Correspondence re: Welzel et al., Cancer Res., 61: 1629–1636.¹

In a recent issue of Cancer Research, Welzel et al. reported the presence of T-nucleotide² insertions in t(11;14) junctions of MCL, in addition to N-nucleotides (1). The authors analyzed both direct and reciprocal BCL-1/J_H and D_H/BCL-1 from 23 positive MCLs on 93 patients. They found that 19% (7 of 37) of the BCL-1/J_H junctions with inserts of more than five nucleotides contained error-prone copies (T-nucleotides) from the surrounding BCL-1 or J_H regions. Hence, in contrast to previous assumptions (2), the formation of t(11;14) junctions might not only involve illegitimate V(D)J-mediated recombination but also a template-dependent, error-prone DNA synthesis. Similarly to t(11;14), the chromosomal translocation t(14;18) (q32;q21), which juxtaposes the BCL-2 gene with the immunoglobulin heavy chain locus, is classically assumed to result from a mistake in the normal mechanism of V(D)J recombination (3). t(14;18) is a hallmark of follicular lymphoma and has also been detected at low frequency in peripheral lymphocytes of healthy individuals (4, 5).

We performed a detailed analysis of de novo insertions in the direct breakpoint of 36 BCL-2/J_H junctions obtained from 12 of 33 healthy individuals for which DNA sequencing was achieved. The presence of T-nucleotides, defined as short sequences of at least five nucleotides inserted in the breakpoint, with sufficient homology to the adjacent flanking sequences to exclude their concomitant presence by chance alone, was investigated using the statistical approach used by Welzel et al (1).

We found highly significant templated nucleotides in 6 of 36 (16%) BCL-2/J_H junctions (Table 1). The homologies were 6–18 nucleotides in length and presented mismatches with the original sequence (mutations, deletions, and insertions) in three cases. Template sequences belonged more frequently to the J_H gene segment (four of six) than the BCL-2 gene (two of six). In one sample (111), the BCL-2/J_H insertion presented highly significant homologies with nucleotides that could either belong to the BCL-2 gene or the J_H gene segment. The BCL-2 11-bp sequence 5′-ACCCAGA-GCCC-3′ was found in reverse complement orientation with one nucleotide insertion. The J_H 11-bp sequence 5′-GGACGTCTGGG-3′ was found in the direct orientation with one nucleotide mutation. The rate of T-nucleotides found in healthy individuals appeared similar to that observed in MCL (16% versus 19%) (1). However, as only direct breakpoints were analyzed in our study, this rate might be underestimated. A paper from the same group reported that >30% of the BCL-2/J_H junctions (reciprocal and direct) found in follicular lymphomas also contained T-nucleotides insertions (6).

Hence, the unusual mechanism that leads to T-nucleotide insertions into BCL-1 and BCL-2/J_H junctions, respectively, found in MCL and follicular lymphoma (1, 6) seems also implicated in the genesis of BCL-2/J_H rearrangements in healthy individuals. Additional investigations on a larger population of subjects will be helpful to characterize the mechanisms and environmental or individual factors that influence the formation of such T-nucleotide insertions.

Roulland et al. have investigated 33 healthy individuals for BCL2/J_H translocations. Analyzing 36 breakpoints, they found 6 (16%) junctions displaying T-nucleotides as described by our group (1, 2). This is the first confirmation of our data by another lab. Roulland et al. have extended our observations in follicular and MCLs to t(14;18) junctions in normal subjects. This indicates that the mechanism of translocation is similar in both instances. We appreciate this interesting work by our French colleagues.

We thank Dr. B. Nadel for providing us the statistical method for the identification of T-nucleotides.