Clonal complexes of carbapenem-resistant Klebsiella pneumoniae recovered from community sewage

CR-Kp

carbapenem-resistant Klebsiella pneumoniae

CC

clonal complex

ST

sequence type

MDR

multidrug-resistant

GNB

Gram-negative bacilli

bla

β-lactamase

STP

sewage treatment plant

RS

raw sewage

TS

treated sewage

UW

upstream water

DW

downstream water

Carbapenem-resistant (CR) Enterobacterales species are a major concern involving patient care and public health decisions. Some isolates in this order, including CR Klebsiella pneumoniae (CR-Kp), may resist nearly all antibiotics, leaving as treatment only less effective and toxic options (Centers for Disease Control 2019). CR-Kp is responsible for life-threatening infections and has been globally recognized as a critical multidrug-resistant (MDR) microorganism. Additionally, CR-Kp strains have been found in different environmental sources including water matrices and soil. Environmental matrices harbor CR-Kp and the risks that this scenario poses to the public health should not be underestimated (Wareth & Neubauer 2021).

It is noticeable that carbapenem resistance rates in K. pneumoniae have steadily increased since the emergence of plasmid-encoded carbapenemase KPC in the first decade of 2000 (Yigit et al. 2001). Despite other carbapenemase genes occurred in K. pneumoniae, such as bla_NDM, bla_IMP, bla_VIM and bla_OXA-48, it is undoubtable that bla_KPC became the most prevalent and highly impacting carbapenemase in CR-Kp (Lee et al. 2016). Strains carrying bla_KPC have been described in all continents, with bla_KPC-2 and bla_KPC-3 being the most common alleles in strains involved in hospital outbreaks in many countries, including Brazil (Woodford et al. 2011; Migliorini et al. 2021). Although bla_KPC is the main carbapenem resistance gene in CR-Kp strains, bla_NDM and bla_OXA-48 also become rapidly disseminated (Lee et al. 2016).

The worldwide emergence of CR-Kp is supported by the occurrence and spread of highly adaptive and successful global clones (Woodford et al. 2011). The sequence type (ST) 258 of CR-Kp is commonly associated with hospital outbreaks, and it has contributed to spread bla genes in the north hemisphere countries since the early 2000s. Of note, ST258 has accounted for most of healthcare-associated infection (HAI) outbreaks linked to bla_KPC-positive CR-Kp strains reported in Germany and the USA (Kitchel et al. 2009; Becker et al. 2018). Likewise, ST11, an ancestral and single-locus variant (SLV) of ST258, is a highly predominant CR-Kp clone in Asia and South America (Wang et al. 2020). In Brazil, most bla_KPC-positive CR-Kp strains in hospital settings belong to ST11 and its SLV ST340 and ST437 (Andrade et al. 2011; Pereira et al. 2013).

The pandemic spread of bla_KPC CR-Kp strains is driven primarily by members from the clonal complex 258 (CC258), which comprises 96 STs, including the international clones ST258; ST11, ST340 and ST437 (belonging to CC11). Additionally, the strains belonging to CC15 (ST15 and ST14) account for about 20% of the outbreaks caused by MDR K. pneumoniae strains (Navon-Venezia et al. 2017). Recently, CC147 was also added to the list of high-risk clones as it was recognized as an emerging cause of HAI outbreaks in different regions of the world (Peirano et al. 2020).

The spread of antibiotic resistance has been recognized by the World Health Organization (WHO) as a significant threat to human health (WHO 2015, 2021). Additionally, the WHO and the International Water Association (IWA) have defined the environmental pathways of resistance transmission and environmental surveillance as priority issues that should integrate One Health surveillance strategies (Sano et al. 2020). Sewage treatment plants (STPs) act as epidemiologic reservoirs contributing for the amplification of CR isolates that ultimately are discharged along with treated effluents (Teban-Man et al. 2021). In this regard, K. pneumoniae strains account for most of the CR isolates that resist sewage treatment and are discharged in the environment (Pereira et al. 2022). Studies worldwide have detected a multitude of pandemic clones of CR-Kp as well as autochthonous clones of epidemiologic relevance in sewage samples (Jin et al. 2018; Surleac et al. 2020).

