Snapshot of Phenotypic and Molecular Virulence and Resistance Profiles in Multidrug-Resistant Strains Isolated in a Tertiary Hospital in Romania

A current major healthcare problem is represented by antibiotic resistance, mainly due to multidrug resistant (MDR) Gram negative bacilli (GNB), because of their extended spread both in hospital facilities and in the community’s environment. The aim of this study was to investigate the virulence traits of Klebsiella pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa MDR, XDR, and PDR strains isolated from various hospitalized patients. These GNB strains were investigated for the presence of soluble virulence factors (VF), such as hemolysins, lecithinase, amylase, lipase, caseinase, gelatinase, and esculin hydrolysis, as well as for the presence of virulence genes encoding for VF involved in adherence (TC, fimH, and fimA), biofilm formation (algD, ecpRAB, mrkA, mrkD, ompA, and epsA), tissue destruction (plcH and plcN), and in toxin production (cnfI, hlyA, hlyD, and exo complex). All P. aeruginosa strains produced hemolysins; 90% produced lecithinase; and 80% harbored algD, plcH, and plcN genes. The esculin hydrolysis was detected in 96.1% of the K. pneumoniae strains, whereas 86% of them were positive for the mrkA gene. All of the A. baumannii strains produced lecithinase and 80% presented the ompA gene. A significant association was found between the number of VF and the XDR strains, regardless of the isolation sources. This study opens new research perspectives related to bacterial fitness and pathogenicity, and it provides new insights regarding the connection between biofilm formation, other virulence factors, and antibiotic resistance.


Introduction
The ability of a bacterial strain to both colonize the host organism and trigger a disease is related to its pathogenic potential, which is enhanced by the expression of one or more virulence factors [1,2]. The clinical presentation of an infectious disease depends on these major properties of a pathogen (pathogenicity and virulence), but also on several host-related factors (such as age, immune status, diet, and other environmental conditions). Different aggressins (enzymes, invasins, toxins, and secretion systems), adhesins (pili, fimbriae, and EPS components), siderophores, and the generation of biofilms are often included among the most studied virulence factors of the potentially pathogenic bacteria. Some virulence genes are chromosomally encoded, i.e., some adhesins, but others are located in plasmids or in other mobile genetic elements. Therefore, most of the AR almost all antibiotic classes [21]. This peculiarity represents a major epidemiological issue, especially in the era of the emergence of new viruses, such as COVID-19, which may infect the human host together with a bacterium. The attention of the researchers falls once again over a series of virulence associated factors, such as OmpA, which are often connected to β-lactam resistance [21].
An important epidemiological feature of these three opportunistic pathogens, which are members of the ESKAPE group, is their ability to colonize and then to cause an infection in hospitalized patients. This process is due to both the presence of resistance and to virulence factors synthesized in a manner dependent on the cellular density, as well as the mechanism of quorum sensing (QS), which protects bacterial cells (favoring their multiplication and host colonization) against antimicrobial substances that are part of the natural non-specific defense mechanisms of the host [22,23].
The aim of our study was to investigate the profiles of both the resistance and virulence of some selected MDR strains belonging to clinically significant opportunistic pathogens such as Klebsiella pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa, often involved in healthcare-associated infections.

Bacterial Strains
Bacterial strains from various biological samples (urine, blood, bile, wound, bronchial aspirate, rectal swab, drainage tube, ascitic fluid, and stool) of hospitalized patients were isolated from units with a high epidemiological risk from the Fundeni Clinical Institute during October 2019-October 2020. All of the strains were identified by an automatic method (BD Phoenix Vitek 2 Compact) and further investigated for antibiotic resistance [24]. The multidrug-resistance profile was established according to the international guidelines [25]. Furthermore, the expression of several virulence factors was investigated in a selected set of 72 MDR, XDR, and PDR strains of K. pneumoniae (n = 51), A. baumannii (n = 10), A. lwoffii (n = 1), and P. aeruginosa (n = 10) with phenotypic and molecular methods. Escherichia coli ATCC 25922 and P. aeruginosa ATCC 27853 were used as the reference strains.

