Genotypic and phenotypic profiles of virulence factors and antimicrobial resistance of Proteus mirabilis isolated from chicken carcasses: potential zoonotic risk

Abstract

Proteus mirabilis is an opportunistic pathogen often associated with a variety of human infections acquired both in the community and in hospitals. In this context, the present work aimed to evaluate the genotypic and phenotypic characteristics of the virulence factors and antimicrobial resistance determinants of 32 P. mirabilis strains isolated from chicken carcasses in a poultry slaughterhouse in the north of the state of Paraná, Brazil, in order to assess a potential zoonotic risk. The isolates presented a variety of virulence genes that contribute to the development of infection in humans. The mrpA, pmfA, atfA (fimbriae), ireA (siderophores receptor), zapA, ptA (Proteases), and hpmA (hemolysin) genes were found in 32 (100%) isolates and ucaA (fimbriae) in 16 (50%). All isolates showed aggregative adherence in HEp-2 cells and formed biofilms. Of all strains, 27 (84.38%) showed cytotoxic effects in Vero cells. Antimicrobial susceptibility was tested using 20 antimicrobials, in which 25 (78.13%) strains were considered multidrug-resistant. The presence of bla_ESBL and bla_ampC genes conferring resistance to β-lactams and qnr to quinolones were also detected in the isolates after presumption in the phenotypic test, in which 7 (21.88%) isolates contained the CTX-M-2 group, 11 (34.38%) contained CIT group and 19 (59.38%) contained qnrD. Therefore, chicken carcasses contaminated with P. mirabilis may pose a health risk to the consumer, as these isolates have a variety of virulence and antimicrobial resistance characteristics that can be found in P. mirabilis strains isolated from human infections.

Introduction

The Brazilian poultry industry has been outstanding in both the national and international markets, with Brazil being considered the second largest producer and leader in the world export of chicken meat. The southern region of Brazil is responsible for the majority of the production of broilers, with Paraná state being the largest national producer and accounting for more than 35% of the exports of all chicken meat produced in the country [1].

In view of the commercialization and relevance of the product throughout the world, it is essential to ensure the microbiological quality of poultry meat. In fact, all processing steps during and after the chicken slaughter can contaminate the meat with bacteria present in the avian microbiota and in the slaughterhouse, especially in equipment that helps in the handling of carcasses, cuts, and by-products [2].

The fact that Proteus spp. is isolated from chicken feces along with other enteric bacteria, such as Escherichia coli, indicates its function as a component of the normal intestinal microbiota [3]. This facilitates the transfer of these bacteria to the slaughter line and cross-contamination, especially in the carcass evisceration process [4].

Some of these bacterial contaminants may remain after carcass processing and survive storage. It is therefore essential for the consumer to be aware that chicken meat can be a source of transmission of microorganisms. Therefore, in order to avoid possible cross-contamination, it is essential for taking care of the disinfection of the handling tools after the preparation of the meat [4].

The isolation of P. mirabilis from food products is documented in some studies, especially in those of animal origin, such as chicken meat, and some of these isolates are considered multidrug-resistant (MDR) [5, 6]. Bacterial resistance is a global public health threat and the excessive use of antimicrobials in poultry accelerates this phenomenon [7]. The problem is predicted to increase considerably in the coming years due to the intensification of the production of foods from animal origin and the non-rational use of antimicrobials [8].

P. mirabilis is a Gram-negative bacterium widely found in the environment and isolated from the intestinal tract of humans and other animals [3]. Although it is considered a commensal, the reports of occasional cases of infection and the description of nosocomial outbreaks point to its opportunistic pathogenic potential. Urinary tract infection (UTI) stands out as the most prevalent P. mirabilis infection, which has the capacity to cause other diseases in humans [9].

