Biomarkers, indicators of a disease or outcome, are used to diagnosis patients, determine prognosis, develop a targeted therapy, and monitor treatment response. The sensitivity and specificity of biomarkers play a key role in targeted therapies. One sensitive biomarker for leukemia is called minimal residual disease (MRD). MRD is when a small number of cancer cells remains in the patient’s blood during or after treatment, leading to cancer relapse. Conventional methods of detecting MRD include flow cytometry to monitor the immune response or digital droplet PCR (ddPCR) to look for genetic signals such as single nucleotide variants (SNVs) or gene fusions. de Boer et al. compared gene fusion detection using RNA-based next generation sequencing (NGS) to the conventionally-used ddPCR method.
The team designed an NGS panel targeting gene fusions commonly found in patients with leukemia and leukemic cell lines. Total RNA from 11 patients was depleted of globin RNA and ribosomal RNA before converting to cDNA libraries. TruSeq®-Compatible Full-length Adapters were added which contain unique molecular identifiers (UMIs) to identify unique reads during sequencing analysis. Fusion genes hybridized to biotin-labeled xGen Lockdown Probes which were pulled down using streptavidin beads (Figure 1). xGen Blocking Oligos prevented hybridization to wild type (non-target) transcripts. Targeted (enriched) transcripts were amplified and sequenced on Illumina sequencers. Sequences were analyzed using samtools. ddPCR was performed on the same patient material. Probes and primers were designed targeting the same fusion genes. RNA was converted to cDNA before performing the ddPCR reaction.
Figure 1. Schematic overview of the NGS MRD fusion gene detection assay. RNA containing a low percentage fusion gene and a high percentage wild-type (WT) transcripts is converted to cDNA and prepared for sequencing by adding adapters. Fusion gene transcripts are hybridized to biotin-labeled probes targeting the junction, and wild-type transcripts and sequence adapters are blocked by unlabeled oligonucleotides (oligo). After hybridization, captured fragments are PCR-amplified and sequenced.
American Association for Clinical Chemistry 2020. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
Results and discussion
MRD detection success is measured in terms of sensitivity. What is the limit of detection? How many cancer cells need to be present to be detected? de Boer et al. maximized sensitivity by using UMIs and decreasing the number of pre- and post-capture PCR cycles. This removed duplicate reads and decreased the error rate. The NGS method proposed by de Boer et al. reached a maximum sensitivity of 1:10,000 cells. The threshold for detecting MRD recommended by the European Luekemia Net is 1:1000 cells, so while the NGS method satisfied that threshold, the sensitivity was exceeded by the traditional ddPCR method’s sensitivity of 1:100,000 cells. de Boer et al. believe sensitivity could be increased by targeting more fusion genes, increasing the amount of input RNA, or further reducing the number of PCR cycles.
The primary advantage of using NGS target capture is that patient samples can be multiplexed, or pooled, to increase experiment efficiency and lower time and cost per sample run. Another advantage is that the sensitivity of the experiment is not impacted by the number of probes, so panels can be expanded to target other genetic aberrations. Probe-based sequencing methods like the one proposed here can cover large regions of interest and are less susceptible to false-negative results due to sequence variations. They also avoid the need for sequence confirmation of aberrations.
Dr Eddy de Boer presented this work at the European Society for Human Genetics. View his talk on the detection of low levels of mutant sequences.