Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Hepatocellular carcinoma (HCC) ranks as the sixth most common neoplasm and the third leading cause of cancer-related death.1 As with many other cancers, HCC detected at an early stage has a better prognosis than the advanced-stage disease, in part due to the relative efficacy of local treatments compared with systemic therapies.1 At present, the serological determination of alpha-fetoprotein (AFP) levels is the only liquid biopsy test used for the detection and surveillance of HCC. However, it has demonstrated serious limitations and most patients are diagnosed in an advanced stage, when current treatments have little effectiveness. Moreover, although patients in early and intermediate stages might benefit from potentially curative treatments, there is a high rate of recurrence.2 Therefore, since the only chance to offer effective treatment is the early detection, there is an urgent need to develop non-invasive and precise methods for early HCC diagnosis and recurrence monitoring.
In this regard, the focus of several recent studies has concentrated on the identification of specific genetic and epigenetic tumour alterations in plasma cell-free DNA (cfDNA). Among the circulating cfDNA, DNA released by the tumour to the peripheral blood has been shown not only to contain the same mutations as the primary tumour cells but also the same epigenetic patterns.3 The most extensively studied epigenetic feature for blood-based cancer biomarker development is DNA methylation, especially the 5-methylcytosine (5-mC) modification at the 5'-cytosine-phosphate-guanine-3' (CpG) dinucleotides. Importantly, DNA methylation patterns are usually tissue- and cancer-type specific. However, in spite of the growing number of reports that show the great capacity of plasma 5-mC biomarkers to detect early-stage cancer patients and to discriminate between cancer types,4 to date, only Septin-9 promoter methylation has received approval from the Food and Drug Administration for its use as a blood-based methylated biomarker for the diagnosis of colon cancer.5
During the DNA demethylation process, the ten–eleven translocation proteins catalyse the oxidation of the 5-mC moiety. Conversion of 5-mC to 5-hydroxymethylcytosine (5-hmC) is the first step in this DNA demethylation pathway. However, 5-hmC is much more than merely an intermediate in 5-mC oxidation, it is also a relatively stable epigenetic mark that plays a relevant role in gene expression regulation.6 7 Importantly, a broad loss of 5-hmC takes place across many types of cancers, bringing forward 5-hmC as a novel epigenetic cancer biomarker.8
Several methods have been developed to map and study 5-hmC in a genome-wide fashion. Bisulfite-based approaches followed by whole-genome sequencing provide the most comprehensive and quantitative information because they are able to detect 5-hmC at single-base resolution.9 However, the oxidation and bisulfite treatment steps required for this type of analysis lead to substantial DNA degradation, and thus the relatively high amounts of starting DNA needed exceed those obtained from blood samples.9 With the goal to overcome this obstacle, in 2010, Song et al developed a bisulfite-free genome-wide 5-hmC sequencing method (5-hmC-Seal) based on selective chemical labelling.7 This 5-hmC-Seal uses β-glucosyltransferase to selectively transfer a chemically modified glucose (6-N3-glucose) onto the hydroxyl group of 5-hmC. The azide group is then chemically modified with biotin for pulling-down the 5-hmC-containing DNA fragments for subsequent sequencing.7 Although this method cannot provide single-base 5-hmC information, it is nowadays the state-of-the-art technology for the study of genome-wide 5-hmC dynamics in cfDNA, because it can be used with limiting amounts of DNA.10 This 5-hmC-Seal approach has allowed to identify specific blood-based 5-hmC signatures in different types of cancer patients, including colorectal, gastric, oesophageal or lung cancer.9 To be noted, the majority of these 5-hmC marker panels achieved superior sensitivity and specificity than the classical non-invasive cancer biomarkers.9
In Gut, Cai et al 11 aim to develop an integrated diagnostic model using the 5-hmC profiles in cfDNA to distinguish early HCC from non-HCC patients. For this purpose, the authors applied the genome-wide 5-hmC-Seal technique to cfDNAs collected from a large cohort of subjects, including 1204 HCC patients, 392 patients with chronic hepatitis B (CHB) virus infection or liver cirrhosis (LC) and 958 healthy individuals and patients with benign liver lesions. The cohort was first divided into training and validation sets, and outcomes from the training set were analysed resulting in 917 candidate genes with differential 5-hmC content in cfDNA between the subgroup of early HCC patients and the groups of CHB/LC patients and control individuals. Subsequently, the authors employed a deep learning algorithm that uncovered a final 5-hmC marker panel, composed by 32 genes, able to segregate early HCC patients from non-HCC, including an accurate segregation from those individuals with CHB or LC. Outcomes were then validated using the internal validation set plus an external cohort. Remarkably, this 5-hmC-based panel not only outperformed the AFP predictions, but was also able to accurately diagnose those early HCC patients that would have been misclassified based on AFP data alone.11
As mentioned above, parallel to early HCC detection, predicting the relapse of a cancer patient after a first complete response treatment remains a formidable challenge in modern medicine. It is tempting to speculate that the 5-hmC diagnostic panel identified by Cai et al 11 might be instrumental for this purpose.
Besides the application of these 5-hmC alterations as cancer biomarkers for blood detection, the understanding of their functional relevance is helping to unravel the multistep and complex tumorigenic process and to identify new druggable targets and therapeutic strategies. In this regard, and in agreement with the literature,6 7 Cai et al 11 show that the 5-hmC profiles were enriched within the gene bodies, specifically at enhancer regions, and associated with gene expression activation. These results place 5-hmC as another piece into the highly complex and coordinated puzzle of epigenetic gene expression regulation and its cancer-associated reprogramming. However, great efforts are still needed to comprehend the dynamic interplay between DNA methylation status and gene transcription, and to distinguish between ‘driver’ and mechanistically irrelevant (‘passenger’) alterations.
Altogether, the work of Cai et al 11 further supports the assessment of epigenetic modifications in cfDNA as a powerful liquid biopsy biomarker for early cancer diagnosis. In particular, this work provides new evidence and highlights the use of the 5-hmC modification for early HCC detection (figure 1). However, technology-wise, the translation into the routine clinical practise of analytical procedures like the one described here is challenging. Methodologies still need to be simplified and standardised for accuracy and reproducibility, and the publication of confirmatory independent studies would be important. Nevertheless, in view of the dismal prognosis of patients with advanced-stage HCC, the development of such diagnostic techniques is essential and should be prioritised.
Contributors Full authorship.
Funding MA is supported by the 'aecc Scientific Foundation' (post-doctoral fellowship); Gobierno de Navarra (2018-055); Instituto de Salud Carlos III (PI19/00613); HEPACARE Project from la Caixa.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Commissioned; internally peer reviewed.