Several virulence factors may be associated with the pathogenicity of P. mirabilis in humans, such as fimbriae mannose-resistant Proteus-like (MR/P), Proteus mirabilis fimbriae (PMF), and uroepithelial cell adhesin (UCA), which allow bacterial-cell adhesion to host cells [9], PtA and ZapA proteases that degrade structural proteins and the immune system, respectively [10, 11], hemolysins HpmA and HlyA, considered toxins for host cells and involved in colonization of the urinary tract [12, 13], siderophores receptor IreA, considered of great importance in the acquisition of Fe⁺³ of the host [14]. In addition, adhesion capacity, formation of biofilms, and cytotoxicity are also important in the infectious process [9, 15].

Additionally, P. mirabilis has been associated with foodborne disease, despite their primary pathogenic role has not been confirmed [16–18], and may be a public health threat to society due to its strong association with a variety of human infectious diseases [19]. Therefore, more research is indicated to discuss and elucidate the zoonotic potential of P. mirabilis.

The present study aimed to characterize P. mirabilis strains isolated from chicken carcasses in the slaughter line, as well as the genotypic and phenotypic characteristics of virulence factors and antimicrobial resistance. To our knowledge, this is the first study from Brazil that explores the general virulence and antimicrobial resistance characteristics of P. mirabilis strains isolated from carcasses of broiler chickens, in order to explore their zoonotic potential.

Materials and methods

Food samples

A total of 42 chicken carcasses were evaluated through sterile swabs scrubbed on the surface of the thighs and abdomen of non-eviscerated chickens in two time periods separated by a 15-day interval. All carcasses were randomly selected in the slaughter line during a sanitary inspection of a poultry slaughterhouse in the north of Paraná, Brazil. The isolates were collected using sterile Swabs and the collected material was transported in Cary-Blair medium (Difco™, USA).

Bacterial isolation and identification

The collected material was transferred to Brain Heart Infusion broth (Difco™, USA) and cultured at 37 °C for 24 h. After bacterial growth, the samples were plated on MacConkey Agar plates (Difco™, USA), and incubated at 37 °C for 24 h. Suspected colonies were confirmed to be P. mirabilis with the biochemical tests EPM, MILi (PROBAC™, BR), and Simmons citrate (Difco™, USA).

Detection of virulence genes

We investigated nine genes associated with the virulence of P. mirabilis in humans. The selected genes included ireA (siderophore receptor), mrpA, ucaA, pmfA, atfA (fimbriae), ptA and zapA (proteases), and hmpA and hlyA (hemolysins), as shown in Table 1. The bacterial DNA was obtained by the boiling extraction method. The polymerase chain reaction (PCR) was performed on a GeneAmp PCR System 9700 thermal cycler (Applied Biosystems™) containing the final volume of 25 μL, composed of 2 mM MgCl₂ (Invitrogen™), 10× buffer (Invitrogen™), 0.2 mM dNTPs (Invitrogen™), 20 pmol Primer Forward, 20 pmol Primer Reverse (Invitrogen™), 1.25 U Taq DNA polymerase (Invitrogen®), bacterial DNA, and Milli-Q ultrapure water (Millipore™). PCR products were size separated by gel electrophoresis on a 2% agarose gel stained with SYBR SAFE (Invitrogen™) and emerged in TBE buffer (89 mM Tris base, 89 mM Boric Acid, 2 mM EDTA, pH 8.3). The molecular marker used was a 1 Kb Ladder (Invitrogen™). The amplicons were observed in a transilluminator with ultraviolet light (Vilbert Loumart™). The primers designed in this study were based on the complete genome sequence of the P. mirabilis strain HI4320 [22] and were used in the PCR following the conditions: 95 °C/5 min, 30 cycles of 95 °C/1 min, hybridization temperature/1 min, 72 °C/1 min, and final extension at 72 °C/7 min. The strain P. mirabilis HI4320 was used as positive control [22] and the reaction without DNA as a negative control.

Table 1.

