Skip to main content
PMC home page
eLife logoLink to eLife
. 2024 Apr 16;12:RP91853. doi: 10.7554/eLife.91853

High-risk Escherichia coli clones that cause neonatal meningitis and association with recrudescent infection

Nguyen Thi Khanh Nhu 1,2,3, Minh-Duy Phan 1,2,3, Steven J Hancock 2,3,†, Kate M Peters 1,2,3, Laura Alvarez-Fraga 2,3,‡, Brian M Forde 3,4, Stacey B Andersen 5, Thyl Miliya 6, Patrick NA Harris 4,7, Scott A Beatson 2,3, Sanmarie Schlebusch 4,7,8, Haakon Bergh 7, Paul Turner 6,9, Annelie Brauner 10, Benita Westerlund-Wikström 11, Adam D Irwin 3,4,12,✉, Mark A Schembri 1,2,3,✉
Editors: Bavesh D Kana13, Bavesh D Kana14
PMCID: PMC11021048  PMID: 38622998

Abstract

Neonatal meningitis is a devastating disease associated with high mortality and neurological sequelae. Escherichia coli is the second most common cause of neonatal meningitis in full-term infants (herein NMEC) and the most common cause of meningitis in preterm neonates. Here, we investigated the genomic relatedness of a collection of 58 NMEC isolates spanning 1974–2020 and isolated from seven different geographic regions. We show NMEC are comprised of diverse sequence types (STs), with ST95 (34.5%) and ST1193 (15.5%) the most common. No single virulence gene profile was conserved in all isolates; however, genes encoding fimbrial adhesins, iron acquisition systems, the K1 capsule, and O antigen types O18, O75, and O2 were most prevalent. Antibiotic resistance genes occurred infrequently in our collection. We also monitored the infection dynamics in three patients that suffered recrudescent invasive infection caused by the original infecting isolate despite appropriate antibiotic treatment based on antibiogram profile and resistance genotype. These patients exhibited severe gut dysbiosis. In one patient, the causative NMEC isolate was also detected in the fecal flora at the time of the second infection episode and after treatment. Thus, although antibiotics are the standard of care for NMEC treatment, our data suggest that failure to eliminate the causative NMEC that resides intestinally can lead to the existence of a refractory reservoir that may seed recrudescent infection.

Research organism: E. coli

Introduction

Neonatal meningitis (NM) is a devastating disease with a mortality rate of 10–15% and severe neurological sequelae including hearing loss, reduced motor skills, and impaired development in 30–50% of cases (Doctor et al., 2001; Stevens et al., 2003; Harvey et al., 1999). The incidence of disease is highest in low-income countries and occurs at a rate of 0.1–6.1/1000 live births (Harvey et al., 1999). Escherichia coli is the second most common cause of NM in full-term infants (herein NMEC), after group B Streptococcus (GBS) (Ouchenir et al., 2017; Gaschignard et al., 2011), and the most common cause of meningitis in preterm neonates (Gaschignard et al., 2011; Basmaci et al., 2015). Together, these two pathogens cause ~60% of all cases, with on average one case of NMEC for every two cases of GBS (May et al., 2005; Holt et al., 2001). In several countries, NM incidence caused by GBS has declined due to maternal intrapartum antibiotic prophylaxis; however, NM incidence caused by E. coli remains the same (May et al., 2005; van der Flier, 2021). Moreover, NMEC is a significant cause of relapsed infections in neonates (Anderson and Gilbert, 1990).

NMEC are categorised genetically based on multi-locus sequence type (ST) or by serotyping based on cell-surface O antigen (O), capsule (K), and flagella (H) antigens. Analysis of NMEC diversity in France revealed ~25% of isolates belong to the ST95 clonal complex (STc95) (Geslain et al., 2019), however, a global picture of NMEC epidemiology is lacking. NMEC possess a limited diversity of serotypes, dominated by O18:K1:H7, O1:K1, O7:K1, O16:K1, O83:K1, and O45:K1:H7, which together account for >70% of NMEC (Sarff et al., 1975; Plainvert et al., 2007; Bidet et al., 2007; Johnson et al., 2002). Notably, ~80% of NMEC express the K1 capsule, a polysaccharide comprising linear homopolymers of α2–8-linked N-acetyl neuraminic acid (Sarff et al., 1975; Robbins et al., 1974). Apart from the K1 capsule, specific NMEC virulence factors are less-well defined, though studies have demonstrated a role for S fimbriae (Prasadarao et al., 1993), the outer membrane protein OmpA (Prasadarao et al., 1996), the endothelial invasin IbeA (Huang et al., 2001), and the cytotoxin necrotising factor CNF1 (Wang and Kim, 2013) in translocation of NMEC across the blood–brain barrier and pathogenesis. A large plasmid encoding colicin V (ColV), colicin Ia bacteriocins, and several virulence genes including iron-chelating siderophore systems has also been strongly associated with NMEC virulence (Peigne et al., 2009).

Despite being the second major NM aetiology, genomic studies on NMEC are lacking, with most reporting single NMEC complete genomes. Here, we present the genomic analyses of a collection of 58 NMEC isolates obtained from seven different geographic regions over 46 years to understand virulence gene content, antibiotic resistance, and genomic diversity. In addition, we provide a complete genome for 18 NMEC isolates representing different STs, serotypes, and virulence gene profiles, thus more than tripling the number of available NMEC genomes that can be used as references in future studies. Three infants in our study suffered recrudescent invasive NMEC infection, and we show this was caused by the same isolate. We further revealed that patients that suffered recrudescent invasive infection had severe gut dysbiosis, and detected the infecting isolate in the intestinal microflora, suggesting NMEC colonisation of the gut provides a reservoir that can seed repeat infection.

Results

Establishment of an NMEC collection from geographically diverse locations

A collection of 52 NMEC isolates cultured from the blood or cerebrospinal fluid (CSF) of neonates with meningitis was established with the addition of six completely sequenced NMEC genomes available on the NCBI database. This yielded a final set of 58 NMEC isolates spanning 1974–2020. The isolates were obtained from seven different geographic locations; Finland (n = 17, 29.3%), Sweden (n = 14, 24.1%), Australia (n = 15, 25.9%), Cambodia (n = 7, 12.1%), USA (n = 3, 5.2%), France (n = 1, 1.7%), and the Netherlands (n = 1, 1.7%).

ST95 and ST1193 are the two major STs of NMEC

Phylogenetic analysis was performed on the 58 NMEC isolates, with an additional eight well-characterised E. coli strains included for referencing (EC958, UTI89, MS7163, CFT073, UMN026, 536, APEC01, and MG1655). The NMEC isolates were diverse, and spanned phylogroups A, B2, C, D, and F; the majority of isolates were from phylogroup B2 (n = 48, 82.8%). Overall, the isolates belonged to 22 STs, of which 15 STs only contained one isolate. ST95 (n = 20, 34.5%) and ST1193 (n = 9, 15.5%) were the two most common NMEC STs (Figure 1, Supplementary file 1). ST95 isolates were obtained over the entire study period, while ST1193 isolates were more recent and only obtained from 2013. Four isolates belonged to ST390 (6.9%), which is part of the STc95. One isolate belonged to a novel ST designated ST11637, which is part of the ST14 clonal complex (STc14) that also contains ST1193 (Figure 1, Supplementary file 1). Isolates from other common phylogroup B2 extra-intestinal pathogenic E. coli (ExPEC) lineages, ST131, ST73, and ST69, as well as several STs associated with environmental sources such as ST48 and ST23, were detected in the collection. However, it is notable that the high incidence of NM associated with ST95 and ST1193 does not reflect the broader high prevalence of major ExPEC clones associated with human infections in the publicly available Enterobase database (Zhou et al., 2020; Figure 1—figure supplement 1), suggesting ST95 and ST1193 exhibit specific virulence features associated with their capacity to cause NM.

Figure 1. Maximum likelihood phylogram displaying the relationship of the NMEC isolates with their associated serotype and virulence factor profile.

Non-NMEC isolates used in the analysis for referencing are italicised. The phylogram was built and recombination regions removed employing Parsnp, using 185,911 core single-nucleotide polymorphisms (SNPs) and NMEC strain IHE3034 as the reference. The scale bar indicates branch lengths in numbers of SNPs. NMEC isolates with available complete genomes are bold-italicised, while NMEC isolates that were completely sequenced in this study are indicated in bold and marked with an asterisk. The NMEC isolates that caused recrudescent invasive infection in this study are indicated in red. Branches are coloured according to phylogroups: orange, phylogroup F; red, phylogroup C; green, phylogroup A; violet, phylogroup D; and blue, phylogroup B2. The presence of specific virulence factors is indicated in dark blue. The phylogeny can be viewed interactively at https://microreact.org/project/oNfA4v16h3tQbqREoYtCXj-high-risk-escherichia-coli-clones-that-cause-neonatal-meningitis.

Figure 1.

Figure 1—figure supplement 1. Number of human-derived E. coli strains from ST95, ST1193, ST38, ST131, ST73, ST10, and ST69 available in the Enterobase database.

Figure 1—figure supplement 1.

Strains were stratified based on their year of isolation, spanning the periods before 2000, 2001–2005, 2006–2010, 2011–2015, and 2016–2022.
Figure 1—figure supplement 2. Antibiotic resistance gene profile of NMEC strains in the collection.

Figure 1—figure supplement 2.

The presence of each resistance gene is denoted by black shading.
Figure 1—figure supplement 3. ST95 NMEC strains contain more virulence factors than ST1193 NMEC strains.

Figure 1—figure supplement 3.

(A) The number of virulence genes (grouped as in Figure 1) for each strain within each sequence type (ST). (B) The number of virulence genes grouped by their functions in ST95 versus ST1193 strains. p-value was calculated using Mann–Whitney two-tailed unpaired test.
Figure 1—figure supplement 4. K1 capsule production in NMEC.

Figure 1—figure supplement 4.

K1 capsule production was detected by ELISA using a monoclonal antibody specific for polysialic acid. Strains with an OD420 >0.133 (mean + 3 standard deviations of a negative control kpsD mutant; dashed line) were considered positive for K1 capsule production. Data points represent independent biological replicates with horizontal lines as the mean.
Open in a new tab

Eighteen NMEC isolates were completely sequenced using complementary long-read Oxford Nanopore Technology, enabling accurate comparison of NMEC genome size, genomic island composition and location, and plasmid and prophage diversity. These isolates spanned the diversity in the collection, representing 11 different STs, including two ST1193 isolates (one with the dominant O75:H5 serotype and one with an unusual O6:H5 serotype), five ST95 isolates with different serotypes, and one isolate from the novel ST11637.

