CXC Chemokines Exhibit Bactericidal Activity against Multidrug-Resistant Gram-Negative Pathogens

Abstract

The continued rise and spread of antimicrobial resistance among bacterial pathogens pose a serious challenge to global health. Countering antimicrobial-resistant pathogens requires a multifaceted effort that includes the discovery of novel therapeutic approaches. Here, we establish the capacity of the human CXC chemokines CXCL9 and CXCL10 to kill multidrug-resistant Gram-negative bacteria, including New Delhi metallo-beta-lactamase-1-producing Klebsiella pneumoniae and colistin-resistant members of the family Enterobacteriaceae that harbor the mobile colistin resistance protein MCR-1 and thus possess phosphoethanolamine-modified lipid A. Colistin-resistant K. pneumoniae isolates affected by genetic mutation of the PmrA/PmrB two-component system, a chromosomally encoded regulator of lipopolysaccharide modification, and containing 4-amino-4-deoxy-l-arabinose-modified lipid A were also found to be susceptible to chemokine-mediated antimicrobial activity. However, loss of PhoP/PhoQ autoregulatory control, caused by disruption of the gene encoding the negative regulator MgrB, limited the bactericidal effects of CXCL9 and CXCL10 in a variable, strain-specific manner. Cumulatively, these findings provide mechanistic insight into chemokine-mediated antimicrobial activity, highlight disparities amongst determinants of colistin resistance, and suggest that chemokine-mediated bactericidal effects merit additional investigation as a therapeutic avenue for treating infections caused by multidrug-resistant pathogens.IMPORTANCE As bacterial pathogens become resistant to multiple antibiotics, the infections they cause become increasingly difficult to treat. Carbapenem antibiotics provide an essential clinical barrier against multidrug-resistant bacteria; however, the dissemination of bacterial enzymes capable of inactivating carbapenems threatens the utility of these important antibiotics. Compounding this concern is the global spread of bacteria invulnerable to colistin, a polymyxin antibiotic considered to be a last line of defense against carbapenem-resistant pathogens. As the effectiveness of existing antibiotics erodes, it is critical to develop innovative antimicrobial therapies. To this end, we demonstrate that the chemokines CXCL9 and CXCL10 kill the most concerning carbapenem- and colistin-resistant pathogens. Our findings provide a unique and timely foundation for therapeutic strategies capable of countering antibiotic-resistant "superbugs."

Keywords: Gram negative; antimicrobial resistance; carbapenem; chemokine; colistin.

FIG 1

Bactericidal effects of CXC chemokines against CRE. (A) Carbapenem-resistant K. pneumoniae isolates and a species-matched control (ATCC 43816) are shown. Multilocus sequence types (MLSTs) and genes encoding the NDM-1 and/or OXA-48 carbapenemases are indicated. Antimicrobial susceptibilities (reported as MIC [µg/ml]) were interpreted as resistant (red) or susceptible (green) in accordance with established breakpoints. The antibiotics tested were ampicillin (AMP), amoxicillin-clavulanic acid (AMC), AMP-sulbactam (SAM), piperacillin-tazobactam (TZP), cefazolin (CFZ), cefoxitin (FOX), cefotaxime (CTX), ceftriaxone (CRO), aztreonam (ATM), cefepime (FEP), ertapenem (ERT), doripenem (DOR), imipenem (IPM), meropenem (MEM), CST, amikacin (AMK), gentamicin (GEN), tobramycin (TOB), ciprofloxacin (CIP), levofloxacin (LVX), tetracycline (TET), tigecycline (TGC), and trimethoprim-sulfamethoxazole (SXT). Bacteria were treated with 48 µg/ml CXCL10 (B) or CXCL9 (C), and survival was measured by CFU determination (limit of detection, 500 CFU/ml; n.d., none detected). Data are expressed as percentages of the respective untreated-control value and represent the mean ± the standard error of the mean (n = 3). *, P < 0.05; **, P < 0.01 (compared to ATCC 43816). (D) CRE isolate BL12125 was treated with increasing concentrations of CXCL10. Data are expressed as percentages of the untreated-control value and represent the mean ± the standard error of the mean (n = 3). ***, P < 0.001 (compared to the untreated control).

FIG 2

Chemokine-mediated antimicrobial activity against mcr-1⁺ bacteria. (A) CST susceptibility testing of E. coli by modified Etest. CST concentrations are in µg/ml. Images are representative of two or three separate tests. Printed CO indicates a CST test strip. (B) Molecular structures and mass-to-charge (m/z) ratios of signature ions identified in E. coli mass spectra. (C) Lipid extracts were analyzed by MALDI-TOF mass spectrometry. The unmodified base peak (m/z 1,796) was observed in all spectra. Major ions at m/z 1,919 and 1,839, corresponding to pEtN-modified lipid A and pEtN-modified lipid A lacking the second phosphate moiety, were detected only in spectra generated from mcr-1⁺ isolates. E. coli isolates were treated with 4 or 12 µg/ml CXCL10 (D) or CXCL9 (E); K. pneumoniae strains were treated with 48 µg/ml CXCL10 (F). Different chemokine concentrations were assayed on the basis of the inherent susceptibilities of these organisms. Survival was measured by CFU determination (limit of detection, 500 CFU/ml; n.d., none detected). Data are expressed as percentages of the respective untreated-control value and represent the mean ± the standard error of the mean (n = 3). ns, not significant.

FIG 3

Effects of chromosomal determinants of CST resistance on the bactericidal activity of CXC chemokines. (A) MLSTs, carbapenemase genes, and antimicrobial susceptibilities of CST-resistant K. pneumoniae isolates are shown. MICs [µg/ml] were interpreted as resistant (red), intermediate (yellow), or susceptible (green) on the basis of established breakpoints. For definitions of abbreviations, see the legend to Fig. 1. Bacteria were treated with 48 µg/ml CXCL10 (B) or CXCL9 (C), and survival was measured by CFU determination. Data are expressed as percentages of the respective untreated-control value and represent the mean ± the standard error of the mean (n = 3). **, P < 0.01; ***, P < 0.001; ns, not significant (compared to ATCC 43816 [panel B] or the untreated control [panel C]). (D) Visualization of mgrB amplicons generated by PCR from the isolates indicated. Amplicon size of intact K. pneumoniae mgrB, 235 bp. Markers indicate 1,000 and 200 bp.

FIG 4

Effect of mgrB complementation on resistance phenotypes. (A) CST susceptibility testing of K. pneumoniae strains by modified E test. CST concentrations are in µg/ml. Images are representative of three separate tests. (B) Molecular structures and m/z values of selected ions identified in K. pneumoniae mass spectra. (C) MALDI-TOF analysis demonstrated the canonical K. pneumoniae lipid A base peak (m/z 1,824). Major ions at 1,955 (m/z 1,824 + l-Ara4N), 1,971 (m/z 1,840 + l-Ara4N), and 1,891 (m/z 1,971 − PO₃) were observed in mass spectra from strains harboring disrupted mgrB. Additional ions observed at m/z 1,840, 2,063, and 2,079 are indicative of lipid A hydroxylation, palmitoylation, or both, respectively. (D) Bacteria were treated with 48 µg/ml CXCL10. Survival was measured by CFU counting (limit of detection, 500 CFU/ml; n.d., none detected). Data are expressed as percentages of the respective untreated-control value and represent the mean ± the standard error of the mean (n = 3). (E) CST susceptibility testing by modified E test. Images are representative of two or three separate tests. (F) Bacteria were treated with 48 µg/ml CXCL10. Survival was measured by CFU counting (limit of detection, 500 CFU/ml; n.d., none detected). Data are expressed as percentages of the respective untreated control and represent the mean ± the standard error of the mean (n = 3).