Genes associated with virulence of P. mirabilis screened by PCR

Genes	Primers sequences (5′➔3′)	(bp)	Hybridization temperatures	References
ireA	(F) ACTACGATAACGAGCGCCAG (R) GCCCTAACTGGGGGAATACG	681	60 °C	This study
mrpA	(F) GAGCCATTCAATTAGGAATAATCCA (R) AGCTCTGTACTTCCTTGTACAGA	648	58 °C	[20]
ucaA	(F) GCTTTTACATCCCCAGCGGT (R) GCTGCATTTGCTGGCTCATC	476	60 °C	This study
pmfA	(F) CAAATTAATCTAGAACCACTC (R) ATTATAGAGGATCCCTTGAAGGTA	617	54 °C	This study
atfA	(F) CATAATTTCTAGACCTGCCCTAGCA (R) CTGCTTGGATCCGTAATTTTTAACG	382	50 °C	[21]
ptA	(F) CCACTGCGATTATCCGCTCT (R) ATCGGCAGAAGTGACAAGCA	686	60 °C	This study
zapA	(F) TATCGTCTCCTTCGCCTCCA (R) TGGCGCAAATACGACTACCA	332	59 °C	This study
hpmA	(F) GTTGAGGGGCGTTATCAAGAGTC (R) GATACTGTTTTGCCCTTTTGTGC	709	55 °C	[12]
hlyA	(F) AACAAGGATAAGCACTGTTCTGGCT (R) ACCATATAAGCGGTCATTCCCGTCA	1177	63 °C	[12]

bp base pairs

Adherence in HEp-2 cells

The adherence capacity of the isolates of P. mirabilis to human larynx carcinoma (HEp-2) cells was analyzed as previously described [23]. Bacterial-cell interactions were assayed for 6 h. HEp-2 cells were grown in 24-well plates with 13 mm coverslips containing 1 mL DMEM (Dulbecco’s Modified Eagle Medium). After the formation of a HEp-2 cell monolayer, the medium was discarded and the plates were washed three times with phosphate-buffered saline (PBS), containing 9.7 mM Na₂HPO₄, 1.25 mM KH₂PO₄, 137.93 mM NaCl, and 2.68 mM KCl. Next, 1 mL DMEM without antibiotics supplemented with 2% fetal bovine serum (FBS) was added to each well. Subsequently, a 40 μL aliquot of the bacterial culture overnight in tryptic soy broth (TSB) was added to each well and the plates were incubated for 3 h at 37 °C. After this step, the wells were washed five times with 1 mL sterile PBS to remove non-adhering bacteria. Once more, 1 mL of DMEM with 2% SFB was added to the wells and incubated for an additional 3 h at 37 °C. At the end of the 6 h period, the wells were washed five times with PBS, then the coverslips were fixed with 100% methanol and stained with May-Grünwald for 5 min and Giemsa for 20 min. The coverslips were observed under a light microscope at 1000x magnification. Cells incubated without bacteria were used as negative control and E. coli E2348/69 [24], 042 [25] and C1845 [26] as positive controls for the patterns of localized, aggregative and diffuse adhesion, respectively.

Cytotoxicity in Vero cell cultures

All 32 bacterial isolates were tested in Vero cells (African green monkey kidney) to evaluate cytotoxic capacity as demonstrated [27], with modifications. The isolates were cultured in 3 mL TSB broth at 37 °C for 18 h under shaking at 180 rpm. After bacterial growth, the samples were centrifuged at 13000 g for 10 min. The supernatants were collected and filtered through a syringe filter with a PVDF membrane (Durapore™) with 0.22-μm pore size and 47-mm diameter. Filtered supernatants were added in four repetitions each to a 96-well polystyrene plate at a dilution of 1:10 and subsequently incubated for 72 h at 37 °C and 5% CO₂.

The cytotoxicity of the isolates was quantified measuring the metabolic activity of Vero cells with the MTT (3-[4,5-dimethyl-thiazol-2-yl]-2,5-diphenyl-tetrazolium bromide) assay [28]. Cells without bacterial supernatant were used as a negative control and E. coli O157:H7 (EDL933) was used as a positive control for cytotoxicity in Vero cells [29]. The isolate was considered highly cytotoxic when 50% or more of cell death were measured in comparison with the negative control.

