Genetic Relatedness, Antibiotic Resistance, and Effect of Silver Nanoparticle on Biofilm Formation by Clostridium perfringens Isolated from Chickens, Pigeons, Camels, and Human Consumers

In this study, we determined the prevalence and toxin types of antibiotic-resistant Clostridium perfringens in chicken, pigeons, camels, and humans. We investigated the inhibitory effects of AgNPs on biofilm formation ability of the isolates and the genetic relatedness of the isolates from various sources determined using RAPD-PCR. Fifty isolates were identified using PCR, and all the isolates were of type A. The cpe and cpb2 genes were detected in 12% and 56% of the isolates, respectively. The effect of AgNPs on biofilm production of six representative isolates indicated that at the highest concentration of AgNPs (100 µg/mL), the inhibition percentages were 80.8–82.8%. The RAPD-PCR patterns of the 50 C. perfringens isolates from various sources revealed 33 profiles and four clusters, and the discriminatory power of RAPD-PCR was high. Multidrug-resistant C. perfringens isolates are predominant in the study area. The inhibition of biofilm formation by C. perfringens isolates was dose-dependent, and RAPD-PCR is a promising method for studying the genetic relatedness between the isolates from various sources. This is the first report of AgNPs’ anti-biofilm activity against C. perfringens from chickens, pigeons, camels, and humans, to the best of our knowledge.


Introduction
Clostridium perfringens is a Gram-positive anaerobic pathogen that inhabits the intestine of various animal species and humans. Necrotic enteritis (NE) is considered the main lesion caused by the organism in chickens and pigeons. The disease is fatal, and the main clinical manifestations include anorexia, depression, decreased growth performance, and reduced

Bacteriological Examination
The intestinal contents were directly obtained in sterile cooked meat broth (CMB; TM MEDIA, Titan Biotech Ltd., Delhi ISO 9,001, India). We aseptically homogenized 25 g chicken meat samples in 225 mL CMB. The human stool samples were directly enriched in CMB tubes. The tubes were anaerobically incubated at 37 • C for 24 h in an anaerobic jar containing gas generating kits (anaeroGen, OXOID Ltd., Hampshire, UK) for enrichment. For isolation of C. perfringens, a loopful from the enriched cultures were streaked onto the surface of reinforced clostridial agar (CM0151, OXOID Ltd., Hampshire, UK), and then anaerobically incubated at 37 • C for 24-48 h in an anaerobic jar containing gas generating kits. Presumptive shiny, pin-headed, and translucent C. perfringens colonies were biochemically identified using hemolysis tests, Gram stains, sugar fermentation, lecithinase tests, nitrate reduction, and motility tests.

Molecular Identification
We used PCR for confirmation of the biochemically suspected isolates. Following the manufacturer's instructions, the extraction of DNA was conducted for 50 isolates using the QIAamp DNA Mini kit (Qiagen GmbH, Hilden, Germany, Catalog no. 51304).

Antimicrobial Susceptibility Testing
The antibiotic susceptibility of the isolates from chicken, pigeons, camels, and human sources was determined with the broth micro-dilution method. The 14 antibiotics used were penicillin (PEN), ampicillin (AMP), amoxicillin (AMX), ampicillin-sulbactam (SAM), clindamycin (CLI), metronidazole (MTZ), vancomycin (VAN), imipenem (IPM), meropenem (MEM), chloramphenicol (CHL), tetracycline (TET), cefotaxime (CTX), cefoxitin (FOX), and ceftriaxone (CRO). The interpretation criteria for the antibiotics were based on EUCAST (2019), except for TET, CTX, FOX, and CRO, which were according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [23]. The minimum inhibitory concentration (MIC) was determined by double-fold dilution of the antimicrobials Vet. Sci. 2022, 9,109 4 of 18 (0.125-256 µg/mL) in Brucella broth (OXOID Ltd., Hampshire, UK), as recommended by CLSI guidelines (CLSI 2011). The dilution was conducted in sterile 96-well flat microplates (0.05 mL 2× antimicrobial/well) using fresh culture from overnight growth on blood agar. The culture was suspended in 5 mL sterile deionized water to achieve 0.5 McFarland turbidity, then 0.1 mL suspension was added to 11 mL Brucella broth. Next, 50 mL of thoroughly mixed suspension was transferred to individual wells of microplates. The MIC was read as the lowest antimicrobial concentration that inhibits visible bacterial growth after anaerobic incubation at 37 • C for 24 h. Growth control, broth sterility, and C. perfringens ATCC 19574 strain were included in each run to evaluate the method's reliability. The multiple antibiotic resistance (MAR) index was determined as the ratio of the number of antibiotics to which C. perfringens isolates exhibited resistance to the number of drugs for which the isolates were examined [24]. Multidrug resistance (MDR) is defined as the resistance of an isolate to at least one agent in three or more antibiotic classes, while extensively drug-resistant isolates (XDR) are defined as isolates resistant to at least one agent in all but two or fewer antimicrobial categories [25].

