Suppression PCR-Based Selective Enrichment Sequencing for Pathogen and Antimicrobial Resistance Detection on Cell-Free DNA in Sepsis-A Targeted, Blood Culture-Independent Approach for Rapid Pathogen and Resistance Diagnostics in Septic Patients

Abstract

Sepsis is a life-threatening syndrome triggered by infection and accompanied by high mortality, with antimicrobial resistances (AMRs) further escalating clinical challenges. The rapid and reliable detection of causative pathogens and AMRs are key factors for fast and appropriate treatment, in order to improve outcomes in septic patients. However, current sepsis diagnostics based on blood culture is limited by low sensitivity and specificity while current molecular approaches fail to enter clinical routine. Therefore, we developed a suppression PCR-based selective enrichment sequencing approach (SUPSETS), providing a molecular method combining multiplex suppression PCR with Nanopore sequencing to identify most common sepsis-causative pathogens and AMRs using plasma cell-free DNA. Applying only 1 mL of plasma, we targeted eight pathogens across three kingdoms and ten AMRs in a proof-of-concept study. SUPSETS was successfully tested in an experimental research study on the first ten clinical samples and revealed comparable results to clinical metagenomics while clearly outperforming blood culture. Several clinically relevant AMRs could be additionally detected. Furthermore, SUPSETS provided first pathogen and AMR-specific sequencing reads within minutes of starting sequencing, thereby potentially decreasing time-to-results to 11-13 h and suggesting diagnostic potential in sepsis.

Keywords: Nanopore sequencing; antimicrobial resistances; cell-free DNA; next-generation sequencing; precision diagnostics; real-time diagnostics; sepsis; suppression PCR.

Figure 1

Principle and workflow of SUPSETS: from blood draw and cfDNA isolation (step 1), suppression adapter ligation (step 2), and multiplex suppression PCR (steps 3–5) to sequencing and data analysis for pathogen and AMR detection (steps 6 and 7). Image created by BioRender.

Figure 2

SUPSETS technical validation on artificial spike-in samples. (A) Technical validation of SUPSETS with single spike-in samples: 4000 GE digested and adapter-ligated genomic pathogen DNA was spiked in 6000 GE digested and adapter-ligated genomic human DNA, and the SUPSETS workflow was performed with a primer mix containing two primers per tested targeted pathogen, one genus-specific Candida primer, and one general 16 S-targeting primer. Single primers were exchanged between experiments to improve performance or added with the expansion of the targeted pathogen panel (Supplementary Table S1). X-axis: Spike-in pathogen per column. Y-axis: Regions targeted by pathogen-specific primers. (B) Reads per primer-related amplicon region for artificial human DNA sample with and without spiked-in target DNA: 4000 GE digested and adapter-ligated genomic DNA for each pathogen and AMR was spiked to 6000 GE digested and adapter-ligated genomic human DNA in the human- & AMR- & pathogen DNA sample, and incubated with a primer mix containing two primers per targeted pathogen, one genus-specific Candida primer, one general 16 S-targeting primer, and sixteen primers targeting genes for antimicrobial resistances. For negative control, no pathogen-derived DNA was included in the same experimental setup, only using 6000 GE digested and adapter-ligated human DNA. The X-axis shows the target regions excluding primers C. alb T2, AAC(6)-Ib, TEM116, ndm1, OXA48, blaCTX-M-15 V2, and blaKPC-2 V2 due to performance lack. The Y-axis depicts the percentage of normalized read counts. For the corresponding primers, see Supplementary Table S1. n.a., not available. Normalized counts = max(coverage)_i/(Σ_(i→j)max(coverage)_i + Σhuman reads); column corrected. i for target region.

Figure 3

Appearance of SUPSETS-detected cfDNA amplicons. The results of SUPSETS workflow for clinical samples FL_01 and FL_03 were mapped against the respective pathogen or AMR reference genome and visualized in IGV browser. For each amplicon, maximum coverages with a coverage greater than two were counted and visualized. The primer binding site is depicted in blue. Mismatches are displayed in multi-color.

Figure 4

Validation on selected clinical samples. Septic patient samples were selected according to their pathogen and AMR profile from the MIRSI study [21] and applied to the SUPSETS workflow. (A) Qualitative analysis comparing NGS data (“ground truth”) to the SUPSETS experiment on MinION Flow Cell and blood culture results. (B) Qualitative analysis comparing NGS data (“ground truth”) to the SUPSETS experiment on MinION Flongle and blood culture results. (C) Processed reads were normalized to filtered mapped reads (Q-score > 9) for each sample for four selected patient samples within one MinION Flongle sequencing experiment. Samples FL_01–04 were sequenced together on one MinION Flongle run for 24 h; FL_04-single indicates a single-sample MinION Flongle sequencing experiment for 3 h. The red vertical indicates the threshold of 0.01% normalized reads. Target amplicons are indicated according to the color legends. (A,B) True positive hits compared to NGS data are indicated in dark blue, false positive signals in light grey, light blue reflects true negative and dark grey false negative results, and white is for no data available. * Igκ reads were taken after mapping against the reference genome (hg38).

Figure 5

Real-time detection of first pathogen and AMR reads. For a retrospective analysis, processed reads are shown at the time they were sequenced. Analysis is shown for two different clinical samples: (A) (FL_04-single) and (B) (FC_01). Arrows with time data indicate the first target hits according to the shape and colored legend, respectively.