Biofilm formation

The biofilm assay was performed on 96-well polystyrene plates as previously described [30] using crystal violet. E. coli 042 (O44:H18) [25] was used as positive control for biofilm formation and TSB Broth (Difco™) only was used as negative control. The absorbance readings (A) were performed in a spectrophotometer at a wavelength of 570 nm. The threshold value of the absorbance (T) confirmed the biofilm formation and was defined as the sum of the arithmetic mean of the negative control (nc) and three times the standard deviation (δ), according to the formula (T = xnc + 3δ).

Susceptibility to antimicrobials

The resistance of all 32 isolates was evaluated with the disc diffusion method in Mueller-Hinton agar, using the antimicrobials as recommended by the Clinical and Laboratory Standards Institute [31]. The antimicrobials (Oxoid™) used were Ampicillin (AMP) 10 μg, amoxicillin + clavulanate (AMC) 20/10 μg, cephalotin (CFL) 30 μg, cefoxitin (CFO) 30 μg, ceftazidime (CAZ) 30 μg, Ceftriaxone (CRO) 30 μg, cefotaxime (CTX) 30 μg, cefepime (CPM) 30 μg, nalidixic Acid (NAL) 30 μg, norfloxacin (NOR) 10 μg, ciprofloxacin (CIP) 5 μg, trimethoprim-sulfamethoxazole (SUT) 1.25/23.75 μg, aztreonam (ATM) 30 μg, chloramphenicol (CLO) 30 μg, gentamicin (GEN) 10 μg, amikacin (AMI) 30 μg, ertapenem (ETP) 10 μg, and imipenem (IPM) 10 μg. Additionally, ceftiofur (CTF) 30 μg, enrofloxacin (ENO) 5 μg were used [32]. The isolate was considered multidrug-resistant when it exhibited resistance to three or more classes of antimicrobials [33]. E. coli ATCC® 25922™ was used for quality control purposes.

Detection of resistance genes

The detection of resistance genes was performed by PCR assay. All isolates that exhibited resistance to cephalosporins in the phenotypic assay were screened for the presence of ESBL-encoding groups of the type CTX-M 1, 2, 8, 9, and 25 [34], TEM and SHV [35], as well as the six plasmid-mediated AmpC-specific families, MOX, FOX, EBC, ACC, DHA, and CIT [36]. The quinolone resistance genes qnrA, qnrB, qnrS [37], and qnrD [38] were also investigated in isolates that exhibited resistance to quinolones in the phenotypic assay.

Results

Genotypic and phenotypic profile of virulence

From a total of 42 chicken carcasses evaluated, 32 were positive for the presence of P. mirabilis. Only one strain was isolated from each positive carcass, totalizing 32 strains. The bacterial isolates studied presented a variety of virulence genes that enable and contribute to the development of infection in humans. The mrpA, pmfA, atfA (fimbriae), zapA, ptA (proteases), hpmA (hemolysin), and ireA (siderophore receptor) genes were detected in all isolates, evidencing a high prevalence of these genes in P. mirabilis isolated from chicken carcasses. The gene ucaA (fimbriae) was the least prevalent among fimbrial genes, being found in 16 (50%) isolates. The hlyA (hemolysin) gene was not detected in any isolate.

The phenotypic characterization of the virulence factors investigated in our study showed that all 32 (100%) isolates expressed a pattern of aggregative adhesion to HEp-2 cells cultures (Fig. 1) and showed a capacity for biofilm formation in polystyrene plates, in which 17 isolates (53.12%) formed a strong biofilm and 15 (46.88%) a very strong biofilm (Fig. 2). In addition to these phenotypic characteristics of virulence, a highly cytotoxic effect on Vero cell culture was measured for a total number of 27 (84.38%) isolates. The genotypic and phenotypic profile of the virulence factors of these strains is presented in Table 2.