Biofilm Formation
The microtiter plate method was employed in evaluating the ability of C. perfringens isolates to form a biofilm [26,27]. Colonies under investigation were incubated overnight on blood agar under anaerobic conditions. Each isolate was adjusted to match McFarland obesity tube No. 0.5 (1.5 × 10 8 CFU/mL) in Brucella broth. Then, 20 mL were distributed in wells of microtiter plates with a flat bottom. Each sample was distributed in three wells containing 180 µL tryptic soy broth supplemented with 1% sterile glucose. The plates were then incubated at 44 • C for six days under anaerobic conditions. The medium was then discarded, and the wells were washed using 200 µL sterile phosphate-buffered saline (PBS) trice to remove non-adherent cells. After drying the plates for 45 min, each well was stained with 110 µL crystal violet solution (0.4%) for 45 min, followed by washing twice using 350 µL distilled water. De-staining was conducted by pipetting 200 µL 95% ethanol in each well for 45 min. Finally, 100 µL of the de-stained solution was transferred to the wells in new sterile microtiter plates, and the amount of crystal violet was measured using an ELISA reader (model: sunrise R4, serial no. 610000079) at OD 620 nm after adjustment to zero of the negative control. The experiment was conducted in triplicate, and the data were represented as mean ± standard deviation. The cut-off value (ODc) was calculated using the formula: ODc = average OD of negative control + (3 × standard deviation of negative control). The OD of each isolate was obtained by the formula: OD = average OD of the isolate-ODc. The data obtained were used to classify the strains as non, weak, moderate, and strong biofilm producers according to the following equations [28]: non-biofilm producer = OD ≤ ODc; weak biofilm producer = ODc < OD ≤ 2 × ODc; moderate biofilm producer = 2 × ODc < OD ≤ 4 × ODc; strong biofilm producer = 4 × ODc < OD [29].