Antibiotic resistance in NMEC

Antibiotic resistance profiling revealed an overall low level of resistance in the collection. The ST1193 isolates contained fluoroquinolone resistance defining mutations in gyrA (S83L D87N) and parC (S80I), as previously described for this lineage (Johnson et al., 2019). In addition, 77.8% of ST1193 isolates (7/9 isolates) also harboured at least one gene conferring resistance to aminoglycosides (aac(3)-IId, aadA5, aph(3″)-Ib, and aph(6)-Id), trimethoprim (dftA17), and sulphonamides (sul1 and sul2) (Figure 1—figure supplement 2). Six out of the seven isolates from Cambodia had more than one antibiotic resistance gene, likely reflecting increased antibiotic resistance rates in this region (Reed et al., 2019). Indeed, in addition to gyrA and parC mutations for fluoroquinolone resistance, CAM-NMEC-6 contains 14 antibiotic resistance genes (including resistance to third-generation cephalosporins and carbapenems) and CAM-NMEC-4 contains 11 antibiotic resistance genes (Figure 1—figure supplement 2).

Virulence factors in NMEC

The isolates exhibited variable distribution of virulence genes previously linked to NMEC pathogenesis. The most prevalent genes were those involved in iron uptake, including the enterobactin (98%), yersiniabactin (98%), aerobactin (62%), and salmochelin (55%) siderophore systems, and the heme receptors chuA (93%) and hma (62%) (Figure 1). Also common were the sitABCD genes encoding an iron/manganese transporter (98%). The presence of fimbrial and afimbrial adhesins was also diverse. The most prevalent adhesins were type 1 fimbriae (100%), mat (ecp) fimbriae (98%), and the fdeC adhesin (98%). Genes encoding P and S fimbriae were detected in 36% and 22% of NMEC isolates, respectively. The most prevalent toxin was the uropathogenic-specific genotoxin usp (83%), which was only found in phylogroup B2 isolates. Other toxin genes encoding the serine protease autotransporters Vat (65% prevalence) and Sat (29%), hemolysin (12%), and cytotoxic necrotising factor-1 (7%) were less prevalent. Additional virulence genes included the aslA arylsulfatase (95%), the iss lipoprotein (76%), and the ibeA invasin (33%). The ColV-plasmid was present in 33% of the isolates (Figure 1, Supplementary file 1). Direct comparison of virulence factors between ST95 and ST1193, the two most dominant NMEC STs, revealed that the ST95 isolates (n = 20) contained significantly more virulence factors than the ST1193 isolates (n = 9); p-value <0.001, Mann–Whitney two-tailed unpaired test (Figure 1—figure supplement 3).

NMEC comprise a dominant K1 capsule type and a limited pool of O and H serotypes

The capsule type of the NMEC isolates was determined by in silico typing. K1 was the dominant capsule type in the collection (43/58 isolates, 74.1%) (Figure 1). Thirty-four of these isolates were available for capsule testing, and we confirmed K1 expression by ELISA in all but two isolates (Figure 1—figure supplement 4). Other capsule types included K2, K5, and K14 (Supplementary file 1). A capsule type could not be resolved for 12 isolates, of which eight did not possess a Group II or III capsule type based on the absence of the conserved kpsD gene (Figure 1; Supplementary file 1).

In silico O antigen (O) and flagella (H) serotypes were also determined. O18 was the most common O type (n = 16, 27.6%), followed by O75 (n = 8, 13.9%) and O2 (n = 7, 12.1%). The most dominant H types were H7 (n = 19, 32.8%), H5 (n = 13, 22.4%), and H4 (n = 9, 15.5%). The most common serotype was O18:H7:K1 (n = 14, 24.1%); these isolates belonged to STc95 (nine ST95, four ST390, and one ST416). The second most common serotype was O75:H5:K1 (n = 8, 13.8%); six isolates from ST1193 possessed this serotype.

NMEC can cause recrudescent invasive infection despite appropriate antibiotic treatment

During 2019 - 2020, three patients from which NMEC isolates were originally cultured suffered recrudescent invasive infection (Figure 1; MS21522, MS21524, and MS22733), providing an opportunity to compare the infecting isolates over time using whole-genome sequencing. In all cases, the infecting E. coli isolates were susceptible to the therapy, which comprised cefotaxime (50 mg/kg/dose 8 hourly), switched to ceftriaxone (100 mg/kg/day) to facilitate home parenteral antibiotic administration. Bacterial culture was performed from blood, CSF, urine, and/or stool during the infection period (Figure 2). These patients were from different regions in Australia.

Figure 2. Infection and treatment profile of patients suffering NM and recrudescent invasive infection.

Figure 2.

Open in a new tab

Indicated is the hospital admission history of patients, together with the timeline of sample collection, identified E. coli isolates and their infection source, and isolate identification based on whole-genome sequencing, metagenomic sequencing (MetaWGS), or fimH amplicon sequencing. Genomic relatedness is indicated based on the number of single-nucleotide polymorphisms (SNPs). The time of admission for the initial episode is indicated as day 0, with subsequent timepoints indicated as days post initial admission. Admission and discharge days are indicated in red and green, respectively.

Patient 1

Patient 1 (0–8 weeks of age) was admitted to the emergency department with fever, respiratory distress, and sepsis. The child was diagnosed with meningitis based on a CSF pleocytosis (>2000 white blood cells [WBCs], low glucose, elevated protein), positive CSF E. coli PCR and a positive blood culture for E. coli (MS21522). Two weeks after completion of a 3-week course of appropriately dosed therapy with third-generation cephalosporins as described above, the child developed similar symptoms of fever and irritability. Lumbar puncture was performed and the CSF culture was positive for E. coli (MS21576). Both the initial blood culture isolate and the relapse CSF isolate were non-susceptible to ciprofloxacin and gentamycin, and whole-genome sequencing revealed they were identical (ST1193 O18:K1:H5; fimH64), with no single-nucleotide polymorphisms (SNPs) nor indels (Figure 2A). Unlike the typical ST1193 O75 serotype (Johnson et al., 2019), this isolate contained a unique O18 serotype. The isolate possessed mutations in gyrA (S83L D87N) and parC (S80I), which explain its resistance to ciprofloxacin, as well as a multidrug resistance IncF plasmid containing genes conferring resistance to aminoglycosides (aac(3)-IId, aadA5, aph(3″)-Ib, and aph(6)-Id), trimethoprim (dfrA17), sulphonamides (sul1 and sul2), and macrolides (mphA) (Figure 1—figure supplement 2). Treatment of the relapse was extended to 6 weeks of intravenous therapy. At follow-up, no anatomical or immunological abnormality was identified and development is normal.

Patient 2

Patient 2 (0–8 weeks of age) presented to the emergency department with a febrile illness. Blood and urine cultures on admission were positive for E. coli. CSF taken 24 hr after treatment revealed pleocytosis (>300 WBCs, >95% polymorphs) but no bacteria were cultured. The patient completed a 3-week course of appropriately dosed antibiotic therapy with third-generation cephalosporins. In the 6-week period after discharge, the child had several short admissions to hospital, but no infection was identified. At 11 weeks post initial infection, the child was readmitted to hospital with high fever. CSF cultures were negative and microscopy was normal, but cultures from blood, urine, and faeces were all positive for E. coli. Whole-genome sequencing revealed that all isolates belonged to ST537 O75:H5 (fimH5; STc14). Pairwise comparison of the recrudescent isolates showed that the urine and fecal isolates were identical to the original isolate, while the blood isolate contained one nonsynonymous SNP in the mdoH gene encoding a glucan biosynthesis glucosyltransferase (T1358G; V453G). This mutation is located in the large cytoplasmic domain of MdoH likely involved in polymerisation of glucose from UDP glucose; the isolate exhibited a mucoid colony morphology suggestive of increased colanic acid production. The isolates did not possess plasmids nor antibiotic resistance genes. The infant experienced recurrent urinary tract infections with E. coli and other urinary pathogens through infancy despite normal urinary tract anatomy. At follow-up, no other history of invasive infection nor identified immunodeficiency were noted, and the child was reported to be developing normally.

Patient 3

Patient 3 (0–8 weeks of age) was admitted to the paediatric intensive care unit with fever and seizures. CSF and blood cultured a fully susceptible E. coli. Two weeks after completing a 4-week course of appropriate therapy with third-generation cephalosporins, the infant was readmitted to hospital with fever and irritability, with further investigation identifying E. coli in CSF, urine, and blood. Three weeks after the completion of the 6-week treatment course, the infant experienced a second relapse, with E. coli isolated from both CSF and blood. Whole-genome sequencing revealed that all isolates were identical and belonged to ST131 O25b:K1:H4 (fimH22). These isolates contained a ColV-virulence plasmid, but did not harbour acquired antibiotic resistance genes. The infant received a further 6-week course of therapy. Extensive imaging studies including repeated magnetic resonance imaging of the head and spine, computed tomography imaging of the head and chest, ultrasound imaging of abdomen and pelvis, and nuclear medicine imaging did not show a congenital malformation or abscess. Immunological work-up did not show a known primary immunodeficiency. At 2 years of age, speech delay is reported but no other developmental abnormality.

