In vitro susceptibility of OXA-48, NDM, VIM and IMP enzyme- producing Klebsiella spp. and Escherichia coli to fosfomycin

Introduction: Infections caused by Carbapenemase-producing Enterobacterales (CPE) are an important public health issue. Intravenous fosfomycin can be considered as an alternative for the treatment of serious infections caused by CPE. In this study, in vitro activity of fosfomycin was investigated among CPE isolates. Methodology: Overall, 158 clinically relevant isolates obtained from 18 hospitals of 13 cities in Turkey with predetermined carbapenemase types were evaluated in the study, including Escherichia coli (n = 19) and Klebsiella spp. (n = 139). In vitro activity of fosfomycin was determined with agar dilution method. Among Klebsiella spp., 104 harbored blaOXA-48, 15 isolates carried both blaOXA-48 and blaNDM; three had both blaOXA-48 and blaVIM and nine isolates had blaNDM alone. Four isolates carried only blaVIM and two isolates harbored blaIMP alone. One isolate co-harbored blaVIM and blaNDM. Among E. coli isolates, blaOXA-48 and blaNDM were carried by 18 and one isolates, respectively. Results: Resistance to fosfomycin was detected in 43.7% of the isolates. Among Klebsiella spp. and E. coli, these rates were 46.8% and 21.1%, respectively. In Klebsiella spp. resistance to fosfomycin was 49.5% in blaOXA-48 carriers; 26.7% in isolates co-harbouring blaOXA-48 and blaNDM and 66.7% in blaNDM carriers. In E. coli, fosfomycin resistance was detected among 16.7% of the blaOXA-48 carriers. Conclusions: High level of fosfomycin resistance in these isolates may be attributable to the fact that these isolates are multidrug resistant. The genetic background of resistance should also be investigated in order to understand the co-occurrence and transfer of resistance among the CPE.


Introduction
Carbapenemase-producing Enterobacterales (CPE) are threating problem and present global health issue, as they are frequently multidrug-resistant to available antibiotics. Infections due to CPE are mostly severe and related to poor outcomes [1]. Clinically and epidemiologically, three classes of carbapenemases are responsible for carbapenem resistance: Ambler class A enzyme (such as KPC type), zinc-dependent Ambler class B Metallo-ß-lactamases (MBLs) (such as VIM, IMP, and NDM types) and Ambler class D carbapenemases (such as OXA-48 type enzymes). Moreover, the prevalence of the types of carbapenemase varies with geographical location. In Turkey, OXA-48 enzyme is the most frequent type of carbapenemase, followed by NDM types [2]. The KPC enzyme is very rare in Turkey and has been detected very recently in a few isolates [3]. There are very few antimicrobial options for the treatment of infections caused by CPE and there is a great need for new alternatives.
Fosfomycin has been available for more than a decade and used orally for uncomplicated urinary tract infections. In some countries, intravenous form (IV) is available for gram-negative serious infections such as sepsis and pneumonia and considered to be effective in some infections caused by CPE [4]. The intravenous form of fosfomycin has just been introduced for clinical use in Turkey, however, studies evaluating in vitro activity of this drug against CPE isolates are lacking. The aim of this study was to determine the in vitro efficacy of fosfomycin against Escherichia coli and Klebsiella spp. with specific carbapenemase types.

Antibiotic Susceptibility Testing
Minimum inhibitory concentration (MIC) of the isolates was determined with agar dilution method for fosfomycin (Abcam, Cambridge, USA) on Mueller-Hinton agar supplemented with glucose-6-phosphate (25 mg/L) (Sigma-Aldrich, St Louis, USA). Briefly, an inoculum of 10 4 CFU/mL was prepared and inoculated to the plates containing fosfomycin concentrations ranging from 0.25 to 128 mg/L using a multipoint inoculator and allowed to dry. Incubation was done at 35°C for 16-20 hours [6]. Results were interpreted following EUCAST v.10.0 guidelines. [7] ATCC 25922 E. coli was used as the quality control strain.

Statistical Analysis
All statistical analyses were performed using the SPSS software, version 23.0 for Windows (IBM Corp., New York, NY). A two-sided p value < 0.05 was considered as significant. Fisher's exact test and Pearson Chi-square were used to compare categorical variables as appropriate.