Anti-Biofilm Activity of AgNPs-H 2 O 2
AgNPs-H 2 O 2 (Top Superpower-vision) was provided as a commercial product by El-Delta Center for Nanosilver Technology, Mansoura, Egypt. The stock solution of the product comprised 45 nm silver nanoparticles (0.00004467 mL/liter) with hydrogen peroxide (50% liter) and natural herbs, i.e., mint (1 mL/liter), at a concentration of 5 mL/liter of water. Then, the product was diluted using Mueller-Hinton broth.
Out of 14 strong biofilm-producing isolates, six XDR representative isolates from each source were chosen for the biofilm inhibition experiment. The anti-biofilm activity of the AgNPs was determined qualitatively using the tube method [30]. In brief, 50 µL overnight culture of the targeted bacteria in LB broth was further diluted to adjust its turbidity according to 0.5 McFarland Standards (5 × 10 5 CFU/mL). The suspension was added to the tubes containing 2 mL sterilized Brucella broth, and these tubes were incubated at 37 • C for 24 h under anaerobic conditions after adding different concentrations of AgNPs (25,50,75, and 100 µg/mL) in separate tubes. Negative control without the bacterial suspension Vet. Sci. 2022, 9, 109 5 of 18 and the positive control left without the addition of AgNPs were also included in the experiment. After incubation, the broth culture was decanted and washed twice with PBS. The inside of the tubes was stained with crystal violet dye (0.1%) for 30 min; the excess dye was decanted and gently washed off using deionized water. The tubes were dried, and biofilm formation ability was determined by observing a thin layer of blue film on the walls of the tubes. For quantitative estimation of biofilm formation, six representative isolates were chosen from strong biofilm producers; the microtiter plate assay was used [26,27]. In this method, 96-well microtiter plates were used. The wells were inoculated with 180 µL Brucella broth and 10 µL culture grown overnight and further diluted to adjust its final concentration to 5 × 10 5 CFU/mL and 10 µL AgNPs (concentrations used were 0-100 µg/mL) and incubated at 37 • C for 24 h under anaerobic conditions [30]. Negative and positive controls were also used in the assay using sterile growth medium only and working solution, respectively. The experiment was conducted in triplicate to evaluate its reproducibility, and the values were expressed as the average of the three independent experiments. The percentage of biofilm inhibition was calculated using the following formula according to Kalishwaralal, BarathManiKanth [30]: 1 − OD620 of cells treated with AgNPs OD620 of non-treated control × 100

Genotyping
The genetic relatedness between C. perfringens isolates from chickens, pigeons, camels, and human patients was evaluated using RAPD-PCR [31] as previously described. The fingerprinting data were transformed into a binary code depending on the presence or absence of each band. A dendrogram was generated by the unweighted pair group method with an arithmetic average (UPGMA) and Ward's hierarchical clustering routine. Cluster analysis and dendrogram construction were conducted using SPSS, version 22 (IBM, Armonk, NY, USA, 2013). The discriminatory powers of both methods were measured using Simpson's index of diversity (D), indicating the average probability that a typing system will assign a different type to two unrelated strains randomly sampled from a population. A D-value of more than 0.9 indicates good differentiation [32] (Hunter, 1990).

Data Analysis
Data were introduced into R software (R Core Team, 2019; version 3.5.3) for data visualization. The R package "Complex-Heatmap" was used to build a heatmap based on virulence genes, biofilm category, and antimicrobials tested in isolates from chicken, pigeon, camel, and human consumers [33].

Postmortem Examination of Intestinal Samples from Birds
Examination of the intestine from chickens showed subclinical NE ( Figure 1A,B). Multifocal pale foci of mucosal necrosis were observed from the serosa. The lumen of the intestine was filled with gas bubbles in the broiler. The intestine was filled with thick, brownish watery exudate in the jejunum ( Figure 1C,D). Examination of the pigeon's intestine exhibited no evidence of enteritis.

Prevalence of C. perfringens in the Examined Samples
Analysis of the obtained samples revealed the isolation of 50 (15.2%) C. perfringens out of the 330 examined samples (Table 1). A higher isolation rate was obtained from pigeon intestinal contents (66.7%). There was no statistically significant difference between the isolation of C. perfringens from intestinal content (20%) and meat of chickens (10%), p = 0.26, and the prevalence of C. perfringens from camels was insignificantly higher in diarrheic feces (18%) compared to raw camel meat (4%) (p = 0.06). In human samples, 4% were positive for C. perfringens. Vet. Sci. 2022, 9, 109 6 of 18

Postmortem Examination of Intestinal Samples from Birds
Examination of the intestine from chickens showed subclinical NE ( Figure 1A,B). Multifocal pale foci of mucosal necrosis were observed from the serosa. The lumen of the intestine was filled with gas bubbles in the broiler. The intestine was filled with thick, brownish watery exudate in the jejunum ( Figure 1C,D). Examination of the pigeon's intestine exhibited no evidence of enteritis.