The gut as a reservoir to seed recrudescent infection

In all three patients that suffered NM and recrudescent invasive infection, the causative isolates were susceptible to third-generation cephalosporins, suggesting the existence of a persistent reservoir that could evade the cidal effect of antibiotic treatment and seed repeat infection. Indeed, the fact that the causative E. coli isolate was detected from a fecal sample at the time of the recrudescent infection in patient 2 (day 77 after initial admission), suggests that NMEC could persist in the gut and cause repeat infection, an observation that has also been reported for uropathogenic E. coli that cause recurrent urinary tract infection (Forde et al., 2019) and acute pyelonephritis in infants (Tullus et al., 1984). Therefore, we retrospectively examined available stored fecal samples from patient 2 at 8- and 12-week follow-up visits post recrudescent infection (days 149 and 174 after initial admission) and patient 3 during treatment and at discharge after the third episode (days 126 and 147 after initial admission) using short-read metagenomic sequencing (Figure 3). Although no fecal samples were available for comparative analysis from either patient prior to antibiotic treatment, we observed a low level of diversity in the composition of the microbiome of both patients, consistent with severe dysbiosis. The microbiome of patient 2 was dominated by Enterobacter (37.4% relative abundance), Achromobacter (23.4% relative abundance), and Bacteroides (22.7% relative abundance) genera at 8-week post recrudescent infection (day 149 after initial admission), and by Bacteroides genera (75.8% relative abundance) at 12-week post recrudescent infection (day 174 after initial admission). The relative abundance of E. coli was 2.05% and 4.1% in each of these samples, respectively, and further analysis using StrainGE (van Dijk et al., 2022) showed that the isolates were most closely matched to the original causative MS21524 isolate. We further employed complementary long-read metagenomic sequencing to analyse the 8-week post relapse infection sample, which enabled construction of a complete E. coli genome that was identical to the causative ST537 (fimH5) isolate (Figures 2 and 3; Supplementary file 3). In the 12-week post recrudescent infection fecal sample from patient 2, amplicon sequencing targeting fimH identified the presence of E. coli with the same fimH type as the causative isolate (fimH5). Thus, two independent analyses of samples taken 4 weeks apart demonstrated the existence of the E. coli ST537 isolate in the intestinal microflora of patient 2. In patient 3, the microbiome was dominated by Enterococcus genera at both timepoints examined (93% and 97.4% relative abundance). We were unable to detect E. coli by fimH amplicon sequencing and the relative abundance of E. coli in these fecal samples was extremely low (<0.01%) based on metagenomic sequencing (Supplementary file 3). The extensive dysbiosis revealed in this patient is likely an outcome of the three rounds of antibiotic treatment.

Figure 3. Relative abundance of bacterial genera (≥0.01%) in the gut microbiome of patient 2 at 8- and 12-week follow-up post relapsed infection (days 149 and 174 after initial admission) (A) and patient 3 during treatment and at discharge after the third episode (days 126 and 147 after initial admission) (B).

Figure 3.

Open in a new tab

Discussion

In this study, we present a genomic analysis of 58 NMEC isolates obtained over 46 years spanning seven different geographic locations and reveal a dominance of ST95 and ST1193. We also provide direct evidence to implicate the gut as a reservoir for recrudescent invasive infection in some patients despite appropriate antibiotic treatment.

The majority of the NMEC isolates in our study belonged to phylogroup B2 (82.8%), an observation consistent with other reports (Wijetunge et al., 2015; Bidet et al., 2007). These isolates were predominantly from two major STs, ST95, and ST1193. ST95 represents a major clonal lineage responsible for urinary tract and bloodstream infections (Manges et al., 2019; Kallonen et al., 2017), and were identified throughout the period of investigation. This lineage was also previously shown to cause ~25% of NM cases in France in the period 2004–2015 (Geslain et al., 2019), demonstrating its enhanced capacity to cause disseminated infection in newborns. ST1193, on the other hand, was first identified in 2012 (Platell et al., 2012), and is the second most common fluoroquinolone-resistant E. coli lineage after ST131 (Johnson et al., 2019; Tchesnokova et al., 2019). ST1193 causing NM was first reported in the USA in 2016 (Nielsen et al., 2018). Here, ST1193 accounted for 15.5% of NMEC isolates, all of which were obtained from 2013, and was the dominant lineage since this time. This is consistent with a report in China that showed ST1193 was the most common NMEC (21.4%), followed by ST95 (17.9%), between 2009 and 2015 (Ding et al., 2021). Concerningly, the ST1193 isolates examined here carry genes encoding several aminoglycoside-modifying enzymes, generating a resistance profile that may lead to the clinical failure of empiric regimens such as ampicillin and gentamicin, a therapeutic combination used in many settings to treat NM and early-onset sepsis (Fleiss et al., 2023; Fuchs et al., 2018). This, in combination with reports of co-resistance to third-generation cephalosporins for some ST1193 isolates (Johnson et al., 2019; Ding et al., 2021), would limit the choice of antibiotic treatments. The dominance of both ST95 and ST1193 in our collection is notable since other widespread E. coli phylogroup B2 lineages such as ST131, ST73, ST69, and ST12 do not cause similar rates of NM disease. We speculate this is due to the prevailing K1 polysialic acid capsule serotype found in ST95 and the newly emerged ST1193 clone (Johnson et al., 2019; Goh et al., 2017) in combination with other virulence factors (Johnson et al., 2002; Wijetunge et al., 2015; Bidet et al., 2007; Figure 4) and the immature immune system of preterm infants. Understanding the risk of these clones, as well as perinatal transmission and antibiotic resistance patterns, may inform the appropriateness of interventions such as maternal screening or antimicrobial prophylaxis.

Figure 4. Summary of key NMEC virulence genes based on genome profiling performed in this study.

Figure 4.

Open in a new tab

Shown are shared virulence genes common to most NMEC, as well as ST95- and ST1193-specific NMEC virulence genes.

© 2024, BioRender Inc

Figure 4 was created using BioRender, and is published under a CC BY-NC-ND 4.0 license. Further reproductions must adhere to the terms of this license.

Although reported rarely, recrudescent invasive E. coli infection in NM patients, including several infants born preterm, has been documented in single study reports (Vissing et al., 2021; Bingen et al., 1993). In these reports, infants received appropriate antibiotic treatment based on antibiogram profiling and no clear clinical risk factors to explain recrudescence were identified, highlighting our limited understanding of NM aetiology. Here, we tracked NMEC recurrence using whole-genome sequencing in three patients that suffered NM and recrudescent invasive infection, and demonstrated that the isolate causing recrudescence was the same as the original causative isolate and susceptible to the initial antibiotic therapy. In one patient (patient 2), we identified the causative isolate in the stool at days 77, 149, and 174 after initial detection in the bloodstream, providing direct evidence of persistence in the gut, and implicating this site as a reservoir to seed repeat infection. This isolate belonged to ST537 (serotype O75:H5) and is from the same clonal complex as ST1193 (i.e. STc14).

This study had several limitations. First, our NMEC collection was restricted to seven geographic regions, a reflection of the difficulty in acquiring isolates causing this disease. Second, we did not have access to a complete set of stool samples spanning pre- and post-treatment in the patients that suffered NM and recrudescent invasive infection. This impacted our capacity to monitor E. coli persistence and evaluate the effect of antibiotic treatment on changes in the microbiome over time. Third, we did not have access to urine or stool samples from the mother of the infants that suffered recrudescence, and thus cannot rule out mother-to-child transmission as a mechanism of reinfection. Finally, we did not have clinical data on the weeks of gestation for all patients, and thus could not compare virulence factors from NMEC isolated from preterm versus term infants. Regardless, our study describes the genomic diversity of NMEC, highlighting ST95 and ST1193 as the most important clonal lineages associated with this devastating disease. Although antibiotics are the standard of care for NMEC treatment, we show that even when appropriate antibiotics are used, in some cases they do not eliminate the causative NMEC that resides intestinally. Together with associated antibiotic-driven dysbiosis, this reveals a need to consider diagnostic and therapeutic interventions to mitigate the risk of recrudescent infection.

Methods

Bacterial isolates

A collection of 52 NMEC isolates obtained from 1974 to 2020 was achieved from Sweden, Finland, Cambodia, and Australia. Isolates were stored in glycerol at −80°C until use. All isolates were cultured in Lysogeny broth. The collection comprised 42 isolates from confirmed meningitis cases (29 cultured from CSF and 13 cultured from blood) and 10 isolates from clinically diagnosed meningitis cases (all cultured from blood) (Supplementary file 1). This collection was complemented by the addition of six completely sequenced NMEC genomes available on the NCBI database, namely strains IHE3034, RS218, S88, NMEC58, MCJCHV-1, and CE10.

DNA extraction, genome sequencing, and analyses

Genome sequencing was performed using paired-end Illumina methodology. Illumina sequencing data were processed by removing adapters and low-quality reads using Trimmomatic v0.36 (Bolger et al., 2014), with a minimum quality score of 10 and minimum read length of 50. Trimmed reads were de novo assembled using SPAdes v3.12.0 (Bankevich et al., 2012) with default parameters. Draft assemblies of the 52 NMEC isolates from this study, together with six complete NMEC genomes and eight complete genomes from other characterised E. coli representing different phylogroups, were subjected to phylogenetic analysis using parsnp v1.5.3 (Treangen et al., 2014). A subset of 18 isolates were additionally sequenced using Oxford Nanopore Technology long-read sequencing (Nanopore). Complete NMEC genomes were achieved using a combination of Illumina short-read and Nanopore long-read data and analysis employing the MicroPIPE tool (Murigneux et al., 2021).

In silico and molecular analyses

Virulence-associated genes, antibiotic resistance genes, plasmids and serotyping were evaluated using ABRicate (RRID:SCR_021093, version 0.8; https://github.com/tseemann/abricate) with built-in databases (Chen et al., 2016; Feldgarden et al., 2019; Carattoli et al., 2014; Ingle et al., 2016), with the percentage nucleotide identity and coverage cut-off set at 90% and 80%, respectively. Capsule typing was performed employing Kaptive (Wyres et al., 2016) using an in-house E. coli capsule database (Goh et al., 2017) and manually checked. Chromosomal point mutations associated with antibiotic resistance were detected using PointFinder (Zankari et al., 2017). FimH amplicon sequencing was performed as previously described (Willner et al., 2014; Chen et al., 2018); allelic variants were identified using FimTyper (Roer et al., 2017).

K1 ELISA

K1 capsule expression was detected by ELISA using an anti-polysialic acid antibody single chain Fv fragment (Nagae et al., 2013) as the primary antibody, anti-His antibody, and alkaline phosphatase anti-mouse IgG as the secondary and tertiary antibodies, respectively; p-nitrophenylphosphate (Sigma) was used as the substrate. Optical density was measured at 420 nm.

Metagenomic sequencing and analyses

Metagenomic sequencing was performed on DNA extracted from fecal samples using the Illumnina NovaSeq6000 platform. Adapters and low-quality reads were trimmed using Trimmomatic v0.36 (Bolger et al., 2014), employing a minimum quality score of 10 and minimum read length of 50. Sequencing reads corresponding to human DNA were discarded by mapping the trimmed reads to the human genome hg38 (accession number GCA_000001405.29) using bowtie2 (Langmead and Salzberg, 2012). Taxonomical profiling was performed with Kraken2 (Wood et al., 2019) followed by Bracken (Lu et al., 2017).

