Mature T-cell non-Hodgkin lymphomas (T-NHLs) are rare in western Europe and in the USA, and their pathogenesis, as well as underlying genetic alterations, is poorly understood 1. One major exception is anaplastic large cell lymphoma (ALCL) that carries translocations involving the ALK gene in chromosome 2p23 in a significant proportion of cases 2. While conventional karyotyping 3–5, comparative genomic hybridization (CGH) 6, 7 and, very recently, a matrix-CGH-based approach 8 have revealed a large number of genetic alterations and genomic imbalances in T-NHL [including recurrent aberrations such as gains in 9q in enteropathy-type T-cell lymphoma (EATCL) 6], recurrent chromosomal translocations appear to be a rare event in peripheral T-NHL, not otherwise specified (PTCL NOS), angio-immunoblastic T-cell lymphoma (AILT), and EATCL. Nevertheless, a novel and recurrent translocation, t(5;9)(q33;q22), fusing the ITK and SYK genes was recently identified in rare cases of PTCL NOS, with predominant involvement of lymphoid follicles and positivity for CD10 and BCL6 9.
Analogous to the immunoglobulin (IgH) receptor loci that are frequently affected by translocations in B-cell lymphomas 10, the T-cell receptor (TCR) gene loci are targeted by chromosomal breakpoints in approximately 30% of precursor T-cell lymphoblastic leukaemias/lymphomas involving various translocation partners 11. Likewise, T-cell prolymphocytic leukaemia (T-PLL) harbours frequent alterations of the TCRα/δ locus, usually caused by an inv(14)(q11q32) 1. Most of these translocations result in the transcriptional deregulation of an oncogene by juxtaposing it to regulatory sequences of the TCR genes, thus playing an important role in the pathogenesis of these disorders. Whereas chromosomal translocations involving the TCRα/δ, TCRβ, and TCRγ loci were initially thought to be absent in mature T-NHL other than T-PLL, recent data suggest that occasional cases of PTCL NOS and AILT do carry breakpoints in one of the TCR loci 12. More specifically, the translocation t(14;19)(q11;q13) involving the TCRα/δ and BCL3 loci was demonstrated in two PTCL NOS and in one AILT recently 13.
To assess the accurate frequency of breakpoints occurring in one of the TCR loci in mature T-NHL, we designed improved fluorescence in situ hybridization (FISH) assays for the TCR loci that are feasible for large-scale analysis of formalin-fixed and paraffin-embedded (FFPE) lymphoma specimens in a tissue microarray (TMA) format. Using this approach, we identified two additional mature T-NHL cases with a chromosomal breakpoint in the TCRα/δ locus. In contrast, no breaks within the TCRβ and TCRγ loci were detected among 245 well-characterized T-cell lymphoma samples, making the rearrangement of TCR loci a very rare event in these neoplasms that occurs in less than 1% of cases.
Materials and methods
Probe design and FISH procedures
BAC clones flanking the TCRα/δ, TCRβ, and TCRγ loci were selected according to previously published literature 12 and their genomic location using the NCBI Map Viewer (http://www.ncbi.nlm.nih.gov/mapview/) (Figure 1 and Table 1).
Table 1. BAC clones and commercially available probes used for the present FISH study
Location to breakpoint
RZPD = German Resource Center for Genome Research.
In order to test for putative translocation partners in cases showing a chromosomal breakpoint in one of the TCR loci, additional FISH studies were performed using probes for BCL11b, HOX11A/TLX1, MYC, BCL3, the TCL cluster, and the IgH locus (Table 1). The BCL11b break-apart probe set was designed using the BACs RP11-543C4, RP11-860P2, RP11-1127D7, and RP11-1082A3.
The combination of BACs RP11-725G5, RP11-164H13, and RP11-185P18 flanking and partly spanning the TCL cluster 12 was kindly provided by R Siebert (Institute of Human Genetics, Kiel, Germany).
Fluorescence in situ hybridization (FISH) on meta- phase preparations and cell suspensions was performed according to a standard protocol 14. FISH on routine formalin-fixed and paraffin-embedded (FFPE) tissue samples was carried out on whole paraffin sections and in a tissue microarray (TMA) format as previously described 15. To establish the cut-off levels for the TCRα/δ, TCRβ, and TCRγ probe sets in FFPE tissue sections, two reactive lymph nodes and ten B-cell lymphomas were used as negative controls. To characterize the cells carrying a specific chromosomal breakpoint further, FICTION was used as described previously 16, with antibodies for CD3 and CD79a.