Evaluation of Antibiotic Susceptibility
The antibiotic susceptibility testing was investigated according to CLSI 2021 (Clinical Laboratory Standard Institute), using disc diffusion and automatic methods (BD Phoenix/Vitek 2 Compact). Phenotypic confirmation for ESBL production was performed through a double-disc synergy test (DDST) (Ceftazidime-Amoxicillin/Clavulanic acid-Ceftriaxone) [26].

Target Gene
Primer Name Primer Sequence Amplicon Size (bp) Primers used for antibiotic resistance genes detection

Evaluation of the Soluble Enzymatic Factors
The virulence phenotypes were assessed by performing enzymatic tests for the expression of some soluble virulence factors, as described elsewhere [27]. Briefly, soluble virulence factors were detected using solid culture media supplemented with 5% sheep blood, 2.5% yolk, 1% Tween 80, 15% soluble casein, 3% gelatine, 1% starch and Fe3+ citrate for the bacterial toxins/enzymes.
The primers used during the PCR reaction are depicted in Table 1.

Statistical Analysis
The statistical analysis was performed with the programs Microsoft Excel 2021, Vassarstats [28], and PSPP 1.6.2 [29]. The obtained frequency values were compared using Fisher Exact Probability Test and chi-square test. The correlations depicted below were established using Pearson' s correlation coefficient in bivariate correlation and a p value less than 0.05 was considered statistically significant.
The tested strains of K. pneumoniae, P. aeruginosa, and Acinetobacter spp. showed high rates of resistance to all of the tested antibiotics (Supplementary Table S1). All of the K. pneumoniae strains were resistant to fluoroquinolones (ciprofloxacin and/or levofloxacin), piperacillin/tazobactam, carbapenems (ertapenem and/or imipenem and/or meropenem), and third-generation cephalosporins (ceftriaxone and cefotaxime); 74.5% were resistant to trimethoprim/sulfamethoxazole; 45.1% of strains were resistant to gentamycin; 29.4% showed resistance to amikacin; and only 17.6% were resistant to ceftazidime/avibactam. The P. aeruginosa strains were resistant to aminoglycosides, fluoro-quinolones, ceftazidime, carbapenems (imipenem and meropenem), while 30% of them displayed resistance to piperacillin/tazobactam. All P. aeruginosa strains were susceptible to ceftazidime/avibactam. The A. baumannii strains were resistant to aminoglycosides, fluoroquinolones, ceftazidime, and carbapenems (imipenem and meropenem), 54.5% were resistant to trimethoprim/sulfamethoxazole, while a lower percent (18.2%) displayed a resistance pattern to ampicillin-sulbactam and minocycline.
The phenotypic virulence analysis of P. aeruginosa revealed the production of haemolysins (100% of the strains), while 90% of the strains isolated from the blood cultures produced lecithinase.
The majority (96.1%) of K. pneumoniae strains hydrolysed esculin producing esculetol, which acted as a pseudo-siderophore, and 25.5% were amylase producers. Only 11.8% of the K. pneumoniae strains were lecithinase producers, while gelatinase was found in 9.8% of them. Here, 31.4% of the K. pneumoniae strains produced lipase. We included, in Table 3, an analysis regarding the probability of encountering some of these VF in several isolation sources. Only caseinase in K. pneumoniae strains was more likely to be source-dependent (p = 0.047). Table 3. Bacterial strains producing soluble enzymatic factors. * N = number of isolates with a virulence factor. Percentages were calculated considering the total number of samples of a specific strain encountered in each isolation source. ** p-values reflected the significance between the differences in distribution of a virulence gene or resistance gene over the source types. Every p-value was obtained using a chi-test and only those values < 0.05 were considered significant. NA = not applicable.
