Ig Heavy Chain Gene Rearrangements in T-Cell Acute Lymphoblastic Leukemia Exhibit Predominant Dh6-19 and Dh7-27 Gene Usage, Can Result in Complete V-D-J Rearrangements, and Are Rare in T-Cell Receptor β Lineage

WHEN REARRANGED Ig and T-cell receptor (TCR) genes were found to be useful clonal leukemia-specific markers, they were initially considered to be lineage-specific.1,2However, the recognition of abnormal TCR gene rearrangements in precursor-B-acute lymphoblastic leukemia (ALL) and in some mature B-cell malignancies together with the finding of rearranged Ig heavy chain (IGH) genes in some T-ALL raised the question of lineage infidelity or promiscuity.2,3 TCR gene rearrangements in precursor-B-ALL were observed in 60% to 80% of patients, indicating that these recombinations should rather be regarded as an ubiquitously occurring (cross-lineage) phenomenon in this type of B-cell malignancy.4-7 In a recent study on 202 precursor-B-ALL patients, we even found cross-lineage TCR gene rearrangements in 93% of the patients.8 The occurrence of cross-lineage rearranged TCR genes is related to the maturation stage, being significantly lower in immature CD10⁻ precursor-B-ALL as compared with CD10⁺ precursor-B-ALL.8,9Analogous to the hierarchical order during early T-cell development, rearrangements and/or deletions in the TCR delta (TCRD) locus occur most frequently (89% of patients), followed by TCR gamma (TCRG) (59%) and TCR beta (TCRB) (35%) gene rearrangements.4-6,8 Furthermore, most TCRB gene rearrangements are restricted to the Jβ2 locus, the majority ofTCRG gene rearrangements involve the Jγ1 gene segments, and a striking predominance of incomplete Vδ2-Dδ3 and Dδ2-Dδ3 rearrangements in the TCRD locus is observed.4-11 

In contrast to the high frequency of cross-lineage TCR gene recombinations in precursor-B-ALL, IGH gene rearrangements in T-ALL are relatively rare. In a meta-analysis of previously reported small patient groups, the prevalence of rearranged IGH genes was estimated at 10% to 15% of lymphoblastic T-cell malignancies.11 All early studies were exclusively based on Southern blotting (SB), and at that time, the configuration ofIGH diversity (D) regions was not completely known. Therefore, it was not possible to reliably discern between incomplete Dh-Jh and complete Vh-(Dh-)Jh gene rearrangements and to identify cross-lineage IGH gene rearrangement patterns, which are characteristic for T-ALL.11,12 The only consistent finding was that a significant proportion of IGH rearrangements in human leukemic T cells involved the Dh7-27 (DQ52) gene segment, which is located immediately upstream of the Jhregion (Fig 1).13 14 

To study cross-lineage IGH gene rearrangements in T-ALL in detail, we analyzed a large group of 118 T-ALL patients by Southern blotting to determine the precise incidence of rearranged IGHgenes. Heteroduplex polymerase chain reaction (PCR) analysis and subsequent sequencing were applied to identify the rearrangedIGH gene segments as well as the junctional regions of the rearrangements. In addition to Vh family-specific primers and a Jh consensus primer, we designed a new set of Dhfamily-specific primers to detect and identify all complete Vh-Jh and incomplete Dh-Jh gene rearrangements.

Peripheral blood (PB) or bone marrow (BM) samples from 118 T-ALL patients (84 children and 34 adults) were obtained at initial diagnosis.7,15 Mononuclear cells (MNC) were isolated from PB or BM samples by Ficoll-Paque centrifugation (density, 1.077 g/mL; Pharmacia, Uppsala, Sweden) and subjected to detailed immunophenotyping according to standard protocols.16,17 To analyze sufficient numbers of CD3⁺ T-ALL (especially TCRγδ⁺T-ALL), we selected T-ALL cell samples based on their CD3/TCR immunophenotype, resulting in 69 CD3⁻ T-ALL (58% of the total series), 25 TCRαβ⁺ T-ALL (21%), and 24 TCRγδ⁺ T-ALL (20%). In an entirely random series of T-ALL, this immunophenotype distribution would approximately be 70%, 20%, and 10%, respectively.18 Although in 2 CD3⁺ T-ALL cases no information about TCR protein expression was available, they were included into the TCRγδ⁺ T-ALL group based on the Southern blot finding of biallelic TCRD rearrangements in both cases.

