Assessing the complementary information from an increased number of biologically relevant features in liquid biopsy-derived RNA-Seq data

Liquid biopsy-derived RNA sequencing (lbRNA-seq) exhibits significant promise for clinic-oriented cancer diagnostics due to its non-invasiveness and ease of repeatability. Despite substantial advancements, obstacles like technical artefacts and process standardisation impede seamless clinical integration. Alongside addressing technical aspects such as normalising fluctuating low-input material and establishing a standardised clinical workflow, the lack of result validation using independent datasets remains a critical factor contributing to the often low reproducibility of liquid biopsy-detected biomarkers. Considering the outlined drawbacks, our objective was to establish a workflow/methodology characterised by: 1. Harness the rich diversity of biological features accessible through lbRNA-seq data, encompassing a holistic range of molecular and functional attributes. These components are seamlessly integrated via a Machine Learning-based Ensemble Classification framework, enabling a unified and comprehensive analysis of the intricate information encoded within the data. 2. Implementing and rigorously benchmarking intra-sample normalisation methods to heighten their relevance within clinical settings. 3. Thoroughly assessing its efficacy across independent test sets to ascertain its robustness and potential utility. Using ten datasets from several studies comprising three different sources of biological material, we first show that while the best-performing normalisation methods depend strongly on the dataset and coupled Machine Learning method, the rather simple Counts Per Million method is generally very robust, showing comparable performance to cross-sample methods. Subsequently, we demonstrate that the innovative biofeature types introduced in this study, such as the Fraction of Canonical Transcript, harbour complementary information. Consequently, their inclusion consistently enhances prediction power compared to models relying solely on gene expression-based biofeatures. Finally, we demonstrate that the workflow is robust on completely independent datasets, generally from different labs and/or different protocols. Taken together, the workflow presented here outperforms generally employed methods in prediction accuracy and may hold potential for clinical diagnostics application due to its specific design.

Brief Report: circRUNX1 as Potential Biomarker for Cancer Recurrence in EGFR Mutation-Positive Surgically Resected NSCLC

Introduction

As recently evidenced by the ADAURA trial, most patients with stages IB to IIIA of resected EGFR-mutant lung adenocarcinoma benefit from osimertinib as adjuvant therapy. Nevertheless, predictive markers of response and recurrence are still an unmet need for more than 10% of these patients. Some circular RNAs (circRNAs) have been reported to play a role in tumor growth and proliferation. In this project, we studied circRNA expression levels in formalin-fixed, paraffin-embedded lung tumor samples to explore their biomarker potential and develop a machine learning (ML)-based signature that could predict the benefit of adjuvant EGFR tyrosine kinase inhibitors in patients with EGFR-mutant NSCLC.

Methods

Patients with surgically resected EGFR mutant-positive, stages I to IIIB NSCLC were recruited from February 2007 to December 2015. Formalin-fixed, paraffin-embedded tumor samples were retrospectively collected from those patients with a follow-up period of more than or equal to 36 months (N = 76). Clinicopathologic features were annotated. Total RNA was purified and quantified prior nCounter processing with our circRNA custom panel. Data analysis and ML were performed taking into consideration circRNA expression levels and recurrence-free survival (RFS). RFS was defined from the day of surgery to the day when recurrence was detected radiologically or the death owing to any cause.

Results

Of the 76 patients with EGFR mutation-positive NSCLC included in the study, 34 relapsed within 3 years after resection. The median age of the relapsing cohort was 71.5 (range: 49-89) years. Most patients were nonsmokers (n = 21; 61.8%) and of female sex (n = 21; 61.8%). Most cases (n = 17; 50%) presented an exon 21 mutation, whereas 15 and four patients had an exon 19 and exon 18 mutation, respectively. Differential expression analysis revealed that circRUNX1, along with circFUT8 and circAASDH, was up-regulated in relapsing patients (p < 0.05 and >2 fold-change). A ML-based circRNA signature predictive of recurrence in patients with EGFR mutation-positive NSCLC, comprising circRUNX1, was developed. Our final model including selected 6-circRNA signature with random forest algorithm was able to classify relapsing patients with an accuracy of 83% and an area under the receiver operating characteristic curve of 0.91.RFS was significantly shorter not only for the subgroup of patients with high versus low circRUNX1 expression but also for the group classified as recurrent by the ML circRNA signature when compared with those classified as nonrecurrent.

Conclusions

Our findings suggest that circRUNX1 and the presented ML-developed signature could be novel tools to predict the benefit of adjuvant EGFR tyrosine kinase inhibitors with regard to RFS in patients with EGFR-mutant NSCLC. The training and validation phases of our ML signature will be conducted including bigger independent cohorts.