Fig. 1.

Aggregative adhesion pattern exhibited by P. mirabilis strain isolated from chicken carcass (magnification × 1000)

Fig. 2.

Intensity of P. mirabilis biofilm isolated from chicken carcasses

Table 2.

Genotypic and phenotypic characteristics of virulence and antimicrobial resistance of P. mirabilis isolated from chicken carcasses

Strains	Virulence genes	A.A	C.V.C	Biofilm	Phenotypic profile of resistance	Resistance genes
LB UEL – A01	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CTF, AMC, CFL, CFO, CAZ, CTX, NAL, ENO, SUT	CIT, qnrD
LB UEL – A02	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	−	+	AMP, CTF, AMC, CFL, CFO, CAZ, CRO, CTX, CPM, NAL, ENO, SUT	CIT, qnrD
LB UEL – A03	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	−	+	AMP, AMC, CFL, CFO, CAZ, CTX, NAL, SUT, GEN	CIT
LB UEL – A04	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CTF, AMC, CFL, CFO, CAZ, CRO, CTX, CPM, NAL, NOR, ENO, CIP, SUT, GEN	CTX-M-2 group, qnrD
LB UEL – A05	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CTF, AMC, CFL, CAZ, CRO, CTX, NAL, ENO, CIP, SUT, ATM, GEN	qnrD
LB UEL – A06	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, NAL, NOR, ENO, CIP, SUT, GEN	qnrD
LB UEL – A07	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, AMC, CFL, NAL, NOR, ENO, CIP, SUT, ATM, GEN	qnrD
LB UEL – A08	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CTF, AMC, CFL, CFO, CAZ, CRO, CTX	CIT
LB UEL – A09	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CTF, AMC, CFL, CFO, CRO, CTX, CPM, NAL, ENO, SUT, ATM, CLO	CTX-M-2 group, CIT, qnrD
LB UEL – A10	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	−	+	AMP, CTF, AMC, CFL, CFO, CRO, CTX, CPM, NAL, NOR, ENO, CIP, SUT	CTX-M-2 group, qnrD
LB UEL – A11	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CTF, CFL, CFO, CRO, CTX, CPM, NAL, NOR, ENO, CIP, SUT	CTX-M-2 group, qnrD
LB UEL – A12	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	CFL, NAL, SUT	Not found
LB UEL – A13	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	−	+	AMP, AMC, CTF, CFL, CFO, CAZ, CRO, CTX, NAL, ENO, SUT, GEN	CIT, qnrD
LB UEL – A14	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	CFL	Not found
LB UEL – A15	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, AMC, CTF, CFL, CFO, CAZ, CRO, CTX, NAL, ENO, SUT, CLO, GEN	CIT, qnrD
LB UEL – A16	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, AMC, CFL, CFO, CAZ, CRO, CTX, NAL, NOR, ENO, CIP, SUT, GEN, AMI	CIT, qnrD
LB UEL – A17	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMC, CFL, CFO, CAZ, CTX, CPM, NAL, ATM	CTX-M-2 group
LB UEL – A18	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA,	+	+	+	AMP, CTF, AMC, CFL, CRO, CTX, CPM, NAL, ENO, CIP, SUT, GEN	CTX-M-2 group, qnrD
LB UEL – A19	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	Not found	Not found
LB UEL – A20	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, NAL, NOR, ENO, CIP, SUT, GEN	qnrD
LB UEL – A21	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CFL, NAL, NOR, ENO, CIP, SUT, CLO, GEN	qnrD
LB UEL – A22	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CTF, AMC, CFL, CFO, CAZ, CRO, CTX, CPM, NAL, NOR, ENO, CIP, SUT, ATM	CTX-M-2 group, CIT, qnrD
LB UEL – A23	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	−	+	NAL, ENO	qnrD
LB UEL – A24	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CFL, NAL, NOR, ENO, CIP, SUT, GEN	qnrD
LB UEL – A25	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CTF, AMC, CFL, CFO, CAZ, CRO, CTX, NAL, ENO, SUT, CLO, GEN	CIT, qnrD
LB UEL – A26	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CFL, CFO, NAL, NOR, ENO, CIP, SUT, GEN	qnrD
LB UEL – A27	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, AMC, CFL, CFO, CAZ, CRO, CTX, NAL, ENO, SUT, GEN	CIT, qnrD
LB UEL – A28	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	CTF, AMC, CFL, CFO, NAL, SUT, GEN, AMI	Not found
LB UEL – A29	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	NAL, ENO, CIP	qnrD
LB UEL – A30	mrpA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, CTF, AMC, CFL, CFO, CRO, CTX, CPM, NAL, GEN	CTX-M-2 group
LB UEL – A31	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	CFL, CFO, NAL, ENO	qnrD
LB UEL – A32	mrpA, ucaA, pmfA, atfA, ireA, ptA, zapA, hpmA	+	+	+	AMP, AMC, CFL, CFO, NAL, SUT, GEN	Not found