DNA was isolated from fresh or frozen MNC fractions as described previously.2 Fifteen micrograms of DNA was digested with the appropriate restriction enzymes (Pharmacia), size-separated in 0.7% agarose gels, and transferred to Nytran-13N nylon membranes (Schleicher and Schuell, Dassel, Germany), as described.2,19,IGH gene configuration was analyzed in the 118 T-ALL patients with the IGHJ6 probe (DAKO Corp, Carpinteria, CA) in Bgl II, BamHI/HindIII,EcoRI, and/or HindIII digests.20 The configuration of the TCRD genes was analyzed in 101 of 118 patients with the TCRDV2, TCRDD2, TCRDJ1, TCRDJ2, TCRDJ3, TCRDRE, and TCRAPJ probes (DAKO Corp) in Bgl II, EcoRI, andHindIII digests.15 

PCR was essentially performed as described previously.7,19In each 50 μL PCR reaction, 50 ng DNA sample, 6.3 pmol of the 5′ and 3′ oligonucleotide primers, and 0.5 U AmpliTaq Gold polymerase (PE Biosystems, Foster City, CA) were used. The sequences of the oligonucleotides used for amplification of complete Vh-Jh gene rearrangements (6 IGH framework-1 Vh-family specific primers, and 1 Jh consensus primer) were published before.21,22 Based on recently published data of germline DNA D-region sequences of the human IGHlocus23 (EMBL accession no. X97051; for the detailed organization of the IGH D-region, see Fig 1), 7 family-specific Dh primers were designed using OLIGO 6.0 software (Dr W. Rychlik, Molecular Biology Insights, Inc, Plymouth, MN) applying previously described guidelines (Table1).24 Oligonucleotide primers of 22 to 24 bp were positioned at least 50 bp upstream of the involved recombination signal sequence (RSS). Secondary structures such as primer dimers and hairpins were avoided, and the melting temperature (Tm) was 68°C ± 3°C. All primers were synthesized on an ABI 392 DNA synthesizer (PE Biosystems) using the solid-phase phosphotriester method.

PCR conditions were initial denaturation for 3 minutes at 92°C, followed by 35 cycles of 45 seconds at 92°C, 90 seconds at 60°C, and 2 minutes at 72°C using a Perkin-Elmer 480 thermal cycler (PE Biosystems). After the last cycle, an additional extension step of 10 minutes at 72°C was performed. Appropriate positive and negative controls were included in all experiments.24 

Heteroduplex analysis of PCR products included denaturation at 94°C for 5 minutes after the final cycle of amplification and subsequent renaturation at 4°C for 60 minutes to induce duplex formation.25 Afterwards, the duplexes were immediately loaded on 6% nondenaturing polyacrylamide gels in 0.5× Tris-boric acid-EDTA (TBE) buffer, run at room temperature, and visualized by ethidium bromide staining.25 A 100-bp DNA ladder (Promega Corp, Madison, WI) was used as size marker.

PCR products found to be clonal by heteroduplex analysis were directly sequenced except for cases in which heteroduplex PCR analysis showed more than two clonal bands, ie, either two homoduplexes, or an additional upper band resulting from extension to downstream Jh segments, or a Dh7-27-Jh1 germline band accompanying a Dh7-27-Jh rearrangement. In such cases, homoduplexes were excised from the polyacrylamide gel and eluted as described before.26 27 The eluted PCR products were either directly sequenced or subjected to second-step PCR with the same primer pair to increase the amount of template for sequence analysis. Sequencing was performed using the dye-terminator cycle sequencing kit with AmpliTaq DNA polymerase FS on an ABI 377 sequencer (PE Biosystems). Briefly, 50 to 200 ng of PCR product and 3.2 pmol primer were used in a 15 μL reaction volume. The cycling protocol was 96°C for 30 seconds, followed by 60°C for 4 minutes for a total of 25 cycles. Each PCR product was sequenced in two directions.