ALL-tRNAseq enables robust tRNA profiling in tissue samples

Transfer RNAs (tRNAs) are small adaptor RNAs essential for mRNA translation. Alterations in the cellular tRNA population can directly affect mRNA decoding rates and translational efficiency during cancer development and progression. To evaluate changes in the composition of the tRNA pool, multiple sequencing approaches have been developed to overcome reverse transcription blocks caused by the stable structures of these molecules and their numerous base modifications. However, it remains unclear whether current sequencing protocols faithfully capture tRNAs existing in cells or tissues. This is specifically challenging for clinical tissue samples that often present variable RNA qualities. For this reason, we developed ALL-tRNAseq, which combines the highly processive MarathonRT and RNA demethylation for the robust assessment of tRNA expression, together with a randomized adapter ligation strategy prior to reverse transcription to assess tRNA fragmentation levels in both cell lines and tissues. Incorporation of tRNA fragments not only informed on sample integrity but also significantly improved tRNA profiling of tissue samples. Our data showed that our profiling strategy effectively improves classification of oncogenic signatures in glioblastoma and diffuse large B-cell lymphoma tissues, particularly for samples presenting higher levels of RNA fragmentation, further highlighting the utility of ALL-tRNAseq for translational research.

Multiplex Analysis of CircRNAs from Plasma Extracellular Vesicle-Enriched Samples for the Detection of Early-Stage Non-Small Cell Lung Cancer

BACKGROUND

The analysis of liquid biopsies brings new opportunities in the precision oncology field. Under this context, extracellular vesicle circular RNAs (EV-circRNAs) have gained interest as biomarkers for lung cancer (LC) detection. However, standardized and robust protocols need to be developed to boost their potential in the clinical setting. Although nCounter has been used for the analysis of other liquid biopsy substrates and biomarkers, it has never been employed for EV-circRNA analysis of LC patients.

METHODS

EVs were isolated from early-stage LC patients (n = 36) and controls (n = 30). Different volumes of plasma, together with different number of pre-amplification cycles, were tested to reach the best nCounter outcome. Differential expression analysis of circRNAs was performed, along with the testing of different machine learning (ML) methods for the development of a prognostic signature for LC.

RESULTS

A combination of 500 μL of plasma input with 10 cycles of pre-amplification was selected for the rest of the study. Eight circRNAs were found upregulated in LC. Further ML analysis selected a 10-circRNA signature able to discriminate LC from controls with AUC ROC of 0.86.

CONCLUSIONS

This study validates the use of the nCounter platform for multiplexed EV-circRNA expression studies in LC patient samples, allowing the development of prognostic signatures.

Selective isolation of extracellular vesicles from minimally processed human plasma as a translational strategy for liquid biopsies

Background

Intercellular communication is mediated by extracellular vesicles (EVs), as they enclose selectively packaged biomolecules that can be horizontally transferred from donor to recipient cells. Because all cells constantly generate and recycle EVs, they provide accurate timed snapshots of individual pathophysiological status. Since blood plasma circulates through the whole body, it is often the biofluid of choice for biomarker detection in EVs. Blood collection is easy and minimally invasive, yet reproducible procedures to obtain pure EV samples from circulating biofluids are still lacking. Here, we addressed central aspects of EV immunoaffinity isolation from simple and complex matrices, such as plasma.

Methods

Cell-generated EV spike-in models were isolated and purified by size-exclusion chromatography, stained with cellular dyes and characterized by nano flow cytometry. Fluorescently-labelled spike-in EVs emerged as reliable, high-throughput and easily measurable readouts, which were employed to optimize our EV immunoprecipitation strategy and evaluate its performance. Plasma-derived EVs were captured and detected using this straightforward protocol, sequentially combining isolation and staining of specific surface markers, such as CD9 or CD41. Multiplexed digital transcript detection data was generated using the Nanostring nCounter platform and evaluated through a dedicated bioinformatics pipeline.

Results

Beads with covalently-conjugated antibodies on their surface outperformed streptavidin-conjugated beads, coated with biotinylated antibodies, in EV immunoprecipitation. Fluorescent EV spike recovery evidenced that target EV subpopulations can be efficiently retrieved from plasma, and that their enrichment is dependent not only on complex matrix composition, but also on the EV surface phenotype. Finally, mRNA profiling experiments proved that distinct EV subpopulations can be captured by directly targeting different surface markers. Furthermore, EVs isolated with anti-CD61 beads enclosed mRNA expression patterns that might be associated to early-stage lung cancer, in contrast with EVs captured through CD9, CD63 or CD81. The differential clinical value carried within each distinct EV subset highlights the advantages of selective isolation.