A.A, aggregative adhesion; C.V.C, cytotoxicity in vero cells; AMP, ampicillin; CTF, ceftiofur; AMC, amoxicillin-clavulanic acid; CFL, cephalotin; CFO, cefoxitin; CAZ, ceftazidime; CRO, ceftriaxone; CTX, cefotaxime; CPM, cefepime; NAL, nalidixic acid; NOR, norfloxacin; ENO, enrofloxacin; CIP, ciprofloxacin; SUT, trimethoprim-sulfamethoxazole; ATM, aztreonam; CLO, chloramphenicol; GEN, gentamicin; AMI, amikacin. + positive; − negative

Genotypic and phenotypic profile of resistance to antimicrobials

Twenty antimicrobials belonging to several classes were tested. Resistance profiles of isolates ranged of strains susceptible to all antimicrobials tested to strains resistant to 15 different antimicrobials. Of all the antimicrobials tested, the isolates presented a higher frequency of resistance (> 50%) to ampicillin, amoxicillin-clavulanic acid, cefotaxime, nalidixic acid, enrofloxacin, trimethoprim-sulfamethoxazole, and gentamicin. The isolates showed greater susceptibility to the antimicrobials aztreonam, chloramphenicol, amikacin, ertapenem, and imipenem, highlighting the last two in which none of the isolates presented resistance. The frequency of antimicrobial resistance of the isolates is shown in Fig. 3. Of all the strains, 25 (78.13%) were MDR, being resistant to three or more classes of antimicrobials.

Fig. 3.

Frequency of antimicrobial resistance of P. mirabilis isolated from chicken carcasses

In relation to the bla_ESBL and AmpC genes responsible for conferring resistance to β-lactam antimicrobials, eight (25%) isolated had the CTX-M-2 group, while 11 (34.38%) had only the CIT group. The plasmidial gene qnrD that confers resistance to quinolones was found in 23 (71.88%) isolates, demonstrating the high prevalence of this gene in P. mirabilis isolated from chicken carcasses. The CTX-M-2 and CIT groups were found together in two isolates, CIT and qnrD in nine, qnrD and CTX-M-2 in four isolates. The CTX-M-2, CIT, and qnrD groups were found together in two strains. (Table 2).

Discussion

The bacterial isolates studied presented a variety of virulence genes that enable and contribute to the development of infection. Our results evidenced a variety of genes responsible for encoding fimbriae (Table 2) that contribute to the UTI in humans, such as fimbriae MR/P, PMF, and UCA [39, 40]. A study conducted by Barbour et al. [41] detected the mrpA gene in all strains of P. mirabilis isolated from chicken and human and identified a high nucleotide similarity between them, suggesting a possible zoonotic potential of these strains. The ambient temperature fimbriae (ATF) do not contribute to UTI, but its expression may allow the permanence of P. mirabilis in the environment [21].