Recently, we reported the predominance of CR-Kp among bla_KPC-or-NDM strains isolated from STPs in the capital city of Brazil, Brasília (Pereira et al. 2022). In the present study, we carried out a further characterization concerning ST, clonal complexes and alleles of bla_KPC and bla_NDM in CR-Kp strains. Additionally, a historical collection of clinical CR-Kp strains was added to the phylogenetic analyses to verify clonality with CR-Kp strains recovered from sewage.

Brasília has a population of some 3 million inhabitants and an area of over 5,800 km² organized into administrative regions which differ from each other in the number of sewer connections and the number of hospitals and economic profile (Table 1). Around 85% of the sewage produced is collected in proper sewerage and treated in 14 public STPs, of which 9 carry out tertiary treatment (Table 1). Thus, STPs serve communities located in well-defined geographical catchment areas.

Three rounds of sewage and receiving water body sampling were carried out in April, July and August of 2017 including 12 STPs in Brasília. A total of 138 samples, including sewage samples (n = 70) and water samples (n = 68), were collected. Sewage samples were collected in the STP directly from the sewage inlet pipe [raw sewage (RS) – n = 35] and directly from the treated sewage outlet pipe [treated sewage (TS) – n = 35]. Water samples were collected in receiving water bodies on spots located 50 m upstream [upstream water (UW) – n = 30] and 50 m downstream [downstream water (DW) – n = 38] from the point where the sewage outlet pipe discharged the treated sewage. With regard to STP-13 and SPT-5, the water body receiving the treated sewage is Lake Paranoá (Supplementary Material, Figure 1). Given the absence of appreciable water streams on the lakeside, the water samples collected from the lake were all considered downstream water samples. Samples were collected in sterile conical tubes (50 mL) during the mornings, preserved at room temperature (20–22 °C) in insulated boxes and sent for culture at the Central Laboratory for Public Health (LACEN-DF) in the same work shift.

Five hundred microliters of each sample were cultured in Luria-Bertani broth supplemented with vancomycin (7.5 mg/L) and ertapenem (2.5 mg/L) at 36.5°C for 24 h. Positive CR cultures were streaked on chromogenic and differential agar (ChromID^® ESBL – bioMérieux, Craponne, France) and incubated at 36.5°C for 24 h for the isolation and presumptive identification of Gram-negative bacilli (GNB). Three colonies of each presumptive bacterial group were isolated per sample limited to a maximum of 10 colonies per sample. Isolates from each presumptive bacterial group (Escherichia coli, KESC [Klebsiella spp., Enterobacter spp., Serratia spp. and Citrobacter spp.], Proteeae and non-fermenting bacilli) were recovered per sample following the order of predominance of the colonies. Colonies were preserved into a semisolid nutrient medium (0.8% agar) (Acumedia Manufacturers, Inc., Lansing, USA) stored in hermetically sealed tubes (3 mL) at room temperature (20–22 °C) and were kept safe from direct light exposure.

The bacterial isolates were recovered and grown on Mueller Hinton agar for 24 h at 37 °C for obtaining isolated colonies. The identification was accomplished using the Vitek MS system [matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) system – bioMérieux, Marcy-L̀Etoile, France] in accordance with the manufacturer's instructions. E. coli strain ATCC™ (American Type Culture Collection) 8739 was used as a positive control. The Myla^® database (bioMérieux, Marcy-L’Etoile, França) was accessed for the identification of bacterial isolates. A confidence level greater than 90% was adopted for K. pneumoniae species assignment. Antimicrobial susceptibility tests were accomplished using the MicroScan WalkAway™ system (Dade, Behring, USA).

Carbapenemase genes were detected by standard polymerase chain reactions (PCRs). Supernatants derived from bacterial suspensions in deionized water and treated by boiling (100 °C for 15 min) were used as the source of DNA template. Primers were used to detect multiple alleles of carbapenemase genes (bla_KPC, bla_NDM, bla_IMP, bla_VIM and bla_OXA-48) (Faria-Junior et al. 2016) and the transposon Tn4401 harboring bla_KPC (Pereira et al. 2022) (Table 2). The thermocycling conditions were as follows: preheating at 92 °C for 1 min, followed by 30 cycles at 94 °C for 1 min, primer-specific annealing temperature (Table 2) for 30 s and 72 °C for 1 min.

