Comparative Analysis of a Carbapenemase-producing ...

1
Comparative Analysis of a Carbapenemase-producing Klebsiella pneumoniae Outbreak in a New York State Healthcare Facility using Multiple Typing Methods Catharine Prussing 1 , Alyssa Sossei 1 , Lauren Dentinger 1 , Oneida Shushe 1 , Emily Snavely 2 , Wolfgang Haas 1 , Pascal Lapierre 1 , Anna Kidney 1 , Navjot Singh 1 , Elizabeth Nazarian 1 , Kimberlee Musser 1 , Kara Mitchell 1 1 Wadsworth Center, NYSDOH, Albany, NY; 2 University of Utah – ARUP Laboratories, Salt Lake City, UT Objective To investigate a large outbreak of carbapenemase-producing carbapenem- resistant Enterobacteriaceae (CP-CRE) at a healthcare facility detected through the AR Laboratory Network using multiple molecular epidemiology typing methods. Methods 44 clinical isolates collected from 28 patients over 6 months All isolates were identified as K. pneumoniae using MALDI-TOF mass spectrometry 28 isolates were resistant to doripenem, imipenem, and/or meropenem by Kirby-Bauer disc diffusion Genomic DNA from cultured isolates was sequenced using the Illumina MiSeq platform Isolates were further characterized using three typing methods: 1. The OpGen Acuitas ® rapid real-time PCR AMR Gene Panel, which can detect five bacterial species and 47 antibiotic resistance (AR) genes. PCR results are uploaded into the Acuitas Lighthouse ® Software, which generates a unique Lighthouse Profile based on the organism and presence/absence of AR genes. 2. The AMR pipeline, an in-house developed whole genome sequencing (WGS) pipeline 1 , which conducts in silico multi-locus sequence typing (MLST) analysis using mlst 2 and identifies AR genes using ABRicate 3 3. The LegioCluster pipeline, an in-house developed WGS pipeline 4 , which assesses isolate relatedness using mutation event (ME) analysis A subset of 14 isolates were sequenced using long-read Oxford Nanopore Technologies MinION sequencing to further characterize the plasmid content and genomic context of AR genes. Genomes were assembled using Unicycler v0.4.8 5 and annotated using prokka v1.14.6 6 and AMRFinderPlus v3.9.3 7 . Plasmids were characterized using MOB-typer v3.0.0 8 , and by querying against the PLSDB plasmid database 9 . Results Conclusions References 1. Prussing et al. 2020. Front Microbiol 11:2007. 2. https://github.com/tseemann/mlst 3. https://github.com/tseemann/abricate 4. Haas et al. 2020. J Clin Microbiol 59:e00967-20. 5. Wick et al. 2017. PLoS Comput Biol 13:1-22. 6. Seemann 2014. Bioinformatics 30:2068-2069. 7. Feldgarden et al. 2019. Antimicrob Agents Chemother 63:e00483-19. 8. Robertson et al. 2020. Microb Genomics 6:e000435. 9. Galata et al. 2018. Nucleic Acids Res 47:D195-D202. Multiple typing methods improved our understanding of the molecular epidemiology of a large and complex outbreak of K. pneumoniae. These methods produced highly concordant results, though the WGS pipelines provided slightly increased resolution compared with the Acuitas ® AMR Gene Panel. MinION sequencing characterized the genomic context of AR genes and identified a plasmid that was shared by isolates that were not closely related by ME analysis. These findings can be used to improve infection control practices in this and other facilities. Figure 1. Comparison of three typing methods for characterization of outbreak isolates. All isolates carried multiple AR genes, and 28 isolates carried the bla KPC-3 carbapenemase gene. All typing methods identified four clusters: ME Cluster 1 (0-48 MEs apart, except for 2 distantly related isolates that differ from all other isolates by ~1,600-5,300 MEs): 18 KPC-positive isolates, ST258, Acuitas Lighthouse ® Profile 33, 34, or 36 ME Cluster 2 (0-15 MEs apart, except for 1 distantly related isolate that differs from all other isolates by ~100-1000 MEs): 8 KPC-positive isolates, ST258, Acuitas Lighthouse ® Profile 37 ME Cluster 3 (0-7 MEs apart): 12 KPC-negative isolates, ST307, Acuitas Lighthouse ® Profile 30 or 31 ME Cluster 4 (1-22 MEs apart): 4 KPC-negative isolates, ST20, Acuitas Lighthouse Profile 32 Two isolates (43 and 44) did not belong to a cluster by any typing method. Figure 2. Gene map and BLAST comparisons of bla KPC-3 -carrying plasmid identified in 5 isolates. MinION sequencing of 14 isolates identified 6 distinct bla KPC-3 -carrying plasmids. A highly similar ~70kb bla KPC-3 -carrying IncR plasmid was identified in Isolates 1, 2, 11, and 13 (all in ME Cluster 1), as well as Isolate 17, which was only distantly related to other isolates in ME Cluster 1. The plasmid from Isolate 17 (Kpn-WC17-006-02) differed from the other bla KPC-3 -carrying plasmids from ME Cluster 1 (Kpn-WC17-004-03, Kpn-WC17- 005-03, Kpn-WC17-013-03, and Kpn-WC17-014-03) by the insertion of the transposase gene ISKpn26 at nucleotide position 22,364-23,640. These plasmids were closely related to other previously sequenced plasmids (e.g. Genbank NZ_CP010396.1 and CP010362.1, both sequenced from K. pneumoniae isolates). Acknowledgements We thank the Wadsworth Center Advanced Genomic Technologies Center for conducting Illumina sequencing. This work was supported by the New York State Department of Health, Cooperative Agreement Number NU50CK000423 funded by the Centers for Disease Control and Prevention (CDC), Cooperative Agreement U60OE000103 funded by CDC through the Association of Public Health Laboratories, and a grant from the Empire State Development.

