Error-Prone DNA Polymerases iota and beta Are Over-Expressed in B-CLL Cells: Correlation with Specific IgVH Point-Mutations and Implication for the Pathogenesis of Intraclonal IgVH Diversification.

A somatic hypermutation (SHM) process is known to be ongoing in IgVH genes of B-CLL cells (aka intraclonal IgVH diversification) introducing non-random point-mutations due to the combination of two events: a positive selection operated by the antigen and/or the activation of a yet poorly investigated SHM machinery. We analysed here the expression of DNA polymerases (pol) in >95% pure CD19+cells from 44 B-CLLs, all characterized for IgVH mutations (<2%,16 cases; >2%, 28 cases) by cloning and sequencing 5–10 transcripts/B-CLL. The investigated DNA pols were pol alpha, delta, epsilon (proof-reading, involved in DNA replication), pol zeta, eta, iota (error-prone, translesion DNA pols) and pol beta (error-prone, involved in base-excision repair, BER). DNA pol expression, investigated by real-time PCR, was reported as relative levels of DNA pol/beta2microglobulin ratio. Expression of DNA pols was also studied in a series of non-CLL B cell lines and in purified naive (CD19+CD27−) and memory (CD19+CD27+) B cells from PB samples of 5 healthy donors. DNA pols iota and beta were the major DNA pols in B-CLLs; they were expressed at median levels which were from 4.7 (pol beta, range 3.6–6.7, p<0.001) to 19.8 (pol iota, range 15.1–28.2, p<0.0000) folds higher than those of the other DNA pols; these latter were all expressed at significantly lower levels than in normal (naive/memory) B cells or non-CLL B cell lines. As compared to non-CLL B cell lines, DNA pols beta and iota were expressed at 86.5 (p=0.00012) and 7.7 (p<0.0000) folds higher levels, respectively. In addition, pol beta was the sole DNA pol expressed in B-CLL cells at higher (2.5-fold) levels than in normal naive/memory B cells (p=0.028). Although the high expression, DNA pol beta and iota half-lives (57 and 83 mins, respectively), evaluated by blocking RNA synthesis by 5.0 mg/ml actinomycin D up to 4 hrs, were comparable to those of the other DNA pols (range 44–83 mins). By analysing the pattern of specific IgVH mutations in 57 transcripts from 10 B-CLLs with >2% IgVH mutations and lacking antigen-selection, we demonstrated significant correlations (r > 0.70) between pol beta expression and IgVH point-mutations targeting the nucleotide A (A->C, r=0.72), T (T->X, r=0.71; T->C, r=0.82) or C (C->X, r=0.70). Conversely, no correlations were detected in 145 transcripts from 18 B-CLLs with >2% IgVH mutations and evidence of antigen-selection, suggesting that, in these cases, the effects of DNA pols could be masked by the positive pressure exerted by the antigen. In B-CLL cells, DNA pol beta expression was also compared with the expression of activation-induced cytidine deaminase (AID) and uracil-DNA glycosylase-2 (UNG2), enzymes involved in the preliminary steps of SHM. Interestingly, pol beta expression levels were significantly higher (p=0.014) in UNG2+ (14/44), as compared to UNG2- (29/44) cases (median relative levels=16.8 vs. 3.9). A similiar trend was also observed by comparing AID+ with AID- B-CLLs (median relative levels=11.2 vs. 6.8). Taken together, these findings suggest a non-canonical SHM process being active in B-CLL cells in which AID/UNG2 create the abasic sites, while the error-prone DNA pol beta (and iota) may be responsible of specific IgVH point-mutations mainly through a process (BER) not necessary implying DNA replication.