In Supplementary Table S2, we show the virulence and resistance profiles of the identified strains of Acinetobacter spp., P. aeruginosa, and K. pneumoniae. No significant association was encountered between the virulence or resistance profiles and the isolation sources.
Next, we searched for an association between a certain virulence pattern and the isolation source. The association of fimA and mrkA genes was more likely to have a sourcerelated distribution (p = 0.03) (see Table 4).
Because most of the analyzed strains were K. pneumoniae, we continued our correlation analysis on them. Thus, after indicating the source of isolation (urine, blood culture and other sources), we used a bivariate correlation tool to find any connection between the numbers of strains belonging to each resistance phenotype (XDR-extensive drug-resistant; MDR-multidrug resistant; PDR-pan drug resistant) and the numbers of strains with VF or with resistance genes. The categories of resistance were established according to international consensus guidelines [25]. Significant associations were found between the numbers of XDR strains and those with two virulence factors, or those with three virulence genes (Figures 1 and 2). A significant correlation was found between the strains with one VF and those with one resistance gene ( Figure 3). according to international consensus guidelines [25]. Significant associations were found between the numbers of XDR strains and those with two virulence factors, or those with three virulence genes (Figures 1 and 2). A significant correlation was found between the strains with one VF and those with one resistance gene ( Figure 3).

Discussion
Currently, multidrug resistance in GNB is intensely studied because of the need for improved strategies for the treatment of infection caused by these opportunistic pathogens. In addition, the mechanisms related to the pathogenesis of MDR GNB are intensely being studied in order to gain a better understanding of their adaptability, fitness, and spreading rates [18,21].
Our research focused on the virulence traits of some selected strains belonging to three MDR, XDR, or PDR microorganisms, in relation to their phenotypic profile of AR. All of the investigated bacterial strains, and especially K. pneumoniae, showed high resistance levels, in accordance with other studies [30,31]. K. pneumoniae strains mostly carried the blaKPC carbapenemase encoding gene, followed by a smaller percentage of blaOXA-48 and blaNDM genes. Among the ESBLs, the most prevalent ones were blaCTX-M, followed by blaSHV and blaTEM. For A. baumannii, two acquired CHLDs encoding genes were revealed in the analyzed strains: blaOXA-23 and blaOXA-24. The same results were found by Gheorghe et al. (2019) in their study [32]. Most of the strains of P. aeruginosa were positive for blaIMP (Supplementary Figure S4) and blaVIM (Supplementary Figure S5).
P. aeruginosa is one of the pathogens that harbor numerous VF involved in cell invasion and toxicity. Most of its VF also are involved in biofilm formation. One of the most important VFs is an exopolysaccharide, alginate, with a role of adhesion to the cellular substrate and a protective role against certain environmental adverse conditions. There-

Discussion
Currently, multidrug resistance in GNB is intensely studied because of the need for improved strategies for the treatment of infection caused by these opportunistic pathogens. In addition, the mechanisms related to the pathogenesis of MDR GNB are intensely being studied in order to gain a better understanding of their adaptability, fitness, and spreading rates [18,21].