Prevalence of C. perfringens in the Examined Samples
Analysis of the obtained samples revealed the isolation of 50 (15.2%) C. perfringens out of the 330 examined samples (Table 1). A higher isolation rate was obtained from pigeon intestinal contents (66.7%). There was no statistically significant difference between the isolation of C. perfringens from intestinal content (20%) and meat of chickens (10%), p

Toxinotyping of C. perfringens Isolates
All isolates were of type A (positive for the cpa gene), and only isolates from camel feces (3, 33.3%), camel meat (1, 50%), and human stools (2, 50%) were positive for the cpe gene (Table 1, Figure 2). The cpb2 gene was identified in 28 (56%) of the isolates from all the sources. None of the isolates were found to harbor beta-, iota-, or epsilon-associated genes.

Antimicrobial Susceptibility Testing
Antimicrobial susceptibility testing of 50 C. perfringens isolates from different sources revealed the resistance of 82-94% of the isolates to penicillin, cefotaxime, cefoxitin, ceftriaxone, clindamycin, and chloramphenicol (Table 2, Figure 2). However, all isolates were susceptible to vancomycin, and 86% exhibited susceptibility to metronidazole. Moreover, 74% of each of the examined isolates were sensitive to ampicillin, amoxicillin, and ampicillin-sulbactam. Multiple drug resistance was observed in 92% of the isolates (46/50), and the MAR index ranged from 0.28 to 0.9, with an average of 0.63 (Table 3).

Biofilm Formation
C. perfringens isolates under investigation were all biofilm producers, of which 14 (28%) were strong biofilm producers with an average OD620 of 0.5037 ± 0.03. All strong biofilmproducing isolates were resistant to at least ten antimicrobials with an MAR index of 0.7 or more, and they were classified as XDR isolates (Table 3). Moderate biofilm producers encountered 60% (30/50) of the isolates, and they were MDR with MAR indices of 0.2-0.6.

Anti-Biofilm Activity of AgNPs-H 2 O 2
The effect of AgNPs on biofilm production of six representative C. perfringens isolates was estimated by the qualitative tube method using various concentrations of AgNPs. Positive results were indicated by the presence of a thin layer of biofilms after staining with the dye. The isolates from chicken, camels, and humans required a concentration of 100 µg/mL to inhibit biofilm formation. In the case of the isolate from pigeons, no biofilm production was observed at 75 µg/mL concentration (Table 4). The quantitative biofilm estimation was conducted using the microtiter plate method, and the inhibition percentage of biofilm production with different AgNP concentrations was investigated. The inhibition of biofilm formation by C. perfringens isolates was dosedependent. At the highest concentration of AgNPs (100 µg/mL), the six examined isolates showed inhibition percentages of 80.8-82.8% (Figure 3 and Supplementary Tables S1 and S2).

Genotyping
The RAPD-PCR patterns of the 50 C. perfringens isolates from different sources were investigated by a single amplification profile. The multiple DNA fragments ranged in size from 300 to 2700 bp. The primer sets manufactured 33 profiles (referred to as R1-R33). The discriminatory power of the RAPD-PCR was calculated using Simpson's index of diversity, and the D-value was 0.9763, which indicated a high discriminatory power. The dendrogram analysis of the examined isolates (n = 50) showed four clusters (Figure 4). The similarity between the human, pigeon, and camel isolates in the same cluster was 100%.

Genotyping
The RAPD-PCR patterns of the 50 C. perfringens isolates from different sources were investigated by a single amplification profile. The multiple DNA fragments ranged in size from 300 to 2700 bp. The primer sets manufactured 33 profiles (referred to as R1-R33). The discriminatory power of the RAPD-PCR was calculated using Simpson's index of diversity, and the D-value was 0.9763, which indicated a high discriminatory power. The dendrogram analysis of the examined isolates (n = 50) showed four clusters (Figure 4). The similarity between the human, pigeon, and camel isolates in the same cluster was 100%.

