Same-Day Simultaneous Diagnosis of Bacterial and Fungal Infections in Clinical Practice by Nanopore Targeted Sequencing

SPECIMENS AND CLINICAL ELECTRONIC RECORD

The study population consisted of 1312 specimens from 1257 patients at Renmin Hospital of Wuhan University, China, whose specimens were sent for pathogen testing between November 2018 and November 2019. Parallel testing by routine culture and NTS was performed. Adequate material was submitted for culture, and surplus material was processed by NTS. All specimens (n = 1312) were used to assess the NTS and culture results. The clinical records of each patient were retrieved and 1005 patients were diagnosed or suspected with infection by clinicians according to clinical symptoms and/or positive results of other clinical tests. These patients were used to evaluate the diagnostic yield and sensitivity of NTS. Patient clinical diagnosis and demographic characteristics were obtained from electronic medical records (26). The study was approved by the Ethics Committee of Renmin Hospital of Wuhan University (WDRY2019-K056).

NTS

All primers used in this study are listed in Table S1, the DNA extraction method of specimen was described in supporting method. Amplification of the 16s rRNA gene and rpoB was performed in a 20 μL reaction system with 8 μL extracted DNA, 2 μL barcoded primer (10 μM), and 10 μL 2×KOD OneTM PCR Master Mix (TOYOBO) using the following procedure: 1 cycle at 98°C for 3 min and 35 cycles at 98°C for 10 s, 55°C for 5 s, and 68°C for 10 s, followed by a final elongation step at 68°C for 5 min. ITS1/2 (fungal internal transcribed spacers 1 and 2, abbreviated as ITS1/2) was first amplified using the same reaction system and PCR procedure using the primer mix without barcode and the PCR product was purified with 0.8× AMpure beads (Beckman Coulter) and eluted in 10 μL Tris-EDTA (TE) buffer. Then, 5 μL eluate was used for barcoding PCR with 5 μL barcoded ITS1/2 primer set (10 μM) and 10 μL 2× Phusion U Multiplex PCR Master Mix using the following procedure: 1 cycle at 98°C for 3 min and 10 cycles at 98°C for 10 s, 55°C for 5 s, and 68°C for 5 s, followed by a final elongation step at 68°C for 5 min. The barcoded products of the 16s rRNA gene, ITS1/2 and rpoB amplification from the same samples were pooled according to mass ratio of 10:3:1. The pooled products from the different samples were mixed equally and used to construct sequencing libraries using the 1D Ligation Kit (SQK-LSK109; Oxford Nanopore). The library was sequenced using Oxford Nanopore MinION or GridION. TE buffer was assayed in each batch as a negative control. The bioinformatic analysis was described in the supplement methods section, several modifications were made following previous studies and samples were tested positive for bacteria or fungi if they met any of the established thresholds (Supporting method). Detected microorganisms were described as critical pathogens, opportunistic pathogens, or typically nonpathogenic commensal microbes according to published literature and clinical guidelines (27) (Table S2).

MOCK COMMUNITY

All 50 clinical strains used to evaluate performance were isolated from clinical specimens in the laboratory or purchased from the American Type Culture Collection (ATCC) (Manassas, VA, USA). Six strains were selected to construct a mock community (Table S3) and blood samples were collected from healthy volunteers with no symptoms of infection in physical examination. The concentration of each strain was determined by calibration curves from cultures.

STATISTICAL ANALYSIS

The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated as previously described by comparing the NTS results with composite final clinical diagnosis (28). A P-value < 0.05 was considered significant. Proportional outcomes were compared using the ² test or Fisher’s exact test. Continuous variables were compared using the Student’s t-test or Wilcoxon signed-rank test. Data analyses were performed using SPSS 22.0.

DESIGN OF NTS FOR SAME-DAY SIMULTANEOUS DIAGNOSIS OF BACTERIAL AND FUNGAL INFECTIONS

Full-length bacterial 16s rRNA gene and fungal ITS1/2 gene were selected as universal marker genes for identification of bacteria and fungi, respectively. For specific markers, as Mycobacterium tuberculosis (MTB) was the most concerned clinical critical pathogen, we chose rpoB for Mycobacterium spp. (29) as the target. The design of NTS primers (Table S1) and the sequencing library preparation were described in the supporting result. For sequencing, we chose a nanopore platform that could sequence long nucleic acid fragments and simultaneously analyze the data-output in real-time. This allowed immediate detection of bacterial and fungal infection by periodical analysis of output sequences using our in-house bioinformatic pipeline (supporting methods). In summary, a NTS detection that targeted the amplification of bacterial and fungal marker genes and that is able to execute nanopore sequencing to identify the pathogens was developed (Fig. 1).

• Download figure
• Open in new tab

Fig. 1.

