isolates in this study were predominantly of the G serotype, consistent with previous studies in China, Europe, America, and Denmark and with P. aeruginosa strains isolated from pets, farm animals, and zoo animals (1,3,5,21,22). Moreover, the O6 (G) serotype was the serotype most frequently isolated from humans (23). However, 10 isolates were untypeable, which suggests that there may be at least 1 new serotype. Serotype G may play a role in mink hemorrhagic pneumonia. Although humans and mink can be infected with the same serotypes, cross-infections occur rarely. Mink hemorrhagic pneumonia caused by P. aeruginosa correlates with the season, feeding environment, and secondary or mixed infections that may be opportunistic (1).

Generally, bacterial biofilms protect the resident organisms from antimicrobial agents and host immune responses. We found 90 isolates that could form biofilms and only 8 that could not. This was consistent with the finding that most isolates exhibited diversity in slime production, swimming, swarming, and twitching. This diversity has previously been reported for P. aeruginosa (24).

Monitoring of antimicrobial susceptibility is necessary for guiding antimicrobial therapy for bacterial diseases. We found most of the isolates to be resistant to 10 antibiotics, and this may be related to the increased use of antibiotics to treat mink bacterial infections (3,22,25). Some of the antibiotics we tested are frequently used in veterinary clinics. In our group of 98 isolates, resistance, especially to the β-lactams, was very serious. Multidrug resistance and cross-resistance were also common, and all our isolates were resistant to oxacillin, ampicillin, penicillin G, amoxicillin, ceftriaxone, cefazolin, cefaclor, sulfonamide, tilmicosin, and tildipirosin. However, 95% of the isolates were still susceptible to gentamicin and polymyxin B, and all were sensitive to amikacin. This is consistent with the lack of use of these antibiotics in mink farms.

Although there are few reports of drug resistance in P. aeruginosa isolates from mink, the global trend is of increasing antimicrobial resistance, and this is especially true for β-lactam antibiotics (3,25). Therefore, an active surveillance program should be established for mink farms to prevent multidrug resistance and reduce the rate of death due to ineffective antibiotic treatment of P. aeruginosa infections.

Genotyping of P. aeruginosa isolates by PFGE has been used to study the epidemiologic aspects of P. aeruginosa infections in mink in China and Denmark (1,3,5). The 56 different genotypes we found showed a high level of genetic diversity, including diversity within the serotype groups. Our results demonstrated that genotyping by PFGE [diversity (D) = 0.983] was more discriminatory than serotyping (D = 0.612) and was consistent with previous observations (1,26). Several isolates showed identical PFGE patterns, and these also had similar serotypes (PAM11 and 12, PAM90 and 95, and PAM85 and 86). These findings suggest a clonal spread to different regions. Other studies (1,3) have revealed that the food chain may play a role in the horizontal transmission of P. aeruginosa causing hemorrhagic pneumonia in mink. The field environment for mink, including their feed (frozen chicken meat and animal viscera), handling equipment, water, and diseased animals, may lead to the occurrence of hemorrhagic pneumonia (1,3). The mechanism responsible for the spread of P. aeruginosa in mink populations needs further study.

In conclusion, new epidemiologic information about P. aeruginosa in mink infections was revealed in this study. A survey of serotype frequency, biofilm formation, slime production, and motility activities (swimming, swarming, and twitching) indicated multidrug resistance, high resistance rates for critical antimicrobials, and major genetic differences among isolates. The findings can assist veterinarians and mink farmers to understand the epidemiologic characteristics and drug resistance of P. aeruginosa. They also provide a theoretical basis for the prevention and treatment of hemorrhagic pneumonia caused by P. aeruginosa in mink.