Abstract

Although colistin is frequently regarded as the antibiotic of last resort in treating carbapenem-resistant Klebsiella pneumoniae, colistin heteroresistance may in part be associated with antibiotic treatment failure. However, we do not know how widespread the colistin heteroresistance is in carbapenem- resistant K. pneumoniae isolates. In this study, we performed colistin disc diffusion assays, E-tests, and population analysis profiling for KPC-2-producing K. pneumoniae isolates to identify colistin heteroresistance. Although no colistin- resistant colonies were detected by the disc diffusion test and E-test, a colistin-resistant subpopulation was identified in population analysis profiling in all colistin-susceptible, KPC-2-producing K. pneumoniae isolates. Colistin-resistant subpopulations were also identified even when isolates had no colistin exposure. The ratio of colistin-resistant subpopulations to the total population increased as the exposure concentration of colistin increased. In in vitro time-kill assays, regrowth was observed in all isolates after 2 h upon exposure to colistin. We identified common amino acid alterations in PhoQ, PhoP, and PmrB in colistin-resistant subpopulations from some isolates, but no substitutions were found in most resistant subpopulations from other isolates. In all colistin-resistant subpopulations, overexpression of PhoQ and PbgP was observed. In this study, we demonstrated that colistin heteroresistance may be common in KPC-2-producing K. pneumoniae isolates, which could not be detected in the disc diffusion method and E-test. Colistin heteroresistance may cause colistin treatment failure in part and may evolve into resistance. Thus, development of more reliable diagnostic methods is required to detect colistin heteroresistance.

• Keywords: Artificial inoculation;Basal stem rot;Disease severity index;Rubber wood block;Vegetative growth measurement