Microbial profiling of wound pathogens in isolates from an Egyptian hospital using a microarray chip

Chronic polymicrobial infections represent diagnostic challenges for both molecular and culture methods. Chronic wounds, inducing chronic pain and reducing the mobility of individuals, have a notable effect on the quality of life. At any given time, 1% of the population is usually affected and, therefore, multiple treatments are also required. Identification of the etiology of such infections facilitates the appropriate treatment. In this study, 20 wound samples were collected from 20 patients with suspected wound infection. The panel of the current assay targets 12 clinically relevant aerobic pathogens, commonly associated with chronic wound infection. Microbial wound infections were detected by both usual polymerase chain reaction (PCR) and subsequently testing using the DNA chip. In the current study, the results of culture-free bacterial identification using the two methods of DNA analyses were compared. By molecular detection using PCR, seven different bacterial species were identified: Citrobacter spp. (100%), Enterobacter spp. (100%), Klebsiella pneumoniae (100%), Pseudomonas aeruginosa (90%), Proteus spp. (80%), Escherichia coli (60%), and Staphylococcus aureus (10%). Mixed microbial infections were detected in all samples indicating four, five, or six different bacteria, identified in each sample. Microarray detection in comparison to PCR indicated 100% matching. These results demonstrate the possibility of fast identification of wound infection pathogens even in a mixed culture in a very short time, which in turn facilitates the proceeding steps for proper treatment.


INTRODUCTION
One of the most common health problems is usually associated with wound infection (Giacometti et al., 2000), commonly induced by the entry of the bacteria through contamination of breached skin. With the same importance as renal and cardiovascular disorders, wounds can also have a similar impact on the quality of life. The Global Wound Care Market report indicated that this sector has reached about 20 billion dollars and is estimated to exceed 25 billion dollars in 2023 worldwide (Weller et al., 2020). For this reason, and the pressure induced on both patients and the medical system, efforts have been made with the aim of managing wound infections and their associated pathological conditions, which in turn could improve the quality of life and increase life expectancy. In recent decades, different approaches have been developed to improve the rate and quality of chronic wound healing (Nosrati et al., 2021). Despite the multiple etiologies of wound infections, bacterial colonization usually characterizes their chronicity (Loesche et al., 2017). Multiple variable organisms concerning wound infection were documented in previous studies (Melling et al., 2001). Most wound culture isolation procedures implement swab cultures, the easiest traumafree way, which avoid hazards of complications on the wound bed (Meyers, 2008). A variety of microbes such as bacteria, fungi, and parasites are associated with wound infection (Bowler et al., 2001). Wound microbiome composition could also be influenced by ecological processes, which in turn induces the emergence of polymicrobial infections exhibiting both synergistic effect and enhanced tolerance to antimicrobials (Baishya and Wakeman, 2019;Kalan and Brennan, 2019). In addition, stable microbial communities are usually correlated with delayed healing (Loesche et al., 2017;Sloan et al., 2019). Moreover, genotypic effects could also be the possible way to explain the recently observed patientspecific immunological responses to the same microbial exposure (Deusenbery et al., 2019). Both Gram-positive and Gram-negative bacteria are usually detected including Escherichia coli, Proteus spp., and Klebsiella spp., in addition to Enterobacter spp. and Staphylococcus aureus, which was also identified (El-Saed et al., 2020;Gupta et al., 2019). However, a challenging bacterium, Pseudomonas aeruginosa, exhibiting resistance to most antibacterial drugs, could also be detected (Bowler et al., 2001). In addition, the presence of anaerobic bacteria at baseline infection could also be associated with worse healing outcomes (Min et al., 2020). In contrast to the intestinal microbiota, normal skin is mainly colonized by only a few taxa including Corynebacterium, coagulase-negative Staphylococci, and Propionibacterium (Scharschmidt and Fischbach, 2013) in which sebaceous sites are dominated by lipophilic Propionibacterium species, while Corynebacterium and Staphylococci are preferentially abundant in moist areas (Costello et al., 2009;Grice et al., 2009;Oh et al., 2014). Wound infections are classified into two major classes, skin and soft tissue infections, which are usually undistinguished upon the progression of infection. In addition, in these infections when untreated, infecting bacteria affect healing and may also produce other signs and symptoms. Moreover, wound infections are considered among the most leading nosocomial infections and are associated with increased morbidity and other medical expenses (Cutting and White, 2004). For assessing and diagnosing infection, the most common signs usually include redness, swelling, increased drainage, and increased pain. Therefore, microbial etiology is one of the fundamental steps required before starting the treatment protocols. Most microbial identification is usually based on the bacterial separation by an enrichment culture, which favors the selection of some microbes, while excluding other species. For this reason, bacterial identification is now shifting away from metabolic biochemical testing toward genetic or molecular identification (Espy et al., 2006;Han et al., 2011;Seng et al., 2010). Molecular technologies have provided worldwide researchers with more rapid and sensitive tools to examine human microbiota, in comparison with the old culturebased testing (Bowler and Davies, 1999;Dowd et al., 2008a). In most studies, universal primers for 16S rRNA genes are used, followed by specific identification approaches such as polymerase chain reaction (PCR) (Hill et al., 2003) and denaturing gradient gel electrophoresis and sequencing (Dowd et al., 2008a(Dowd et al., , 2008bHill et al., 2003;Price et al., 2009). In addition, these methods exhibited successfulness in the detection of bacteria in burn wounds (Pirnay et al., 2000), blood (Rothman et al., 2002;Wellinghausen et al., 2004;Yang et al., 2002), cerebrospinal fluid (Poppert et al., 2005), joint fluid (Yang et al., 2008), and heart tissue (Breitkopf et al., 2005). Moreover, in most cases, each experiment can be executed in not more than a few hours (Ecker et al., 2010). In the same respect, for developing new therapeutic strategies, the use of these techniques speeds up the early detection of infecting microorganisms assisting clinicians and pathologists to find out the appropriate treatment of hospitalized patients.
For the detection of pathogens, DNA probes are immobilized in microarray experiments, in which more than one copy of each target gene could be included (Cannon et al., 2010;Shen et al., 2015). However, the detection of wound-associated bacterial pathogens was rarely described in previous studies. The aim of this study is the detection of wound pathogens in mixed samples and their application in samples from Egyptian hospitals.