Multilocus sequence typing (MLST) was performed in accordance with the Institut Pasteur scheme using seven housekeeping genes [rpoB (beta-subunit of RNA polymerase), gapA (glyceraldehyde 3-phosphate dehydrogenase); mdh (malate dehydrogenase); pgi (phosphoglucose isomerase); phoE (phosphorine E); infB (translation initiation factor 2) and tonB (periplasmic energy transducer)]. Primers and the ST assignment obeyed Institut Pasteur's protocols (https://bigsdb.pasteur.fr/klebsiella/klebsiella.html). BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems™, Foster City, CA, USA) was employed in sequencing reactions and amplicons were analyzed with ABI-Prism 3500 Genetic Analyzer (Applied Biosystems™, Carlsbad, CA, USA). Clonal complexes (CCs) were defined as groups of two or more independent isolates that shared identical alleles at six loci (single-locus variants, SLV). Minimum spanning trees were constructed with PHYLOViZ software (version 2.0) applying the goeBURST algorithm based on the 7-loci MLST Pasteur scheme.

In order to infer phylogenetic positioning of environmental CR-Kp strains, a historical collection of 52 clinical CR-Kp strains recovered in 13 hospitals (nine public and four private hospitals) and previously characterized (Silva et al. 2021) were included in the phylogenetic analysis (Supplementary Material, Table 1). MLST loci were concatenated into 2,121-base-long sequences (rpoB-gap-pgi-phoE-infB) (Gadagkar et al. 2005). The evolutionary distances were computed using the maximum composite likelihood method and are in the units of the number of base substitutions per site. Evolutionary analyses were conducted in MEGA X, and dendrograms were edited using the web-based software iTOL (https://itol.embl.de/).

Analyses were performed using the IBM^® SPSS^® Statistics software (version 20). Non-parametric analyses were carried out with Fisher's exact test (two-sided). Trends in contingency tables were assessed with the Mantel–Haenszel χ² statistic test for the overall degree of association between row and column variables.

All of the 154 CR-Kp strains were tested for carbapenemase genes with 143 (92.85%) tested positive for at least one gene. The bla_KPC gene was detected in 98 strains (63.6%) followed by bla_NDM, detected in 54 (35.1%) strains. Nine strains (5.8%) were simultaneously positive for both bla_KPC and bla_NDM genes. All of the bla_KPC-positive CR-Kp strains also tested positive for the transposon Tn4401. The genes bla_VIM, bla_IMP and bla_OXA-48 were not detected in CR-Kp strains. CR-Kp strains harboring bla_KPC or bla_NDM (bla_KPC-or-NDM strains) were equally distributed (p > 0.05) across the analyzed samples (RS, TS and DW) (Table 3).

bla_KPC-or-NDM CR-Kp strains were recovered from all but one STP (STP-9). Additionally, bla_KPC-or-NDM strains were unevenly recovered from STPs, as only 3 out of 11 SPTs (STP-2, STP-12 and STP-13) gathered 41.5% (64/154) of the bla_KPC-or-NDM strains. In Brasília, STPs serve areas differing in the volume of sewage produced, number of household sewer connections, number of hospitals and hospitalized patients (hospitalizations) as well as in farming activities (Table 1). The distribution of strains as a function of STP service profiles showed that the occurrence of bla_KPC-or-NDM strains was statistically associated with the number of district hospitals, number of hospitalizations and number of sewer connections in the regions serviced by STPs as well as with the volume of treated sewage (Table 4). The frequency of bla_KPC-or-NDM strains was higher as the number of district hospital, hospitalizations and sewer connections increased, clearly showing positive associations with these parameters (Table 4).

Forty-six percent (66/143) of bla_KPC-or-NDM strains were isolated from ST and DW samples (Table 5). bla_KPC-or-NDM strains were detected in TS and DW samples in 5 out of the 12 (45%) STPs, showing that the strains escape from sewage treatment and remain alive in receiving water bodies (Table 5). The distribution of bla_KPC-or-NDM strains in RS and TS samples was significantly different (p < 0.05) when comparing secondary and tertiary STPs (Supplementary Material, Table 2). TS in tertiary STPs showed a lower frequency for the isolation of bla_KPC-or-NDM strains. This fact suggests that tertiary STPs perform better in restraining multiresistant bacteria, despite the environment discharge of bla_KPC-or-NDM strains.