Besides fimbrial genes, a high prevalence of genes zapA, ptA (proteases), hpmA (hemolysin), and ireA (siderophores receptor) was also found in our isolates (Table 2), evidencing that chicken carcasses may contain P. mirabilis with a diversity of virulence genes. In relation to the genes coding hemolysins, all isolates had hpmA, whereas hlyA was not detected in any of these. These findings are in agreement with other studies, which report a higher prevalence of hpmA than hlyA in P. mirabilis [12, 42]. Few studies involving research on virulence factors in P. mirabilis strains of food and infections are reported. This is the first study that shows evidence for the presence of the virulence genes pmfA, ucaA, atfA, hpmA, zapA, ptA, and ireA in P. mirabilis isolated from chicken carcasses.

We have observed that most of the virulence factors investigated are related to an extraintestinal pathogenic profile, such as the extraintestinal pathogenic Escherichia coli (ExPEC). This clue may be supported by the fact that our isolates possessed a diversity of virulence factors that are commonly found in ExPEC, such as fimbriae, hemolysins, proteases, and the iron uptake system [43–45]. Other studies also report a high prevalence of virulence factors in E. coli isolated from chicken carcasses and highlight the pathogenic potential of these strains in developing extraintestinal infections in humans [46, 47]. This hypothesis is strongly supported by other works [48–50], as well as a study conducted by Vincent et al. [51], which reports an intimate similarity between strains of E. coli isolated from chicken meat and UTIs.

We believe that chicken meat can be contaminated with P. mirabilis in the slaughter line during the carcass evisceration process or by cross contamination. It is also believed that these strains can be transmitted to consumers, colonize the human intestinal tract, and possibly develop extraintestinal infections, such as UTIs, similarly to what it is believed to occur with E. coli from food sources [48–51].

The colonization of the epithelium is important for P. mirabilis uropathogenicity. This ability can be demonstrated by experiments using different cell lines [52]. The aggregative adherence pattern has been previously reported in uropathogenic P. mirabilis [20]. It is known that the expression of the MR\P fimbriae contributes to this adhesion pattern [20]. This evidence supports our results since all the isolates had the mrpA gene and expressed the aggregative adherence pattern. This is the first report of P. mirabilis isolated from chicken carcasses expressing the aggregative adherence pattern in HEp-2 cells. This pattern of adhesion has also been demonstrated in E. coli isolated from extraintestinal infections [43, 45].

The cytotoxic effect is also considered important in the infectious process, especially in uropathogenecity, and was exhibited by most of our isolates (84.38%). There is evidence that the hemolysin HpmA is responsible for most of the cytotoxicity in renal cell lines since the isogenic mutant strains for this virulence factor were significantly less cytotoxic than the wild type strains [53]. Despite all of our isolates have possessed the hpmA gene, some of them have not expressed cytotoxic effect in Vero cells which may be due to the non-expression of this gene in vitro. However, their expression may occur in vivo and contribute to the infectious process.

Foodborne biofilm-forming bacteria pose serious risks to human health and have become increasingly frequent in the food industry, opportunizing the dissemination of biofilm-forming strains, mainly through cross-contamination [54, 55]. Our results are alarming, considering that all the strains isolated from chicken carcasses in a slaughterhouse in the north of Paraná formed a strong or very strong biofilm. In the host, biofilm formation protects bacteria from the immune system and antimicrobials, as well as contribute to the persistence of infection [56]. Biofilms in P. mirabilis are well studied, they are mainly associated with urinary catheters, which can block urinary flow and cause ascending infection, pyelonephritis, and possible sepsis [57]. It is important to note that some of our isolates, such as the strain LB UEL - A07, expressed a biofilm at the intensity similar to the positive control EAEC 042, a diarrheagenic E. coli very well known to form a very strong biofilm [25] (Fig. 2).

The isolation of bacteria that pose a threat to the health of the consumer becomes more aggravating when these, besides having factors of virulence, also present resistance to the antimicrobials, mainly when these are used for the treatment of infections in humans [58]. Our results in Fig. 3 show, that even though many antimicrobials are prohibited on poultry farms, a high frequency of P. mirabilis MDR was found, in which 25 (78.13%) isolates were resistant to three or more classes of antimicrobials [33].