Conclusions

This EV isolation protocol facilitated the extraction of clinically useful information from plasma. Compatible with common downstream analytics, it is a readily implementable research tool, tailored to provide a truly translational solution in routine clinical workflows, fostering the inclusion of EVs in novel liquid biopsy settings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40364-022-00404-1.

sRNAbench and sRNAtoolbox 2022 update: accurate miRNA and sncRNA profiling for model and non-model organisms

The NCBI Sequence Read Archive currently hosts microRNA sequencing data for over 800 different species, evidencing the existence of a broad taxonomic distribution in the field of small RNA research. Simultaneously, the number of samples per miRNA-seq study continues to increase resulting in a vast amount of data that requires accurate, fast and user-friendly analysis methods. Since the previous release of sRNAtoolbox in 2019, 55 000 sRNAbench jobs have been submitted which has motivated many improvements in its usability and the scope of the underlying annotation database. With this update, users can upload an unlimited number of samples or import them from Google Drive, Dropbox or URLs. Micro- and small RNA profiling can now be carried out using high-confidence Metazoan and plant specific databases, MirGeneDB and PmiREN respectively, together with genome assemblies and libraries from 441 Ensembl species. The new results page includes straightforward sample annotation to allow downstream differential expression analysis with sRNAde. Unassigned reads can also be explored by means of a new tool that performs mapping to microbial references, which can reveal contamination events or biologically meaningful findings as we describe in the example. sRNAtoolbox is available at: https://arn.ugr.es/srnatoolbox/.

Digital multiplexed analysis of circular RNAs in FFPE and fresh non‐small cell lung cancer specimens

Although many studies highlight the implication of circular RNAs (circRNAs) in carcinogenesis and tumor progression, their potential as cancer biomarkers has not yet been fully explored in the clinic due to the limitations of current quantification methods. Here, we report the use of the nCounter platform as a valid technology for the analysis of circRNA expression patterns in non‐small cell lung cancer (NSCLC) specimens. Under this context, our custom‐made circRNA panel was able to detect circRNA expression both in NSCLC cells and formalin‐fixed paraffin‐embedded (FFPE) tissues. CircFUT8 was overexpressed in NSCLC, contrasting with circEPB41L2, circBNC2, and circSOX13 downregulation even at the early stages of the disease. Machine learning (ML) approaches from different paradigms allowed discrimination of NSCLC from nontumor controls (NTCs) with an 8‐circRNA signature. An additional 4‐circRNA signature was able to classify early‐stage NSCLC samples from NTC, reaching a maximum area under the ROC curve (AUC) of 0.981. Our results not only present two circRNA signatures with diagnosis potential but also introduce nCounter processing following ML as a feasible protocol for the study and development of circRNA signatures for NSCLC.

The Analysis of Platelet-Derived circRNA Repertoire as Potential Diagnostic Biomarker for Non-Small Cell Lung Cancer

Tumor-educated Platelets (TEPs) have emerged as rich biosources of cancer-related RNA profiles in liquid biopsies applicable for cancer detection. Although human blood platelets have been found to be enriched in circular RNA (circRNA), no studies have investigated the potential of circRNA as platelet-derived biomarkers for cancer. In this proof-of-concept study, we examine whether the circRNA signature of blood platelets can be used as a liquid biopsy biomarker for the detection of non-small cell lung cancer (NSCLC). We analyzed the total RNA, extracted from the platelet samples collected from NSCLC patients and asymptomatic individuals, using RNA sequencing (RNA-Seq). Identification and quantification of known and novel circRNAs were performed using the accurate CircRNA finder suite (ACFS), followed by the differential transcript expression analysis using a modified version of our thromboSeq software. Out of 4732 detected circRNAs, we identified 411 circRNAs that are significantly (p-value < 0.05) differentially expressed between asymptomatic individuals and NSCLC patients. Using the false discovery rate (FDR) of 0.05 as cutoff, we selected the nuclear receptor-interacting protein 1 (NRIP1) circRNA (circNRIP1) as a potential biomarker candidate for further validation by reverse transcription-quantitative PCR (RT-qPCR). This analysis was performed on an independent cohort of platelet samples. The RT-qPCR results confirmed the RNA-Seq data analysis, with significant downregulation of circNRIP1 in platelets derived from NSCLC patients. Our findings suggest that circRNAs found in blood platelets may hold diagnostic biomarkers potential for the detection of NSCLC using liquid biopsies.