Discussion
C. perfringens is known as a major public health risk, causing diseases in animals and humans; it also causes economic losses, especially in poultry flocks [34]. This study investigated the prevalence of C. perfringens in various sources of animal and human origin. Toxin typing, antibiotic resistance profile, genotyping, biofilm formation ability, and the effect of silver nanoparticles on biofilm were also evaluated. The source of chicken meat contamination is mainly the intestinal contents during slaughtering and processing [4].

Discussion
C. perfringens is known as a major public health risk, causing diseases in animals and humans; it also causes economic losses, especially in poultry flocks [34]. This study investigated the prevalence of C. perfringens in various sources of animal and human origin. Toxin typing, antibiotic resistance profile, genotyping, biofilm formation ability, and the effect of silver nanoparticles on biofilm were also evaluated. The source of chicken meat contamination is mainly the intestinal contents during slaughtering and processing [4]. Fifty C. perfringens isolates were recovered from all the examined sources, and all the isolates were of type A. In our study, the isolation rate of C. perfringens was higher in pigeon intestines (66.7%), followed by chicken intestines (20%). The high prevalence of C. perfringens in chickens and pigeons could be attributed to the colonization of the organism in the intestinal tract early in poultry life, even from hatchery [35].
A study in Egypt reported that 41.7% of broiler chickens were infected with C. perfringens, of which 35.4% were asymptomatic [36]. In China, 23.1% of live poultry from farms and markets was positive, while 13.6% of fresh chicken meat was positive [4]. In Jordan, 43.2% of broiler chicken flocks were positive [37]. C. perfringens in poultry has acute clinical or sub-clinical forms. The acute disease leads to a high mortality rate and accounts for 1% of deaths per day [38]. The subclinical form is characterized by damage to the intestinal mucosa, which results in decreased digestion and absorption, reduced weight gain, and increased feed conversion ratio [39]. Although healthy birds usually contain less than 10 5 colony-forming units (CFU) of C. perfringens per gram of digesta, NE is produced due to the proliferation of pathogenic C. perfringens strains to reach 10 7 and 10 9 CFU per gram of digesta [40].
Regarding pigeons, scarce studies reported the prevalence of C. perfringens; for instance, Rahman, Sharma [41] reported an isolation rate of 33.3%. Despite the high isolation rate in pigeon intestine, no NE lesions were observed. However, the birds under investigation had symptoms of diarrhea and growth retardation. This could be explained by the presence of another microbial pathogen, which in turn decreased the immunity of the pigeons, thus causing the high isolation rate of C. perfringens [42].
All C. perfringens toxin types produce alpha-toxin; however, higher amounts of the toxin are produced by type A than the other types [43]. Necrotic enteritis and the subclinical form of C. perfringens infection in poultry are caused by C. perfringens type A, and to a lesser extent type C [44]. The predominance of type A from NE cases has been reported in different studies [4,37,45]. The main source of infection for chickens in farms and outlets is crossinfection via feces; therefore, proper hygienic measures, including disposal of contaminated bedding, are crucial to control C. perfringens infections in farms and markets [4].
None of the chicken or pigeon isolates harbored the cpe gene, whereas the cpb2 gene was identified in 73.3% (11/15) of isolates of chicken origin and 20% (4/20) of isolates of pigeon origin. Accordingly, Gharaibeh, A1 Rifai [37] reported the isolation of C. perfringens type A isolates from chicken flocks suffering from enteritis, and the isolates were not positive for enterotoxin-associated genes. In Jordan, all isolates recovered from broiler chickens with enteritis were classified as type A and non-enterotoxin producers [37]. Similar results were also reported in Finland [46] and Sweden [44]. Beta2 toxins have been associated with enteric diseases in humans and other animals and NE in birds [47]. This explains our findings where cpb2 gene was recovered from all examined sources with variable percentages.