Long-read metagenomic sequencing

Long-read metagenomic sequencing was performed on DNA extracted from a fecal sample. A HiFi gDNA library was prepared using the SMRTbell Express Template Prep Kit 2.0 (PacBio, 100-938-900) according to the low input protocol (PacBio, PN 101-730-400 Version 06 [June 2020]). As the sample DNA was already fragmented with a tight peak (mode size 9.4 kb), no shearing was performed; the sample was concentrated using Ampure PB beads (PacBio, PCB-100-265-900) and used directly as input into library preparation. The entire quantity of purified DNA (360 ng) was used to make the library as follows. The DNA was treated to remove single-stranded overhangs, followed by a DNA damage repair reaction and an end-repair/A-tailing reaction. Overhang barcoded adapters were ligated to the A-tailed library fragments, followed by a nuclease treatment to remove damaged library fragments, and then purification with AMPure PB beads. The library was size-selected to remove fragments <3 kb using AMPure PB beads. The final purified, size-selected library was quantified on the Qubit fluorometer using the Qubit dsDNA HS assay kit (Invitrogen, Q32854) to assess concentration, and run on the Agilent Femto Pulse using the 55 kb BAC Analysis Kit (Agilent, FP-1003-0275) to assess fragment size distribution.

Sequencing was performed using the PacBio Sequel II (software/chemistry v10.1). The library pool was prepared for sequencing according to the SMRT Link (v10.1) sample setup calculator, following the standard protocol for Diffusion loading with Ampure PB bead purification, using Sequencing Primer v5, Sequel II Binding Kit v2.2, and the Sequel II DNA Internal Control v1. Adaptive loading was utilised, with nominated on-plate loading concentration of 80 pM. The polymerase-bound library was sequenced on 1 SMRT Cell with a 30-hr movie time plus a 2-hr pre-extension using the Sequel II Sequencing 2.0 Kit (PacBio, 101-820-200) and SMRT Cell 8M (PacBio, 101-389-001).

After sequencing, the data were processed to generate CCS reads and demultiplex samples using the default settings of the CCS with Demultiplexing application in SMRT Link (v10.1). The demultiplexed reads were assembled de novo using Hifiasm (Cheng et al., 2021). Assembled contigs were subject to taxonomic profiling using kraken2 (Wood et al., 2019) and fastANI (Jain et al., 2018). Contigs taxonomically assigned as E. coli were subjected to in silico sequence typing using MLST (version 2.11) (https://github.com/tseemann/mlst; Seemann, 2022) and mlst profiles from PubMLST (Jolley and Maiden, 2010).

Acknowledgements

The authors would like to thank Michelle Bauer for technical expertise and the laboratories contributing the isolates, Pathology Queensland and Mater Pathology. At the time of the study SS was affiliated with Mater Pathology, South Brisbane, Australia.

Funding Statement

The funders had no role in study design, data collection, and interpretation, or the decision to submit the work for publication. For the purpose of Open Access, the authors have applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission.

Contributor Information

Adam D Irwin, Email: a.irwin@uq.edu.au.

Mark A Schembri, Email: m.schembri@uq.edu.au.

Bavesh D Kana, University of the Witwatersrand, South Africa.

Bavesh D Kana, University of the Witwatersrand, South Africa.

Funding Information

This paper was supported by the following grants:

  • National Health and Medical Research Council APP1181958 to Nguyen Thi Khanh Nhu, Minh-Duy Phan, Mark A Schembri.

  • National Health and Medical Research Council GNT1197743 to Adam D Irwin.

  • Queensland Children's Hospital Foundation 50270 to Patrick NA Harris, Scott A Beatson, Sanmarie Schlebusch, Adam D Irwin, Mark A Schembri.

  • Australian Infectious Diseases Research Centre to Nguyen Thi Khanh Nhu, Adam D Irwin, Mark A Schembri.

  • Wellcome Trust 10.35802/220211 to Paul Turner.

  • National Health and Medical Research Council APP2001431 to Nguyen Thi Khanh Nhu, Minh-Duy Phan, Mark A Schembri.

Additional information

Competing interests

No competing interests declared.

Author contributions

Formal analysis, Funding acquisition, Investigation, Writing – original draft, Writing – review and editing.

Formal analysis, Funding acquisition, Investigation, Writing – review and editing.

Investigation, Writing – review and editing.

Investigation, Writing – review and editing.

Investigation, Writing – review and editing.

Investigation, Writing – review and editing.

Investigation, Writing – review and editing.

Investigation, Writing – review and editing.

Resources, Funding acquisition, Writing – review and editing.

Resources, Funding acquisition, Writing – review and editing.

Resources, Funding acquisition, Writing – review and editing.

Resources, Writing – review and editing.

Resources, Funding acquisition, Writing – review and editing.

Conceptualization, Resources, Writing – review and editing.

Conceptualization, Resources, Writing – review and editing.

Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Writing – original draft, Writing – review and editing.

Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Writing – original draft, Writing – review and editing.

Ethics

The study received ethical approval from the Children's Health Queensland Human Research Ethics Committee (LNR/18/QCHQ/45045), ratified by The University of Queensland (2019000752). The study was deemed as low risk with waiver of informed consent and consent to publish. The Children's Health Queensland Human Research Ethics Committee approved an amendment on 27 July 2021 to report the clinical details included in the case series. Written informed consent was obtained from the carers of each of the infants.

Additional files

Supplementary file 1. Isolates used in this study.
elife-91853-supp1.xlsx (32.5KB, xlsx)
Supplementary file 2. Completely sequenced NMEC isolates.
elife-91853-supp2.xlsx (32.7KB, xlsx)
Supplementary file 3. Metagenomic sequence analysis.
elife-91853-supp3.xlsx (63.7KB, xlsx)
Supplementary file 4. Accession numbers of strains sequenced in the study.
elife-91853-supp4.xlsx (16.6KB, xlsx)

Data availability

Genome sequence data have been deposited in the Sequence Read Archive under the BioProjects PRJNA757133 and PRJNA893826. Sample accession numbers are listed in Supplementary file 4.

The following datasets were generated:

Phan M-D, Hancock SJ, Peters KM, Alvarez-Fraga L, Forde BM, Andersen SB, Miliya T, Harris PNA, Beatson SA, Schlebusch S, Bergh H, Turner P, Brauner A, Westerlund-Wikström B, Irwin AD, Schembri MA, Nhu NTK. 2024. Neonatal meningitis Escherichia coli. NCBI BioProject. PRJNA757133

Phan M-D, Hancock SJ, Peters KM, Alvarez-Fraga L, Forde BM, Andersen SB, Miliya T, Harris PNA, Beatson SA, Schlebusch S, Bergh H, Turner P, Brauner A, Westerlund-Wikström B, Irwin AD, Schembri MA, Nhu NTK. 2024. Culture-independent long read metagenomic diagnostics for genomic surveillance and infection control of pathogenic bacteria in clinical settings. NCBI BioProject. PRJNA893826

