Prediction of ceftriaxone MIC inNeisseria gonorrhoeae usingDNA microarray technology and regression analysis

Abstract

Decreased susceptibility of Neisseria gonorrhoeae to extended-spectrum cephalosporins is a major concern. Elucidation of the phenotypic and genetic characteristics of such isolates is a priority task. We developed a method for predicting the N. gonorrhoeae ceftriaxone susceptibility level (MICcro) by identifying genetic determinants of resistance using low-density hydrogel microarrays and a regression equation. A training dataset, containing 5631 isolates from the Pathogenwatch database and 181 isolates obtained in the Russian Federation during 2018-19, was used to build a regression model. The regression equation was tested on 14 WHO reference strains. Ceftriaxone resistance determinants for the 448 evaluated clinical isolates collected in Russia were identified using microarray analysis, and MICcro values were calculated using the regression equation and compared with those measured by the serial dilution method. The regression equation for calculating MICcro values included 20 chromosomal resistance determinants. The greatest contributions to the increase in MICcro were shown to be PBP2: Ala-501→Pro, Ala-311→Val, Gly-545→Ser substitutions, Asp(345-346) insertion; and PorB: Gly-120→Arg substitution. The substitutions PBP2: Ala-501→Thr/Val, PorB: Gly-120→Asn/Asp/Lys and PBP1: Leu-421→Pro had weaker effects. For 94.4% of the isolates in the evaluation set, the predicted MICcro was within one doubling dilution of the experimentally determined MICcro. No ceftriaxone-resistant isolates were identified in the analysed samples from Russia, and no interpretative errors were detected in the MICcro calculations. The developed strategy for predicting ceftriaxone MIC can be used for the continuous surveillance of known and emerging resistant N. gonorrhoeae isolates.