Vh, Dh, and Jh segments were identified using DNAPLOT software (W. Müller, H-H. Althaus, University of Cologne, Cologne, Germany) by searching for homology with all known human germline Vh, Dh, and Jh sequences obtained from the VBASE directory of human Ig genes (http://www.mrc-cpe.cam.ac.uk/imt-doc/).28 For alignments of D segments in Vh-Dh-Jh or Dh-Dh-Jh rearrangements, it was necessary to have at least 10 consecutive matching nucleotides.23Palindromic (P-region) nucleotides (maximally 2) generated during the joining process were recognized as being palindromic to the juxtaposed nucleotides of an untrimmed rearranged gene segment.29Extensive N-regions (nucleotides that cannot be assigned to V, D, J gene segments, or P-regions) were analyzed in more detail by comparing them to the most recent update of GenBank using the BLAST sequence similarity-searching tool (National Center for Biotechnology Information; http://www.ncbi.nlm.nih.gov/BLAST/).30 

IGH gene rearrangements were found in 22% (26/118) of T-ALL patients and were equally distributed between the different age groups, ie, 23% (19/84) of children and 21% (7/34) of adults. In the majority of cases, this concerned monoallelic rearrangements (69% [18/26]), and in 31% (8/26) biallelic rearrangements were observed. In 2 of these patients, SB analysis showed weak bands, most probably derived from subclones. Cross-lineage IGH gene rearrangements were found in 19% (13/69) of CD3⁻ T-ALL and in 50% of TCRγδ⁺ T-ALL (12/24), whereas only a single TCRαβ⁺ T-ALL (4% [1/25]) displayed a rearrangedIGH gene on one allele.

Because CD3⁻ T-ALL theoretically represent precursor stages of both TCRαβ⁺ and TCRγδ⁺ T-ALL, we analyzed the configuration of the TCRD genes in 57 CD3⁻ T-ALL cases and used it as an additional marker for further subdivision of this group.15,18 TheTCRD gene configuration of each allele can potentially pass three consecutive stages: germline, rearrangement, and deletion.15 18 Except for 1 case, all IGH gene rearrangements were found in patients with at least one TCRDgene rearrangement; in this single CD3⁻ T-ALL, a monoallelic IGH gene rearrangement was observed in combination with deletion of both TCRD alleles.

Based on the above-described results, we analyzed the association of cross-lineage IGH gene rearrangements and TCRD gene configuration for 101 of the 118 T-ALLs. Rearrangements in the IGHlocus appeared to be almost exclusively associated with TCRDgene rearrangements. They were evenly distributed between cases with 1 rearranged and 1 germline TCRD allele (1 of 3 cases [33%]), cases with biallelic TCRD rearrangements (15 of 47 cases [32%]), and cases with 1 rearranged and 1 deleted TCRDallele (7 of 22 cases [32%]). Remarkably, only 1 of 26 cases with biallelically deleted TCRD genes (4%) displayed a cross-lineage IGH gene rearrangement.

Taking these data together, cross-lineage IGH gene rearrangements occurred in only 5% (2/38) of αβ-lineage T-ALL, ie, either TCRαβ⁺ T-ALL (n = 25) or CD3⁻T-ALL with biallelic TCRD deletion (n = 13).

Detailed PCR analysis of the IGH locus in the 26 patients with Southern blot documented IGH gene rearrangements was based on 13 primer combinations covering the vast majority of complete Vh-(Dh-)Jh joinings and potentially all incomplete Dh-Jh rearrangements. Heteroduplex PCR analysis showed a total of 39 clonal homoduplexes, reflecting 7 (18%) complete Vh-(Dh)-Jh joinings and 32 (82%) incomplete Dh-Jh rearrangements. Complete Vh-(Dh)-Jh rearrangements were found in 4 CD3⁻ T-ALL and 2 TCRγδ⁺ T-ALL patients, with 1 TCRγδ⁺ T-ALL showing biallelic Vh-(Dh-)Jh joinings. Heteroduplex PCR analysis for incomplete Dh7-27-Jh rearrangements is shown in Fig 2.