Design of NTS for simultaneous detection of bacterial and fungal infection. Three marker gene databases, corresponding to 16s rRNA, ITS1/2, and rpoB, were constructed from the sequences from the NCBI database. The results of multiplex sequence alignment of each database was used to analyze the conservative region and degeneration of each marker gene base. The blue primers were used in previous studies and selected as “start primers” for the design of other “additional primers” which is labeled by other colors. The red vertical bar on the primer indicates degenerate base. The “start primers” and “additional primers” for each marker gene was mixed at the molar ratio of 3:1. The mixed primers were then used to amplify the marker gene, individually. The PCR products of the 16s rRNA, ITS1/2, and rpoB from the same sample were then pooled at a mass ratio of 10:3:1. This pooled product was used for library preparation, nanopore sequencing and bioinformatic analysis.

PERFORMANCE OF NTS BASED ON STANDARD STRAINS AND MOCK COMMUNITIES

To verify the identification accuracy, NTS was tested by individual analyses of 50 common clinical strains isolated from clinical specimens in the laboratory or purchased from ATCC. In total, 100% and 92% of the tested isolates were identified correctly at the genus and species level, respectively (Table S4). Considering clinical samples, such as sputum, that usually contain high abundance of commensal microbes and that the pathogens may have relatively lower abundance, in order to investigate the sensitivity and sequencing data required for detection the low-abundance microbes, a mock community (three bacteria and three fungi) was constructed to simulate a complex clinical sample and the concentration of each strain was determined by calibration curves from cultures (Table S3). Considering that different levels of background human DNA could potentially affect sensitivity, the performance of NTS was tested using healthy human blood spiked with varied concentrations of the mock community. NTS detected each species at all six pathogen densities, and the least abundant fungi and bacteria, with 25 colony-forming units (CFU) in a community of 1000 CFU, were consistently detected in 12 repeats (Fig. 2A, B). Therefore, the limit of detection was determined to be 25 CFU in 12/12 positive replicates (Table S5).

• Download figure
• Open in new tab

Fig. 2.

Proficiency test of NTS with a mock community. Detection of the least abundant bacteria (A) and fungi (B) in the mock community. Moraxella catarrhalis and Candida tropicalis each accounted for 5% of the mock community. Three different densities from 25 to 150 CFU/mL were tested with 12 repeats. Each spot represents a repeat detection. (C) Comparison of the RPM of each strain in the mock community in a saline background with that in human blood. The correlation between the detection of bacteria (C) and fungi (D) with sequencing data used for the analysis. Dashed line indicates the positive threshold between the sample and control used as the limit of detection for pathogens in this study. Abbreviations: NTS, nanopore targeted sequencing; CFU, colony-forming units; RPM, aligned reads of per million sequence reads.

To determine a proper sequencing time (corresponding to sequencing data) that is required for confident detection of different amounts of microbes in community, a sequencing library contained those 48 samples was sequenced on a chip for 12 h. The most abundant bacteria or fungi, accounting for 30% of the mock community, were detected in all 12 repeats using total 400,000 sequencing reads, which required approximately 30 min. Meanwhile the least abundant bacteria or fungi, accounting for 5% of the mock community, were identified using total 2,000,000 reads, which required approximately 4 h (Fig. 2C, D). This indicated that 4 h sequencing was enough for detection of all pathogens and that a sequencing library with fewer samples could reduce the sequencing time. Considering the diversity of clinical samples, in which pathogens may account for >30% or <5% of the microbial community, the pathogen detection time can be achieved <30-minute or require > 4 h of sequencing time.

COMPARISON OF NTS AND CULTURE METHOD IN CLINICAL PRACTICE

Upon verification of the performance of NTS using the mock community, 1312 clinical specimens from 1257 patients were parallelly analyzed by NTS and culture, including 301 and 1011 specimens from sterile (low commensal microbes background) and non-sterile settings (high commensal microbes background), respectively.

Using the culture-based results as a reference, among 289 culture-positive cases, NTS detected pathogens in 97.58% cases (n = 282/289), including identical findings to culture results in 52.94% cases (n = 153/289), detection of additional potential pathogens other than the culture results in 30.44% of cases (n = 88/289), different species of same genus detected between NTS and culture (6.92%, n = 20/289) and different pathogens detected between two methods (7.27%, n = 21/289) (Fig. 3A). Considering the different growth rates of bacteria and fungi, 84 of the 88 cases of additional pathogens detected by NTS were diagnosed as bacterial or fungal infections only by culture methods, in which additional critical pathogens (n = 2), additional opportunistic fungi or bacteria (n = 67), and potential coinfection with opportunistic fungi and bacteria (n = 19) were detected (Fig. 3A). The NPV of NTS (culture as reference) was 98.88%, which indicated that NTS was able to detect most of the pathogens identified by culture (Table 1). The pathogens detected by culture methods were missed by NTS in 2.42% cases (n = 7/289), in which high background of commensal microbes may impact the detection of pathogen with low abundance.

• Download figure
• Open in new tab

Fig. 3.

Comparison of NTS and culture method in real clinical practice. (A) Evaluation of concordance between NTS and culture results. Detected microbes were defined as critical or opportunistic bacteria or fungi according to previous studies and clinical guidelines. (B) Positive rates of NTS for each kind of specimen compared to that of culture results using the clinical diagnosis of each patient as reference. Abbreviations: NTS, nanopore targeted sequencing; BALF, bronchoalveolar lavage fluid; CSF, cerebrospinal fluid.