References

  1. Anderson SG, Gilbert GL. Neonatal gram negative meningitis: a 10-year review, with reference to outcome and relapse of infection. Journal of Paediatrics and Child Health. 1990;26:212–216. doi: 10.1111/j.1440-1754.1990.tb02432.x. [DOI] [PubMed] [Google Scholar]
  2. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, Lesin VM, Nikolenko SI, Pham S, Prjibelski AD, Pyshkin AV, Sirotkin AV, Vyahhi N, Tesler G, Alekseyev MA, Pevzner PA. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. Journal of Computational Biology. 2012;19:455–477. doi: 10.1089/cmb.2012.0021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Basmaci R, Bonacorsi S, Bidet P, Biran V, Aujard Y, Bingen E, Béchet S, Cohen R, Levy C. Escherichia coli meningitis features in 325 children from 2001 to 2013 in france. Clinical Infectious Diseases. 2015;61:779–786. doi: 10.1093/cid/civ367. [DOI] [PubMed] [Google Scholar]
  4. Bidet P, Mahjoub-Messai F, Blanco J, Blanco J, Dehem M, Aujard Y, Bingen E, Bonacorsi S. Combined multilocus sequence typing and O serogrouping distinguishes Escherichia coli subtypes associated with infant urosepsis and/or meningitis. The Journal of Infectious Diseases. 2007;196:297–303. doi: 10.1086/518897. [DOI] [PubMed] [Google Scholar]
  5. Bingen E, Cavé H, Aujard Y, Lambert-Zechovsky N, Desjardins P, Elion J, Denamur E. Molecular analysis of multiply recurrent meningitis due to Escherichia coli K1 in an infant. Clinical Infectious Diseases. 1993;16:82–85. doi: 10.1093/clinids/16.1.82. [DOI] [PubMed] [Google Scholar]
  6. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30:2114–2120. doi: 10.1093/bioinformatics/btu170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Carattoli A, Zankari E, García-Fernández A, Voldby Larsen M, Lund O, Villa L, Møller Aarestrup F, Hasman H. In silico detection and typing of plasmids using PlasmidFinder and plasmid multilocus sequence typing. Antimicrobial Agents and Chemotherapy. 2014;58:3895–3903. doi: 10.1128/AAC.02412-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chen L, Zheng D, Liu B, Yang J, Jin Q. VFDB 2016: hierarchical and refined dataset for big data analysis--10 years on. Nucleic Acids Research. 2016;44:D694–D697. doi: 10.1093/nar/gkv1239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chen Z, Phan M-D, Bates LJ, Peters KM, Mukerjee C, Moore KH, Schembri MA. The urinary microbiome in patients with refractory urge incontinence and recurrent urinary tract infection. International Urogynecology Journal. 2018;29:1775–1782. doi: 10.1007/s00192-018-3679-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cheng H, Concepcion GT, Feng X, Zhang H, Li H. Haplotype-resolved de novo assembly using phased assembly graphs with hifiasm. Nature Methods. 2021;18:170–175. doi: 10.1038/s41592-020-01056-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Ding Y, Zhang J, Yao K, Gao W, Wang Y. Molecular characteristics of the new emerging global clone ST1193 among clinical isolates of Escherichia coli from neonatal invasive infections in China. European Journal of Clinical Microbiology & Infectious Diseases. 2021;40:833–840. doi: 10.1007/s10096-020-04079-0. [DOI] [PubMed] [Google Scholar]
  12. Doctor BA, Newman N, Minich NM, Taylor HG, Fanaroff AA, Hack M. Clinical outcomes of neonatal meningitis in very-low birth-weight infants. Clinical Pediatrics. 2001;40:473–480. doi: 10.1177/000992280104000901. [DOI] [PubMed] [Google Scholar]
  13. Feldgarden M, Brover V, Haft DH, Prasad AB, Slotta DJ, Tolstoy I, Tyson GH, Zhao S, Hsu CH, McDermott PF, Tadesse DA, Morales C, Simmons M, Tillman G, Wasilenko J, Folster JP, Klimke W. Using the NCBI AMRFinder Tool to Determine Antimicrobial Resistance Genotype-Phenotype Correlations Within a Collection of NARMS Isolates. bioRxiv. 2019 doi: 10.1101/550707. [DOI]
  14. Fleiss N, Schwabenbauer K, Randis TM, Polin RA. What’s new in the management of neonatal early-onset sepsis? Archives of Disease in Childhood. Fetal and Neonatal Edition. 2023;108:10–14. doi: 10.1136/archdischild-2021-323532. [DOI] [PubMed] [Google Scholar]
  15. Forde BM, Roberts LW, Phan M-D, Peters KM, Fleming BA, Russell CW, Lenherr SM, Myers JB, Barker AP, Fisher MA, Chong T-M, Yin W-F, Chan K-G, Schembri MA, Mulvey MA, Beatson SA. Population dynamics of an Escherichia coli ST131 lineage during recurrent urinary tract infection. Nature Communications. 2019;10:3643. doi: 10.1038/s41467-019-11571-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fuchs A, Bielicki J, Mathur S, Sharland M, Van Den Anker JN. Reviewing the WHO guidelines for antibiotic use for sepsis in neonates and children. Paediatrics and International Child Health. 2018;38:S3–S15. doi: 10.1080/20469047.2017.1408738. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Gaschignard J, Levy C, Romain O, Cohen R, Bingen E, Aujard Y, Boileau P. Neonatal bacterial meningitis: 444 cases in 7 years. The Pediatric Infectious Disease Journal. 2011;30:212–217. doi: 10.1097/inf.0b013e3181fab1e7. [DOI] [PubMed] [Google Scholar]
  18. Geslain G, Birgy A, Adiba S, Magnan M, Courroux C, Levy C, Cohen R, Bidet P, Bonacorsi S. Genome sequencing of strains of the most prevalent clonal group of O1:K1:H7 Escherichia coli that causes neonatal meningitis in France. BMC Microbiology. 2019;19:17. doi: 10.1186/s12866-018-1376-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Goh KGK, Phan M-D, Forde BM, Chong TM, Yin W-F, Chan K-G, Ulett GC, Sweet MJ, Beatson SA, Schembri MA. Genome-wide discovery of genes required for capsule production by uropathogenic Escherichia coli. mBio. 2017;8:e01558-17. doi: 10.1128/mBio.01558-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Harvey D, Holt DE, Bedford H. Bacterial meningitis in the newborn: a prospective study of mortality and morbidity. Seminars in Perinatology. 1999;23:218–225. doi: 10.1016/s0146-0005(99)80066-4. [DOI] [PubMed] [Google Scholar]
  21. Holt DE, Halket S, de Louvois J, Harvey D. Neonatal meningitis in England and Wales: 10 years on. Archives of Disease in Childhood. Fetal and Neonatal Edition. 2001;84:F85–F89. doi: 10.1136/fn.84.2.f85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Huang SH, Chen YH, Kong G, Chen SH, Besemer J, Borodovsky M, Jong A. A novel genetic island of meningitic Escherichia coli K1 containing the ibeA invasion gene (GimA): functional annotation and carbon-source-regulated invasion of human brain microvascular endothelial cells. Functional & Integrative Genomics. 2001;1:312–322. doi: 10.1007/s101420100039. [DOI] [PubMed] [Google Scholar]
  23. Ingle DJ, Valcanis M, Kuzevski A, Tauschek M, Inouye M, Stinear T, Levine MM, Robins-Browne RM, Holt KE. In silico serotyping of E. coli from short read data identifies limited novel O-loci but extensive diversity of O:H serotype combinations within and between pathogenic lineages. Microbial Genomics. 2016;2:e000064. doi: 10.1099/mgen.0.000064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Jain C, Rodriguez-R LM, Phillippy AM, Konstantinidis KT, Aluru S. High throughput ANI analysis of 90K prokaryotic genomes reveals clear species boundaries. Nature Communications. 2018;9:5114. doi: 10.1038/s41467-018-07641-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Johnson JR, Oswald E, O’Bryan TT, Kuskowski MA, Spanjaard L. Phylogenetic distribution of virulence‐associated genes among Escherichia coli isolates associated with neonatal bacterial meningitis in the netherlands. The Journal of Infectious Diseases. 2002;185:774–784. doi: 10.1086/339343. [DOI] [PubMed] [Google Scholar]
  26. Johnson TJ, Elnekave E, Miller EA, Munoz-Aguayo J, Flores Figueroa C, Johnston B, Nielson DW, Logue CM, Johnson JR. Phylogenomic analysis of extraintestinal pathogenic Escherichia coli sequence type 1193, an emerging multidrug-resistant clonal group. Antimicrobial Agents and Chemotherapy. 2019;63:e01913-18. doi: 10.1128/AAC.01913-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jolley KA, Maiden MCJ. BIGSdb: Scalable analysis of bacterial genome variation at the population level. BMC Bioinformatics. 2010;11:595. doi: 10.1186/1471-2105-11-595. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kallonen T, Brodrick HJ, Harris SR, Corander J, Brown NM, Martin V, Peacock SJ, Parkhill J. Systematic longitudinal survey of invasive Escherichia coli in England demonstrates a stable population structure only transiently disturbed by the emergence of ST131. Genome Research. 2017;27:1437–1449. doi: 10.1101/gr.216606.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nature Methods. 2012;9:357–359. doi: 10.1038/nmeth.1923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lu J, Breitwieser FP, Thielen P, Salzberg SL. Bracken: estimating species abundance in metagenomics data. PeerJ Computer Science. 2017;3:e104. doi: 10.7717/peerj-cs.104. [DOI] [Google Scholar]
  31. Manges AR, Geum HM, Guo A, Edens TJ, Fibke CD, Pitout JDD. Global extraintestinal pathogenic Escherichia coli (ExPEC) lineages. Clinical Microbiology Reviews. 2019;32:e00135-18. doi: 10.1128/CMR.00135-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. May M, Daley AJ, Donath S, Isaacs D, Australasian Study Group for Neonatal Infections Early onset neonatal meningitis in Australia and New Zealand, 1992-2002. Archives of Disease in Childhood. Fetal and Neonatal Edition. 2005;90:F324–F327. doi: 10.1136/adc.2004.066134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Murigneux V, Roberts LW, Forde BM, Phan M-D, Nhu NTK, Irwin AD, Harris PNA, Paterson DL, Schembri MA, Whiley DM, Beatson SA. MicroPIPE: validating an end-to-end workflow for high-quality complete bacterial genome construction. BMC Genomics. 2021;22:474. doi: 10.1186/s12864-021-07767-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nagae M, Ikeda A, Hane M, Hanashima S, Kitajima K, Sato C, Yamaguchi Y. Crystal structure of anti-polysialic acid antibody single chain Fv fragment complexed with octasialic acid: insight into the binding preference for polysialic acid. The Journal of Biological Chemistry. 2013;288:33784–33796. doi: 10.1074/jbc.M113.496224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Nielsen DW, Ricker N, Barbieri NL, Wynn JL, Gómez-Duarte OG, Iqbal J, Nolan LK, Allen HK, Logue CM. Complete genome sequence of the multidrug-resistant neonatal meningitis Escherichia coli serotype O75:H5:K1 strain mcjchv-1 (NMEC-O75) Microbiology Resource Announcements. 2018;7:10. doi: 10.1128/MRA.01043-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ouchenir L, Renaud C, Khan S, Bitnun A, Boisvert A-A, McDonald J, Bowes J, Brophy J, Barton M, Ting J, Roberts A, Hawkes M, Robinson JL. The epidemiology, management, and outcomes of bacterial meningitis in infants. Pediatrics. 2017;140:e20170476. doi: 10.1542/peds.2017-0476. [DOI] [PubMed] [Google Scholar]
  37. Peigne C, Bidet P, Mahjoub-Messai F, Plainvert C, Barbe V, Médigue C, Frapy E, Nassif X, Denamur E, Bingen E, Bonacorsi S. The plasmid of Escherichia coli strain S88 (O45:K1:H7) that causes neonatal meningitis is closely related to avian pathogenic E. coli plasmids and is associated with high-level bacteremia in a neonatal rat meningitis model. Infection and Immunity. 2009;77:2272–2284. doi: 10.1128/IAI.01333-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Plainvert C, Bidet P, Peigne C, Barbe V, Médigue C, Denamur E, Bingen E, Bonacorsi S. A new O-antigen gene cluster has A key role in the virulence of the Escherichia coli meningitis clone O45:K1:H7. Journal of Bacteriology. 2007;189:8528–8536. doi: 10.1128/JB.01013-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Platell JL, Trott DJ, Johnson JR, Heisig P, Heisig A, Clabots CR, Johnston B, Cobbold RN. Prominence of an O75 clonal group (clonal complex 14) among non-ST131 fluoroquinolone-resistant Escherichia coli causing extraintestinal infections in humans and dogs in Australia. Antimicrobial Agents and Chemotherapy. 2012;56:3898–3904. doi: 10.1128/AAC.06120-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Prasadarao NV, Wass CA, Hacker J, Jann K, Kim KS. Adhesion of S-fimbriated Escherichia coli to brain glycolipids mediated by sfaA gene-encoded protein of S-fimbriae. The Journal of Biological Chemistry. 1993;268:10356–10363. [PubMed] [Google Scholar]
  41. Prasadarao NV, Wass CA, Weiser JN, Stins MF, Huang SH, Kim KS. Outer membrane protein A of Escherichia coli contributes to invasion of brain microvascular endothelial cells. Infection and Immunity. 1996;64:146–153. doi: 10.1128/iai.64.1.146-153.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Reed TAN, Krang S, Miliya T, Townell N, Letchford J, Bun S, Sar B, Osbjer K, Seng S, Chou M, By Y, Vanchinsuren L, Nov V, Chau D, Phe T, de Lauzanne A, Ly S, Turner P, Cambodia Technical Working Group on Antimicrobial Resistance Antimicrobial resistance in Cambodia: a review. International Journal of Infectious Diseases. 2019;85:98–107. doi: 10.1016/j.ijid.2019.05.036. [DOI] [PubMed] [Google Scholar]
  43. Robbins JB, McCracken GH, Gotschlich EC, Orskov F, Orskov I, Hanson LA. Escherichia coli K1 capsular polysaccharide associated with neonatal meningitis. The New England Journal of Medicine. 1974;290:1216–1220. doi: 10.1056/NEJM197405302902202. [DOI] [PubMed] [Google Scholar]
  44. Roer L, Tchesnokova V, Allesøe R, Muradova M, Chattopadhyay S, Ahrenfeldt J, Thomsen MCF, Lund O, Hansen F, Hammerum AM, Sokurenko E, Hasman H. Development of a web tool for Escherichia coli subtyping based on fimH alleles. Journal of Clinical Microbiology. 2017;55:2538–2543. doi: 10.1128/JCM.00737-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Sarff LD, McCracken GH, Schiffer MS, Glode MP, Robbins JB, Orskov I, Orskov F. Epidemiology of Escherichia coli K1 in healthy and diseased newborns. Lancet. 1975;1:1099–1104. doi: 10.1016/s0140-6736(75)92496-4. [DOI] [PubMed] [Google Scholar]
  46. Seemann T. Mlst. 157c519GitHub. 2022 https://github.com/tseemann/mlst
  47. Stevens JP, Eames M, Kent A, Halket S, Holt D, Harvey D. Long term outcome of neonatal meningitis. Archives of Disease in Childhood. Fetal and Neonatal Edition. 2003;88:F179–F184. doi: 10.1136/fn.88.3.f179. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Tchesnokova VL, Rechkina E, Larson L, Ferrier K, Weaver JL, Schroeder DW, She R, Butler-Wu SM, Aguero-Rosenfeld ME, Zerr D, Fang FC, Ralston J, Riddell K, Scholes D, Weissman S, Parker K, Spellberg B, Johnson JR, Sokurenko EV. Rapid and extensive expansion in the united states of a new multidrug-resistant Escherichia coli clonal group, sequence type 1193. Clinical Infectious Diseases. 2019;68:334–337. doi: 10.1093/cid/ciy525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Treangen TJ, Ondov BD, Koren S, Phillippy AM. The harvest suite for rapid core-genome alignment and visualization of thousands of intraspecific microbial genomes. Genome Biology. 2014;15:524. doi: 10.1186/s13059-014-0524-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Tullus K, Hörlin K, Svenson SB, Källenius G. Epidemic outbreaks of acute pyelonephritis caused by nosocomial spread of P fimbriated Escherichia coli in children. The Journal of Infectious Diseases. 1984;150:728–736. doi: 10.1093/infdis/150.5.728. [DOI] [PubMed] [Google Scholar]
  51. van der Flier M. Neonatal meningitis: small babies, big problem. The Lancet Child & Adolescent Health. 2021;5:386–387. doi: 10.1016/S2352-4642(21)00092-4. [DOI] [PubMed] [Google Scholar]
  52. van Dijk LR, Walker BJ, Straub TJ, Worby CJ, Grote A, Schreiber HL, Anyansi C, Pickering AJ, Hultgren SJ, Manson AL, Abeel T, Earl AM. StrainGE: a toolkit to track and characterize low-abundance strains in complex microbial communities. Genome Biology. 2022;23:74. doi: 10.1186/s13059-022-02630-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Vissing NH, Mønster MB, Nordly S, Dayani GK, Heedegaard SS, Knudsen JD, Nygaard U. Relapse of neonatal Escherichia coli meningitis: did we miss something at first? Children. 2021;8:126. doi: 10.3390/children8020126. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Wang MH, Kim KS. Cytotoxic necrotizing factor 1 contributes to Escherichia coli meningitis. Toxins. 2013;5:2270–2280. doi: 10.3390/toxins5112270. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Wijetunge DSS, Gongati S, DebRoy C, Kim KS, Couraud PO, Romero IA, Weksler B, Kariyawasam S. Characterizing the pathotype of neonatal meningitis causing Escherichia coli (NMEC) BMC Microbiology. 2015;15:211. doi: 10.1186/s12866-015-0547-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Willner D, Low S, Steen JA, George N, Nimmo GR, Schembri MA, Hugenholtz P. Single clinical isolates from acute uncomplicated urinary tract infections are representative of dominant in situ populations. mBio. 2014;5:e01064-13. doi: 10.1128/mBio.01064-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Wood DE, Lu J, Langmead B. Improved metagenomic analysis with Kraken 2. Genome Biology. 2019;20:257. doi: 10.1186/s13059-019-1891-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Wyres KL, Wick RR, Gorrie C, Jenney A, Follador R, Thomson NR, Holt KE. Identification of Klebsiella capsule synthesis loci from whole genome data. Microbial Genomics. 2016;2:e000102. doi: 10.1099/mgen.0.000102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Zankari E, Allesøe R, Joensen KG, Cavaco LM, Lund O, Aarestrup FM. PointFinder: a novel web tool for WGS-based detection of antimicrobial resistance associated with chromosomal point mutations in bacterial pathogens. The Journal of Antimicrobial Chemotherapy. 2017;72:2764–2768. doi: 10.1093/jac/dkx217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Zhou Z, Alikhan NF, Mohamed K, Fan Y, Agama Study G, Achtman M. The EnteroBase user’s guide, with case studies on Salmonella transmissions, Yersinia pestis phylogeny, and Escherichia core genomic diversity. Genome Research. 2020;30:138–152. doi: 10.1101/gr.251678.119. [DOI] [PMC free article] [PubMed] [Google Scholar]

