An unequivocal superbug: PDR Klebsiella pneumoniae with an arsenal of resistance and virulence factor genes

Introduction: Infections caused by extensively-drug resistant (XDR) and pan-drug resistant (PDR) Klebsiella pneumoniae represent an emerging threat due to the high associated mortality. This study aimed to characterize two carbapenem resistant K. pneumoniae strains from the same patient, the first being PDR (referred to as IMP 1078b) and the second being XDR (referred to as IMP 1078s) isolated from the same patient. Methodology: Antimicrobial susceptibility testing was done for the 2 K. pneumoniae isolates, followed by carbapenem/β-lactamase inhibitor combination assay, and fitness cost against cefepime and meropenem. Then, whole-genome sequence analysis was performed to decipher the molecular mechanisms behind the high level of resistance recorded in both isolates. Finally, qRT-PCR was done for β-lactam resistant genes. Results: This is the first report about a K. pneumoniae isolate harboring 47 antimicrobial resistance genes and having type IV pilli (Yersinia) and the fimbrial adherence determinant Stb (Salmonella) as virulence factors. Further analysis on both isolates are discussed within the article. Conclusions: The co-existence of a high number of antimicrobial resistant (AMR) genes and virulence factor genes may lead to a life threatening invasive and untreatable infection.


Introduction
Healthcare associated infections, such as pneumonia, catheter associated blood stream infections, surgical site infections, and catheter associated urinary tract infections caused by resistant Gram-negative bacteria such as Klebsiella pneumoniae are increasing worldwide [1]. Their burden is particularly heavy in the critically ill patients where there is an association between infection with such multidrug-resistant (MDR) organisms and poor outcomes [2,3]. Carbapenems are the mainstay of treatment for infections with extendedspectrum beta-lactamase producing (ESBL) K. pneumoniae [4]. However, some strains have acquired resistance against these antibiotics, leaving colistin as the only treatment option [5].
Infections caused by antibiotic resistant bacteria are increasing worldwide. Each year the U.S. reports more than 2.8 million infections with antibiotic resistant bacteria. This led the Centers for Disease Control and Prevention (CDC) to publish an antibiotic resistance threat report in 2019, classifying carbapenem-resistant Enterobacteriaceae (CRE) as an urgent threat [6]. In Lebanon and the Middle East North Africa (MENA) region, the rates of Gram-negative resistance are very high [7][8][9]. With the raging conflicts in the MENA region, especially the Syrian conflict and due to the transfer of patients from field hospitals at the Lebanese-Syrian border to hospitals within Lebanon, there has been a noticeable increase in the rates of Gram-negative resistant organisms in peripheral towns and later in central Lebanese hospitals. Here we report a case of two K. pneumoniae strains that were recovered from a patient who was initially hospitalized at a peripheral hospital at the Lebanese-Syrian border. We therefore aimed at testing the susceptibility of the isolates against a battery of antibiotics used in clinical settings and determining phenotypic and genotypic mechanisms of resistance of these isolates using whole genome sequencing.
Following a motorcycle accident, a 22 year-old Lebanese man sustained severe trauma in his hometown close to the Lebanese-Syrian border. Unconscious, he was taken initially to a peripheral hospital where he was intubated for mechanical ventilation and a central line was inserted and admitted to the ICU. After 6 days in the other hospital, during which he received piperacillin-tazobactam, vancomycin and dexamethasone, he was transferred to a tertiary care hospital in Lebanon for continuity of care. Upon admission, he was afebrile and comatose. Workup revealed a subarachnoid hemorrhage and brain contusions with surrounding edema and multiple closed fractures of the extremities, chest subcutaneous emphysema, small pneumopericardium and pneumomediastinum. The old central line was discontinued and a new one inserted. Cultures from blood, urine and deep tracheal aspirates were taken in addition to skin screening as per standard screening protocols for ICU transfers at our institution. Those cultures later grew carbapenem-resistant Klebsiella pneumoniae from the deep tracheal aspirate (DTA) and the skin (sensitive to tigecycline, intermediate to colistin and fosfomycin). The following day the patient developed a high-grade fever. He was started on piperacillin-tazobactam and vancomycin after removing the newly inserted central line and sending appropriate cultures. He remained febrile for several days with evidence of a left lower lobe pneumonia prompting changes to the antibiotics he was receiving based on the initial screening cultures: the DTA culture grew carbapenem-resistant K. pneumoniae and E.coli both sensitive to tigecycline, with the K. pneumoniae being intermediate to colistin and resistant to fosfomycin. A week following admission, he was persistently febrile, therefore new blood cultures were taken and the sample from the central line grew K. pneumoniae after 16 hours (sensitive to tigecycline, resistant to colistin and fosfomycin). The peripheral blood cultures remained negative. The patient received inhaled colistin and later inhaled amikacin; carbapenems were discontinued as the minimum inhibitory concentrations (MICs) to these agents were all greater than 32. Despite the infection with panresistant organisms, our patient's condition improved. The bacteremia was related to the central line and it resolved as soon as the line was discontinued, which is essential with Gram-negative rod (GNR) central lineassociated bloodstream infections (CLABSI). He became afebrile with marked neurologic and clinical recovery and was extubated, and transferred to the regular floor, with eventual discharge home. The Supplementary Table1 lists all the different cultures and results. The Supplementary Figure1 shows the timeline of different antibiotic administration.

