Reduced Ceftazidime-Avibactam Susceptibility in KPC-Producing Klebsiella pneumoniae From Patients Without Ceftazidime-Avibactam Use History – A Multicenter Study in China

KPC-producing Klebsiella pneumoniae (KPC-KP) is the most widely spread carbapenem-resistant Enterobacteriaceae (CRE) in China. Avibactam is a novel non-β-lactam β-lactamase inhibitor which is highly active against KPC. Recently, ceftazidime-avibactam (CAZ-AVI) was approved for clinical treatment in China. Here we conducted a retrospective study to examine the antimicrobial susceptibility of CAZ-AVI prior to its usage in China, and evaluated the potential to develop resistance in KPC-KP. CAZ-AVI MICs were tested in 347 KPC-KP isolates collected from patients with no prior treatment with this combination from six medical centers in China. Almost all isolates (n = 346; 99.7%) were CAZ-AVI-susceptible, with only 12 (3.5%) which showed reduced susceptibility (MIC ≥ 4/4 μg/ml) or resistance. The 12 isolates belong to ST11 and half of them carry virulence genes. In comparison to susceptible isolates, these isolates demonstrated higher blaKPC–2 copy numbers and expressions, and demonstrated higher frequency of developing CAZ-AVI resistance.


INTRODUCTION
With the widely use of carbapenem antibiotics, carbapenem-resistant Enterobacteriaceae (CRE) have been increasingly detected worldwide. The production of carbapenemases is the leading cause of carbapenem resistance in CRE. Klebsiella pneumoniae carbapenemase (KPC) is currently the most widely spread carbapenemase in the world, including China (Nordmann and Poirel, 2014), while K. pneumoniae is the main clinical species producing KPC (Zhang et al., 2017). KPC-type β-lactamases could hydrolyze carbapenems and almost all β-lactam antibiotics, and traditional β-lactamase inhibitors have limited effects on KPC (Papp-Wallace et al., 2010). In addition, isolates producing KPC are commonly resistant to many other clinical agents (Pollett et al., 2014) due to the co-expression of several resistant determinants. Novel treatments for infections caused by KPC-producing K. pneumoniae (KPC-KP) are in urgent need.
In this study, we examined the CAZ-AVI susceptibility of KPC-KP from patients with no prior CAZ-AVI treatment history in China and evaluated the potential of isolates with reduced susceptibility to this combination to develop resistance through in vitro selection experiments.  (Weisburg et al., 1991).

Determination of CAZ-AVI Minimal Inhibitory Concentrations
The CAZ-AVI minimal inhibitory concentrations (MICs) were determined using the broth microdilution method recommended by CLSI (Clinical and Laboratory Standards Institute, 2018). Escherichia coli ATCC 25922 was used as quality control strain (Clinical and Laboratory Standards Institute, 2018). MICs were interpreted according to CLSI breakpoints (Clinical and Laboratory Standards Institute, 2018). A CAZ-AVI MIC of ≥4/4 µg/ml was used as the cut-off of reduced susceptibility to CAZ-AVI for this study (Shen et al., 2017).

Collection of Clinical Information
The clinical information including city, age range, isolation date, clinical department, sample, medical condition and outcome of the patients from whom the carbapenem-resistant K. pneumoniae was isolated was collected using EPIINFO software based on the medical records.

Multilocus Sequence Typing (MLST)
Multilocus sequence typing (MLST) was conducted to investigate the genetic relationships of the isolates with reduced CAZ-AVI susceptibility. PCR followed by Sanger sequencing was used to detect conserved housekeeping including gapA, infB, mdh, pgi, phoE, rpoB, and tonB (Diancourt et al., 2005). Allelic profiles and sequence types (STs) were determined using the K. pneumoniae MLST database 1 .

Pulsed-Field Gel Electrophoresis (PFGE)
The clonal relatedness between the isolates with reduced CAZ-AVI susceptibility was investigated by pulsed-field gel

Investigation of the Capsular Types and Virulence Genes
Multiplex PCR-II analysis was applied to investigate the capsular types including K1, K2, KL64, KL47 (Yu et al., 2018), and multiplex PCR-III analysis was performed to detect four virulence genes previously described on the pLVPK virulence plasmid namely rmpA, rmpA2, iroN, and iutA (Yu et al., 2018) in the isolates with reduced CAZ-AVI susceptibility.