eLife assessment

Bavesh D Kana 1

This valuable study presents findings characterising the genomic features of E. coli isolated from neonatal meningitis from seven countries, and documents bacterial persistence and reinfection in two case studies. The genomic analyses are solid, although the inclusion of a larger number of isolates from more diverse geographies would have strengthened the generalisability of findings. The work will be of interest to people involved in the management of neonatal meningitis patients, and those studying E. coli epidemiology, diversity, and pathogenesis.

Reviewer #1 (Public Review):

Anonymous

Summary:

This study uses whole genome sequencing to characterise the population structure and genetic diversity of a collection of 58 isolates of E. coli associated with neonatal meningitis (NMEC) from seven countries, including 52 isolates that the authors sequenced themselves and a further 6 publicly available genome sequences. Additionally, the study used sequencing to investigate three case studies of apparent relapse. The data show that in all three cases, the relapse was caused by the same NMEC strain as the initial infection. In two cases they also found evidence for gut persistence of the NMEC strain, which may act as a reservoir for persistence and reinfection in neonates. This finding is of clinical importance as it suggests that decolonisation of the gut could be helpful in preventing relapse of meningitis in NMEC patients.

Strengths:

The study presents complete genome sequences for n=18 diverse isolates, which will serve as useful references for future studies of NMEC. The genomic analyses are high quality, the population genomic analyses are comprehensive and the case study investigations are convincing. The full data set (including phylogenetic tree, annotated with source, lineage and virulence factor information) are publicly available in interactive form via the MicroReact platform.

Weaknesses:

The NMEC collection described in the study includes isolates from just seven countries. The majority (n=51/58, 88%) are from high-income countries in Europe, Australia or North America; the rest are from Cambodia (n=7, 12%). Therefore it is not clear how well the results reflect the global diversity of NMEC, nor the populations of NMEC affecting the most populous regions.

The virulence factors section highlights several potentially interesting genes that are present at apparently high frequency in the NMEC genomes; however without knowing their frequency in the broader E. coli population it is hard to know the significance of this.

eLife. 2024 Apr 16;12:RP91853. doi: 10.7554/eLife.91853.3.sa2

Author response

Nhu Nguyen 1, Minh-Duy Phan 2, Steven J Hancock 3, Kate M Peters 4, Laura Alvarez-Fraga 5, Brian M Forde 6, Stacey B Andersen 7, Thyl Miliya 8, Patrick Harris 9, Scott Beatson 10, Sanmarie Schlebusch 11, Haakon Bergh 12, Paul Turner 13, Annelie Brauner 14, Benita Westerlund-Wikström 15, Adam D Irwin 16, Mark A Schembri 17

The following is the authors’ response to the original reviews.

eLife assessment

This study presents valuable findings characterising the genomic features of E. coli isolated from neonatal meningitis from seven countries, and documents bacterial persistence and reinfection in two case studies. The genomic analyses are solid, although the inclusion of a larger number of isolates from more diverse geographies would have strengthened the generalisability of findings. The work will be of interest to people involved in the management of neonatal meningitis patients, and those studying E. coli epidemiology, diversity, and pathogenesis.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This study uses whole genome sequencing to characterise the population structure and genetic diversity of a collection of 58 isolates of E. coli associated with neonatal meningitis (NMEC) from seven countries, including 52 isolates that the authors sequenced themselves and a further 6 publicly available genome sequences. Additionally, the study used sequencing to investigate three case studies of apparent relapse. The data show that in all three cases, the relapse was caused by the same NMEC strain as the initial infection. In two cases they also found evidence for gut persistence of the NMEC strain, which may act as a reservoir for persistence and reinfection in neonates. This finding is of clinical importance as it suggests that decolonisation of the gut could be helpful in preventing relapse of meningitis in NMEC patients.

Strengths:

The study presents complete genome sequences for n=18 diverse isolates, which will serve as useful references for future studies of NMEC. The genomic analyses are high quality, the population genomic analyses are comprehensive and the case study investigations are convincing.

We agree

Weaknesses:

The NMEC collection described in the study includes isolates from just seven countries. The majority (n=51/58, 88%) are from high-income countries in Europe, Australia, or North America; the rest are from Cambodia (n=7, 12%). Therefore it is not clear how well the results reflect the global diversity of NMEC, nor the populations of NMEC affecting the most populous regions.

The virulence factors section highlights several potentially interesting genes that are present at apparently high frequency in the NMEC genomes; however, without knowing their frequency in the broader E. coli population it is hard to know the significance of this.

We acknowledged the limitations of our NMEC collection in the Discussion. We agree the prevalence of virulence factors in our collection is interesting. The limited size of our collection prevented further evaluation of the prevalence of these virulence factors in a broader E. coli population.

Reviewer #2 (Public Review):

Summary:

In this work, the authors present a robust genomic dataset profiling 58 isolates of neonatal meningitis-causing E. coli (NMEC), the largest such cohort to be profiled to date. The authors provide genomic information on virulence and antibiotic resistance genomic markers, as well as serotype and capsule information. They go on to probe three cases in which infants presented with recurrent febrile infection and meningitis and provide evidence indicating that the original isolate is likely causing the second infection and that an asymptomatic reservoir exists in the gut. Accompanying these results, the authors demonstrate that gut dysbiosis coincides with the meningitis.