A total of 47 CR-Kp strains were submitted to the MLST assay and the sequencing of bla carbapenemase genes in order to uncover the diversity of genotypes as well as the occurrence of clinically relevant clones in sewage strains (Table 6). It was possible to assign 11 different STs to the CR-Kp strains. Half of the strains (n = 23) were assigned to CC11, most of which belonging to ST11 (n = 13–28.2%) and ST340 (n = 7–15.2%). Other STs were detected in more than one strain: ST15 (n = 3–6.5%), ST784 (n = 2–4.3%) and ST2703 (n = 2–4.3%).

Table 6

Allelic profile, ST and CC found in CR-Kp isolates recovered from STPs

^aAsterisk (*) followed by a number indicates single locus variant and the variant allele (1- gapA; 2- infB; 3- ndh; 4- pgi; 5- phoE; 6- rpoB; and 7- tonB). Double asterisks (**) indicate double locus variant. NA, non-amplified locus.

View Large

With regard to alleles of carbapenemase genes, 22 (47.8%) CR-Kp strains harbored bla_KPC-2, 10 (21.7%) strains harbored bla_NDM-1 and 1 strain harbored bla_KPC-3 (Table 6). The most frequent genotype (CC-bla allele) detected was CC11-bla_KPC2 (19/47 strains – 42%), which was found in strains isolated from 8 out of the 12 STPs (Table 6). It is noteworthy that 10 (21.7%) CR-Kp strains host simultaneously bla_KPC-2 and bla_NDM-1 (bla_KPC-2+NDM-1). bla_KPC-2+NDM-1 combination was detected in CR-Kp strains belonging to epidemic clones CC11 and CC17 as well as CC2703, CC735 and CC1859. Among eight strains recovered from TS samples and submitted to sequencing, genotypes CC11-bla_KPC-2 and CC2703-bla_NDM-1 were detected in five CR-Kp strains (Table 6).

Six major CCs formed by central STs (11, 15, 17, 147, 1642 and 2703) and their respective locus variants were identified, with five of them (11, 15, 17, 147 and 2703) shared by hospital and sewage CR-Kp strains. With regard to strains belonging to CC11 (n = 39), 23 strains (58%) were isolated from clinical samples recovered in hospital settings and 16 strains (41%) were isolated from sewage samples. Such even distribution of clinical and sewage samples into CCs was also observed for strains from CC15 (4 vs. 4), CC17 (2 vs. 3) and CC147 (3 vs. 2). Differently, CR-Kp strains from CC2703 were predominantly detected in clinical samples (13/16 –81%). The CC1642 was only assigned to sewage strains.

Concerning carbapenemase genes, bla_KPC and bla_NDM were unevenly detected in clonal complexes CC11 and CC2703. CR-Kp strains belonging to the CC11 accounted for 76% (36/47) of the bla_KPC-positive strains. In contrast, CC2703 accounted for 44% (16/36) of the bla_NDM-positive strains. bla_KPC+NDM-positive strains were identified in the CCs 11, 17, 1642 and 2703 (Figure 2).

K. pneumoniae is a leading cause of HAI, often caused by multidrug-resistant strains. Additionally, K. pneumoniae has multiple environmental reservoirs displaying a complex epidemiology that must be addressed from the perspective of a ‘One Health’ approach (Ludden et al. 2020; Wareth & Neubauer 2021). In England, a study with 87 K. pneumoniae strains recovered from different sources showed STs shared by clinical (human stool) and environmental strains (STP) as well as shared by strains from clinical samples (human stool) and livestock samples (Ludden et al. 2020). In addition to environmental resistance, studies have shown the ineffectiveness of STPs in eliminating CR-Kp from sewage, leading to the downstream contamination of water bodies, which contributes to the spread of antimicrobial resistance into the community (Ludden et al. 2020; Loudermilk et al. 2022).

