MUTATIONS IN THE TP53 GENE SHOW FEATURES OF SOMATIC HYPERMUTATION PROCESS WITH PROMINENT DIFFERENCE BETWEEN IGHV MUTATED AND UNMUTATED CHRONIC LYMPHOCYTIC LEUKEMIA

Konference: 2015 20th Congress of the European Hematology Association - účast ČR

Kategorie: Maligní lymfomy a leukémie

Téma: Poster

Číslo abstraktu: P587

Autoři: Mgr. Nikola Tom; RNDr. Jitka Malčíková, Ph.D.; Mgr. Lenka Radová, Ph.D.; Mgr. Barbara Kantorová; Mgr. Filip Pardy; Mgr. Šárka Pavlová, PhD; Mgr. Karol Pál; MUDr. Mgr. Marek Mráz, Ph.D.; MvDr. Boris Tichý, Ph.D.; prof. MUDr. Michael Doubek, Ph.D.; MUDr. Yvona Brychtová, Ph.D.; Mgr. Karla Plevová, Ph.D.; prof. MUDr. Jiří Mayer, CSc.; Doc. MUDr. Martin Trbušek, PhD; prof. RNDr. Šárka Pospíšilová, Ph.D.

Background

Introducing somatic hypermutations (SHM) into genes coding for heavy and light chains of immunoglobulins (IGHV, IGLV) is a physiological process indispensable during antibody maturation. SHM introduction is a two-step process, which in the first step involves deamination of cytosine to uracil by an enzyme Activation-Induced (Cytidine) Deaminase (AID). During the second step, uracil and potentially also flanking nucleotides are removed and error-prone DNA polymerases (mainly polymerase eta) are recruited to fill the gap. Off-targeting of AID may result in mutations in non-immunoglobulin genes, including tumor suppressor gene TP53



Aims
We explored the TP53 mutation patterns with regards to SHM features emphasizing the differences between mutations occurring in IGHV unmutated (U-CLL) and IGHV mutated (M-CLL) groups of patients.

Methods
In order to reduce the selection bias, we used the set of TP53 mutations detected using ultra-deep Next Generation Sequencing (NGS) with high sensitivity (0.2%) allowing to consider the minor subclonal mutations with less prominent selective advantage. For NGS analysis, Illumina Miseq platform has been used with average coverage 34788; (range 1674–177021). For variant detection, an in house bioinformatics pipeline was established. Statistical evaluation was performed by shearwater algorithm computing Bayes classifiers based on a betabinomial model. Only point substitutions were taken into account. Altogether, 464 TP53 mutations in 73 high risk CLL patients were analyzed. 121 mutations were found in M-CLL cases and 343 mutations were detected in U-CLL patients. The high number of mutations should be attributed to the high number of minor subclonal mutations detected mainly in patients in relapse after treatment. In 60% of patients, results from repeated examinations were included (2-4 examinations per patient; each mutation detected in more than one consecutive sample was considered only once). 72.6% of patients received treatment before the first or during consecutive analyses. 

Results
U-CLL showed a higher proportion of mutations in C:G pairs comparing to M-CLL (66% vs. 51% of all mutations; P=0.003). Out of these, G:C>A:T substitutions, the primary result of AID cytidine deamination, were the prevalent events observed in U-CLL but not M-CLL (59.6% vs. 40.3% of all C:G mutations; P=0.009). Focusing on the AID-targeted sequence motif RGYW/WRCY, no difference between the two subgroups was observed (2.5% vs. 3.1%; P=ns). However, we found a significant overrepresentation of mutations in GNW motif in U-CLL (14.58% of all mutations vs. 6.61% in M-CLL; P=0.025). GNW is a strand-biased motif derived from RGYW most frequently targeted by AID in IGLV genes. On the other hand, TP53mutations detected in M-CLL cases showed the features of targeting by polymerase eta: (i) frequent targeting of A:T pairs (49% vs. 34% in U-CLL; P=0.003); with prominent strand bias favoring A over T (13.8 fold vs. 3.1 fold in U-CLL; P=0.002; (ii) prevalence of mutations in WA/TW motifs (40.5% vs. 23.91% in U-CLL; P<0.0001) with strand bias favoring WA observed in both groups, but extremely prominent in M-CLL (23.5 fold in M-CLL vs. 4.37 fold in U-CLL; P= 0.014).

Summary

We documented the significantly different patterns of TP53 mutations in U-CLL vs. M-CLL. Mutations detected in U-CLL showed the features corresponding to the first step of SHM process: deamination by AID with strand bias possibly attributable to AID activity on the transcribed strand. In contrast, spectra of mutations detected in M-CLL cases suggested the more prominent involvement of polymerase eta during the second SHM step.

Supported by CZ.1.07/2.3.00/30.0009, CZ.1.05/1.1.00/02.0068, NT13519, NT13493, NT11218, MUNI/A/1180/2014, NGS-PTL/2012-2015/no.306242, MSMT-CR (2013-2015, no.7E13008)



Keyword(s): Chronic lymphocytic leukemia, Somatic hypermutation, Tumor suppressor

www.eha.org

Datum přednesení příspěvku: 13. 6. 2015