Quantitative Real-Time PCR (qRT-PCR) and bla KPC Promoter Region Sequencing
A total of 16 isolates, including 8 isolates with reduced susceptibility and 8 randomly selected susceptible isolates, were examined by quantitative real-time PCR (qRT-PCR) to assess the bla KPC−2 copy numbers relative to an internal K. pneumoniae housekeeping gene, rpoB, as previously described (Kitchel et al., 2010). The expression of bla KPC−2 in the same isolate set was measured using DNA-free RNA preparations by qRT-PCR as described previously (Ruzin et al., 2005;Doumith et al., 2009). The statistical software used in this study was Prism 5 (Graph Pad Software). In addition, the bla KPC promoter regions from 12 isolates with reduced CAZ-AVI susceptibility and the 8 randomly selected susceptible isolates were amplified and sequenced using the primers (KPCpro-F, 5 -AACGGTCGTATCAGCGACAT-3 and KPCpro-R, 5 -CGAGTTTAGCGAATGGTTCC-3 ), covering the previously described promoter regions in KPC-KP isolates from China (Huang et al., 2019).

In vitro Selection Testing
All isolates underwent qRT-PCR detection, except for the CAZ-AVI (MIC 16/4 µg/ml) resistant isolate (n = 15), were subject to in vitro CAZ-AVI selective pressure testing, using a previously described method (Rodriguez-Villodres et al., 2020). The 15 strains included 7 isolates with reduced susceptibility (MIC ≥ 4/4 µg/ml) and 8 randomly selected susceptible isolates (see above). In brief, in vitro selection was performed by inoculation of ∼10 8 cfu in 2-ml LB broth containing CAZ-AVI at the 0.5 × MICs followed by incubation for 24 h. This procedure was repeated daily, each time doubling the CAZ-AVI concentration up to a maximum of 8/4 µg/ml. Selected colonies collected at different in vitro selection antibiotic levels were subject to CAZ-AVI susceptibility testing. The isolates with increased MICs ≥ 16/4 µg/ml after CAZ-AVI selection were determined as selected CAZ-AVI resistance.   We then used CAZ-AVI MIC of ≥4/4 µg/ml as the cut-off of reduced susceptibility to CAZ-AVI for this study (Shen et al., 2017). A total of 12 KPC-KP were found to have a CAZ-AVI MIC value ≥ 4/4 µg/ml with high level resistance to carbapenems (meropenem MICs ≥ 256 µg/ml) ( Table 1).

Clinical Information
The 12 isolates were obtained from distinct patients aged between 0 and 65 years old. They were from Beijing (n = 6), Kunming (n = 3), Guangzhou (n = 2), and Suzhou (n = 1). Four out of six Beijing isolates were collected from patients in ICU wards within the same month. In addition, most of these patients received empirical antibiotic treatments before these isolation of strains. Among them, carbapenems were the most commonly used antibiotic, followed by piperacillin-tazobactam ( Table 3).

Detection of ESBL and Virulence Genes and Capsular Genotyping
These 12 isolates contain ESBLs genes including bla CTX−M−65 (n = 9), bla CTX−M−14 (n = 3) and bla SHV−12 (n = 8). In addition, the results of multiplex PCR showed that 6 isolates from Beijing were positive for the four virulence genes, namely rmpA, rmpA2, iroN, and iutA. Eight isolates belonged to capsular type KL64 and the other 4 isolates were from KL47 ( Table 4).

MLST Sequence Types and PFGE Patterns
Multilocus sequence typing results showed that all 12 isolates belonged to ST11. Furthermore, these 12 isolates were investigated by PFGE. As shown in Figure 1, the 6 isolates from Beijing shared the same PFGE pattern. In addition, 2 of the 3 isolates from Kunming were highly homologous (>90%, dice similarity coefficient). The other isolates demonstrated different pulsotypes.

Outer Membrance Porin Gene Sequence Analysis
Sequencing of the outer membrane porin genes ompK35 and ompK36 showed that all 12 isolates contain a mutant OmpK35, with a premature stop codon at amino acid position 63, as well as a mutant OmpK36, due to the glycine and aspartic acid duplication at amino acid 134 (134-135 GD insertion). However, the OmpK35 and OmpK36 gene sequences of 8 randomly selected susceptible K. pneumoniae ST11 isolates showed the same genotypes (OmpK35 stop codon and OmpK36 134-135 GD insertion) as those of the 12 isolates with reduced CAZ/AVI susceptibility.

Mechanism of Reduced CAZ-AVI Susceptibility
Polymerase chain reaction detection and Sanger sequencing showed that the 12 isolates all harbored wild-type bla KPC−2 . Examination of the bla KPC−2 promoter regions failed to identify any mutations in comparison to the sequences from the susceptible strains. The results of qRT-PCR showed that the relative bla KPC−2 copy numbers in the reduced susceptibility group were significantly higher than those in the susceptibility group (2.6-fold, P = 0.0004) (Figure 2A). In addition, the relative expressions of bla KPC−2 in the reduced susceptibility group were 3.9-fold higher than those in the susceptibility group (P = 0.0034) ( Figure 2B).

In vitro Selection Testing
The in vitro selection experiments showed that 4 of the 7 isolates with reduced susceptibility developed resistance to CAZ-AVI at the selection concentration of 2/4 µg/ml, and all 7 isolates developed resistance to CAZ-AVI with MICs ranging from 32/4 to 256/4 µg/ml when the selection concentration reached 4/4 µg/ml. By contrast, no isolates were developed to be resistance to CAZ-AVI in the susceptibility group, and all eight isolates stopped growing when the selection concentration reached 8/4 µg/ml.

DISCUSSION
At present time, KPCs are the most common carbapenemases identified worldwide especially in K. pneumoniae from China (Nordmann and Poirel, 2014). Due to the limited therapies, the detection rate of KPC-KP in China demonstrated a continuous upward trend (Hu et al., 2016). In this study, CAZ-AVI showed potent in vitro activities against KPC-KP, which agreed with previous studies (de Jonge et al., 2016;Spiliopoulou et al., 2020). However, ∼3% isolates displayed reduced susceptibility to CAZ-AVI (MIC ≥ 4/4 µg/ml) despite they were obtained from patients without previous CAZ-AVI treatment history. Notably, these isolates also showed high level resistance to carbapenems. In 2018, a fatal outbreak caused by ST11 KPC-KP with acquisition of a pLVPK-like virulence plasmid that increased the virulence of these isolates was reported (Gu et al., 2018). In our study, half of the isolates with reduced CAZ-AVI susceptibility contained several known virulence genes, suggesting that these isolates may have increased virulence, which should be closely monitored.
Since the previously reported gene mutations (e.g., D179Y) associated to CAZ-AVI resistance (Giddins et al., 2018) weren't found in the bla KPC−2 genes, this study suggested that mechanisms other than bla KPC−2 gene mutation were underlying the reduced CAZ-AVI susceptibility among these isolates. OmpK35/36 defects had previously been reported to lead to reduced susceptibility or resistance to CAZ-AVI in KPC-KP (Nelson et al., 2017). In this study, our results showed 12 isolates with reduced CAZ-AVI susceptibility contained OmpK35 and OmpK36 gene mutations, however, the same gene mutations were also found in the susceptible strains, which suggested that OmpK35/36 defects may only partially contribute to the reduced CAZ-AVI susceptibility among those 12 strains, while additional mechanisms may be involved. Our results demonstrated that the bla KPC−2 copy numbers and expressions in the reduced susceptibility group were significantly higher than those in the susceptibility group. We therefore suspected that the reduced CAZ-AVI susceptibility in these strains was likely due to the higher bla KPC−2 copy numbers and gene expressions, in combination to the OmpK35/36 defects. The results were consistent with some previously published studies (Shen et al., 2017;Zhang et al., 2020). The higher copy numbers of bla KPC−2 may potentially result from the high copy numbers of the plasmids carrying bla KPC (Roth et al., 2011) and/or a duplication of mobile genetic elements associated with bla KPC (Coppi et al., 2020). Since the examination of the bla KPC−2 promoter regions failed to identify any mutations, the higher bla KPC gene expression may potentially be affected by the higher copy numbers of bla KPC or other gene regulatory mechanisms. Further studies, including whole genome sequences, are needed to explore the molecular mechanisms underlying the higher bla KPC−2 copy numbers and gene expressions among those strains. Shields et al. (2017) has reported on the acquisition of CAZ-AVI resistance among ST258 KPC-KP during treatment in the United States. After that, (Raisanen et al., 2019) isolated a ST39 KPC-KP that was resistant to CAZ-AVI after this combination treatment. In China, ST11 KPC-KP is commonly prevalent (Chen et al., 2014;Zhang et al., 2017). In this study, the ST11 KPC-KP with reduced susceptibility were more prone to develop CAZ-AVI resistance compared to susceptible isolates under the pressure of CAZ-AVI exposing.
Taken together, our study demonstrated that CAZ-AVI has potent in vitro activities against KPC-KP in China and highlighted the clinical significance of the isolates with reduced susceptibility to CAZ-AVI isolated from patients without previous CAZ-AVI treatment history. Our results suggested that the optimal clinical usage of CAZ-AVI should be guided by the in vitro susceptibility results in order to prevent selection resistance.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

ETHICS STATEMENT
This study was approved by the institutional review board (IRB) of The Second Affiliated Hospital of Soochow University. The clinical isolates were retrospectively collected, and patient data were not included in this study, therefore the need for written informed consent was waived by the IRB.