Genomic Profile of Chronic Lymphocytic Leukemia in Korea Identified by Targeted Sequencing

Abstract

Chronic lymphocytic leukemia (CLL) is extremely rare in Asian countries and there has been one report on genetic changes for 5 genes (TP53, SF3B1, NOTCH1, MYD88, and BIRC3) by Sanger sequencing in Chinese CLL. Yet studies of CLL in Asian countries using Next generation sequencing have not been reported. We aimed to characterize the genomic profiles of Korean CLL and to find out ethnic differences in somatic mutations with prognostic implications. We performed targeted sequencing for 87 gene panel using next-generation sequencing along with G-banding and fluorescent in situ hybridization (FISH) for chromosome 12, 13q14.3 deletion, 17p13 deletion, and 11q22 deletion. Overall, 36 out of 48 patients (75%) harbored at least one mutation and mean number of mutation per patient was 1.6 (range 0-6). Aberrant karyotypes were observed in 30.4% by G-banding and 66.7% by FISH. Most recurrent mutation (>10% frequency) was ATM (20.8%) followed by TP53 (14.6%), SF3B1 (10.4%), KLHL6 (8.3%), and BCOR (6.25%). Mutations of MYD88 was associated with moderate adverse prognosis by multiple comparisons (P = 0.055). Mutation frequencies of MYD88, SAMHD1, EGR2, DDX3X, ZMYM3, and MED12 showed similar incidence with Caucasians, while mutation frequencies of ATM, TP53, KLHL6, BCOR and CDKN2A tend to be higher in Koreans than in Caucasians. Especially, ATM mutation showed 1.5 fold higher incidence than Caucasians, while mutation frequencies of SF3B1, NOTCH1, CHD2 and POT1 tend to be lower in Koreans than in Caucasians. However, mutation frequencies between Caucasians and Koreans were not significantly different statistically, probably due to low number of patients. Collectively, mutational profile and adverse prognostic genes in Korean CLL were different from those of Caucasians, suggesting an ethnic difference, while profile of cytogenetic aberrations was similar to those of Caucasians.

Fig 1. Summary of the variant-filtering pipeline.

Fig 2. Diagrams for mutations observed in more than 5% of the cases.

Genes include (A) ATM, (B) TP53, (C) SF3B1, (D) KLHL6, (E) BCOR, (F) LAMB4, and (G) NOTCH1 (Transcript ID: ATM, NM_000051; TP53, NM_001126114; SF3B1, NM_024582; KLHL6, NM_130446; BCOR, NM_001123383; LAMB4, NM_007356; NOTCH1, NM_017617).

Fig 3. Genomic landscape of the Korean patients.

(A) Frequency of gene mutations broken down by CLL stage. (B) Number of mutations according to the Binet stage. (C) Number of mutation sites in each gene that occur in more than 5% of the population.

Fig 4. Kaplan-Meier survival curves of patients who transformed to Richter’s syndrome.

Fig 5. Correlation between various gene mutations and between gene mutation and cytogenetics.

Correlation coefficients with raw P<0.05 were presented as colored boxes. Applying FDR correction for multiple comparison, there were no statistically significant correaltions.

Fig 6. Kaplan-Meier survival curves according to cytogenetic abnormality.

Fig 7. Kaplan-Meier survival curves according to abnormalities detected by FISH.

Fig 8. Kaplan-Meier survival curves according to the number (n< 3, n≥3) of somatic mutations.

Fig 9. Kaplan-Meier survival curves in patients with somatic mutations.

(A) TP53, (B) MYD88, and (C) ATM gene mutation or 11q22 deletion in Korean CLL. After applying multiple comparison by using FDR, only MYD88 showed a tendency for adverse prognosis (P = 0.055).

Fig 10. Disease-free survival curves for patients with mutations.

(A) TP53 and (B) MYD88 genes After applying multiple comparison by using FDR, TP53 and MYD88 showed a tendency for shorter disease-free survival (P = 0.054, both).

Fig 11. Hazard ratios with 95% confidence intervals for overall survival for each gene variant.

Data of gene mutations which are shown only in one patient are not shown. Raw P-values are shown in this figure. After applying FDR for multiple comparison, MYD88 mutation is the only statistically significant gene which had high hazard ratio (P = 0.045).

Fig 12. Comparison of mutation frequencies (%) between Caucasians and Koreans.

Frequencies of mutation in Caucasian were calculated, based on the data of Landau et al. (2015) and Puente et al.(2015).