Transcript of Comparative Analysis of a Carbapenemase-producing ...

Comparative Analysis of a Carbapenemase-producing Klebsiella pneumoniae Outbreak in a New York State Healthcare Facility using Multiple Typing Methods Catharine Prussing1, Alyssa Sossei1, Lauren Dentinger1, Oneida Shushe1, Emily Snavely2, Wolfgang Haas1, Pascal Lapierre1, Anna Kidney1, Navjot Singh1,Elizabeth Nazarian1, Kimberlee Musser1, Kara Mitchell11Wadsworth Center, NYSDOH, Albany, NY; 2University of Utah – ARUP Laboratories, Salt Lake City, UT

Objective

To investigate a large outbreak of carbapenemase-producing carbapenem-resistant Enterobacteriaceae (CP-CRE) at a healthcare facility detected through the AR Laboratory Network using multiple molecular epidemiology typing methods.

Methods

• 44 clinical isolates collected from 28 patients over 6 months• All isolates were identified as K. pneumoniae using

MALDI-TOF mass spectrometry• 28 isolates were resistant to doripenem, imipenem,

and/or meropenem by Kirby-Bauer disc diffusion• Genomic DNA from cultured isolates was sequenced using

the Illumina MiSeq platform• Isolates were further characterized using three typing

methods:1. The OpGen Acuitas® rapid real-time PCR AMR Gene

Panel, which can detect five bacterial species and 47 antibiotic resistance (AR) genes. PCR results are uploaded into the Acuitas Lighthouse® Software, which generates a unique Lighthouse Profile based on the organism and presence/absence of AR genes.

2. The AMR pipeline, an in-house developed whole genome sequencing (WGS) pipeline1, which conducts in silico multi-locus sequence typing (MLST) analysis using mlst2 and identifies AR genes using ABRicate3

3. The LegioCluster pipeline, an in-house developed WGS pipeline4, which assesses isolate relatedness using mutation event (ME) analysis

• A subset of 14 isolates were sequenced using long-read Oxford Nanopore Technologies MinION sequencing to further characterize the plasmid content and genomic context of AR genes.• Genomes were assembled using Unicycler v0.4.85 and

annotated using prokka v1.14.66 and AMRFinderPlusv3.9.37.