Our research focused on the virulence traits of some selected strains belonging to three MDR, XDR, or PDR microorganisms, in relation to their phenotypic profile of AR. All of the investigated bacterial strains, and especially K. pneumoniae, showed high resistance levels, in accordance with other studies [30,31]. K. pneumoniae strains mostly carried the bla KPC carbapenemase encoding gene, followed by a smaller percentage of bla OXA-48 and bla NDM genes. Among the ESBLs, the most prevalent ones were bla CTX-M , followed by bla SHV and bla TEM . For A. baumannii, two acquired CHLDs encoding genes were revealed in the analyzed strains: bla OXA-23 and bla OXA-24 . The same results were found by Gheorghe et al. (2019) in their study [32]. Most of the strains of P. aeruginosa were positive for bla IMP (Supplementary Figure S4) and bla VIM (Supplementary Figure S5). P. aeruginosa is one of the pathogens that harbor numerous VF involved in cell invasion and toxicity. Most of its VF also are involved in biofilm formation. One of the most important VFs is an exopolysaccharide, alginate, with a role of adhesion to the cellular substrate and a protective role against certain environmental adverse conditions. Therefore, its presence is required for the survival of the microorganism and its multiplication within microbial biofilms [33][34][35]. Among VF, the presence of flagella and pili, exotoxin A, protease, alginate production, biofilm formation, and secretion systems III and IV determine the high pathogenicity of P. aeruginosa strains [36]. These factors seem not to act alone, but in the presence of other enzymes involved in pathogenicity [37]; these soluble components identified in our study were exoenzymes, such as hemolysins, caseinase, and lecithinase. Apart from gelatinase, we could also find lipase, esculinase, and amylase in strains of P. aeruginosa. These virulence factors were also identified in previous studies performed on such strains. Even if amylase, caseinase, and lipase are not literally real virulence factors, but metabolic enzymes (encoded by chromosomal genes), their presence in strains from the Enterobacterales and Pseudomonadaceae families facilitates adaptability inside the host and even the survival of these bacteria in the external environment, thus increasing the risk of transmission. This is a very important trait among the P. aeruginosa strains, which can pass and infect other hosts, thus creating a chain of events that lead to the spread of MDR, with a major impact on the health of the community [38].
LasA was described as the first virulence gene found in P. aeruginosa with an elastase and protease activity [39], while LasB was described as a metalloenzyme secreted by pathogenic strains of P. aeruginosa. A major role of LasB is currently designated in bacterial intercellular communication, and this metalloenzyme is also reported in 90% of the urine specimens [37,[40][41][42]. A study conducted in Teheran, between June 2013 and June 2014, on pediatric patients with cystic fibrosis, revealed an increased prevalence rate in virulence genes in P. aeruginosa, of over 63.1% for toxA and 95.4% for lasB [43]. Most of the isolated strains in this study had a combination between many virulence genes, especially between plcH, plcN, exoS, exo U, and algD ( Supplementary Figures S6-S8). The phenotypic expression of these genes consists of the destruction of tissues (plcH, plcN, exoS, and exoU) or in the cellular adhesion and formation of biofilms (algD) [44]. All these features promote the spreading process and the complex development of biofilms, which, in turn, are substantial factors in the development of multidrug resistance [45]. These intricate relations could be a subject for further study.
The virulence genes are different among the intestinal colonizing K. pneumoniae strains, compared with those from other sources [37]. This might be influenced by the fact that several environmental characteristics influence the adherence and survival of these strains, thus leading to the development of complex virulence systems [46]. Emphasizing that K. pneumoniae represents one of the most important opportunistic pathogens, this microorganism is frequently involved in urinary and respiratory tract infections, as well as sepsis [47]. The pathogenicity and virulence of K. pneumoniae strains are mainly due to the presence of the capsule, with the most common strains being those with the capsule type K1, K2, K5, K54, and K57, as well as K20 (which has two virulence factors, rmpA and wcaG) [48,49]. It was observed that the virulence gene wcaG is involved in the production of the capsular fucose, associated with K1 and K54 strains, while the rmpA gene was identified in strains of Klebsiella spp. type K2 [50][51][52]. However, the presence of a capsule was not the only contributor to virulence, with several studies underlining the association of various virulence factors in the persistence of this microorganism [53]. In our study, the majority of K. pneumoniae strains were positive for more than two virulence factors.