Methodology
Ethical approval was not required as clinical isolates were collected and stored as part of routine clinical care. Clinical isolates and patient records/information were anonymous and de-identified prior to analysis.

Identification of the isolates
The recovered isolates in culture were identified using the Matrix-Assisted Laser Desorption/Ionization Time of Flight (MALDI-TOF) system (Bruker Daltonik, GmbH, Bremen, Germany) with a score of green flags.

Broth Micro-dilution assay
Broth microdilution was done against 19 different antibiotics from different families. Serial dilution took place between columns 1 and 11 to have concentration ranging from 2048 μg/mL to 2 μg/mL. Half of the wells in column 12 were used as a positive control and the other half as a negative control. For each isolate a bacterial suspension of 0.5 MacFarland was prepared, followed by dilution to reach a concentration of 5 × 10 6 CPU/mL. This was followed by adding 10 µL of the latter into all the well between columns 1-11, and in the positive control designated wells in column 12, ending with a final volume of 100 µL in all the wells. The plate was then placed in the incubator at 37 °C for 18 hours after which the negative control was checked to ensure the absence of contamination. The positive control was checked to ensure that the bacterial suspension was properly prepared, and growth took place. Wells 1-11 were checked for bacterial growth, the well preceding the first well with bacterial growth, was referred as the well containing the MIC. Experiments were run in duplicates for each bacterial isolate. The results were interpreted according to the CLSI M100 guideline [10].

Disk Diffusion
The experiment was performed using the Kirby-Bauer technique. For each isolate, a bacterial suspension equivalent to 0.5 MacFarland was prepared. Then it was subcultured on a round Mueller-Hinton agar plate, in all the directions, to ensure that the bacterial suspension covered all the plate using a sterile swab. The plate was left for around 10 minutes closed on the bench, followed by the addition of the 24 tested antibiotics (8 per plate). The plate was then incubated at 37 °C for 18-24 hours after which the zone of inhibition diameters were measured and the results were interpreted according to the CLSI M100 guideline [10].

Fitness Cost assay
The tested isolates were first subcultured on MacConkey agar and incubated at 37 °C for 18-24 hours. The next day, a loop full of each bacterial isolate was transferred into 10 mL of sterile cation adjusted Mueller-Hinton broth and incubated at 37 °C for 18-24 hours. Then, the turbid inoculated broth of each isolate was diluted at a 1:1000 ratio. The latter was then transferred into 4 distinct wells (200 µL each) of a 96 well microtiter plate. The replication rate of each tested isolate was measured using a densitometer (OD 600 nm) for 16 hours with reads at 30 minutes intervals. The results were then averaged, normalized, and plotted against the K. pneumoniae (DSM ® 30104) [11].