In the present study, 7 out of 12 STPs surveyed had treated sewage samples positive for CR-Kp strains bearing bla_KPC-or-NDM, and 5 out of these STPs also had downstream water samples positive for CR-Kp bearing bla_KPC-or-NDM. Since 2010s, data from waste-based epidemiological studies have revealed the spreading of high-risk clones of CR-Kp strains (CC11 – ST340 and ST347) in Brazilian urban rivers (Oliveira et al. 2014). Additionally, the role of STPs impacted by hospital wastewater in spreading CR-Kp isolates in urban water matrices has also been reported (Picão et al. 2013). However, studies addressing the clonal structure of CR-Kp isolates recovered from sewage-impacted environments and hospital settings remain scarce in Brazil. We showed that international high-risk clones of CR-Kp strains bearing bla_KPC-or-NDM are shared by isolates recovered from sewage and clinical samples, including CC11 (ST11 and ST340), CC147 and ST15. Other studies have also shown a predominance of CC11, especially of the ST340, among CR-Kp strains recovered from clinical samples in hospital settings in Brazil (Monteiro et al. 2019; Tolentino et al. 2019). To the best of our knowledge, this is the first report on the high-risk clone CC147 harboring bla_KPC-or-NDM in Brazil (Peirano et al. 2020). In addition, several uncommon STs have been detected in CR-Kp strains isolated from HAI in Brazil, revealing the diversity of STs found in some regions of the country (Nakamura-Silva et al. 2022). Now, we have detected the underreported ST2703 in CR-Kp strains recovered from sewage and clinical samples. ST2703 is a relevant clone in local epidemiology, accounting for 25% of HAI caused by CR-Kp strains, being outnumbered only by ST11 strains (Supplementary Material, Table 1).

Our data corroborate the environmental resilience of the CR-Kp strains, as clinical and environmental CR-Kp strains were clustered in common clades despite having been isolated on different occasions over a 3-year period (2015–2017). Furthermore, clonality between strains recovered from clinical samples, hospital settings and wastewater has been demonstrated even in collections of Klebsiella strains historically sampled over 5 years (Loudermilk et al. 2022).

The high transmissibility of the Tn4401 transposon remains the primary mechanism for the spread of bla_KPC in K. pneumoniae and in other GNB species (Reyes et al. 2020; Pereira et al. 2022). In the current study, all of the bla_KPC-positive CR-Kp strains tested positive for the Tn4401 transposon. Our data support the role of Tn4401 as the main genetic element for the spread of bla_KPC in environmental CR-Kp strains.

STPs are considered epidemiological reservoirs that support the environmental spread of multidrug-resistant bacteria via treated effluents, regardless of the level of sewage treatment (Picão et al. 2013; Loudermilk et al. 2022). In the USA, CR Enterobacterales strains have been recovered from effluents in 42% of STPs reporting disinfection (tertiary treatment) with chlorination and in 12% of STPs reporting disinfection using ultraviolet radiation (Mathys et al. 2019). Likewise, our data showed that CR-Kp strains were recovered from 3 out of 7 (42%) STPs carrying out the tertiary level of sewage treatment.

Despite scientific evidence that the environment may play a significant role in spreading antibiotic resistance, especially in low-resource settings, the exact role of the environment in exposing humans to resistant bacteria is still unclear (Sano et al. 2020). Health implications of recreational exposure to CR-GNB also remain uncertain and are scarcely explored (Leonard et al. 2022). Nevertheless, the ingestion of water during recreational activities is recognized as an exposure route for asymptomatic colonization of humans (Leonard et al. 2022). Understanding the threat that high-risk clones of CR-Kp pose and the mechanisms by which people are exposed to these clones can support strategies to prevent antibiotic resistance spreading and community-acquired resistant infections.

CR-Kp strains are detected in sewage samples mainly in areas impacted by hospital activities and remain viable in water bodies downstream from STPs despite sewage treatment. Additionally, strains belonging to high-risk clones [ST11 and ST340 (CC11); ST15; and ST17] and locally relevant clones (ST2703) of CR-Kp were detected in sewage samples and showed clonality with strains recovered from patients in hospital settings. Strains with genotypes CC11-bla_KPC-2 and CC2703-bla_NDM-1 were detected in treated sewage samples. Thus, we conclude that STPs impacted by hospital activities contribute to spreading clinically relevant clones of CR-Kp strains in superficial water matrices. Additionally, sewage surveillance on STPs may help to understand local epidemiology of CR-Kp strains and to address the environmental risks to the public health.

This work was supported by the Fundação de Apoio à Pesquisa do Distrito Federal (FAP-DF) with grant no. 193.000.713/2016.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.