On the other hand, many VF contribute to a series of mechanisms that have a complex pathogenic response, such as biofilm formation. In this study, we emphasized that most of the encountered VF in all three mainly researched microorganisms were involved in biofilm formation (such as mrkA, ecpRAB, and algD) (Supplementary Figures S9 and S10) and adherence (fimA) (Supplementary Figures S11-S14). We also identified several correlations between the number of virulence genes and antibiotic resistance, depicted as a repetitive pattern in all of the isolation sites. However, the combination between two or more virulence genes was not newly discovered among MDR K. pneumoniae. The association between the ecp complex and the fim system was, for instance, observed in other strains of K. pneumoniae isolated from intra-hospital infections [54]. This probably brought a new mechanism of adaptability and biofilm formation for these bacterial strains, in order to increase their survival rate under any circumstances. However, the observed effect might be perilous for the health of the entire community because it might lead to the spread of MDR organisms. In our case, because of the increased number of genes involved in biofilm formation, all correlations need further study because of several interference factors that contribute to the achievement of multidrug-resistant genes/mechanisms. There are studies that emphasize the connection between biofilm formation and antibiotic resistance through various mechanisms [55]. Therefore, this subject needs further analysis.
In bacterial biofilms, tolerance to different classes of antibiotics is due to the different mechanisms involved, such as, the presence of the ndvB and brlR genes for the aminoglycosides, the efflux pumps for the polymyxin E, and porins (MexAB-oprM and MexCD-oprJ) for beta-lactams. Antibiotic tolerance is also given by the impermeability of the matrix secreted by the bacterial cells inside the biofilm. In fluoroquinolones, antibiotic tolerance appears because of the decrease in oxygen concentration, the low metabolic activity/dormancy state, or the activity of the efflux pumps [56]. A major part in biofilm formation is also played by the QS complex [39]. Three QS systems are known in P. aeruginosa: the most important is LuxI/LuxR, which controls the expression of virulence factors, but a major part is also played by the third non-LuxI/LuxR-type system called the Pseudomonas quinolone signal system (PQS). In the first phase, LuxI/LasI synthesizes the 3OC12HSL lactone detected by the cytoplasmic LuxR homologue LasR. The LasR-3OC12HSL complex activates the transcription of target genes, including those encoding VF, such as elastase, proteases, and exotoxin A. One of the targets of LasR-3OC12HSL is lasI; another target is a second luxI homolog called rhlI (synthesizes a second AI, butanoyl homoserine lactone-C4HSL). At high concentrations, this AI binds to RhlR, a second LuxR homolog, RhlR-C4HSL, which activates target genes, including those encoding elastase, proteases, pyocyanin, and siderophores. Among its targets is rhlI, which leads to autoinduction of the second QS system [56]. It is, therefore, already known that biofilms are a result of complex interactions both in Klebsiella pneumoniae and in P. aeruginosa.
Another pathogen that is also an important biofilm producer is A. baumannii [57], an opportunistic pathogen with major implications in HAIs, responsible for outbreaks and severe infections especially in patients with different immunodeficiencies. Although poor in the production of soluble VF (in our study, almost all of the strains produced only one VF-ompA) (see Supplementary Figure S15), it proved to be an important nosocomial pathogen due to the formation of biofilms (especially on medical devices) and due to its MDR, XDR, or PDR phenotypes [30,32,58,59]. Compared with the soluble VF reported for other Gram-negative bacteria, a small number of these factors have been identified for A. baumannii, in comparison with K. pneumoniae and P. aeruginosa. In a study performed by Zeighami et al. (2019) on a significant number of A. baumannii strains, it was observed that most of the strains that carried virulence genes were involved in bacterial biofilm development [60]. Among them, 98% simultaneously showed several virulence genes; the highest prevalence was represented by csuE (100%) and ompA (81%) genes. The severity of infections produced by pathogenic strains of A. baumannii and P. aeruginosa affecting both the respiratory system (pulmonary ventilated patients) and the teguments (patients with a compromised epidermal barrier) was due to the presence of QS mechanisms [59][60][61][62]. Gheorghe et al. also reported only strains positives for the ompA gene [32].