Carbapenem/β-Lactamase Inhibitor Combination assay
Following the MIC determination of both isolates against carbapenems, Meropenem/ β-lactamase inhibitor combinations experiment was performed by adding fixed concentrations of the inhibitors to the experimental wells of a standard antimicrobial broth microdilution assay. We followed CLSI guidelines in this assay. However, minor modifications to broth volumes were made in order to accommodate for the presence of the β-lactamase inhibitors (βLIs) while keeping the concentrations of the meropenem and bacterial suspensions in accordance with CLSI recommendations. For isolates harboring bla OXA -type carbapenemases, Avibactam (MedChem Express, Monmouth Junction, NJ, United States) was used as the βLI at a fixed concentration of 4 μg/mL. However, for isolates that harbored bla NDM , ethylenediaminetetraacetic acid calcium disodium salt (calcium-EDTA) (Sigma R, St. Louis, MO, United States) was used as the βLI at a fixed concentration of 32 μg/mL. In addition, both isolates were tested against both βLIs at their aforementioned fixed concentrations without the addition of meropenem in order to rule out any antibacterial activity exhibited by the inhibitors on the tested isolates. The MICs of the 4 tested isolates were interpreted according to the CLSI M100 guideline [10]. Escherichia coli 1176 (harbors bla NDM-1 only) and E. coli 57 (harbors bla OXA-48 only) were used as a control in the experiment [12].

Whole Genome Sequencing (WGS)
To prepare whole-genome sequencing libraries, the cryopreserved stocks were grown on MacConkey agar. Genomic DNA was extracted using standard methods (Qiagen, Valencia, CA), and NexteraXT libraries were prepared using the manufacturer's protocols (Illumina, San Diego, CA) and sequenced on an Illumina HiSeq 4000, 2 × 150 bp.

Screening results
Two K. pneumoniae isolates were recovered from the patient. The cultures led to the isolation of a K. pneumoniae isolate from the blood (IMP 1078b). Moreover, the skin screening led to the isolation of the second K. pneumoniae isolate (IMP 1078s). The 2 isolates were identified using MALDI-TOF mass spectrometry and later confirmed by WGS.

Antibiotics Susceptibility Testing
The antibiotic susceptibility testing results done by both broth micro-dilution assay (Table 1) and Kirby-Bauer technique ( Table 2) showed that the IMP 1078s is XDR since it was resistant to all the tested antibiotics except for colistin. However, the IMP 1078b is PDR since the isolate was resistant to all the tested antibiotics.
A plethora of virulence genes were detected in both isolates, including: Type I and III fimbriae, serum resistance loci, anti-phagocytic genes, iron acquisition system (Salmochelin and Yersiniabactin), rcsAB gene, and acrAB efflux pump gene (Table 4). Furthermore, we hereby report the first type IV pilli (Yersinia) pilW and the fimbrial adherence determinant Stb (Salmonella) in K. pneumoniae.