Sequence analysis confirmed monoclonality in 7 complete Vh-(Dh-)Jh rearrangements, which involved 7 different gene segments from 4 families: Vh1-3, Vh1-69, Vh3-13, Vh3-23, Vh3-33, Vh4-4, and Vh6-1. None of the rearrangements was potentially functional. Six sequences were out-of-frame joinings, whereas the single rearrangement with an in-frame Vh-Jh contained a stop codon in the junctional region.

Sequence analysis of the junctional regions of complete Vh-(Dh-)Jh recombinations allowed identification of a D segment in 5 of 7 joinings. Moreover, 1 incomplete rearrangement was of the Dh-Dh-Jhtype. The frequencies of different Dh family members found among the 38 identified Dh sequences in the complete and incomplete cross-lineage IGH gene rearrangements are summarized in Table 2. Usage of the Dh6 family was most prominent (47%), with the Dh6-19 gene segment being preferentially used (45% of all identified Dhsequences; Table 3). The second most frequently used Dh segment was Dh7-27 (21%). Seven rearrangements (18%) contained Dh segments of the Dh1 family, whereas 5 other rearrangements used various segments of the remaining 4 Dh families (Table 2). Taken together, only 5 of the 38 rearranged Dh segments belonged to the most upstream part of the Dh region, whereas all other 33 Dh gene segments (87%) belonged to the most downstream part of the Dh region (Table 2). No relationship between age and Dh gene segment usage was observed.

The sizes of the Dh-Jh junctional regions ranged from 0 to maximally 32 nucleotides, with an average of 7.6 nucleotides. Three of 37 Dh-Jh junctions (8%) did not have any randomly inserted N-region nucleotides. P-nucleotides, indicating the absence of deletion, were present in 7 Dh-Jh joinings (19%).

The frequencies of different Jh gene segments in IGHgene rearrangements in T-ALL are summarized in Table 4. The Jh4 gene segment was found most frequently in approximately one third of joinings, followed by Jh6 in 20% of cases. The remaining 4 Jh segments were almost equally used, each comprising 10% to 15% of the rearrangements.

SB analysis and heteroduplex PCR analysis showed fully concordant results in 20 of 26 cases with cross-lineage IGH gene rearrangements in T-ALL. In 1 case with a single rearranged band on SB, we were not able to amplify the clonal rearrangement with the applied primer sets. In the remaining 5 cases, the number of clonal PCR-detected homoduplexes was higher than the number of rearranged bands in SB, which may suggest the presence of minor subclones undetectable by SB. In 1 of these seemingly oligoclonal cases, the identified incomplete and complete rearrangements (1 and 2 in Table 3) shared the same Dh-Jh fragment, suggesting ongoing Vh to Dh-Jh joining. In 4 other patients, the detected rearrangements were not related in their used gene segments. This may reflect secondary Dh-Jh joining, with concomitant deletion of a pre-existing Dh-Jhrearrangement.

Based on the combined SB/PCR results, we found evidence for IGHoligoclonality at diagnosis in 27% (7/26) of T-ALL patients with this type of cross-lineage recombination. This includes the above-mentioned 5 cases with the higher number of PCR-detected homoduplexes than the number of rearranged bands in SB, and 2 additional PCR-positive cases with weak bands on SB analysis, apparently derived from subclones.20 

We investigated a large group of 118 T-ALL patients for the presence of cross-lineage IGH gene rearrangements. Based on SB analysis, we identified such rearrangements in 22% of T-ALL, which is slightly higher than previously reported.11,12 The vast majority (82%) of IGH gene rearrangements in T-ALL concerned incomplete Dh-Jh joinings. However, complete Vh-(Dh-)Jh recombinations were also documented in our group of T-ALL patients. The usage of Vhgene segments was seemingly random and not limited to the ones most proximal to the Jh cluster. Nevertheless, 6 of 7 involved Vh gene segments were derived from the proximal 3′ portion of the IGH locus, a pattern that is seen in first trimester fetal Vh-(Dh)-Jhrearrangements.31 None of the seven Vh-(Dh-)Jh rearrangements was potentially functional. To our knowledge, this is the first extensive evidence for the occurrence of clonal complete IGH gene rearrangements in T-ALL. They were previously reported in a single case of T-lymphoblastic non-Hodgkin’s lymphoma, which is the lymphomatous counterpart of T-ALL.32 