In our study, C. perfringens type A was recovered from 10% of retail chicken meat samples. In Belgium, 100% of C. perfringens isolates from chicken meat were of type A [48], which suggests that cpa might be a universal gene in C. perfringens isolated from meat samples [49]. In Korea, 11.1% [50] and 33% [12] of retail chicken meat samples were positive for C. perfringens, while in Jordan, 43% of chicken meat samples were positive [37]. Chicken and ground chicken meat examined in Japan for C. perfringens contamination showed that 97% and 100% of the samples were positive, respectively [51]. The cpe gene, which encodes the enterotoxin, is reported in less than 5% of C. perfringens isolates [52]. All our chicken isolates were cpe-negative. In accordance, none of the isolates from chicken and beef meat samples were positive for other toxin encoding genes other than the cpa gene [12]. However, in the USA and Japan, only 1-3% of C. perfringens isolates from chicken harbored the cpe gene [10,53]. Another study in Japan reported 0.5-0.7% cpe-positive isolates from beef samples [10].
The high detection rate of the cpa gene in C. perfringens isolates could be attributable to the presence of the gene on the chromosome. While, cpb, etx, and iap are located on plasmids, the cpe could be found either on chromosomes or plasmids [7]. Therefore, acquisition or loss of plasmids might have a role in changing toxin types [49]. In our study, the absence of genes except cpa could be explained by the loss of mobile genetic elements, and this is following a previous study [12]. Furthermore, the carriage of the cpe gene was reported to vary; in food poisoning isolates, it is on the chromosome, while it is on the transferable plasmids in isolates from other gastrointestinal (GI) diseases, such as antibiotic-associated diarrhea [51].
The first report on beta2 toxin and its encoding gene (cpb2) was in C. perfringens type C isolates from necrotizing enterocolitis in a piglet [54]. No significant homologies were observed between the amino acid sequence of cpb2 and cpb from the beta toxin [55]. The beta2 toxin was found to possess a weaker cytotoxic activity than the beta toxin, despite having similar biological activity [54]. However, a possible association was reported between enteric disease and the presence of C. perfringens isolates carrying the cpb2 gene [56,57]. In this study, the cpb2 gene was identified from C. perfringens strains isolated from chickens, pigeons, camels, and human sources. Our results conform with previously published studies that identified the cpb2 gene in isolates from chickens [56,58], humans [47], and camels [58]. A recent study in Egypt documented a high frequency of the cpb2 gene from C. perfringens isolates in various sources, including chickens and humans [59]. In the USA, all isolates from chicken and human patients were of type A, and the cpb2 gene was isolated from 93% and 65% of chicken and human isolates, respectively [60]. An investigation of food poisoning outbreaks showed that undercooked meat contaminated with C. perfringens resulted in the survival and growth of the organism, which spread to other servings, causing food poisoning in consumers [61]. Nearly 70% of C. perfringens food poisoning outbreaks and 20% of all non-foodborne gastrointestinal diseases are caused by enterotoxigenic C. perfringens [5]. In Korea, 33% of chicken meat samples were positive for C. perfringens and all isolates were of type A, with positive results for only the cpa gene and harmful for the cpe, cpb, etx, iap, and netB genes [12].
Fewer studies have documented the isolation of C. perfringens from camels; in our study, 18% of diarrheic feces from camels were identified, and all nine isolates were of type A, of which seven (77.8%) and three (33.3%) were positive for cpb2 and cpe genes, respectively. According to Fayez, Elsohaby [62], the isolation of C. perfringens from diarrheic calves and adults was 73.7%, and most of the isolates were of type A. Moreover, other studies have reported that C. perfringens type A is the predominant type from camels and other animals [63,64]. Scarce information has been reported regarding the occurrence of C. perfringens in camel meat. Our results showed that two (4%) retail camel meat samples were contaminated, and one isolate was positive for the cpe gene, while both isolates harbored the cpb2 gene. These findings show the potential of camel meat to serve as a source of C. perfringens. The presence of cpb2 gene is correlated with gastrointestinal diseases in humans and other animals. The two isolates from camel meat in our study were XDR and strong biofilm producers, thus showing the potential of these isolates to cause foodborne gastroenteritis in human consumers. Previous studies have reported the isolation of C. perfringens from retail meat from various animals [10,65]. These results exhibit the urgent need for high sanitary conditions during the slaughter of animals and during the retail of meat at markets to minimize the risk of foodborne intoxication caused by C. perfringens [65].