Numerous recent studies carried out in Romania concerning AR had a major impact, both clinically and especially epidemiologically, due to the high spreading capacity of the resistance genes among Enterobacterales [63]. One of the most recent studies (Tîrgu-Mures , , 2018) performed on 19 strains of carbapenemase producing Enterobacterales showed that the resistance genes detected by PCR were bla KPC , bla OXA-48 , and bla NDM [64]. Gavriliu et al. (2016) mentioned, in a retrospective study conducted in "INBI Matei Bals , ", a lower resistance level of K. pneumoniae strains isolated from blood cultures, especially to fluoroquinolones, aminoglycosides, and carbapenems, and to third generation cephalosporins, compared with our country's results reported in EARS-Net [65].
In Romania, Gheorghe et al. (2014), reported for the first time the presence of the bla VIM-4 gene in P. aeruginosa clinical strains (involved in acquired resistance mediated by class I integrons) [66]. A year later, Lixandru et al. described for the first time the presence of the bla VIM-1 and bla KPC-2 genes in strains of K. pneumoniae, highlighting the urgent need for measures to combat and prevent intrahospital infections [67]. Two years later, , mentioned for the first time the presence of the bla OXA-72 gene in A. baumannii strains, responsible for plasmid-mediated resistance through plasmid pAB120 [68].
Only a few papers have been published in Romania on the topic of antibiotic resistance and virulence; for instance, Delcaru et al. (2017) investigated these associated features in UTI strains; their study revealed a high ability of the UTI strains to adhere to HEp-2 cells. The Gram-negative UTI strains produced pore-forming toxins (hemolysins, lecithinase, and lipase), proteases, amylase, and siderophore-like molecules resulted from the process of esculin hydrolysis. In their opinion, searching for the correlation between resistance and virulence factors could provide information useful for anticipation of the clinical evolution and risk of UTI recurrence [22].
Our study depicted numerous carbapenem and ESBL-encoding genes, especially in K. pneumoniae, most of them being encoded by β-lactamases of class A (KPC, TEM, SHV), B (NDM) and D (OXA-48). This complex mixture was reflected in the resistance profile of the identified strains, most of them being classified as either PDR or XDR. These strains also had several VF, most of them being involved in biofilm formation or tissue destruction. However, despite very careful selection of the analyzed strains, the major limitation of our study was the reduced number of isolates (being selected strains with epidemiologically important resistance profiles), which might influence the statistical analyses.
Studies similar to this one can help gain a better understanding of the adaptability of a pathogen to the host environment. Finally, a good adaptation is as a result of a well-balanced distribution of energy resources necessary for growth, multiplication, and colonization; virulence factors expressed in a cell-density-dependent manner; and the pathogen's survival in opposition with innate and adaptive immune system effectors, eventually with antibiotics too. This network of processes modulate the reproductive fitness of the pathogen and influence the community-acquired infections and HAIs evolution; therefore, gaining a better understanding of them will help with the development and implementation of appropriate prevention, diagnosis, and treatment actions [69].

Conclusions
Our study concerning the resistance and virulence profile of carefully selected MDR, XDR, and PDR strains of Gram-negative bacilli from the ESKAPE group (Klebsiella pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa) revealed a high repertoire of ARGs and VMs. The most frequently expressed virulence factors were those involved in host cell/tissue destruction and biofilm formation. Such phenotypes are of interest due to their fitness, adaptability, and capacity of dissemination, both in the environment and in the host. This study, which focused on the genotype−phenotype association of resistance and virulence (less approached in Romania), could contribute to a better understanding of bacterial pathogenesis in patients from intensive care units (where biofilm formers and antibiotic-resistant strains spread). The results of this study show that it is imperative to continue monitoring antibiotic resistance, which is often associated with virulence factors. The careful surveillance of both the environmental and host conditions in which the bacteria/microorganisms localize themselves, grow, and colonize is also of great importance because it helps in establishing the pathogenic characteristics for a better prevention and treatment policy regarding healthcare-associated infections.