Molecular Characterization of Baseline Enterobacterales and Pseudomonas aeruginosa Isolates from a Phase 3 Nosocomial Pneumonia (ASPECT-NP) Clinical Trial

We evaluated β-lactam-resistant baseline Enterobacterales species and Pseudomonas aeruginosa lower respiratory tract isolates collected during the ASPECT-NP phase 3 clinical trial evaluating the safety and efficacy of ceftolozane-tazobactam compared with meropenem for the treatment of nosocomial pneumonia in ventilated adults. Isolates were subjected to whole-genome sequencing, quantitative real-time PCR for quantification of the expression levels of β-lactamase and efflux pump genes, and Western blot analysis for the detection of OprD (P. aeruginosa only). ABSTRACT We reviewed β-lactam-resistant baseline Enterobacterales species and Pseudomonas aeruginosa lower respiratory tract isolates collected during the ASPECT-NP phase 3 clinical trial that evaluated the safety and efficacy of ceftolozane-tazobactam compared with meropenem for the treatment of nosocomial pneumonia in ventilated adults. Isolates were subjected to whole-genome sequencing, real-time PCR for the quantification of the expression levels of β-lactamase and efflux pump genes, and Western blot analysis for the detection of OprD (P. aeruginosa only). Extended-spectrum β-lactamase (ESBL) genes were detected in 168 of 262 Enterobacterales isolates, and among these, blaCTX-M-15 was the most common, detected in 125 isolates. Sixty-one Enterobacterales isolates carried genes encoding carbapenemases, while 33 isolates did not carry ESBLs or carbapenemases. Carbapenemase-producing isolates carried mainly NDM and OXA-48 variants, with ceftolozane-tazobactam MIC values ranging from 4 to 128 µg/ml. Most ceftolozane-tazobactam-nonsusceptible Enterobacterales isolates that did not carry carbapenemases were Klebsiella pneumoniae isolates that exhibited disrupted OmpK35, specific mutations in OmpK36, and, in some isolates, elevated expression of blaCTX-M-15. Among 89 P. aeruginosa isolates, carbapenemases and ESBL-encoding genes were observed in 12 and 22 isolates, respectively. P. aeruginosa isolates without acquired β-lactamases displaying elevated expression of AmpC (14 isolates), elevated expression of efflux pumps (11 isolates), and/or a decrease or loss of OprD (22 isolates) were susceptible to ceftolozane-tazobactam. Ceftolozane-tazobactam was active against >75% of the Enterobacterales isolates from the ASPECT-NP trial that did not carry carbapenemases. K. pneumoniae strains resistant to ceftolozane-tazobactam might represent a challenge for treatment due to their multiple resistance mechanisms. Ceftolozane-tazobactam was among the agents that displayed the greatest activity against P. aeruginosa isolates. (This study has been registered at ClinicalTrials.gov under registration no. NCT02070757.)

organisms (5), with one study showing an equal prevalence for these two groups (6). Staphylococcus aureus is the leading cause of nosocomial pneumonia, followed by Pseudomonas aeruginosa and Klebsiella/Enterobacter species. The occurrence of multidrug-resistant Gram-negative organisms in nosocomial pneumonia differs according to the hospital and geographic region (5, 7), but P. aeruginosa isolates are intrinsically resistant to multiple antimicrobial agents, and Klebsiella and Enterobacter species have been shown to have increasing rates of resistance to several antimicrobial agents; therefore, the therapeutic options for treating nosocomial pneumonia caused by Gramnegative organisms are limited (8,9).
In a comparison of European and U.S. guidelines for the management of nosocomial pneumonia infections (10), a common recommendation is that antimicrobial therapy be initiated promptly to decrease the mortality rate and cost associated with these infections. In a report describing agents that can be used to treat nosocomial pneumonia, Bassetti et al. (11) included ceftolozane-tazobactam. Ceftolozane-tazobactam demonstrates good epithelial lining fluid (ELF) penetration and is active against P. aeruginosa and Enterobacterales, which are among the main causes of nosocomial pneumonia (12).