Fitness cost results
To assess the fitness cost of harboring AMR genes on the clinical K. pneumoniae isolates, growth kinetics assays was performed. K. pneumoniae DSM 30104 was used as a control strain, as it is the Wild Type. The growth rates did not vary significantly for IMP 1078b (p = 0.9946) nor IMP 1078s (p = 0.1860) when compared to K. pneumoniae DSM ( Figure 2). These observations were made when each of the three isolates were grown in unmodified LB broth. To assess the effect of meropenem and cefepime on the fitness cost of both isolates, we evaluated the effect of exposure of the bacteria to antibiotics on the fitness cost of the clinical isolates and the wild type strain. We used the breakpoints of the selected antibiotics, 4 μg/mL for meropenem and 16 μg/mL for cefepime, as the concentration to grow the bacteria. We witnessed that for IMP 1078b ( Figure 3A) and IMP 1078s ( Figure 3B), the growth rates did not change significantly when comparing the division of the bacteria incubated with meropenem to that of the bacteria grown in broth alone (p = 0.2510 and p = 0.7728 respectively). Furthermore, as seen in Figure 4A and Figure 4B, similar results could be observed for the growth rates of these isolates when incubated with or without Cefepime (p = 0.3107 and p = 0.8985 for IMP 1078b and IMP 1078s respectively).   Aminoglycoside Trimethoprim Sulfonamide Sulfonamide Macrolide, Lincosamide and Streptogramin B + + bla CTX-M-14b Beta-lactam + + bla CTX-M- 15 Beta-lactam + + bla  Beta-lactam + + bla  Beta-lactam + + bla  Beta-lactam + + bla SHV  Additionally, the growth rate was slightly improved for both isolates when incubated with either antibiotics, when matched with its unchanged control. Moreover, the efficiency of both antibiotics was supported by visualizing the growth rates of K. pneumoniae DSM 30104 grown in intact broth or media containing either meropenem or cefepime. As noticed in Figures 3C and  4C, the growth rates of K. pneumoniae DSM decreased significantly when incubated with either meropenem or cefepime respectively (p < 0.0001).

Inhibitors
There are multiple mechanisms for resistance in CRE. Our aim is to understand the mechanisms by which our isolates escape the action of carbapenems.

Both
K. pneumoniae isolates expressed carbapenemases: Class B Metallo-β-lactamases (bla NDM-5 ) and class D β-lactamases (bla OXA-48 and bla OXA-9 ). In order to show which enzyme plays the major role in carbapenem resistance in these isolates, each class of carbapenemase was inhibited and the effect on meropenem MICs was recorded. Calcium-EDTA inhibits class B Metallo-β-lactamases by chelating their zinc ions, while avibactam obstructs the action of class D β-lactamases via acylation of their serine. As seen in Table 5, the MIC of Meropenem for both isolates was 256 μg/mL. However, the combination of Meropenem with EDTA significantly dropped the MIC to 64 μg/mL for IMP 1078b and to 32 μg/mL for IMP 1078s. Interestingly, when adding   avibactam alone, the MIC of meropenem remained constant (256 μg/mL) for both isolates. On the other hand, the MIC of meropenem declined to 4 μg/mL and 16 μg/mL for IMP 1078 b and IMP 1078s respectively when combining both Ca-EDTA and avibactam. Taken together, these data indicate that bla NDM-5 represents the main enzyme that these isolates use to hydrolyze carbapenems. E. coli 1176 and E. coli 57 are clinical isolates that express solely one carbapenemase: bla NDM-1 and bla OXA-48 respectively. These bacteria were used as control strains to check the efficiency of EDTA and Avibactam in combination with Meropenem.

Discussion
We report the highest number of AMR genes ever detected in a K. pneumoniae isolate. These 47 AMR genes encode for resistance to all the antimicrobial agents used in clinical setting, except for colistin. Colistin resistance in IMP 1078b may be the reason behind the 1% difference between the genomes of both isolates. Colistin resistance is occasionally caused by the acquisition of the mcr gene [13], which is not the case in our isolate. Other ways that K. pneumoniae can acquire resistance to colistin are: (i) the lpxM gene that leads to the formation of hexa-acylated lipid A by encoding an enzyme that is involved in the addition of the myristoyl group to lipid A [14], and (ii) LPS modification due to the inactivation of mgrB, upregulation of the PhoP/PhoQ signalling system, activation of the PmrA-regulated pmrHFIJKLM operon, and the presence of ArnB [15]. Moreover, several studies were done at the country level that led to the detection of several β-lactamase genes in K. pneumoniae, such as: blaCTX-M-15 [16][17][18], bla TEM-1 [17,18], bla SHV-28 [17], bla OXA-1 [16][17][18], bla OXA-48 [16,18,19], bla NDM-1 [16][17][18], and bla NDM-7 [20]. However, we hereby report the first K. pneumoniae isolate harboring bla NDM-5 , bla OXA-48 , and bla OXA-9 , bla CTX-M-15 , bla CTX-M-14b , 7 different bla SHV genes, and 2 bla TEM genes at the same time. Seventy-two virulence genes were detected in each isolate encoding for 15 different virulence factors. We are the first to report type IV pilli (Yersinia) and the fimbrial adherence determinant Stb (Salmonella) in K. pneumoniae. StbA, stbB, stbC, and stbD are genes encoding for the fimbrial adherence determinant Stb (Salmonella) present in our 2 isolates. Stb are a type 1 fimbriae and are, in addition to 7 other clusters, one of the most abundant fimbrial clusters in the genome of Salmonella spp. [21]. They function by binding to the intestinal epithelial cells or by participating in the colonization of avian or mammalian intestines [22]. Furthermore, type IV pilli (Yersinia) are multifunctional surface structures that function in: biofilm formation, adhesion to host cells and other surfaces, cellular invasion, formation of bacterial aggregates or microcolonies, DNA and phage uptake, electron transfer, and twitching or gliding motility [23]. The cost of harboring AMR genes is believed to reduce bacterial fitness, especially in the absence of antibiotic pressure [24,25]. However, multiple studies have already refuted this hypothesis for Enterobacterales that acquired either carbapenemases [26] or extended-spectrum β-lactamases [27]. Our hypervirulent isolates which garnered an incredible sum of resistance genes further support this evidence. Even in the absence of antibiotics, both IMP 1078b and IMP 1078s did not show a significant decrease in growth rates. However, both isolates demonstrated a slightly enhanced fitness in the presence of antibiotics used. A reason behind the lack of a burden on the growth in these isolates might be from the resistance-conferring plasmids themselves. IncHI family of plasmids such as IncHI1B which is present in both our isolates, can carry a gene called histone-like nucleoid-structuring protein or H-NS [28]. This pleotropic regulator has been shown to regulate a multitude of pathogenicity factors in K. pneumoniae, as well as reducing the fitness cost of plasmid acquisition [29]. The latter effect might be a consequence of the abilities of H-NS: it allows the entry of plasmids to the host with a minimal change on global gene expression patterns, then integrates other plasmidic genes into the established gene expression regulation networks [28]. Thus, H-NS might be allowing plasmid-encoded resistance genes to be expressed constitutively without being affected by antibiotic exposure. This assumption is based on our RT-PCR data ( Supplementary Figures 2-4), where the Although the action of avibactam on bla OXA-48 has been repeatedly proven [30,31], no study has directly linked avibactam to an inhibitory activity on bla OXA-9 . This hypothesis is also supported by the variation of effect of avibactam on Class D carbapenemases [32]. The current report focused on the phenotypic and molecular characterization of two clinical Klebsiella pneumoniae isolates recovered from a patient at a tertiary care Lebanese hospital. Both isolates demonstrated resistance to a wide range of antibiotics. This resistance is encoded by an overabundance of AMR genes. Additionally, the presence of the H-NS factor capable of reducing the burden imposed by the plasmid acquisition and facilitating its conjugable transfer increases the risk of nosocomial outbreaks related to these isolates. Moreover, the co-existence of a high number of AMR genes and virulence factors may lead to a life-threatening invasive K. pneumoniae infection. Despite infection with highly resistant organisms, our patient recovered and did not succumb to the bloodstream infection. In fact, in-vitro observations do not correlate always with the real life experience and the most resistant organism may not always be the most virulent.