Of the 1023 specimens identified as culture-negative, potential pathogens were detected in 39.78% cases (n = 407/1023) by NTS, especially including 63 critical pathogens (Fig. 3A). In order to evaluate the accuracy of the NTS results of these cases, GeneXpert (a Food and Drug Administration-approved assay based on specific PCR of rpoB in MTB) analysis was conducted in 27/34 MTB-positive patients determined by NTS, and 100% concordance was observed; the other 7 patients were confirmed by acid-fast staining (n = 2) or transferred to a specialized hospital according to the clinical symptoms without further confirmation (n = 5) (Table S6).

According to the clinical diagnosis record around the time of the parallel NTS and culture test, 1005/1257 patients were diagnosed or suspected as infection by clinicians according to the clinical symptoms and/or positive results of other clinical tests. However, culture only reported 289 positive cases, which indicated that several pathogens could have been missed and could not to be a reference for comprehensive assess of NTS. Thus, clinical diagnosis was used to evaluate the performance of NTS in infection diagnosis and practicability in clinical application. The sensitivity of NTS (diagnose as reference) were 58.7% (n = 590/1005), which was >36% higher than that for culture-based methods (22.1%, n = 222/1005, P < 0.01), and 52 more potential coinfections with bacteria and fungi were detected by NTS (Table 1). For slow-growing, fastidious pathogens, NTS exhibited a significantly higher positive detection rate compared to that of culture methods, particularly for MTB (34 vs. 4), Aspergillus (80 vs. 20), Pneumocystis jirovecii (20 vs. 0), and Nocardia (9 vs. 3) (Table 1). Moreover, considering the antibiotic pre-exposure in nearly all patients from the Departments of Respiration and Neurology, no colonies, including fast-growing oral flora and colonizers, grew after a 72-h incubation period in 44/401 and 55/62 specimens, among which NTS showed 40.9% (n = 18/44) and 21.0% (n = 12/55) positive results. In those NTS-positive samples, 11/18 (61.1%) and 10/12 (83.3%) were culturable pathogens, such as Streptococcus aureus (Table S7).

Specimen volume, types of pathogens, abundance of colonizing microorganisms, and concentration of human DNA backgrounds in different clinical specimens existed to a large diversity. To assess the impact of different types of specimens on the efficiency of NTS testing, we evaluated the sensitivity of NTS in different specimens. The positive rate of NTS was higher than that of culture methods, particularly in the aqueous humor, which usually has limited specimen volume (60.0% vs. 10.0%, P < 0.05), and bronchoalveolar lavage fluid (BALF), which usually contains high abundance of commensal microbes (56.4% vs. 14.8%, P < 0.05), except for superficial secretions (P = 0.08) in which the main pathogens were fast-growing bacteria, such as Streptococcus aureus (Fig. 3B). These results indicated that detection sensitivity of NTS was less impacted by the type of specimen and that NTS was suitable for the infection detection of various specimens.

For all specimens, considering the complexity of clinical samples and more sequencing data may improve sensitivity, the default total sequencing time was defined as 8 h; the time from sample collection to the generation of the pathogen report was within 24 h for NTS (n = 1312), which was significantly shorter than the estimated median time of 72 h for culture methods (P < 0.001) (Fig. 4). Notably, real-time information was generated during sequencing, and a pre-report can be obtained if the data meet the positive criteria, further reducing the turnaround time. Briefly, in many emergency cases, 10 min sequencing data was sufficient to identify the pathogens and targeted antibiotics can be administrated by clinicians on the same day of detection without waiting for the completion of an 8 h sequencing.

• Download figure
• Open in new tab

Fig. 4.

Schematic illustration of the total duration of NTS in clinical practice, including pretreatment, DNA extraction, targeted amplification and library construction, sequencing, bioinformatic analysis, and generation of the real-time report.

EFFECT OF NTS ON ANTIBIOTIC MANAGEMENT

To assess the impact of pathogen detection by NTS on antibiotic management, we investigated the patients’ electronic records. Empirical antimicrobial agents were initiated before NTS in 93.9% (31/33) of NTS-positive samples in the Department of Respiratory ICU. Based on the NTS results, 51.52% (17/33) of patients were switched to targeted antibiotics (n = 7) or different antibiotics covering the same antibacterial spectrum (n = 10). NTS results did not trigger modification of antibiotic therapy in 42.42% (14/33) of patients; among these patients, 85.71% (12/14) maintained the original anti-infection regimen, which already covered the pathogen detected by NTS, and 14.29% (2/14) were discharged from hospital without subsequent therapy (Table S8). Four patients were evaluated by continuous NTS assay during the treatment process. In four cases, the pathogens detected in the first NTS assay and triggering antibiotic modification were eliminated according to the results of the second NTS assay, resulting in the suspension or de-escalation of targeted antibiotics. In one case, other pathogen were detected in the second NTS, which may have been transient colonizers with the potential to induce a nosocomial infection (Table 2).