• Plasmids were characterized using MOB-typer v3.0.08, and by querying against the PLSDB plasmid database9.

Results

Conclusions

References1. Prussing et al. 2020. Front Microbiol 11:2007.2. https://github.com/tseemann/mlst3. https://github.com/tseemann/abricate4. Haas et al. 2020. J Clin Microbiol 59:e00967-20.5. Wick et al. 2017. PLoS Comput Biol 13:1-22.6. Seemann 2014. Bioinformatics 30:2068-2069.7. Feldgarden et al. 2019. Antimicrob Agents Chemother 63:e00483-19.8. Robertson et al. 2020. Microb Genomics 6:e000435.9. Galata et al. 2018. Nucleic Acids Res 47:D195-D202.

• Multiple typing methods improved our understanding of the molecular epidemiology of a large and complex outbreak of K. pneumoniae.

• These methods produced highly concordant results, though the WGS pipelines provided slightly increased resolution compared with the Acuitas®

AMR Gene Panel.• MinION sequencing characterized the genomic context of AR genes and

identified a plasmid that was shared by isolates that were not closely related by ME analysis.

• These findings can be used to improve infection control practices in this and other facilities.

Figure 1. Comparison of three typing methods for characterization of outbreak isolates.• All isolates carried multiple AR genes, and 28 isolates carried the blaKPC-3 carbapenemase gene.• All typing methods identified four clusters:

• ME Cluster 1 (0-48 MEs apart, except for 2 distantly related isolates that differ from all other isolates by ~1,600-5,300 MEs): 18 KPC-positive isolates, ST258, Acuitas Lighthouse® Profile 33, 34, or 36

• ME Cluster 2 (0-15 MEs apart, except for 1 distantly related isolate that differs from all other isolates by ~100-1000 MEs): 8 KPC-positive isolates, ST258, Acuitas Lighthouse® Profile 37

• ME Cluster 3 (0-7 MEs apart): 12 KPC-negative isolates, ST307, Acuitas Lighthouse® Profile 30 or 31• ME Cluster 4 (1-22 MEs apart): 4 KPC-negative isolates, ST20, Acuitas Lighthouse Profile 32

• Two isolates (43 and 44) did not belong to a cluster by any typing method.

Figure 2. Gene map and BLAST comparisons of blaKPC-3-carrying plasmid identified in 5 isolates.• MinION sequencing of 14 isolates identified 6 distinct blaKPC-3-carrying plasmids.• A highly similar ~70kb blaKPC-3-carrying IncR plasmid was identified in Isolates 1,

2, 11, and 13 (all in ME Cluster 1), as well as Isolate 17, which was only distantly related to other isolates in ME Cluster 1.• The plasmid from Isolate 17 (Kpn-WC17-006-02) differed from the other

blaKPC-3-carrying plasmids from ME Cluster 1 (Kpn-WC17-004-03, Kpn-WC17-005-03, Kpn-WC17-013-03, and Kpn-WC17-014-03) by the insertion of the transposase gene ISKpn26 at nucleotide position 22,364-23,640.

• These plasmids were closely related to other previously sequenced plasmids (e.g. Genbank NZ_CP010396.1 and CP010362.1, both sequenced from K. pneumoniae isolates).

Acknowledgements

• We thank the Wadsworth Center Advanced Genomic Technologies Center for conducting Illumina sequencing. • This work was supported by the New York State Department of Health, Cooperative Agreement Number NU50CK000423

funded by the Centers for Disease Control and Prevention (CDC), Cooperative Agreement U60OE000103 funded by CDC through the Association of Public Health Laboratories, and a grant from the Empire State Development.