Strengths:

The genomics work is meticulously done, utilizing long-read sequencing.

The cohort of isolates is the largest to be sampled to date.

The findings are significant, illuminating the presence of a gut reservoir in infants with repeating infection.

We agree

Weaknesses:

Although the cohort of isolates is large, there is no global representation, entirely omitting Africa and the Americas. This is acknowledged by the group in the discussion, however, it would make the study much more compelling if there was global representation.

We agree. In the Discussion we state this is likely a reflection of the difficulty in acquiring isolates causing neonatal meningitis, in particular from countries with limited microbiology and pathology resources.

Reviewer #3 (Public Review):

Summary:

In this manuscript, Schembri et al performed a molecular analysis by WGS of 52 E. coli strains identified as "causing neonatal meningitis" from several countries and isolated from 1974 to 2020. Sequence types, virulence genes content as well as antibiotic-resistant genes are depicted. In the second part, they also described three cases of relapse and analysed their respective strains as well as the microbiome of three neonates during their relapse. For one patient the same E. coli strain was found in blood and stool (this patient had no meningitis). For two patients microbiome analysis revealed a severe dysbiosis.

Major comments:

Although the authors announce in their title that they study E. coli that cause neonatal meningitis and in methods stipulate that they had a collection of 52 NMEC, we found in Supplementary Table 1, 29 strains (therefore most of the strains) isolated from blood and not CSF. This is a major limitation since only strains isolated from CSF can be designated with certainty as NMEC even if a pleiocytose is observed in the CSF. A very troubling data is the description of patient two with a relapse infection. As stated in the text line 225, CSF microscopy was normal and culture was negative for this patient! Therefore it is clear that patient without meningitis has been included in this study.

We have reviewed the clinical data for our 52 NMEC isolates, noting that for some of the older Finish isolates we relied on previous publications. This data is shown in Table S1. To address the Reviewer’s comment, we have added the following text to the methods section (new text underlined).

‘The collection comprised 42 isolates from confirmed meningitis cases (29 cultured from CSF and 13 cultured from blood) and 10 isolates from clinically diagnosed meningitis cases (all cultured from blood).’

Patient 2 was initially diagnosed with meningitis based on a positive blood culture in the presence of CSF pleocytosis (>300 WBCs, >95% polymorphs). We understand there may be some confusion with reference to a relapsed infection, which we now more accurately describe as recrudescent invasive infection in the revised manuscript.

Another major limitation (not stated in the discussion) is the absence of clinical information on neonates especially the weeks of gestation. It is well known that the risk of infection is dramatically increased in preterm neonates due to their immature immunity. Therefore E. coli causing infection in preterm neonates are not comparable to those causing infection in term neonates notably in their virulence gene content. Indeed, it is mentioned that at least eight strains did not possess a capsule, we can speculate that neonates were preterm, but this information is lacking. The ages of neonates are also lacking. The possible source of infection is not mentioned, notably urinary tract infection. This may have also an impact on the content of VF.

We agree. In the Discussion we now note the following (new text underlined):

‘… we did not have clinical data on the weeks of gestation for all patients, and thus could not compare virulence factors from NMEC isolated from preterm versus term infants.’

Submission to Medrxiv, a requirement for review of our manuscript at eLife, necessitated the removal of some patient identifying information, including precise age and detailed medical history.

Sequence analysis reveals the predominance of ST95 and ST1193 in this collection. The high incidence of ST95 is not surprising and well previously described, therefore, the concluding sentence line 132 indicating that ST95 E. coli should exhibit specific virulence features associated with their capacity to cause NM does not add anything. On the contrary, the high incidence of ST1193 is of interest and should have been discussed more in detail. Which specific virulence factors do they harbor? Any hypothesis explaining their emergence in neonates?

We compared the virulence factors of ST95 and ST1193 and summarized this information in Figure 4. We also discussed how the K1 polysialic acid capsule in ST95 and ST1193 could contribute to the emergence of these STs in NM. Specifically, we stated the following: ‘We speculate this is due to the prevailing K1 polysialic acid capsule serotype found in ST95 and the newly emerged ST1193 clone [22, 37] in combination with other virulence factors [15, 28, 29] (Figure 4) and the immature immune system of preterm infants.’

In the paragraph depicted the VF it is only stated that ST95 contained significantly more VF than the ST1193 strains. And so what? By the way "significantly" is not documented: n=?, p=?

We compared the prevalence of known virulence factors between ST95 and ST1193, and showed that ST95 strains in our collection contained significantly more virulence factors than the ST1193 strains. The P-value and the statistical test used were included in Supplementary Figure 3. To address the reviewers concern, we have now also added this to the main manuscript text as follows (new text underlined):

‘Direct comparison of virulence factors between ST95 and ST1193, the two most dominant NMEC STs, revealed that the ST95 isolates (n = 20) contained significantly more virulence factors than the ST1193 isolates (n=9), p-value < 0.001, Mann-Whitney two-tailed unpaired test (Supplementary Table 1, Supplementary Figure 3).’

The complete sequence of 18 strains is not clear. Results of Supplementary Table 2 are presented in the text and are not discussed.

NMEC isolates that were completely sequenced in this study are indicated in bold and marked with an asterisk in Figure 1. This information is indicated in the figure legend and was provided in the original submission. All information regarding genomic island composition and location, virulence genes and plasmid and prophage diversity is included in Supplementary Table 2. This information is highly descriptive and thus we elected not to include it as text in the main manuscript.

46 years is a very long time for such a small number of strains, making it difficult to put forward epidemiological or evolutionary theories. In the analysis of antibiotic resistance, there are no ESBLs. However, Ding's article (reference 34) and other authors showed that ESBLs are emerging in E. coli neonatal infection. These strains are a major threat that should be studied, unfortunately, the authors haven't had the opportunity to characterize such strains in their manuscript.

We agree 46 years is a long time-span. The study by Ding et al examined 56 isolates comprised of 25 different STs isolated in China from 2009-2015, with ST1193 (n=12) and ST95 (n=10) the most common. Our study examined 58 isolates comprised of 22 different STs isolated in seven different geographic regions from 1974-2020, with ST1193 (n=9) and ST95 (n=20) the most common. Thus, despite differences in the geographic regions from which isolates in the two studies were sourced, there are similarities in the most common STs identified. The fact that we observed less antibiotic resistance, including a lack of ESBL genes, in ST1193 is likely due to the different regions from which the isolates were sourced. We acknowledged and discussed the potential of ST1193 harbouring multidrug resistance including ESBLs in our manuscript as follows:

‘Concerningly, the ST1193 strains examined here carry genes encoding several aminoglycoside-modifying enzymes, generating a resistance profile that may lead to the clinical failure of empiric regimens such as ampicillin and gentamicin, a therapeutic combination used in many settings to treat NM and early-onset sepsis [35, 36]. This, in combination with reports of co-resistance to third-generation cephalosporins for some ST1193 strains [22, 34], would limit the choice of antibiotic treatment.’

Second part of the manuscript:

The three patients who relapsed had a late neonatal infection (> 3 days) with respective ages of 6 days, 7 weeks, and 3 weeks. We do not know whether they are former preterm newborns (no term specified) or whether they have received antibiotics in the meantime.

As noted above, patient ages were not disclosed to comply with submission to Medrxiv, a requirement for review of our manuscript at eLife.

Patient 1: Although this patient had a pleiocytose in CSF, the culture was negative which is surprising and no explanation is provided. Therefore, the diagnosis of meningitis is not certain. Pleiocytose without meningitis has been previously described in neonates with severe sepsis. Line 215: no immunological abnormalities were identified (no details are given).

We respectfully disagree with the reviewer. The diagnosis of meningitis is made unequivocally by the presence of a clearly abnormal CSF microscopy (2430 WBCs) and an invasive E. coli from blood culture. This does not seem controversial to the authors. We had believed it unnecessary to include this corroborative evidence, but have added the following to support our assertion:

‘The child was diagnosed with meningitis based on a cerebrospinal fluid (CSF) pleocytosis (>2000 white blood cells; WBCs, low glucose, elevated protein), positive CSF E. coli PCR and a positive blood culture for E. coli (MS21522).’

On the contrary, the authors are surprised by the statement that CSF pleocytosis occurs in neonatal sepsis ‘without meningitis’ and do not know of any definitions of neonatal meningitis that are not tied to the presence of a CSF pleocytosis. Furthermore, the later isolation of E. coli from the CSF during the relapsed infection re-enforces the initial diagnosis.

Patient 2: This patient had a recurrence of bacteremia without meningitis (line 225: CSF microscopy was normal and culture negative!). This case should be deleted.

In a similar vein to the previous comment, we respectfully assert that this patient has clear evidence of meningitis (330 WBCs in the CSF, taken 24h after initiation of antibiotic treatment). In this case, molecular testing was not performed as, under the principle of diagnostic stewardship, it was not considered necessary by the clinical microbiologists and treating clinicians following the culture of E. coli in the bloodstream. We agree that this is not a case of recurrent meningitis, but our intention was to highlight the recrudescence of an invasive infection (urinary sepsis requiring admission to hospital and intravenous antibiotics) which we hypothesise has arisen from the intestinal reservoir. We did not state that all patients suffered from relapsed meningitis.

Despite this, to address this reviewers concern, we have changed all reference to ‘relapsed infection’ to now read ‘recrudescent invasive infection’ in the revised manuscript.

Patient 3: This patient had two relapses which is exceptional and may suggest the existence of a congenital malformation or a neurological complication such as abscess or empyema therefore, "imaging studies" should be detailed.

This patient underwent extensive imaging investigation to rule out a hidden source. This included repeated MRI imaging of head and spine, CT imaging of head and chest, USS imaging of abdomen and pelvis and nuclear medicine imaging to detect a subtle meningeal defect and CSF leak. All tests were normal, and no abscess or empyema found.

We have modified the text to include this information:

Text in original submission: ‘Imaging studies and immunological work-up were normal.’

New text in revised manuscript (underlined): ‘Extensive imaging studies including repeated MRI imaging of the head and spine, CT imaging of the head and chest, ultrasound imaging of abdomen and pelvis, and nuclear medicine imaging did not show a congenital malformation or abscess. Immunological work-up did not show a known primary immunodeficiency. At two years of age, speech delay is reported but no other developmental abnormality.’