ASPECT-NP (ClinicalTrials registration no. NCT02070757) was a phase 3, randomized, double-blind multicenter study to evaluate the efficacy and safety of ceftolozane-tazobactam administered at 3 g every 8 h (q8h) versus meropenem at 1 g q8h for 8 to 14 days for the treatment of HAP and ventilator-associated pneumonia (VAP) in ventilated adults. ASPECT-NP was recently completed and met all primary and key secondary endpoints. Ceftolozane-tazobactam received approval from the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA) for the treatment of nosocomial pneumonia, including VAP (13).
In the present study, we characterized the resistance mechanisms of baseline lower respiratory tract Enterobacterales and P. aeruginosa isolates from patients enrolled in the ASPECT-NP trial. Our evaluation included an analysis of the outer membrane protein (OMP) sequences for Klebsiella pneumoniae isolates displaying ceftolozane-tazobactam MIC values higher than those of isolates of the same species that carried the same b-lactamases. We performed an in-depth analysis of a selected subset, quantifying the expression of several genes involved in b-lactam resistance in order to understand the impact of the resistance mechanisms on these differences.
Enterobacterales isolates. The most common b-lactamase gene detected among the Enterobacterales isolates was bla CTX-M-15 (175 isolates overall; 125 isolates without carbapenemases). bla CTX-M-15 was observed mainly in combination with other b-lactamase genes (123/175 isolates). The gene encoding CTX-M-15 was detected in 133 of the 162 K. pneumoniae isolates, 27 of the 39 E. coli isolates, and 15 isolates belonging to other Enterobacteriaceae species (Table 1).
Among 33 Enterobacterales isolates not carrying a resistance mechanism against carbapenems and/or broad-spectrum cephalosporins, 31 were inhibited by ceftolozane-tazobactam at #8mg/ml, and 75.8% of these isolates were considered susceptible to this antimicrobial combination according to CLSI breakpoints (Fig. 1). Either these isolates carried other b-lactamase enzymes that had a narrow spectrum or an undetermined spectrum of activity, or they were negative for the presence of acquired b-lactamases.
Analysis of K. pneumoniae OMPs and transcriptional levels of b-lactamases and efflux pumps. Sequences of the genes encoding outer membrane proteins (OMPs) of the 101 K. pneumoniae isolates with genes encoding ESBLs and no carbapenemases were analyzed. These isolates included all 65 K. pneumoniae isolates for which ceftolozane-tazobactam MICs were $4mg/ml (nonsusceptible). For comparison purposes, 36 isolates for which ceftolozane-tazobactam MIC values ranged from 0.25mg/ml to 2mg/ml were evaluated.
Fifty isolates displayed early protein termination due to the presence of stop codons on OmpK35. Only two (5.5%) of these isolates were susceptible to ceftolozane-tazobactam (MIC results, 1 or 2 mg/ml). Isolates for which ceftolozane-tazobactam MIC values were 0.25 mg/ml or 0.5 mg/ml did not have such alterations. OmpK35 disruptions were observed in 48 of the 65 ceftolozane-tazobactam-nonsusceptible isolates (73.8%). Isolates with higher ceftolozane-tazobactam MIC results (.32 mg/ml) were more likely to have these alterations (29 of 32 isolates [90.6%]).
Among 101 isolates, only 12 did not have insertions or deletions in the OmpK36 sequence. These 12 isolates included 8 ceftolozane-tazobactam-susceptible isolates. Only two isolates had ceftolozane-tazobactam MIC values of 4 mg/ml or 8 mg/ml. Among the OmpK36 alterations, a few combinations were more common among ceftolozane-tazobactam-resistant isolates. For example, all the isolates in this set displayed the insertion of a glycine and an aspartate at position 134 (G134_D135insDG) and exhibited ceftolozane-tazobactam MIC values of $64 mg/ml. The insertion of an arginine at position 304 (E304_G305insR) and a 3-amino-acid insertion at position 183 plus a single-amino-acid insertion at position 227 (A183_T184insLSP, N227_L228insK) were  (Table 2). OmpK37 alterations were observed among 40 of the 65 K. pneumoniae isolates for which MICs were $4 mg/ml. However, these alterations were noted in isolates for which ceftolozane-tazobactam MIC values ranged from 4 to .128 mg/ml, and they did not seem to confer resistance alone or in combination with other factors.
A set of 14 isolates displaying ceftolozane-tazobactam MICs of $1 mg/ml that did not carry carbapenemases was evaluated for the expression of acquired b-lactamases and the AcrAB-TolC efflux system. These 14 isolates were then compared to 4 isolates with lower ceftolozane-tazobactam MIC results that carried the same enzymes. Only two isolates for which ceftolozane-tazobactam MICs were $1 mg/ml did not have bla CTX-M-15 expression levels $3-fold higher than those of the control isolates. bla TEM-1 expression was elevated to 5.8-fold for one isolate. Two isolates for which ceftolozanetazobactam MICs were $128 mg/ml had an elevated expression of bla OXA-1_OXA-30 (3.4and 4.9-fold) in addition to higher levels of bla CTX-M-15 . For all isolates, the expression of acrA, bla SHV , and bla CTX-M-14 was similar to that of the control.

b-Lactam Resistance in Phase 3 Clinical Trial Isolates Antimicrobial Agents and Chemotherapy
A total of 12 isolates carried carbapenemase-encoding genes that included genes from the VIM and GES families. Ceftolozane-tazobactam MIC values for these isolates were elevated, ranging from 8 mg/ml to .128 mg/ml; however, ceftolozane-tazobactam MIC results were lower (8 or 16 mg/ml) for isolates carrying bla GES-5 . Finally, four P. aeruginosa isolates carried only penicillinases encoded by bla OXA-10 , bla OXA-2 (two isolates), or bla CARB-2 . The isolate carrying bla OXA-10 displayed OprD loss and a ceftolozane-tazobactam MIC of .128 mg/ml. For the remaining isolates carrying penicillinases, ceftolozanetazobactam MICs were 0.5 to 2 mg/ml.
Among the 51 P. aeruginosa isolates that did not carry acquired b-lactamases, ceftolozane-tazobactam inhibited 50 isolates at #4 mg/ml and all isolates at #8 mg/ml. Ceftolozane-tazobactam-susceptible isolates included eight isolates displaying OprD decrease/loss alone, two with elevated AmpC expression alone, eight with both OprD loss and elevated AmpC expression, and nine with elevated MexCD-OprJ expression, including five with elevated MexCD-OprJ expression alone, two with OprD loss, and one with AmpC overexpression. Additionally, 2 isolates carried multiple resistance mechanisms and 21 isolates without resistance mechanisms were inhibited by ceftolozane-tazobactam at #4 mg/ml. One isolate for which the ceftolozane-tazobactam MIC was 8 mg/ml had an elevated AmpC expression and OprD loss.

DISCUSSION
The isolates evaluated in this study, which constituted a unique collection of lower respiratory tract isolates from patients enrolled in the ASPECT-NP clinical trial, exhibited reduced rates of susceptibility to most antimicrobial agents tested. These results highlight the difficulty of selecting an appropriate antimicrobial regimen for patients with nosocomial pneumonia and also contribute to our understanding of the resistance mechanisms among isolates evaluated in such trials. Ceftolozane-tazobactam was among the most active agents for ESBL-producing Enterobacterales and P. aeruginosa and, as expected from the susceptibility patterns of these isolates, several resistance mechanisms were involved.
Most Enterobacterales isolates in this study carried acquired b-lactamase genes. As observed with surveillance data and other recent clinical trial data sets (14)(15)(16)(17), bla CTX-M-15 was the most common b-lactamase-encoding gene overall among the isolates displaying cephalosporin resistance and susceptibility to carbapenems. The gene encoding CTX-M-15 is often detected with other b-lactamase genes. This combination often confers high levels of resistance to ceftazidime, ceftriaxone, cefepime, and aztreonam. Ceftolozane-tazobactam was very active against E. coli and other non-K. pneumoniae species carrying bla CTX-M-15 and other ESBL genes, but this combination was not as active against K. pneumoniae isolates carrying the same enzymes, prompting us to evaluate additional resistance mechanisms among these isolates.
An analysis of the outer membrane protein sequences of K. pneumoniae isolates with elevated ceftolozane-tazobactam MIC values carrying ESBL genes without carbapenemases demonstrated that a combination of OmpK35 deletion and specific alterations in OmpK36 could play a role in highly elevated ceftolozane-tazobactam MIC values. Several studies demonstrate that OmpK36 amino acid substitutions, deletions, and insertions in the PEFDG motif of the L3 region modify the pore size or conformation, jeopardizing the penetration of carbapenems and cephalosporins (18)(19)(20)(21). These changes have not been demonstrated to decrease the penetration of ceftolozane or tazobactam, but our results indicate that disruption of OmpK35 plus the OmpK36 GD insertion at position 134 (position 119 in the mature protein) can increase the ceftolozane-tazobactam MIC results for isolates producing CTX-M-15 enzymes. In addition to the changes in OMPs, 12 of the 14 isolates that were evaluated for the expression of acquired b-lactamases displayed modest increases in bla CTX-M-15 expression levels. These isolates, which caused serious infections in patients enrolled in the clinical trial, can pose a challenge for treatment with various b-lactam agents and have not been widely described in the literature.
VanScoy and collaborators evaluated the efficacy of ceftolozane-tazobactam against a set of E. coli constructs expressing different levels of CTX-M-15. They concluded that the amount of b-lactamase expressed did not impact the in vivo efficacy of ceftolozane-tazobactam (22). However, the penetration of the antimicrobial agent was not impaired for the isogenic isolates tested. In the K. pneumoniae isolates from this study, impaired membrane permeability in conjunction with elevated expression of CTX-M-15 appeared to lead to elevated ceftolozane-tazobactam MIC results. This finding is corroborated by Nicolas-Chanoine and collaborators (23). They observed that elevated ceftolozane-tazobactam MIC values for K. pneumoniae were caused by b-lactamases that are not inhibited by tazobactam and/or by decreased membrane permeability. In contrast to the clinical trial isolates from our study, the isolates they described had increased efflux pump expression.
With regard to P. aeruginosa, several isolates carried acquired b-lactamases. Ceftolozane-tazobactam displayed activity against some of these isolates, including those carrying genes encoding GES and CTX-M-2. Among P. aeruginosa isolates displaying b-lactam resistance mechanisms caused by chromosomal mutations, ceftolozanetazobactam was the most active b-lactam against isolates overexpressing AmpC, MexAB-OprM, or MexCD-OprN. In addition, ceftolozane-tazobactam was as active as meropenem against isolates with an elevated expression of MexXY-OprM. These isolates had multiple mutation-driven resistance mechanisms, and the comparator agents showed limited activity against them. The in vitro activity of ceftolozane-tazobactam has been evaluated against isolates well characterized for mutation-driven resistance mechanisms and was stable against isolates carrying single or multiple resistance mechanisms, whereas other b-lactam agents were affected by an overproduction of AmpC, OprD loss, and efflux pump overexpression (24).
P. aeruginosa isolates are intrinsically less susceptible than other species to many antimicrobial agents. This species can acquire mutations and resistance mechanisms that encode resistance to agents displaying antipseudomonal activity. Ceftolozanetazobactam is an important addition to the armamentarium for treating P. aeruginosa infections, including nosocomial pneumonia. The empirical use of ceftolozane-tazobactam was reported in the literature prior to the approval of this combination for the treatment of nosocomial pneumonia (25,26). Shortly after the approval of ceftolozane-tazobactam for the treatment of complicated urinary tract and intra-abdominal infections, Gelfand and Cleveland (27) reported the successful treatment of three meropenem-resistant P. aeruginosa isolates with ceftolozane-tazobactam. They emphasized the shortage of effective therapeutic options to treat these infections and highlighted the need for new agents active against infections resulting from multidrug-resistant P. aeruginosa.
As part of the Program to Assess Ceftolozane-Tazobactam Susceptibility (PACTS), we evaluated the activities of ceftolozane-tazobactam and comparator agents against 2,362 Enterobacterales and 1,576 P. aeruginosa clinical isolates collected from respiratory samples of patients hospitalized with pneumonia in 31 U.S. hospitals over a 2-year period. In that study, ceftolozane-tazobactam displayed activity against Enterobacterales isolates producing the most common ESBLs and was also active against P. aeruginosa isolates that were multidrug resistant or extensively drug resistant (28). The isolates collected in the present study differ from our previous surveillance data in that they were collected at different sites and regions and thus might display different resistance patterns. Some K. pneumoniae clinical isolates collected during the ASPECT-NP trial had multiple resistance mechanisms. Additionally, P. aeruginosa isolates from this study harbored acquired b-lactamase genes that are not common in the United States, highlighting the importance of understanding the local epidemiology of resistance mechanisms when one is selecting antimicrobial therapies.
This study revealed an intricate combination of resistance mechanisms leading to elevated ceftolozane-tazobactam MIC values for K. pneumoniae isolates carrying the same b-lactamases as isolates with lower MIC results that carry enzymes inhibited by tazobactam. These isolates were not uncommon in this collection, and the understanding of their resistance mechanisms and epidemiology might be important for the choice of treatment for serious infections such as nosocomial pneumonia, in which the initiation of appropriate therapy choices is decisive for successful outcomes.

MATERIALS AND METHODS
Bacterial isolates and inclusion criteria. We evaluated 352 lower respiratory tract isolates collected from 279 patients enrolled in the ASPECT-NP clinical trial. These isolates included 262 Enterobacterales organisms for which ceftriaxone, aztreonam, ceftazidime, imipenem, or meropenem MIC results were $2 mg/ml and/or ceftolozane-tazobactam MICs were $4 mg/ml. Imipenem was not used in the criteria for Proteus mirabilis or indole-positive Proteeae due to the intrinsically elevated MIC values of this agent. Isolates meeting these criteria were screened for the presence of b-lactamases. Additionally, the transcriptional levels of AmpC were determed for all Escherichia coli, Serratia sp., Citrobacter sp., and Enterobacter sp. isolates that met the MIC-based screening criteria.
The 89 P. aeruginosa isolates analyzed exhibited one or more of the following: (i) ceftazidime MIC values of $16 mg/ml, (ii) imipenem and/or meropenem MIC values of $4 mg/ml, and (iii) ceftolozane-tazobactam MIC values of $8 mg/ml. Isolates meeting these criteria were screened for the presence of b-lactamase-encoding genes and had the transcriptional levels of chromosomally encoded AmpC and the efflux pumps MexAB-OprM, MexCD-OprJ, MexEF-OprN, and MexXY-OprM determined. P. aeruginosa isolates for which imipenem and/or meropenem MIC values were $4 mg/ml were subjected to Western blot analysis in order to detect the presence, absence, or decreased expression of OprD.
Whole-genome sequencing. Total genomic DNA was extracted and used as input material for library construction. DNA libraries were prepared using the Nextera XT library construction protocol and index kit (Illumina, San Diego, CA, USA) and were sequenced on a MiSeq sequencer (Illumina). Sequencing analysis was performed after de novo assembly (17), and specific matches were generated for each sample with the criteria of .94% identity and 40% minimum coverage length. Mutations were considered present when .50% of the sequence reads allowed for base calling. Sequences displaying 100.0% homology with the reference sequences were named according to the reference. Genes with homology of ,100.0% were designated with the suffix "-like" after the gene showing the closest homology.
Outer membrane protein sequences (ompK35, ompK36, and ompK37) of K. pneumoniae isolates were analyzed by comparison with reference sequences of known b-lactam-susceptible isolates.
Gene expression assays. As described previously, gene expression levels were evaluated using quantitative real-time PCR of high-quality RNA samples (24). The endogenous reference gene used was rpsL for all species, with the exception of Serratia marcescens, for which 16S rRNA was used, and K. pneumoniae, for which gyrA was used.
Transcriptional levels were considered elevated when a 10-fold increase over the level for the control isolate from the same bacterial species was noted and moderate when a 5-fold increase was noted.
OprD detection. The phenotypic detection of OprD was assessed by Western blotting as described previously (24). P. aeruginosa PAO1 and two OprD-downregulated laboratory constructs were used as positive and negative controls, respectively.
Data availability. The OMP sequences of the K. pneumoniae isolates analyzed in this study have been deposited in the NCBI Protein database under accession numbers WP_008805872.1 for ompK35, WP_008803977.1 for ompK36, and WP_012541954.1 for ompK37.