Specimens collection and processing
Hospitalized immunocompetent symptomatic patients, almost all adults suffering from wound infection, were included in this study. Control samples were taken from nonhospitalized individuals and those free from any symptoms of wound infection. Samples were collected from 20 different patients admitted to one of the Mansoura hospitals in Egypt. This study was approved and funded by the Competitive Funding Projects Postgraduate Research and Cultural Affairs Sector, Mansoura University. The experimental protocol conducted in this study was approved by the Research Ethics Committee, Faculty of Pharmacy, Mansoura University, and conducted in accordance with the Declaration of Helsinki involving the use and handling of human subjects.

Extraction of genomic DNA and DNA from Standard Cultures
Different bacterial cultures obtained from wound samples were prepared in LB broth, followed by incubation aerobically at 37°C and observation for the presence of visible microbial growth. Genomic DNAs were extracted from 37°C 24 hours cultures of wound samples using the genomic QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany), according to the manufacturer's instructions. By using a Nanodrop instrument (OPTIZEN NanoQ, Mecasys, Daejon, South Korea), the concentration of gDNA was determined. Purified DNAs were stored at −80°C.

PCR Amplification and Labeling to Get Hybridization Targets
By starting from isolated gDNA as a template, nearly full-length 16S rRNA was amplified in a PCR reaction. Two different biotin-labeled primers were used: forward primer 27F (5′-AGAGTTTGATCCTGGCTCAG) and reverse primer 1492R (5′-GGTTACCTTGTTACGACTT). Each PCR reaction is composed of 25 μl containing 2 μl of template DNA, 0.5 μM of each primer, 1.5 mM MgCl 2 , 0.2 mM dNTPs, 1 U Taq polymerase (Dream Taq Green DNA Polymerase, Fermentas), and to the final volume nuclease-free water. PCR reactions were carried out using Cycler 003 PCR Machine [A & E Lab (UK)] with 2 minutes initial denaturation at 94°C, followed by 35 cycles of heating at 94°C for 30 seconds, cooling to 52°C for 30 seconds in an annealing step, and heating at 72°C for 30 seconds in an extension step, and a final extension at 72°C for 10 minutes. The 1% agarose gel and electrophoresis-separated PCR products were visualized by staining with ethidium bromide and UV light exposure.

Strain-Specific Genes PCR Test
As described previously, amplification of both genomic DNA and standard DNA was carried out using primers (Biosearch Technologies) listed in seconds, then cooling to annealing temp (as listed in Table 1) for 30 seconds, and heating at 72°C for 30. Finally, each PCR reaction was ended with a final extension step at 72°C for 10 minutes.

Oligonucleotide Probes Design and synthesis
Segments of 20-30 mer oligonucleotides with similar lengths, melting, and GC content with the optimum BLAST hit were selected after comparison (http://www.ncbi.nlm.nih.gov/ BLAST/) with all available sequences in the GenBank database. All oligonucleotides used in this study were listed in Table 1.

Oligo Aarray Printing
In a 384-well printing plate, using ArrayIt spotting buffer, each printing oligo solution was prepared to a final concentration of 50 pmol μl −1 . Spotting was carried out onto microarray slides in triplicate for each probe (Scienion, Berlin, Germany) using the SpotBot microarrayer (ArrayIt, Sunnyvale, CA) spotting machine (Fig. 1).

Microarray Testing
Samples were labeled and then resuspended in a hybridization buffer (Scienion, Berlin, Germany). The method was carried out according to the recommendations of the manufacturer (Scienion, Berlin, Germany). In a final volume of 16, 8 μl of the PCR products was mixed with 8 μl of the hybridization buffer. The prehybridization step was carried out using 20 μl of the prehybridization solution (Scienion, Berlin, Germany). Labeled samples were then applied to a 42°C prewarmed microarray slide. Hybridizations were carried out in a hybridization station (ArrayIt, Sunnyvale, CA) using the hybridization mix, previously boiled for 2 minutes and cooled on ice. Hybridization was carried out by incubation at 42°C for 12 hours. Finally, at room temperature, using buffers I, II, and III (Scienion, Berlin, Germany), washing was carried out. Microarray staining was carried out by a streptavidin biotin color development system (Fermentas, Waltham, MA) using the manufacturer's instructions. For documentation of the results, images were acquired using the ArrayIt Microarray Scanner (ArrayIt, Sunnyvale, CA). Signal intensities were recorded using the Spotware software after subtracting the local background values from the per-sample median.

Results of bacterial identification
Twenty wound samples were analyzed by molecular testing. Overall, the most common organisms identified using molecular PCR testing were Citrobacter spp. (100%) in 20 samples, Enterobacter spp. (100%) in 20 samples, and K. pneumoniae (100%) in 20 samples, followed by P. aeruginosa (90%) in 18 samples, Proteus spp. in 16 samples (80%), and E. coli (60%) in 12 samples. However, S. aureus was rarely identified in two samples (10%). Other bacterial taxa such as E. faecalis, H. influenzae, and Mycobacterium spp. could not be detected in any isolate. Control samples were free from the tested pathogens. Concerning a mixed microbial infection, a total of eight wounds were found to have five different bacterial taxa; in addition, 10 of the total samples (50%) had six bacterial taxa. However, two of the total wounds (10%) had only four bacterial taxa (Table 2).

Microarray Testing of clinical samples
In this study, by DNA microarray testing, the proportions of bacteria identified from the 20 positive samples were Citrobacter spp. (100%), Enterobacter spp. (100%), and K. pneumoniae (100%), followed by P. aeruginosa (90%), Proteus (80%), and E. coli (60%). However, S. aureus was rarely identified (10%). A mixed bacterial infection could be identified in all different samples as four, five, or six different bacteria were identified in each sample as shown in Figure 2. No difference could be observed when compared to the results obtained by PCR (Table 3).

DISCUSSION
Microbial molecular diagnostic tests, in comparison with traditional culture methods, enable more patient care improvement. However, it is also necessary to continue to include all these introduced newer methods. In addition, it has been proven that culture-based methodologies in bacterial detection are markedly less sensitive to DNA detection methodologies (Han et al., 2011). Moreover, the results obtained using these traditional culture-based methods are usually associated with misidentification reflecting overestimation of the relative presence of easily cultured and identifiable microbes (Davies et al., 2004).

Organism
No According to the recent reports, the results obtained using PCR and 16S sequencing commonly used have demonstrated the close correlation between most of these molecular methods. As previously documented, different molecular methods such as realtime, uniplex, and multiplex PCR, especially in mixed samples, have enhanced both the speed and the sensitivity of microbial detection (Gebert et al., 2008;Lehmann et al., 2008). However, multiplex PCR, in each reaction, can only detect a limited number of pathogens (Edin et al., 2015;Gadsby et al., 2015). For these reasons, when dealing with samples containing a mixed culture or mostly environmental samples, a comprehensive view of microbial communities is required. In the same respect, a better solution is provided by DNA microarrays, especially in mixed microbial infections.
Microarray testing is usually based on the panel of taxa selection, detection capacity, and probe coverage, which may differ between sequence targets. Microarray-detecting oligonucleotides are mainly designed based on the available sequence data and the variation between target organisms exhibiting reduced power for identifying other different organisms (Cannon et al., 2010;Shen et al., 2015).
As previously reported (Costello et al., 2009;Grice and Segre, 2011;Grice et al., 2009;Oh et al., 2014), in sequencing surveys of healthy adults, dry, moist, and sebaceous microenvironmental changes usually affect the relative abundance of bacterial taxa. However, a lower bacterial diversity is usually detected (Scharschmidt and Fischbach, 2013), as the normal skin is usually dominated by only a few taxa, mainly lipophilic Propionibacterium species of the phylum Actinobacteria thriving sebaceous sites, while Staphylococcus and Corynebacterium species of the phyla Firmicutes and Actinobacteria are preferentially abundant in moist areas (Costello et al., 2009;Grice et al., 2009). Interestingly, some members of the skin microbiota such as Staphylococcus epidermidis are nonpathogenic; however, when reaching the blood circulation, serious systemic diseases in some individuals can be caused (Blum and Rodvold, 1987). In this study, to distinguish some major bacterial species, frequently causing wound-associated infections including E. faecalis, Enterobacter spp., K. pneumoniae, H. influenzae, E. coli, S. pyogenes, Clostridium spp., S. aureus, Mycobacterium spp., Citrobacter spp., P. aeruginosa, and Proteus spp., 12 oligonucleotide probes, complementary to 16S PCR products, were selected for detection based on several species-specific regions of each pathogen.
In samples obtained in the current study, by DNA microarray testing, all samples (100%) were found to be positive and identified mainly as Gram-negative bacteria, which is consistent with some previous studies in Saudi Arabia (Alkaaki et al., 2019;Al-Mulhim et al., 2014) and other countries (Allegranzi et al., 2011). The percentages of identified bacteria were Citrobacter spp. (100%), Enterobacter spp., (100%), and K. pneumoniae (100%), followed by P. aeruginosa (90%), Proteus (80%), and E. coli (60%). Similarly, in a previous study, the same bacterial taxa were detected, but at lower percentages, 56.7%, 6.7%, and 3.3% for E. coli, Klebsiella spp., and Enterobacter spp., respectively (Adwan et al., 2016). In addition, an increased level of Pseudomonas species (43%) was detected in a previous study (Gupta et al., 2019), compared to K. pneumoniae and Acinetobacter baumannii, which were detected at the lower prevalence of 28% and 14.83%, respectively. Moreover, other studies in a tertiary care hospital in India reported the predominance of P. aeruginosa, followed by K. pneumoniae and A. baumannii (Dash et al., 2013;Singh et al., 2003). Interestingly, the wound microbiome, in which Pseudomonas dominates, exhibited increased temporal stability against multiantibiotic therapy in comparison to staphylococcal infections (Tipton et al., 2017). However, other different results were obtained in another study on surgical site infections (Negi et al., 2015), as the commonest organism identified was S. aureus (50.4%), followed by E. coli (23.02%), P. aeruginosa (7.9%), and Citrobacter spp. (7.9%). In addition, in another study , in wound samples, A. baumannii was mainly detected (23%) in 28 samples. In this current study, S. aureus was rarely identified in 10% of total isolates. A similar observation was previously reported, as S. aureus was rarely detected (two wound samples) in one study using a microarray  and in another study on combat wound colonization (Brown et al., 2011). On the contrary, a higher prevalence of S. aureus (30%) could be detected in one previous study (Adwan et al., 2016). In addition, in another study, the most commonly isolated organism was S. aureus, followed by coagulase-negative Staphylococci (CNS), Enterococcus spp., and E. coli, while the pathogen detection was mainly dependent on the surgical procedure (Owens and Stoessel, 2008). Moreover, in another study, S. aureus was the most frequent pathogen detected (22.8%), while other pathogens, namely, P. aeruginosa (20.1%), E. coli (12.2%), Klebsiella spp., (12.2%), Enterobacter spp., (7.7%), and Enterococcus spp. (5.9%), were identified at lower percentages (El-Saed et al., 2020).
Interestingly, in previous studies, not so much data were found concerning the detection of mixed wound infection. In this current study, four, five, or six different bacteria were identified in each sample, representing mixed microbial infections. K. pneumoniae, P. aeruginosa, Enterobacter spp., Proteus spp., and Citrobacter spp. were the most identified mixed infections.

CONCLUSION
In the present study, by PCR detection, Citrobacter spp., Enterobacter spp., K. pneumoniae, P. aeruginosa, Proteus spp., and E. coli were mainly detected. In addition, pathogen detection with both PCR and the DNA chip indicated 100% matching. Moreover, the use of a DNA chip in a mixed culture facilitates the fast identification of wound-infecting pathogens and the proper treatments strategies.

CONFLICTS OF INTEREST
The author declares that there are no conflicts of interest regarding the publication of this paper.

FUNDING
This study was funded by the Competitive Funding Projects Postgraduate Research and Cultural Affairs Sector, Mansoura University.
There were no financial interests related to the material in the manuscript.