Results
In this study, CPE isolates from thirteen different cities of Turkey, collected during 2014-2017, were included to maintain geographic diversity. Among them, the most common carbapenemase producing gene in Klebsiella spp. was bla OXA-48 which was found in 123 isolates (75.5%). In 105 isolates, bla OXA-48 was found alone, whereas it was co-harboured with bla NDM and bla VIM in fifteen and three isolates, respectively. Eighteen of E. coli isolates (94.7%) harboured bla  ; and one (5.3%) carried bla NDM alone.
In vitro activity of fosfomycin was evaluated against E. coli and Klebsiella spp. harboring bla OXA-48 , bla NDM , bla VIM and bla IMP type carbapenemase genes in this study. Among all isolates, in-vitro resistance to fosfomycin was 43.7%. These rates were 46.8% and 21.1% in Klebsiella spp. and E. coli, respectively. As in other reports, fosfomycin resistance was significantly higher in Klebsiella spp. compared to E. coli (p = 0.047). Among 139 Klebsiella spp., 52 of 105 bla OXA-48 gene carriers, four of fifteen bla OXA-48 +bla NDM cocarriers, six of nine bla NDM carriers, one of four bla VIM carriers and two of three bla VIM + bla OXA-48 co-carriers were resistant to fosfomycin. In contrast, only three of eighteen bla OXA-48 carriers and the only bla NDM carrier were resistant to fosfomycin among E. coli. (Table 1).

Discussion
In this study, we evaluated in vitro activity of fosfomycin against E. coli and Klebsiella spp. harbouring bla OXA-48 , bla NDM , bla VIM and bla IMP type carbapenemase genes. In Turkey, there is an endemicity of bla OXA-48 carbapenemase however, other types of carbapenemases have also been reported recently [3]. The high rates of bla OXA-48 carrying E. coli and Klebsiella spp. isolates (%89.2) in this study may be attributable to its endemic potential in Turkey.
Intravenous form of fosfomycin was recently introduced into market for clinical use and started to be administered as an option for the treatment of infections caused by CPE in our country. To date, limited number of studies evaluated in vitro activity of fosfomycin against CPE. In Asia, fosfomycin resistance rates of CPE are relatively high. In a study from China, the authors reported a high resistance rate (60.8%) against fosfomycin in KPC-producing K. pneumoniae and an isolate harbouring both blaKPC and fosA3 genes in the same plasmid were found [8] . In a study from Singapore, 1 of 3 bla OXA-48 and 2 of 4 bla IMP harbouring K. pneumoniae isolates were resistant to fosfomycin [9]. However, in America and Europe, resistance rates of fosfomycin remains low among Enterobacterales. A study from USA, fosfomycin resistance was reported as 7% in KPC-producing K. pneumonia [10]. Similarly, fosfomycin resistance was detected in 5.1% of carbapenemase producers in a study from Greece [11]. A recent study from Turkey reported 76.1% fosfomycin resistance among CPE, which were slightly higher in bla OXA-48 carriers (78.2%) [12]. In our study, resistance to fosfomycin was 46.8% and 21.1% in Klebsiella spp. and E. coli, respectively. These high rates are concerning for the treatment of CPE, especially in bla OXA-48 carriers which are endemic in Turkey.
In our study, in Klebsiella spp, harbouring bla OXA-48 alone or in combination, fosfomycin resistance was 47.2%. Among nine bla NDM harbouring Klebsiella spp. fosfomycin resistance was 66.7%. In E. coli however, 16.7% of the isolates carrying bla OXA-48 had MIC values in the resistant range.
Fosfomycin resistance is usually obtained through two main genetic mechanisms: chromosomal mutations and plasmid-encoded resistance, inactivating the antibiotic molecule, which is due to drug modifying enzymes. The most concerning plasmid-mediated resistance determinants against fosfomycin have been characterized as three main types (fosA, fosB, fosC) and their subtypes in Enterobacterales. Plasmid mediated resistance determinants can be horizontally transferred and are usually related with multidrug resistance. Among them, fosA3 is the most common and widely distributed type in Asia, and recently in Europe [13,14]. Co-carriage of fosfomycin resistance genes and carbapenemase coding genes has also been reported [14].
In our study, resistance to fosfomycin was significantly higher in carbapenemase-producing Klebsiella spp. compared to E. coli isolates. However, since the fosfomycin resistance mechanism is mostly related to chromosomal mutations in Klebsiella spp., the genetic background of resistance should also be investigated in order to understand the difference and the co-occurrence and transfer of resistance [15].

Conclusion
The OXA-48-type and Metallo-β-lactamase type carbapenemases are the most prevalent genes in carbapenemase-producing Enterobacterales in Turkey. In this study, we found a high level of in vitro resistance to fosfomycin among Klebsiella spp. isolates harbouring these carbapenemase genes. However, in vitro resistance to fosfomycin among E. coli was significantly lower than Klebsiella spp. isolates. We conclude that fosfomycin can be considered in combination therapies caused by carbapenemaseproducing E. coli or Klebsiella spp. when MIC of fosfomycin is in the therapeutic range. Surveillance studies with fosfomycin should be performed in each center as the frequency of various carbapenemase genes may differ according to geographical location. Furthermore, plasmid-mediated fosfomycin resistance is reported from different geographical locations including Europe. This mechanism of resistance should be investigated in further studies.