Methylome: tumoral ID

da | Giu 15, 2019 | Biologia Molecolare, Cancer

Liquid biopsy consists in an alternative option to traditional biopsy enabling a discovery of tumoral markers and the study of the tumoral progression. The biopsy is a non-invasive technique because it could carry on by tissues and by fluid samples like blood, spit or urine (liquid biopsy). The tumoral progression is monitored trough the detection of cell-free DNA (cfDNA). In this study a joined team of researchers from Toronto and New York, revealed the importance of methylome for the cancer’s identification, in fact they demonstrated that there is a direct correlation between cancer type and methylation pattern [1]. They used liquid biopsy in order to take plasma from blood and detect cfDNA released from dead cancer cells. In this work they used two different techniques: RRBS and cfMeDIP-SEQ. RRBS also called Reduced Representation Bisulfite Sequencing, is a high-throughput technique used for the detection of DNA methylation involving enzyme digestion and bisulfite treatment in order to convert non methylated cytosine into uracil and sequence methylated regions.

In this work they analysed 7 types of cancer. At the beginning they focused only on PDAC (pancreatic adenocarcinoma) for which they took a pool consisting of 24 patients, extracting normal and tumoral tissues analysed by RRBS and cell free DNA from plasma, analysed by cfMeDIP-SEQ. As control they used 24 healthy individuals matched with patients by age and sex.The result of this analysis was the identification of 14716 DMR (differentially methylated regions) in PDAC cases and controls; this data plotted on a heat map showed the different entity of methylation between healthy people and patients.

 

Figure 1.

Researchers used cfMeDIP-Seq as alternative technique among RRBS because bisulfite sequencing is too expensive, gains a low amount of information and leads to a huge degradation of the DNA [2]. The MeDIP-Seq [3] applied to cfDNA is a sensitive, robust and cheap method. Starting from original cell-free-DNA, after end-repairing, A-tailing and adapter ligation, they added filler DNA from Phago-lambda (to improve the methylated region detection’s sensitivity) and Arabidopsis Thaliana DNA as unmethylated control overcoming 100 ng of DNA’s amount. Then, DNA was denaturated by heat and methylated regions were bound to 5-methylcytidine antibodies on magnetic beads, and after immunoprecipitation, were isolated. Recovery of methylated regions was evaluated testing A. Thaliana recovery, authors reached a 99% specificity of analysis, then they carried on some cycles of amplification with qPCR obtaining a library of amplification. Finally, they sequenced the library after the selection of fragments size and the implementation of BioAnalyzer (Fig.1).

Data gained from both methods (RRBS and cfMeDIP-Seq) were concordant so cfMeDIP-Seq can be considered as a good alternative method to RRBS. Thanks to the implementation of statistical methods, they tested the goodness of PBMC and normal tissue as controls. cfMeDIP-Seq was applied on a cohort of 189 samples from 7 different types of cancer and one healthy group (breast cancer, colorectal cancer, lung cancer, renal cancer, leukemia and bladder cancer). Statistical programs representing overlapping between methylation pattern of case and controls, allowed to distinguish univocal methylation pattern peculiar for each cancer type. The classification accuracy for every class has been evaluated calculating AUROC (from ROC curves [4]) underlying a high valuation accuracy per PDAC, controls and AML. High values of AUROC were found both in early and late stage of PDAC e LUC, underlying the high probability of detection in both stages and distinguish them, in fact AUROC values were all between 0,91 and 0,96. The informativity for each cancer type of plasma derived DMR in the cohort of 189 plasma samples was evaluated by t-SNE (t-distributed stochastic neighbour embedding) plot constructing bi-dimensional and three-dimensional plots. They constructed three three-dimensional plots, two for data coming from two databases: TCGA (The Cancer Genome Atlas) and COSMIC (Catalogue of Somatic Mutations In Cancer), and one from plasma cfDNA data obtained by the authors. Representations showed optimal clustering of database data, and a good clustering for the authors data.

Authors developed a robust, sensitive and bisulfite-free methodology for tumoral types identification based on methylation patterns in cfDNA. This work is a good starting point for the creation of a new, non-invasive and reliable diagnostic technique, but need other improvements and validations using independent data sets for every cancer type.

References

  1. Aravanis, A. M., Lee, M. & Klausner, R. D. Next-generation sequencing of circulating tumor DNA for early cancer detection. Cell 168, 571–574 (2017).
  2. Grunau, C., Clark, S. J. & Rosenthal, A. Bisulfte genomic sequencing: systematic investigation of critical experimental parameters. Nucleic AcidsRes. 29, E65 (2001).
  1. Newman, A. M. et al. An ultrasensitive method for quantitating circulating tumor DNA with broad patient coverage. Nat. Med. 20, 548–554 (2014).
  2. Michiels, S., Koscielny, S. & Hill, C. Prediction of cancer outcome with microarrays: a multiple random validation strategy. Lancet 365, 488–492 (2005).

Matteo Uda

Master Industrial Biotechnology student

Riccardo Pedraza

Master Industrial Biotechnology student