The authors suggest a link between intestinal dysbiosis and relapse in three patients. However, the fecal microbiomes of patients without relapse were not analysed, so no comparison is possible.Moreover, dysbiosis after several weeks of antibiotic treatment in a patient hospitalized for a long time is not unexpected. Therefore, it's impossible to make any assumption or draw any conclusion. This part of the manuscript is purely descriptive. Finally, the authors should be more prudent when they state in line 289 "we also provide direct evidence to implicate the gut as a reservoir [...] antibiotic treatment". Indeed the gut colonization of the mothers with the same strain may be also a reservoir (as stated in the discussion line 336). Finally, the authors do not discuss the potential role of ceftriaxone vs cefotaxime in the dysbiosis observed. Ceftriaxone may have a major impact on the microbiota due to its digestive elimination.

We addressed the limitations of our study in the Discussion, including that we did not have access to urine or stool samples from the mother of the infants that suffered recrudescence, and thus cannot rule out mother-to-child transmission as a mechanism of reinfection. We have now added that we did not have clinical data on the weeks of gestation for all patients, and thus could not compare virulence factors from NMEC isolated from preterm versus term infants. The limitations of our study are summarised as follows in the Discussion (new text underlined):

‘This study had several limitations. First, our NMEC strain collection was restricted to seven geographic regions, a reflection of the difficulty in acquiring strains causing this disease. Second, we did not have access to a complete set of stool samples spanning pre- and post-treatment in the patients that suffered NM and recrudescent invasive infection. This impacted our capacity to monitor E. coli persistence and evaluate the effect of antibiotic treatment on changes in the microbiome over time. Third, we did not have access to urine or stool samples from the mother of the infants that suffered recrudescence, and thus cannot rule out mother-to-child transmission as a mechanism of reinfection. Finally, we did not have clinical data on the weeks of gestation for all patients, and thus could not compare virulence factors from NMEC isolated from preterm versus term infants.’

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

It would be useful to mention the sample size (number of genomes analysed, n=58) in the abstract to give readers a sense of the scale of the analysis.

We have added the number of genomes in the abstract as suggested (new text underlined).

‘Here we investigated the genomic relatedness of a collection of 58 NMEC strains spanning 1974-2020 and isolated from seven different geographic regions.’

The term 'strain' is used throughout, it would be clearer to use 'isolates' to describe the biological material and 'genomes' when the unit being referred to is genome sequences. For example, lines 108-111 use 'strain' to mean the collection of 52 isolates but also uses 'strain' to mean the collection of 58 genomes including those of the 52 isolates that the authors sequenced plus a further 6 genomes of isolates that they do not have in their isolate collection.

We have changed the term ‘strain’ to ‘isolate’ or ‘genome’ as suggested.

Figure 1 (annotated phylogeny) is hard to read and interpret, as so much data is presented. It would assist readers if the authors could provide an interactive form of the phylogeny and metadata/genomic feature data discussed in the text, e.g. using microreact.org, so that details can be explored more easily.

This is an excellent suggestion, and we created a project on microreact.org. This information has been added to the Figure 1 legend.

https://microreact.org/project/oNfA4v16h3tQbqREoYtCXj-high-risk-escherichia-coli-clones-that-cause-neonatal-meningitis-and-association-with-recrudescent-infection.

It would be useful to provide information on the frequency and/or distribution of the virulence factors in the broader E. coli population, to provide context for readers and to better understand the importance/significance of the high frequency of the reported virulence factors within NMEC.

As noted above, we agree the prevalence of virulence factors in our collection is interesting. We discussed the prevalence of these virulence factors in our collection, and the detailed data is presented in Table S1. However, we also note a limitation in our study is the number of isolates, and thus we would prefer to avoid evaluation of the prevalence of these virulence factors in the context of a broader E. coli population. There are other studies that have examined NMEC virulence factors in the past; some examples are noted below, and we have now referenced these in our manuscript (note Ref 15 was suggested by Reviewer 3 in a comment below; PMID: 11920295).

Ref 15: Johnson JR, Oswald E, O'Bryan TT, Kuskowski MA, Spanjaard L. Phylogenetic distribution of virulence-associated genes among Escherichia coli isolates associated with neonatal bacterial meningitis in the Netherlands. J Infect Dis 2002; 185(6): 774-84.

Ref 28: Wijetunge DS, Gongati S, DebRoy C, et al. Characterizing the pathotype of neonatal meningitis causing Escherichia coli (NMEC). BMC Microbiol 2015; 15: 211.

Ref 29: Bidet P, Mahjoub-Messai F, Blanco J, et al. Combined Multilocus Sequence Typing and O Serogrouping Distinguishes Escherichia coli Subtypes Associated with Infant Urosepsis and/or Meningitis. J Infect Dis. 2007; 196(2):297-303.

I suggest avoiding the term 'global' to describe the collection, given that there are only seven countries included in the collection and two of the most populous continents (Africa and South America) are not represented at all.

We agree, and now refer to our collection as ‘an NMEC strain collection from geographically diverse locations.’

Reviewer #2 (Recommendations For The Authors):

This is a suggestion regarding discussion/food for thought: This study sheds information on genomic features and indicates the presence of a reservoir in the infected infant. Previous studies have demonstrated the presence of a reservoir in the vaginas of women with recurrent UTIs. Is there any information as to whether the mothers of these infants, especially the three with recrudescent infection, had a UTI or recurrent UTI in their life? It may be worthwhile discussing the potential of testing for E. coli in expecting mothers, if they have a history of UTI.

We do not have such data, and as indicated above we note this as a limitation of our study.

It is unclear as written in the main text, as to whether all three cases of recrudescent infection come from the same geographical location. It would be easier to have this information in the corresponding main text, in addition to the supplement.

The three cases of recrudescent invasive infection were from 3 different locations. We have added the information as following (new text underlined):

‘These patients were from different regions in Australia.’

Reviewer #3 (Recommendations For The Authors):

Line 48 and 67 change the word "devasting".

Changed as suggested.

Line 49 second most in full-term infants.

Changed as suggested.

Line 56 delete the sentence "antibiotic resistance genes occurred infrequently".

We changed the sentence, which now reads (new text underlined):

‘Antibiotic resistance genes occurred infrequently in our collection’.

Line 76 reference 10 is inappropriate.

Reference 10 reported that 5/24 infants treated for neonatal Gram-negative bacillary meningitis over a 10-year period had a relapse of meningitis after the initial course of treatment. Four of the isolates that caused these relapsed infections were E. coli.

To address the reviewers concern, we have altered the text as follows (new text underlined):

‘Moreover, NMEC is an important cause of relapsed infections in neonates [10]’.

Line 83 several references related to serotypes are missing, notably doi.org/10.1086/339343.

We have added this reference.

Line 171 significantly? n=?, p=?

The numbers and P-value were provided in the Supplementary Figure 3 legend. We have now added these to the text as follows:

‘Direct comparison of virulence factors between ST95 and ST1193, the two most dominant NMEC STs, revealed that the ST95 isolates (n = 20) contained significantly more virulence factors than the ST1193 isolates (n = 9); P-value < 0.001, Mann-Whitney two-tailed unpaired test (Supplementary Table 1, Supplementary Figure 3).”

Figure 4 is not necessary.

We respectfully disagree. Figure 4 provides an illustrative comparison of virulence factors between the two most dominant NMEC sequence types, ST95 and ST1193. We believe this will be informative for many readers.

Line 311 "We speculate....of preterm infants" This sentence does not add anything to the discussion.

We respectfully disagree and have kept the sentence. This reflects our opinion.

Line 320 "clear clinical risk factors to explain... ». Term of neonates is missing.

Updated as follows (new text underlined):

‘Although reported rarely, recrudescent invasive E. coli infection in NM patients, including several infants born pre-term, has been documented in single study reports [39, 40]. In these reports, infants received appropriate antibiotic treatment based on antibiogram profiling and no clear clinical risk factors to explain recrudescence were identified, highlighting our limited understanding of NM aetiology.’

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Data Citations

    1. Phan M-D, Hancock SJ, Peters KM, Alvarez-Fraga L, Forde BM, Andersen SB, Miliya T, Harris PNA, Beatson SA, Schlebusch S, Bergh H, Turner P, Brauner A, Westerlund-Wikström B, Irwin AD, Schembri MA, Nhu NTK. 2024. Neonatal meningitis Escherichia coli. NCBI BioProject. PRJNA757133 [DOI] [PMC free article] [PubMed]
    2. Phan M-D, Hancock SJ, Peters KM, Alvarez-Fraga L, Forde BM, Andersen SB, Miliya T, Harris PNA, Beatson SA, Schlebusch S, Bergh H, Turner P, Brauner A, Westerlund-Wikström B, Irwin AD, Schembri MA, Nhu NTK. 2024. Culture-independent long read metagenomic diagnostics for genomic surveillance and infection control of pathogenic bacteria in clinical settings. NCBI BioProject. PRJNA893826

    Supplementary Materials

    Supplementary file 1. Isolates used in this study.
    elife-91853-supp1.xlsx (32.5KB, xlsx)
    Supplementary file 2. Completely sequenced NMEC isolates.
    elife-91853-supp2.xlsx (32.7KB, xlsx)
    Supplementary file 3. Metagenomic sequence analysis.
    elife-91853-supp3.xlsx (63.7KB, xlsx)
    Supplementary file 4. Accession numbers of strains sequenced in the study.
    elife-91853-supp4.xlsx (16.6KB, xlsx)

    Data Availability Statement

    Genome sequence data have been deposited in the Sequence Read Archive under the BioProjects PRJNA757133 and PRJNA893826. Sample accession numbers are listed in Supplementary file 4.

    The following datasets were generated:

    Phan M-D, Hancock SJ, Peters KM, Alvarez-Fraga L, Forde BM, Andersen SB, Miliya T, Harris PNA, Beatson SA, Schlebusch S, Bergh H, Turner P, Brauner A, Westerlund-Wikström B, Irwin AD, Schembri MA, Nhu NTK. 2024. Neonatal meningitis Escherichia coli. NCBI BioProject. PRJNA757133

    Phan M-D, Hancock SJ, Peters KM, Alvarez-Fraga L, Forde BM, Andersen SB, Miliya T, Harris PNA, Beatson SA, Schlebusch S, Bergh H, Turner P, Brauner A, Westerlund-Wikström B, Irwin AD, Schembri MA, Nhu NTK. 2024. Culture-independent long read metagenomic diagnostics for genomic surveillance and infection control of pathogenic bacteria in clinical settings. NCBI BioProject. PRJNA893826

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES