A total of 220 K. pneumoniae strains were collected in Renji hospital from January 2015 to December 2016. 32.3% (71/220) isolates were HMKP, 38.2% (84/220) were CRKP and 3.6% (8/220) were CR-HMKP. The strains were mainly isolated from sputum (CRKP 53.6% vs HMKP 38% vs CR-HMKP 75%), blood (CRKP 13.1% vs HMKP 21.1%, vs CR-HMKP 0%), liver abscesses (CRKP 0% vs HMKP 5.6% vs CR-HMKP12.5%). Community-associated infections were mainly involved in HMKP (35.2%) than CRKP (21.4%) and CR-HMKP (12.5%), while CRKP and CR-HMKP were mostly detected in health-care setting ((87.5%, 78.6%) compared with HMKP (64.8%).
Antimicrobial Susceptibility Testing
Notably, the prevalence of CRKP and CR-HMKP strains exhibiting resistance to carbapenem antimicrobial was higher than that of HMKP (Fig. 1). All of the HMKP were sensitive to the three agents while all the CRKP and CR-HMKP were resistant to them. For HMKP, CRKP and CR-HMKP, MIC50 /90 of imipenem was 0.25/0.5, 128/128 and 128/256,MIC50/90 of meropenem was 0.125/0.125, 128/256 and 128/256, MIC50 /90 of ertapenem was 0.125/0.125, 256/256, 256/256.
Serotyping, MLST profiles and virulence-associated genes
To identify the serotype of 71 HMKP isolates, PCR was performed using primers for K1, K2, K5, K20, K54 and K57 serotypes. A total of 53 (74.6%) were positive for K1, K2, K20 and K57. Twenty-four (33.8%) isolates displayed the K1 serotype, 17 (23.9%) isolates showed K2 serotype, while 6 (8.4%) isolates were K20 and 6 (8.4%) isolates were K57. However, no K5 or K54 serotype was found. MLST analysis identified 22 STs among the 71 HMKP. The most prevalent ST in HMKP was ST23 (26.8%, 19/71), followed by ST65 (12.7%，9/71), ST11(5.6%，4/71) and ST86 (4.2%，3/71). There was strong association between STs and K serotypes. Sixteen of 24 (66.7%) K1 isolates belonged to ST23. Among the 17 K2 isolates, 58.8%(10/17) were ST65 and 23.5% (4/17) were ST375. Virulence genes rmpA, magA and iutA were found in 88.7% (63/71), 35.2% (25/71) and 84.5% (60/71) HMKP isolates respectively. Notably, all the magA positive HMKP were K1 serotype.
For 84 CRKP, 4.8% (4/84) isolates showed serotype K2 and 3.6% (3/84) were serotype K20 and the others (91.6%) were K-nontypable. MLST analysis showed 84.5% (71/84) were ST11, followed by ST15 (4.76%, 4/84). Furthermore, another 9 isolates displayed 6 different STs, including ST15, ST1869, ST323, ST45, ST709, ST722 and ST332. 1.2% (1/84) rmpA, 0% magA and 2.4% (2/84) iutA were positive among CRKP isolates. 85.7% (72/84) were blaKPC-2 positive.
All the 8 CR-HMKP isolates produced KPC-2 carbapenemase and belonged to ST11 by MLST while they were all K nontypable. Virulence genes rmpA, magA and iutA were found in 37.5% (3/8), 0 and 25% (2/8) CR-HMKP isolates.
Seven respective HMKP, CRKP and CR-HMKP isolates were randomly selected to perform serum killing to compare the ability of anti-phagocytosis. Survival rates of HMKP, CRKP and CR-HMKP isolates were evaluated in the presence of normal human serum over a 2-h period (Fig. 2). HMKP isolates exhibited markedly higher survival rates in the presence of normal human serum compared with CRKP and CR-HMKP.
Mouse Lethality Assay
Mouse lethality assay showed that the LD50 of all the 7 CR-HMKP and CRKP strains was > 106 CFU, which indicated very low virulence while LD50 of HMKP was less than 5 × 104 CFU. Furthermore, survival test indicated that mice treated with CRKP and CR-HMKP significantly enhanced survival compared to that for HMKP treated mice (Fig. 3a). Finally, mice sacrifice their life using 100% CO2 for CFU in blood and liver. As results, CRKP and CR-HMKP treated mice exhibited reduced bacterial burden (CFU) in blood and liver compared to HMKP-treated mice (Fig. 3b and c).