The presence of MDR P. mirabilis and producers of ESBLs isolated from chicken meat has been reported in a few studies [6, 59]. The high prevalence of CTX-M-2 and CIT groups has been reported in E. coli isolated from chicken carcasses in Paraná, Brazil [46]. In addition, a study in the UK reported the detection of the CTX-M-2 group in E. coli isolated from chicken meat imported from Brazil [60]. Our study evidenced a high prevalence of the qnrD gene in P. mirabilis isolated from chicken carcasses, which has been commonly found in genera of the Proteeae tribe, such as Proteus, Providencia, and Morganella [61].

It is important to highlight the considerable rate of resistance to quinolones and cephalosporins of the third and fourth generation found in this study (Fig. 3), since these antimicrobials are frequently used in the treatment of human infections [62, 63]. Similar results were also found in a study in China [6]. Currently, a variety of human infectious diseases are currently treated with quinolones, including abdominal infections, chronic bronchitis, community-acquired pneumonia, nosocomial infections, sexually transmitted infections, skin infections, and urinary tract infections [62]. This is the first report of groups CTX-M-2, CIT, and qnrD in P. mirabilis isolated from chicken carcasses in Brazil.

Therefore, chicken meat can be a source of transmission and dissemination of P. mirabilis strains resistant to antimicrobials and carriers plasmid-mediated resistance genes to the community. Thus, it is essential that chicken producers are cautious about the prophylactic use of antimicrobial agents in poultry production, mainly as a food additive [8]. The study of resistant bacteria and plasmid-mediated resistance is important mainly because infections caused by resistant strains are more related to high morbidity and mortality rates than susceptible strains [64].

In the present study, it was demonstrated that chicken carcasses of a slaughterhouse in the north of Paraná are a potential risk for consumer health. As shown in Table 2, they carried P. mirabilis strains with a great diversity of virulence factors and antimicrobial resistance determinants, both phenotypically and genotypically. These findings become more aggravating when we consider the possibility of these strains colonizing the intestinal tract of the consumers and horizontally transfer the plasmid genes associated with antimicrobial resistance to the commensal strains. Our results are alarming, since the state of Paraná is responsible for most of the chicken meat exports from Brazil [1].

Although few reports, P. mirabilis has been associated with an outbreak of gastroenteritis and vomiting in a study that found the identical P. mirabilis strain in a diarrheic stool sample and vomit of a patient affected with food poisoning [17]. These associations were strongly sustained during an outbreak caused by the consumption of pork cooked in China, in which an identical strain was isolated from the hospitalized individuals, kitchen handler, waiter, and contaminated food waste, proving the clonal relation of the strains by Pulsed-field gel electrophoresis (PFGE) [16]. However, more studies need to be done to elucidate whether P. mirabilis can be considered the primary pathogen of food poisoning and which virulence factors contribute to it.

In addition, P. mirabilis has also been associated with pneumonia [65] and sepsis and infection of the central nervous system [66]. Therefore, the high frequency of P. mirabilis recovered from chicken carcasses in our study is alarming, considering that chicken meat may be responsible for disseminating P. mirabilis with virulence potential and MDR via the food chain.

Conclusions

In conclusion, P. mirabilis isolated from chicken carcasses have a variety of virulence factors and are resistant to multiple antimicrobials used in the treatment of human infections. Therefore, it is important that the consumer is aware of the risk when handling and preparing chicken meat, as well as performing a correct disinfection of the kitchen tools after the preparation of this meat, in order to avoid contact, cross-contamination, and possible food poisoning with this pathogen. In addition, we believe that the major risk is the colonization of the intestinal tract by these strains and further the development of extraintestinal diseases, such as UTIs.

Funding

This study was financed by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior Brasil (CAPES) - Finance Code 001.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflicts of interest.