In both the Vh-(Dh-)Jh and Dh-Jh rearrangements, we found a strikingly preferential usage of the more downstream Dh gene segments (87% of identified Dh sequences), especially Dh6-19 (45%) and Dh7-27 (21%). The Dh6-19 gene segment is one of the gene segments that were recently discovered thanks to complete sequencing of the Dh region.23Retrospective analysis of nearly 900 IGH junctional regions of B-lineage cells showed Dh6-19 involvement in approximately 5% of the rearrangements, which is significantly higher than would be expected on a random basis.23 The same holds true for 8 other Dh segments, but 6 of them were not found in our T-ALL patients. The preferential usage of Dh6-19 could not be explained by a more optimal RSS at the 3′ end in comparison to other Dh segments, either. Further studies are needed to define whether this gene segment is also preferentially rearranged in precursor-B cells or whether this finding only relates to the cross-lineage phenomenon in T-ALL. In contrast to Dh6-19, earlier reports indicated that the Dh7-27 gene segment is involved in a significant proportion of IGH gene rearrangements in T-ALL.13,14 Interestingly, we also found this gene segment in a cross-lineage IGH gene rearrangement of a TCRαβ⁺ T-prolymphocytic leukemia (Fig 2). The Dh7-27 gene segment is also preferentially used by fetal B cells but is rarely observed in adult BM and PB.33-36 In the analysis of nearly 900 IGH junctional regions of B-lineage cells, Dh7-27 was found in only 0.5% of the rearrangements.23 Because Dh7-27 consists of only 11 nucleotides, it is more difficult to identify this gene segment in a junctional region after moderate trimming during V(D)J recombination, if stringent assignment criteria are used.

The analysis of Jh segment usage showed a more frequent use of Jh4 and Jh6 gene segments, which is also the case in normal and leukemic B-lineage cells (Table 4).35-37However, in our T-ALL group, a significant proportion of joinings involved the most upstream Jh1 and Jh2 gene segments (28%), which are rarely used by B lymphocytes (∼1%) and B-cell precursors (3% to 4%).35-37 In conclusion, the high frequency of incomplete Dh-Jh rearrangements together with the frequent usage of the more downstream Dhgene segments and the most upstream Jh1 and Jh2 gene segments suggest a predominance of the most immature IGHrearrangements in T-ALL. This particular Dh-Jhrearrangement pattern appears to be nonrandom and is not comparable with any known stage of B-cell ontogeny or B-cell differentiation.31 

Because the types of preferential IGH gene rearrangements have now been identified, it would be relatively easy to screen T-ALL patients for the presence of cross-lineage IGH gene rearrangements and apply them as PCR targets for monitoring of minimal residual disease (MRD) in T-ALL. Three primer combinations (Dh1, Dh6, and Dh7 in combination with a Jh consensus primer) can identify 85% of incomplete Dh-Jh rearrangements in T-ALL. However, we observed oligoclonality in the IGH locus in 27% of T-ALL patients with rearranged IGH genes. In 1 case, we found evidence for continuing Vh to Dh-Jh recombination, whereas Vh replacements have been described previously during disease progression of a T-lymphoblastic lymphoma.32Secondary rearrangements via continuing Vh to Dh-Jh joining, Vh gene replacements, and de novo IGH gene rearrangements have been reported for precursor-B-ALL.38-40 These processes may lead to the emergence of clones with secondary IGH rearrangements. In a previous study, we compared the IGH gene rearrangement patterns between diagnosis and relapse in 40 ALL patients and found that at least one major IGH rearrangement was stable in most cases.41 Therefore, cross-lineage IGH gene rearrangements might be useful as supplementary MRD target in addition to leukemia-specific TCRG and TCRD gene rearrangements and TAL1 gene deletions.24 42 