Healthy people carry less than 10 5 C. perfringens CFU/g feces, while diarrheic patients may carry 10 6 or more CFU/g [66]. The intestinal carriage of C. perfringens may be diverse, and some healthy subjects may also act as potential reservoirs of more than one C. perfringens strain [6]. In this study, C. perfringens isolates were recovered from 4% of human stool from diarrheic patients and were positive for cpa and cpb2, while two were positive for the enterotoxin-associated cpe gene. In Ireland, C. perfringens was isolated from 7.6% of fecal samples from elderly subjects [67].
Yadav, Das [68] identified 43.8% C. perfringens type A isolates from human diarrheal cases, of which 49.1% were found to have only the cpa gene, and 35.1% were positive for cpa and cpb2. The diversity of the study population and the methods of detection could be the reason for the higher carriage of C. perfringens compared to this study [6].
C. perfringens isolates in this study showed resistance to penicillin, tetracycline, cefotaxime, cefoxitin, ceftriaxone, clindamycin, and chloramphenicol. This occurred following a study on C. perfringens from camels and herders; the findings indicated less effective antimicrobial activity of penicillin and cephems [62]. The use of penicillin and tetracycline as growth promotors and for the control of Gram-positive bacterial infections may explain the resistance reported in our study to these drugs [69]. In contrast, penicillin and tetracyclines exhibited sufficient activity against C. perfringens isolates from chicken and human isolates [37,59]. Clindamycin was reported to be effective against C. perfringens isolates [70]. However, our results showed a high resistance rate of 94%, which is comparable with other studies [71,72]. The low resistance of the isolates to ampicillin, amoxicillin, and ampicillin-sulbactam coincides with another study in Egypt [59]. C. perfringens isolates were resistant to amoxicillin, ampicillin, cefotaxime, tetracycline, and clindamycin at rates of 16.1%, 29.8%, 11.3%, 63.7%, and 70.8%, respectively [4]. Another study reported that the most resistant antimicrobial agent for the C. perfringens isolates was tetracycline (33/33, 100%), followed by imipenem (24/33, 72.7%), and 27 of the 33 strains (81.8%) were multiple drug-resistant, exhibiting resistance to at least three classes of antimicrobials [12].
Multidrug resistance was observed in 92% of the isolates, of which 26% were XDR, and the MAR indices ranged from 0.28 to 0.9 with an average of 0.63. These findings conform with previous studies that showed the wide spread of MDR and XDR C. perfringens isolates due to the uncontrolled use of these drugs as growth promoters in animals and for therapeutic uses in humans and animals, which has caused increased resistance [59,62]. It has been shown that MAR higher than 0.2 could be due to contamination from high-risk sources, such as humans and farm animals frequently exposed to antibiotics, thus resulting in potential risk to consumers [73].
Biofilm-forming microbes are responsible for causing 65-80% of infections because they can resist environmental and physical conditions and antibiotic treatment [74]. Although most studies focus on biofilm production and methods of reduction in single species, biofilms in nature mostly comprise multiple species, where inter-species interactions can affect the development, structure, and function of these communities [75]. These interactions could be in the forms of genetic and metabolic exchange and signaling to occur between microorganisms in biofilm communities; however, high throughput and high-resolution methods are needed to reveal these interactions [76]. Such interactions could be either competitive or cooperative, resulting in impacts on maturation, physiology, antimicrobial resistance, and virulence of these communities [76].
All our isolates were biofilm producers, and 28% produced strong biofilms, conforming with another study in Egypt [59]. Biofilms produced by C. perfringens isolates contribute to developing gastrointestinal diarrheal diseases in humans and animals due to the protective effect of biofilms against antibiotic treatment [26]. Several studies have reported the ability of C. perfringens isolates from different sources to produce biofilms [26,27,50,59]. Therefore, to control infections caused by biofilm-producing microbes, biofilm inhibitors should be used.