Abstract 466: Differential expression of circRNAs allows discrimination of NSCLC from cancer-free lung specimens using the nCounter platform

Introduction: Lung cancer mortality ranks as the highest of the cancer-related deaths. Studies on tumor profiling at the genomic, transcriptomic and proteomic level has soared in the last decade improving overall survival of these patients by shaping the current targeted therapies; however, further investigation of novel biomarkers for an early diagnosis still remains imperative.circRNAs are a class of tissue-specific stable structures that control mammalian transcription. Their aberrant expression plays an important role in carcinogenesis and tumor progression which profiles them as valuable biomarkers; Conversely, its potential has not been fully explored in lung cancer due to several limitations of current circRNA quantification methods that prevent their clinical implementation.The nCounter technology allows for quantitative and qualitative assessment of up to 800 targets providing an accurate and factual perspective of expression levels. To our knowledge, thru this study we stand as the first on assessing circRNA differential expression with this platform for lung cancer detection both in FFPE specimens and cell lines providing preliminary evidence of their differential expression in lung cancer.

Methods: Cells were cultured under standard conditions until harvested. RNA was isolated by using Allprep DNA/RNA/miRNA universal kit (Qiagen). FFPE lung tissue samples (n=28; 18 NSCLC,10 non-cancer) were retrospectively collected and micro-dissected. RNA was isolated with High Pure FFPET RNA isolation kit (Roche) and quantified by Qubit (Thermo Fisher).Overnight hybridization and posterior nCounter FLEX processing were performed following NanoString protocol for nCounter Elements. Expression analysis was carried out based on a tailored panel of 85 circRNAs related to the biology of the disease.

Results: FFPE lung tissues revealed a cluster of differentially expressed circRNAs that allow distinction of lung cancer versus control. circFOXP1, circEPB41L2 and circBNC2 ranked as the most downregulated circRNAs, whereas circCHD9, circAASDH, circRUNX1 and circCHST15 led the catalog of most upregulated in cancer specimens. circRNA expression of A549, H2228, H3122, PC9, H1666, and HOP-62 cells was compared to the CCL-171 fibroblast cell line. circEPB41L2 and circFOXP1 were also confirmed distinctly downregulated in cancer cell lines.

Conclusion: This study presents for the first time the use of the differential expression of circRNAs in FFPE tissues for lung cancer discrimination using the nCounter platform. While more samples are currently being collected to increase the statistical power of the study, these results pave the way for the developing of future circRNA-based nCounter tests for lung cancer diagnosis. Further experiments using epithelial cells as control will be carried out and results will be pertinently updated at the time of the meeting.

Citation Format: Carlos Pedraz-Valdunciel, Miguel Ángel Molina-Vila, Stavros Giannoukakos, Nicolas Potie, Ruth Roman-Llado, Jill Bracht, Martyna Filipska, Masaoki Ito, Ana Gimenez-Capitán, Cristina Aguado-Esteban, Sarah Warren, Chung-Ying Huang, Trever Bivona, Rafael Rosell. Differential expression of circRNAs allows discrimination of NSCLC from cancer-free lung specimens using the nCounter platform [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 466.

mirnaQC: a webserver for comparative quality control of miRNA-seq data

Although miRNA-seq is extensively used in many different fields, its quality control is frequently restricted to a PhredScore-based filter. Other important quality related aspects like microRNA yield, the fraction of putative degradation products (such as rRNA fragments) or the percentage of adapter-dimers are hard to assess using absolute thresholds. Here we present mirnaQC, a webserver that relies on 34 quality parameters to assist in miRNA-seq quality control. To improve their interpretability, quality attributes are ranked using a reference distribution obtained from over 36 000 publicly available miRNA-seq datasets. Accepted input formats include FASTQ and SRA accessions. The results page contains several sections that deal with putative technical artefacts related to library preparation, sequencing, contamination or yield. Different visualisations, including PCA and heatmaps, are available to help users identify underlying issues. Finally, we show the usefulness of this approach by analysing two publicly available datasets and discussing the different quality issues that can be detected using mirnaQC.

sRNAbench and sRNAtoolbox 2019: intuitive fast small RNA profiling and differential expression

Since the original publication of sRNAtoolbox in 2015, small RNA research experienced notable advances in different directions. New protocols for small RNA sequencing have become available to address important issues such as adapter ligation bias, PCR amplification artefacts or to include internal controls such as spike-in sequences. New microRNA reference databases were developed with different foci, either prioritizing accuracy (low number of false positives) or completeness (low number of false negatives). Additionally, other small RNA molecules as well as microRNA sequence and length variants (isomiRs) have continued to gain importance. Finally, the number of microRNA sequencing studies deposited in GEO nearly triplicated from 2014 (280) to 2018 (764). These developments imply that fast and easy-to-use tools for expression profiling and subsequent downstream analysis of miRNA-seq data are essential to many researchers. Key features in this sRNAtoolbox release include addition of all major RNA library preparation protocols to sRNAbench and improvements in sRNAde, a tool that summarizes several aspects of small RNA sequencing studies including the detection of consensus differential expression. A special emphasis was put on the user-friendliness of the tools, for instance sRNAbench now supports parallel launching of several jobs to improve reproducibility and user time efficiency.