Multi-Instrumental Analysis Toward Exploring the Diabetic Foot Infection Microbiota

The polymicrobial nature of diabetic foot infection (DFI) makes accurate identification of the DFI microbiota, including rapid detection of drug resistance, challenging. Therefore, the main objective of this study was to apply matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI TOF MS) technique accompanied by multiply culture conditions to determine the microbial patterns of DFIs, as well as to assess the occurrence of drug resistance among Gram-negative bacterial isolates considered a significant cause of the multidrug resistance spread. Furthermore, the results were compared with those obtained using molecular techniques (16S rDNA sequencing, multiplex PCR targeting drug resistance genes) and conventional antibiotic resistance detection methods (Etest strips). The applied MALDI-based method revealed that, by far, most of the infections were polymicrobial (97%) and involved many Gram-positive and -negative bacterial species—19 genera and 16 families in total, mostly Enterobacteriaceae (24.3%), Staphylococcaceae (20.7%), and Enterococcaceae (19.8%). MALDI drug-resistance assay was characterized by higher rate of extended-spectrum beta-lactamases (ESBLs) and carbapenemases producers compared to the reference methods (respectively 31% and 10% compared to 21% and 2%) and revealed that both the incidence of drug resistance and the species composition of DFI were dependent on the antibiotic therapy used. MALDI approach included antibiotic resistance assay and multiply culture conditions provides microbial identification at the level of DNA sequencing, allow isolation of both common (eg. Enterococcus faecalis) and rare (such as Myroides odoratimimus) bacterial species, and is effective in detecting antibiotic-resistance, especially those of particular interest—ESBLs and carbapenemases. Supplementary Information The online version contains supplementary material available at 10.1007/s00284-023-03384-z.


Introduction
Chronic wounds are one of the most devastating impairments related to diabetes among which diabetic foot infection (DFI) represents the most frequent and serious disorder [1]. Considering the growing number of diabetic (ca. 420 million patients so far) DFI is currently considered as a predominant trigger for lower extremity amputations worldwide [2]. Accurate deciphering of the infection causative agent determines taking effective treatment, however, the reliable identification of the DFI microbial patterns in diabetic patients is still challenging due to usually the polymicrobial nature of the infection [3]. Moreover, there is a growing need for reports on the DFI microbial compositions in specific geographical regions to provide local treatment guidelines [4].
Up to date, the DFI diagnosis mostly relies on the traditional culture method and phenotypic identification of the grown colonies or the application of molecular techniques, such as the 16S rDNA PCR amplification and sequencing [5]. While the former is time-consuming and presents limited identification accuracy, the latter is characterized by high sensitivity, discriminatory power, and allows skip the culturing step, however, it requires a well-equipped laboratory and highly qualified staff, which increases the costs of analysis [6]. More recently, matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI TOF MS) seems to be a good solution for mentioned limitations, since fast and cost-effective bacterial identification by MALDI TOF MS along with the multiplication of culture conditions provides high identification accuracy in relatively short time-to-results and opens the possibility for further investigation of biological features including antimicrobial susceptibility of isolates [7]. It was proved in the work Złoch et al. [8] where the accuracy of the reflected microbial patterns of the swab samples derived from DFI patients via MALDI technique significantly improved when the multiply culture conditions was applied.
Although clinicians should avoid antibiotic therapy that is unnecessary, nevertheless, successful treatment of DFIs calls for the administration of appropriate antibiotics [9]. Facing an increased rate of isolation of antibiotic-resistant pathogens, which can be observed in the past few decades, the selection of the effective antimicrobial therapy of DFI becomes more challenging than ever before [2]. In particular, this applies to the constantly growing number of multidrugresistant (MDR) Gram-negative species such as extendedspectrum beta-lactamase (ESBL), carbapenemase-producing Enterobacterales, or MDR Pseudomonas aeruginosa [9]. The frequency of occurrence of MDR Gram-negative pathogens in different geographical area and treatment centers is widely variable, however, current data showing that besides developing countries, this problem is increasingly affecting European and other developed countries [2]. In view of this, reliable and fast detection of antibiotic-resistant Gram-negative pathogens becomes crucial for preventing the failure of the DFI treatment and further spreading of MDR [10].
Many different methods of detecting antibiotic resistance/ susceptibility have found application in routine clinical practice. The most commonly used techniques are phenotypic tests such as disc diffusion or combined disc inhibitory tests, gradient minimal inhibitory concentration (MIC) strips (Etest), the Carba NP tests or most recently the modified carbapenem inactivation method (mCIM) [11,12]. Despite their low costs and simplicity, their major drawbacks are the need for additional incubation which extends time-to-result and relative low specificity and selectivity [13]. Application of molecular methods such as targeted PCR assays including multiplex ones is more specific and selective as well as enables receive results faster (even < 4 h) and detect several different resistance genes in a single run [14,15]. However, the use of the molecular approach is very often limited, especially in developing countries, due to high costs of the analysis, need for highly trained staff, or access to commercial databases and dedicated equipment [6].
In the last few years, special attention has been paid to the utilization of rapid biochemical assays based on antibiotic hydrolyzing activity detection, which opens the possibility of obtaining accurate and reliable results in a simple, fast, and cheap way [16]. In light of this, the MALDI TOF MS technique turned out to be the most promising tool facilitating the indication of a wide range of β-lactamases including clinically relevant cephalosporinases and carbapenemases [17]. Several publications have demonstrated the feasibility of the MALDI TOF MS technique [18][19][20]. Moreover, in the recent work Złoch et al. [21] authors pointed out that the MALDI could be also used to partially classify the class of the carbapenemase present in the sample or as a fast surrogate of standard MIC assay in case of metallo-β-lactamases producing strains (MBL).
The main goal of this study was to apply the MALDI microbial identification technique with the previously established multiple culture conditions to decipher microbial patterns of the swab samples collected from DFI patients treated in the Provincial Polyclinical Hospital in Toruń. (Poland). The accuracy of the MALDI TOF MS identification was evaluated by referring to the sequencing results of the 16S rDNA region. Moreover, as Gram-negative bacteria are considered the primary cause of the multidrug resistance spreading among DFI-suffering patients, the analysis of the frequency of the occurring resistance against different classes of beta-lactams, including cephalosporins and carbapenems using different approaches (MBT STAR BL assay, Etest, multiplex PCR) was performed for the sake of choosing the most accurate one.

Clinical Samples
During studies analyzed clinical specimens derived from 31 patients (42-

Bacteria Isolation and Culturing Technique
For bacteria isolation, serial dilution method (10 -1 -10 -3 ) in sterile peptone water (Sigma Aldrich, Steinheim, Germany) was applied. After defrosting, samples were thoroughly vortexed and then 0.5 mL was transferred into the test tube containing 4.5 mL of sterile peptone water (Sigma Aldrich, Germany) and again vortexed (first dilution-10 -1 ). 100 μL of each dilution was plated onto 5 different culture media previously selected as the most useful in DFI bacteria recovery [8]: Tryptic Soy Agar (TSA; Sigma Aldrich, Steinheim, Germany), Columbia Blood Agar (BLA; Oxoid, Basingstoke, Great Britain), CHROMagar Orientation (CHRA; GRASO Biotech, Starogard Gdański, Poland), Glucose Bromocresol Purple Agar (BCP; Sigma Aldrich, Steinheim, Germany), and Vancomycin Resistant Enterococci Agar (VRE; Oxoid, Basingstoke, Great Britain). All media were in the form of ready-to-use powders except for BLA, which was prepared by adding defibrinated sheep blood (GRASO Biotech, Starogard Gdański, Poland) to the sterilized and dissolved Colombia blood agar base to the final concentration 5% (v/v). Bacterial cultures were incubated at 37 °C for 24 h and then single colonies characterized by different morphological features were selected to obtain pure cultures using the streak plate method on the same media (incubation at 37 °C for 18-24 h).

Identification of Bacterial Isolates Using MALDI TOF MS Technique
For bacteria identification used protein extracts obtained applying formic acid/acetonitrile method according to the MALDI Biotyper protocol (Bruker Daltonik GmbH, Bremen, Germany). 1 microbial loop (10 µl) of fresh biomass was suspended in 300 µl of sterile deionized water, mixed, and then viable bacterial cells were inactivated by adding 900 μL of 96% ethyl alcohol. After vortexing, the resulted bacterial suspension was centrifuged (1300 rev/min, 5 min), the supernatant was discarded and the remaining cell pellet was dried using a vacuum centrifuge at room temperature. Consequently, to the cell pellet 10 µl formic acid (FA) and 10 µl acetonitrile (ACN) was added and mixed. The obtained extract was centrifuged (13,000 rev/min, 5 min.) and 1 µl of supernatant was transferred onto a MALDI MTP 384 ground steel target sample spot (Bruker Daltonik GmbH, Germany). After air-drying, the sample spot was overlaid with 1 µl of MALDI matrix solution-10 mg/mL α-cyano-4-hydroxycinnamic acid (HCCA; Sigma Aldrich, Switzerland) solution prepared in standard solvent solution (50% ACN, 47.5% water and 2.5% trifluoroacetic acid).
Bacterial protein extracts were analyzed using an ultraf-leXtreme MALDI TOF mass spectrometer (Bruker Daltonik GmbH, Bremen, Germany) equipped with the smartbeam-II laser-positive mode. Spectra were collected manually using manufacturer software, flexControl (parameters: 500 shots in-one-single spectra to frequency 2500, m/z range = 2000-20,000, acceleration voltage = 25 kV, global attenuator offset = 20% and attenuator offset = 34% and its range = 34%, laser power = 40%), and subjected to smoothing using the Savistsky-Golay method (width 2 m/z, cycles 10) and baseline corrections using the TopHat algorithm (signal to noise threshold 2; peak detection algorithm-centroid) followed by calibration using the Bruker's Bacterial Test Standard (BTS; Bruker Daltonik GmbH, Bremen, Germany) in quadratic mode via manufacturer software, flexAnalysis. Each sample was measured in quadruplicate (two spots per samples measured in twice). Validated mass spectra were used for bacterial identification via MALDI Biotyper 3.0 Platform (Bruker Daltonik GmbH, Bremen, Germany) based on the both raw spectra (RAW) and Main Spectra (MSP).

Identification of Bacterial Isolates Using 16S rDNA Sequencing
First of all, the total bacterial genomic DNA were obtained using E.Z.N.A.® Bacterial DNA Kit (Omega Bio-tek, Norcross, US) from overnight bacterial cultures (37 °C) following extraction protocol with Glass Beads S supplied by the manufacturer. The extracted bacterial DNA were used for amplification of the 16S rDNA region via PCR technique using universal bacterial primers 27F (5-AGA GTT TGATC-MTGG CTC AG-3) and 1492R (5-GGT TAC CTT GTT ACG ACT T-3), thermostable Taq DNA polymerase (Qiagen, Hilden, Germany), Mastercycler pro S thermocycler (Eppendorf AG, Hamburg, Germany), and PCR program established in the earlier work [22]. The efficiency of the obtained PCR products and purity were studied quantitatively by spectrophotometry as well as gel electrophoresis in 1% agarose. Subsequently, PCR products were send to Genomed (Warsaw, Poland) and were sequenced via the Sanger dideoxy method using the same primers, contigs were assembled via BioEdit Sequences Alignment Editor ver. 7.2.5 [23], and consensus sequences were compared with references sequences in rRNA/ITS databases of the National Center for Biotechnology Information via the BLAST algorithm (https:// blast. ncbi. nlm. nih. gov/ Blast. cgi? PAGE_ TYPE= Blast Search). The DNA sequences determined in this study were submitted to GenBank, and accession numbers are given in the Table S1. The evolutionary tree of identified bacterial strain was inferred based on the Neighbor-Joining and Maximum Composite Likelihood method using MEGA7 software [24] and was visualized using Interactive Tree of Life (iTOL) v 6.5.4 [25].

Microbial Pattern of the DFI Samples Obtained via MALDI TOF MS
Applied MALDI protocol revealed that far most of the DFI patients suffered from polymicrobial infections-97% (Fig. 1a). Moreover, half of the infected wounds contained four or more different microbial species simultaneously -52%. Over two-thirds of the wound samples were occupied by both Gram-positive and -negative species (71%), while another 26% by only Gram-positive (Fig. 1b). In two cases (3%) only Gram-negative species have been found.

16S rDNA Sequencing Results
We performed sequencing of the 16S rDNA region to evaluate the correctness of the MALDI identifications. In the results, 99 bacterial species have been identified-93% with species confidence (Supplementary Table S1). For 12 isolates, we do not get sequencing results due to strain loss during passaging (8 isolates) or the lack of a specific PCR product-the case of all P. aeruginosa strains. In the case of two isolates -Klebsiella oxytoca DFI-13 and Citrobacter freundii DFI-21-the analysis of the 16S rDNA region did not allow for a reliable determination of the species. In 7 cases, discrepancies between MALDI and 16S rDNA sequencing results have been noted-all of these concerned closely related species: E. coli/S. dysenteriae, E. cloacae/hormaechei, C. freundii/braakii, E. cloacae/P. agglomerans, P. vulgaris/terrae, and Acinetobacter haemolyticus/gyllenbergii. All in all, obtained MALDI identification was characterized by 98% genus and 93% species confidence, referring to the DNA sequencing results. Phylogenetic relationship of the identified bacterial isolates is presented on the Fig. 2b.    Table 3). The most frequent resistance against ampicillin was recorded -among 71% of patients and 52% of the isolates. Subsequently, over half of the patients (52%) showed the presence of ESBL strains covering common Enterobacterales species, such as E. coli, C. freundii, K. oxytoca, M. morganii, and P. vulgaris, as well as rarely founded in DFI samples-Flavobacteriaceae (M. odoratimimus) and Sphingobacteriaceae (S. multivorum) members. Considering ESBLs, for all of them, hydrolytic activity against cefotaxime was detected, while in the case of ceftriaxone and ceftazidime percentage of positive results among resistant strains dropped to 38.5 and 15.4%, respectively. Only for one strain -C. koseri DFI-28, the performed MBT STAR BL assay gave unclear results -values between 0.2 and 0.4. Referring to other antibiotics, 29% of strains (mainly E. coli and P. mirabilis) were resistant to piperacillin, while only 10% (P. aeruginosa DFI-5, K. oxytoca DFI-8, M. odoratimimus DFI-22, and S. multivorum DFI-25) showed enzymatic activity against carbapenems (imipenem or meropenem). The frequency of the occurrence of the specific drug resistance had been decreasing in the following order: ampicillin > ESBL > piperacillin > carbapenems. Considering results of the Etest strips, performed tests revealed higher percentage of ampicillin and piperacillin resistant strains compared to the MALDI ones -67% and 50% Gramnegative strains, respectively. In the case of ESBL and carbapenemases-producing bacteria (CPB), the opposite observation had been noted-the share of ESBL strains was lower by 10% and for CPB by 8%. The number of unclear (undefined) results also significantly increases when Etest strips are used, which mostly refers to ESBL detection-cefotaxime 6/42 (14%) and other cephalosporins: ceftazidime 4/42 (10%), cefepime 3/42 (7%), and ceftriaxone 1/42 (2%). Unlike to MALDI results, the highest number of ESBL activity was detected using ceftriaxone which covered two-thirds of all positive strains based on  the Etest strips test. In the case of other cephalosporins, the percentage of positive results ranged from 33 to 42% mostly due to the high rate of undefined results. Regarding specific antibiotics, the far most significant number of discrepancies between the two methods used were noted for cefotaxime -16, which resulted from a higher rate of negative and undefined outcomes of Etests. Contrary to this, ampicillin hydrolysis detection was characterized by the highest complies level-~ 76%. Regarding bacterial species, the most different results were observed in the case of P. aeruginosa strains and Citrobacter and Acinetobacter genera members. Numerous discrepancies were also noted for M. odoratimimus and S. multivorum, which could be associated with their high enzymatic activity leading to undefined results in the Etests.
Screening for antibiotic-resistance genes revealed that only ten strains possessed one or more of the beta-lactamase-encoding genes tested (Table 4). They all refer to ESBLs representing Ambler class A beta-lactamases-TEM, SHV, or CTX-M. Most strains carrying resistance genes were E. coli strains (7) that most frequently had TEM (5 isolates). Regarding other genes, SHV occurred three times -in K. oxytoca DFI-8, E. coli DFI-17, and K. pneumoniae DFI-21, while CTX-M-9 2 times -E. coli DFI-4 and -5. Interestingly, only in 4 cases were ESBL genes detected for strains that demonstrated resistance according to MALDI or Etest assays -E. coli DFI-4 possesses CTX-M-9, E. coli DFI-5 with TEM + CTX-M-9 as well as K. oxytoca DFI-8 and E. coli DFI-17 -both with SHV like genes. The rest strains with detected resistant genes demonstrated only activity against ampicillin or piperacillin and mostly had TEM -5 out of 6 isolates. Despite the detection of carbapenemase activity in 4 strains, PCR analyses did not reveal the presence of any of the analyzed genes encoding carbapenemases, that is, VIM, IMP, NDM, GIM, KPC, GES, or OXA-48.

Impact of the Antibiotic Therapy on the Microbiological Outcome of the DFI Patients
According to medical history, 10 DFI patients were not subjected to any antibiotic treatment; nine were receiving lincosamides (clindamycin), five beta-lactams, two fluoroquinolones, and five were receiving multi-antibiotic treatment (combo). Comparison of the species composition of swab samples according to the antimicrobial treatment used revealed differences in the type of microbial species detected that were considered unique to each patient group (Fig. 4a). Venn diagram analysis showed that when a particular type of antibiotic was used, the species composition of the DFI samples changed. The number of species highlighted varied depending on the type of antibiotic used. The highest number of unique microbial species (8) was found in combined antibiotic treatment, including two Candida species, and in patients treated with lincosamides-5. In contrast, samples from patients treated with beta-lactams and fluoroquinolones had fewer specific species, 2 and 1, respectively, but this observation may be due to the lower representation of these groups. Two bacteria, E. coli, and E. faecalis, appeared to be species common to all groups of DFI patients.
The analysis showed that the antibiotic treatment also affected the number of microbial species presents simultaneously in the sample; samples from treated patients had a higher percentage of samples with four or more microbial species-57% versus 44% (Fig. 4b). Such phenomenon was particularly evident in samples from patients on combination antibiotic therapy, where the percentage of samples containing > 3 species in the sample reached 80%. Considering the prevalence of a particular type of bacteria, antibiotic therapy generally increased Gram-positive bacteria, except for lincosamides, where the opposite trend was observed. A tremendous increase in Gram-positive bacteria was observed with beta-lactams-80% compared to 52% in the untreated group. Samples from treated patients were generally characterized by a higher ESBL share -44% compared to 31% in the non-treated patients, but a slightly lower percentage of the ampicillin and piperacillin-resistant strains occurrence. The highest percentage of drug-resistant Gram-negative bacteria was observed in DFI patients receiving combination antibiotic therapy, with 75% resistant to ESBLs, ampicillin, and piperacillin. Green-positive results, purple-negative results, blue-unclear results, CTX-cefotaxime, CRO-ceftriaxone, CAZ-ceftazidime, CEP-cefepime, IMI-imipenem, MER-meropenem, AMP-ampicillin, PIP-piperacillin, ESBL-extended spectrum β-lactamases, Carb.-carbapenemases Table 4 List of strains in which the tested resistance genes were detected using established multiplex PCR protocols with marked resistance according to MBT STAR BL and Etest

Discussion
Optimal treatment of the infection depends on accurately identifying the microorganisms present and applying appropriate antimicrobial treatment. Failure to adequately treat infection in diabetic foot ulcers leads to progressive tissue damage, impaired wound healing, and serious complications [28]. Clinical practice of DFI diagnosis has relied chiefly on cultivation-dependent methods, which show bias towards microorganisms that thrive under isolation procedures and can grow well on laboratory culture media. Therefore, they often overlook slow-growing, fastidious, anaerobic, and unknown pathogens, which delays the appropriate treatment [29]. In the literature, we can find numerous examples of studies using the traditional culture method, where 46-85% of DFI cases were monobacterial [30,31] with only a minority being polymicrobial infections [32].
On the other hand, there are also papers in which authors have indicated that DFIs are more often the result of polymicrobial infection with complex bacterial communities (microbiome) that impede wound healing [33]. More recent advances in molecular biology technologies have helped to overcome obstacles accompanying traditional methods providing new insights into the bacterial diversity of DFI and have confirmed that chronic wounds, including diabetic foot ulcers, have a polymicrobial nature instead of being colonized by a single species [34,35]. Price et al. [36] found that culture-based method revealed only nine bacterial families compared to 44 denoted using 16S rRNA sequencing which may be the reason for the high prevalence of monomicrobial infections detected by traditional culture. Nevertheless, it should be noticed that molecular approaches are limited by amplification biases, namely, by the primer choice affecting the amplification efficiency of different microbial phyla, as well as by the quality of extracted DNA which depends on the microbial taxa [37]. The optimal culture conditions selected in our previous study [8] enabled revealing a large diversity of bacterial families involved in the development of DFI including rarely detected in DFI Pasteurellaceae, Fig. 4 Effect of the antibiotic treatment on the microbial profiles of the swab samples of the DFI patients. a-Venn diagram showing unique/common bacterial species depending on the antibiotic type used. b-influence of the treatment on the number of species per patient, Gram-type ratio as well as occurrence of the antibiotic resist-ance (ESBL-extended-spectrum beta-lactamase, AMP-ampicillin resistance, PIP-piperacillin resistance). Resistance against carbapenems was not presented due to poor representation among investigated samples Sphingobacteriaceae, Flavobacteriaceae, Planococaceae, and Peptoniphilaceae.
Contrary to popular belief that cultures have a high false-negative rate and lack full representation of the total microbial population in wounds, especially in terms of the pathogenic burden [38,39], culture-based methods can still play an essential role in patient management providing that modern culturomics approach with rapid microbial identification via MALDI technique is applied. As shown in our study, the simultaneous use of culture media sets of different types (universal, selective/differentiating) allows the isolation of fast-growing bacterial species as well as fastidious ones represented both Gram-positive and -negative type of bacteria in short time-to-results. Furthermore, the MALDI technique application assured high identification confidence by comparing species level with 16S rDNA sequencing -93%. To date, only two papers have been published in this field regarding DFI research-our previous work concerning selection culture conditions [8]. and work Jneid et al. [40]. In the second case, authors found a high prevalence of polymicrobial infections (88.3%) and high biodiversity (53 known and 19 unknown bacterial species). In addition, the culture conditions used allowed the isolation of species commonly found in DFI (mostly S. aureus, Enterococcus faecalis, Enterobacter cloacae) as well as the rarest species, such as anaerobic Finegoldia magna. Both studies mentioned above have proven that culturomics does work as a solution to address the limitations of conventional culturing, that is, increase the throughput of identifications and species coverage as well as play a complementary role concerning molecular methods in the exploration of complex microbiota in DFIs.
Revealed high frequency of polymicrobial infections is of utmost importance for patient management since microbial interactions may synergize the pathogenic potential of one or other microorganism, hampering their eradication and further controlling chronic wounds [38]. Liu et al. [4] hypothesized that individual bacterial species may not be able to maintain a pathogenic biofilm independently. However, pathogenic biofilm formation may occur in a symbiotic polymicrobial community in the DFU. Therefore, although most of our studies, Staphylococcus spp., and Corynebacterium spp., are considered part of healthy skin's normal microbiota, they may contribute to a pathogenic community of DFI. Several studies have highlighted the importance of CoNS and Corynebacterium spp. as potential pathogens of DFI and stress their importance concerning chronic wounds, especially in the case of patients with impaired immune responses such as diabetes Our studies revealed a significant share of bacteria belonging to Enterobacteriaceae -24.3% of all identified bacterial families and correlated with other culture-based studies that reported a high incidence of Enterobacteriaceae members in moderate to severe diabetic foot ulcers [38,41]. The predominance of the Enterobacteriaceae family has recently been reported as the largest group of aerobic Gram-negative rods in DFIs [42]. A shift towards the presence of enteric types of bacteria in the recurrent wound may be a result of self-colonization from another body site, e.g., gastrointestinal tracts, which produced a corresponding decline in wound healing since many of such Gram-negative isolates may also be multidrugresistant which makes them very difficult to eradicate with antibiotic therapy. Indeed, drug resistance analysis among isolated Gram-negative bacterial strains showed a relatively high prevalence of ESBL (52% of isolates and 31% of patients) and carbapenemase-producing bacteria-19% of all Gram-negative isolates. Because the increasing severity of prevention and treatment of diabetic foot ulcer infections is associated with high rates of detection of multidrug-resistant bacteria, it is crucial to focus on assessing risk factors for infection with multidrug-resistant bacteria to find more effective treatments [42,43]. Yan et al. [43] during the analysis of risk factors for multidrug-resistant organisms in diabetic foot infection among 180 patients from the Hospital of Jiangnan University (Wuxi Area), noted that 104 of all 182 isolated strains were multidrug-resistant bacteria (66 strains of Gram-negative bacteria and 38 strains of Gram-positive bacteria). In addition, the authors noted that antimicrobial use in the past 3 months was associated with multidrugresistant bacterial infections in patients with diabetic foot ulcers (P < 0.05).
The microbial load, diversity, and presence of pathogenic organisms in the DFU are known to change in response to antibiotic treatment [44]. It has been noted that the wound microbiota of patients treated with antibiotics is significantly different from that of untreated patients. However, no clear distinction has been made between problematic bioburden and benign colonization, which would be clinically relevant to antibiotic treatment decisions [44,45]. Our results showed that the use of antibiotic therapy by patients induces changes in the microbial composition and frequency of species in the wound microbiota, including the gram-type ratio and the frequency of drug resistance. The proportion of Gram-negative bacteria decreased while the antibiotic resistance rate increased. This observation of combination antibiotic therapy was remarkably accurate, indicating the highest number of unique microbial species and the highest ratio of drug-resistant strains. The resistance rate of E. coli was the highest among Gram-negative bacteria, consistent with previous reports. This founding may also be related to the fact that E. coli was also reported as the most common Gram-negative bacterium, as in many other reports, e.g. Tascini et al. [46].
Infection with MDR bacteria in DFU reduces the clinical effect of antibiotic therapy. Our study indicates that empirical antibiotic therapy for DFI should pay particular attention to the risk assessment of Gram-negative bacteria infection, where the susceptibility pattern of Gram-negative bacteria should be regularly monitored in DFI. Nevertheless, many clinics rely on traditional culture and conventional biochemical tests, like strip test, that underestimates wound flora and may lead to inappropriate antibiotics prescribed in up to 45% of cases [47]. The excessive or inappropriate use of antibiotics not only results in ineffective treatment but also aggravates the worldwide crisis of antibiotic resistance [38]. This problem could be solved by using the latest, more adequate, and rapid drug resistance assays, such as those based on the MALDI TOF MS technique. Our study showed that using the MBT STAR BL assay increased the percentage of ESBL-and carbapenemase-producing bacteria detected compared to Etest strips. Additionally, the use of MBT STAR BL is accompanied by a significantly lower rate of unclear results. More and more researchers, including Noster et al. [48], emphasize that the MALDI technique is increasingly embraced for detecting antimicrobial resistance and will likely become an essential part of the routine laboratory soon. Although the disadvantage of this approach is that it only detects resistance conferred by hydrolysis of the target antibiotic, it has high sensitivity and specificity (98-100% and 97-100%, respectively), as well as a relatively short turnaround time-usually 30 min for typical Enterobacterales, and even shorter if an appropriate protocol modification is used, as demonstrated by Złoch et al. [21]. As our studies showed, the application of the MALDI approach could be more feasible for routine drug resistance detection among DFI isolates than molecular technique, such as multiplex PCR reactions, since the latter required expanded knowledge about the taxonomical affiliation of the isolates for suitable primers set designing.

Conclusions
Reliable deciphering of the composition of the wound microbiome in patients with DFI is crucial for subsequent effective therapy. Given the large number of microbial species that may be involved in the development of infection, especially in moderate to severe chronic wounds, practical diagnostic tools should be characterized by accurate identification, short time-to-results as well as the ability to rapidly detect drug resistance in the face of a growing global problem such as MDR bacteria. Such criteria are met by fast MALDI identification combined with multiple culture conditions and rapid detection of antibiotic resistance via MBT STAR BL assay. As demonstrated in our study, this method provides identification information at a level comparable to that obtained from DNA sequencing, allows the isolation of both common bacterial species and those considered rare, including fastidious ones, and is effective in antibiotic detection, especially for that of particular concern like ESBLs and carbapenemases. Application of this technique may help to understand the role of the complex microbiota in the development of DFI in the context of the antibiotic therapy used by patients and its impact on the development of drug resistance. Moreover, our results indicate that culture-based methods can still be essential to routine clinical diagnosis, providing the clinician with relevant information in a reasonable time.

Conflicts of interest
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Ethical Approval
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Nicolaus Copernicus University in Toruń, Collegium Medicum in Bydgoszcz (KB 68/2019, 29 January 2019).

Consent to Participate
Informed consent was obtained from all individual participants included in the study.

Consent fo Publish Not applicable.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.