The chemical, biological, and physical complexity and dynamics of EPs resulted in poor susceptibility to antimicrobial agents [15]. The extracellular polymeric substances either bind to antimicrobials, delaying their diffusion, or they chemically react with antimicrobials, causing inactivation [77]. Therefore, nanoparticles have been used as alternatives for biofilm eradication. Nanoparticles play an important role as carriers of EPS matrix disruptors due to their intrinsic high surface area to volume ratio, providing a platform for the development of materials with a wide spectrum of mechanical, chemical, electrical, and magnetic properties [78].
The use of NPs in inhibiting biofilms has been previously documented for various bacteria [79][80][81]. This study indicated the potential of AgNPs to inhibit biofilm production at concentrations of 75 and 100 µg/mL, and the percentage of inhibition was 80.8-82.8% at the highest AgNP concentrations. The inhibitory effect of AgNPs was previously reported by Siddique, Aslam [81] on biofilm production by Klebsiella pneumoniae isolates. The results revealed 23-86% inhibition in the presence of different concentrations of AgNPs due to the disruption of the EPS matrix. Moreover, other studies reported 89% and 75% biofilm inhibition in Staphylococcus aureus and Escherichia coli, respectively [82]. Studies also demonstrated that nanosized silver (at the size of 100 nM) resulted in a 95% reduction in biofilm production by Pseudomonas aeruginosa due to the disruption of the EPS matrix [83].
Tracing the source and route of transmission of C. perfringens type A strains causing food poisoning is significant in epidemiological investigations. RAPD-PCR is one of the effective methods used for studying the genetic relationship of C. perfringens isolates from various sources [20]. Our results indicated that 50 C. perfringens isolates recovered from various sources were grouped in four clusters. One of the clusters included isolates from humans, pigeons, and camels with 100% similarity, showing a common source of the isolates. Similar results were reported by Afshari, Jamshidi [20], who obtained C. perfringens isolates from human stool, poultry meat, and minced meat in the same cluster. The discriminatory power of RAPD-PCR in our study was 0.9763, which was a high discrimination of the reaction to the isolates from various sources. Our isolates of chicken and pigeon origin were genetically clustered more closely to each other, and this was consistent with a previous study [84].
The control of biofilm-producing bacteria especially in the food chain depends mainly on the identification of their levels and tracing their sources. Bacterial biofilms possess high resistance to commercial antimicrobial agents, which necessitate the need for finding control alternatives. The inhibitory action of AgNPs-H 2 O 2 on bacterial biofilms make it a promising wide-spectrum antibacterial agent. Because of the ecofriendly nature of the product, it can be safely used in the food industry, protecting human health. Further studies are required to find more antimicrobial alternatives for food safety.

Conclusions
In conclusion, MDR and XDR C. perfringens type A are predominant in chickens, pigeons, camels, and humans in the study area. The potential of the isolates to manufacture illness is predicted because of the presence of cpe and cpb2 genes. The anti-biofilm activity of AgNPs is described in the study against MDR and XDR C. perfringens isolates. This is the first report of AgNPs anti-biofilm activity against C. perfringens in chickens, pigeons, camels, and humans to the best of our knowledge. The RAPD-PCR method seems promising for the epidemiological investigation of foodborne diseases caused by C. perfringens and for contamination source tracking in the field of food hygiene and industry.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/vetsci9030109/s1, Table S1: title. Average OD620 of biofilm production by Clostridium perfringens in the presence of different AgNPs concentrations. Table S2: Percent inhibition of biofilm formation by various concentrations of AgNPs against Clostridium perfringens. The absorbance was measured at 620 nm for the quantification of biofilm formation. Institutional Review Board Statement: Informed verbal/written consent for participation in the study was obtained from all participants, and it was approved by the Committee of Animal Welfare and Research Ethics, Faculty of Veterinary Medicine, Zagazig University, Egypt (protocol no. 118/2019).