Cross-lineage IGH gene rearrangements occurred most frequently in TCRγδ⁺ T-ALL (50% of patients) and in 20% of CD3⁻ T-ALL, but we found them in a single case of TCRαβ⁺ T-ALL (4%). Moreover, there was a striking association between the presence of rearranged TCRD genes and the occurrence of IGH recombination. A similar association was previously found for cross-lineage gene rearrangements in acute myeloid leukemia, suggesting that TCRD and IGHgenes are concomitantly accessible for V(D)J recombinase in early hematopoietic precursors.43 Furthermore, in the genotypically most mature T-ALL subgroup with biallelic TCRDgene deletions, an IGH gene rearrangement was observed in only 1 of 26 cases. Because IGH gene rearrangements are also rare in mature T-cell malignancies (∼5%),11 this suggests that the IGH locus may be accessible to V(D)J recombinase activity only in cells at earlier stages of T-cell differentiation. In this context, TCRγδ⁺ T-ALL should be regarded as cells that branched off T-cell development at an early stage of completion ofTCRD and TCRG gene rearrangement processes, when the recombinase activity is still retained.18 

Incomplete Dh-Jh rearrangements are one of the earliest events during normal B-cell development and are already found in CD34⁺/CD19⁻/CD10⁺precursor cells.44,45 Most of the more mature CD34⁺/CD19⁺/CD10⁺ B-lineage precursors contain at least 1 Dh-Jh rearranged allele and frequently also complete Vh-(Dh-)Jhrearrangements.45,46 One could therefore speculate that cross-lineage IGH gene rearrangements in T-ALL might reflect malignant transformation of a thymocyte derived from a CD34⁺/CD19⁻/CD10⁺ precursor cell with rearranged IGH genes. This idea may be supported by a murine model, in which IGH rearrangements were only found at an intermediate stage of thymocyte development.47Nevertheless, this phenomenon was not observed in normal human thymocytes.44 Furthermore, the absence of D-J-Cμ transcripts in fetal human thymocytes depleted of CD34⁺- and/or CD19⁺-bearing cells suggest that Dh-Jh rearrangements in humans may be restricted to normal B-lineage differentiation.44,48 It has also been suggested that IGH gene rearrangements in T-cell precursors may be an aberrant event directing cells into apoptotic pathway, unless they become immortalized by malignant transformation.3 An alternative explanation could be that cross-lineage IGH gene rearrangements in T-ALL are postoncogenic events resulting from the ongoing activity of the common B- and T-cell V(D)J recombinase system on accessible gene loci.3,49,50 We favor the last hypothesis, because this would explain the virtual absence ofIGH gene rearrangements in normal thymocytes and mature T-cell malignancies on the one hand and the presence of IGHoligoclonality and secondary Vh-(Dh-)Jh rearrangements in T-ALL on the other hand.41 The virtual absence of IGHgene rearrangements in αβ-lineage T-ALL with biallelic TCRDgene deletions suggests that the recombination system is not active in this more mature type of T-ALL. Apparently, the recombinational activity is switched off as soon as the rearrangement and deletion processes in the TCRA/TCRD locus are completed.

The authors thank Drs K. Hählen, I.M. Appel, R.M. Egeler, F.G.A.J. Hakvoort-Cammel, W.J.D. Hofhuis, G.E. van Zanen, B. Löwenberg, P. Sonneveld, G. J. Ossenkoppele, G.J. Schuurhuis, G. Verhoef, Ph. Vandekerckhove, M. Stul, E.J. Petersen, A.W. Dekker, D. Campana, J.C. Kluin-Nelemans, W-D. Ludwig, and C.E. van der Schoot and the Dutch Childhood Leukemia Study Group (DCLSG) for kindly providing T-ALL cell samples. Board members of the DCLSG are I.M. Appel, H. van den Berg, J.P.M. Bökkerink, M.C.A. Bruin, J.J. Groot-Loonen, S.S.N. de Graaf, K. Hählen, P.M. Hoogerbrugge, W.A. Kamps, F.A.E. Nabben, J.A. Rammeloo, T. Revesz, A.Y.N. Schouten-van Meeteren, A.J.P. Veerman, M. van Weel-Sipman, and R.S. Weening. We are grateful to Prof Dr R. Benner, Prof Dr D. Sońta-Jakimczyk, and Dr H Hooijkaas for their continuous support and to Dr E. Moreau and I.L.M. Wolvers-Tettero for technical assistance, T.M. van Os for preparation of the figures, and A.D. Korpershoek for secretarial support.