To test the newly designed FISH assays, three human T-cell lymphoblastic lymphoma-derived cell lines (HSB-2, MOLT-3, and KE-37) and 21 primary lymphoma samples were selected. KE-37 harbours a chromosomal breakpoint at the 14q11 locus; HSB-2 carries a translocation, t(1;7)(p34;q35); and MOLT-3 is cytogenetically characterized by the chromosomal translocation t(7;7)(p15;q11).
Among 20 cases of T-cell prolymphocytic leukaemia (T-PLL) selected from the files of the Institute of Pathology, University of Würzburg, three cases carried an inv(14)(q11q32) based on previous karyotypic analysis. A precursor T-cell lymphoblastic leukaemia/lymphoma that was previously shown to harbour a chromosomal break in the TCRγ locus 12 was kindly provided by R Siebert (Kiel).
Two hundred and forty-five mature T-cell lymphomas, referred to the Institute of Pathology, University of Würzburg, Germany and the Department of Pathology, University of Vienna, Austria, were selected for this study [82 peripheral T-cell lymphomas, not otherwise specified (PTCL NOS) including 20 cases of the lympho-epithelioid cell variant of peripheral T-cell lymphoma (Lennert lymphoma); 41 systemic anaplastic large cell lymphomas (ALCLs; 14 ALK-positive and 27 ALK-negative cases); seven primary cutaneous ALCLs, 29 angio-immunoblastic T-cell lymphomas (AITLs); and 86 intestinal (enteropathy-associated) T-cell lymphomas (EATCLs)]. All samples were classified according to the WHO criteria 1. Ethics approval was obtained from the Ethics Committee, University of Würzburg.
Immunohistochemical characterization of the T-cell lymphomas was performed using a broad panel of antibodies according to standard protocols 17.
Following hybridization of FFPE control samples, the cut-offs were set at 12.2% for the TCRα/δ, 10.2% for the TCRβ, and 9.9% for TCRγ probes, respectively.
A chromosomal breakpoint in the TCRα/δ locus was detected in the KE-37 cell line that harboured segregated FISH signals in 92% of the cells (not shown) and in 13 out of 19 T-PLL patient samples showing split signals in 23–76% of cells (Figure 1a), including three cases with a cytogenetically proven translocation.
A chromosomal breakpoint in the TCRγ locus was detected in 27% of the cells of the precursor T-cell lymphoblastic leukaemia/lymphoma with a reported TCRγ translocation (Figure 1b), but not in the MOLT-3 cell line. Eighty-seven per cent of the cells of the HSB-2 cell line carried a split signal when hybridized with the probe set for the TCRβ locus (Figure 1c).
Two T-cell lymphomas harbour chromosomal breakpoints in the TCRα/δ locus
Two of the 245 mature T-cell lymphomas (0.8%) included in the study showed evidence of a chromosomal breakpoint affecting the TCRα/δ locus, whereas no alterations were observed in the TCRβ and/or TCRγ loci. The first sample was an enlarged inguinal lymph node from a 56-year-old female patient. The normal lymph-node architecture was effaced and a diffuse infiltrate was observed consisting of small to medium-sized lymphocytes with moderate nuclear irregularities and occasional clear cell features. Clusters of epithelioid histiocytes were present in large numbers throughout the lymph node (Figure 2a). Immunophenotypically, tumour cells expressed the T-cell markers CD2, CD3, CD5, and CD8, and the cytotoxic molecules TIA-1 and granzyme B, while CD4, CD56, CD57, and CD30 were not expressed (Figures 2b and 2c). On the molecular level, clonal rearrangement of the TCRγ locus could be demonstrated by PCR. Taken together, the diagnosis of the lympho-epithelioid cell variant of peripheral T-cell lymphoma (Lennert lymphoma) was established. This case showed evidence of a chromosomal break within the TCRα/δ genes in 35% of the cells (Figures 2d and 2e). Subsequent hybridizations identified a breakpoint affecting the BCL3 gene in approximately the same percentage of cells, suggesting the presence of the translocation t(14;19) in the tumour cells juxtaposing the BCL3 gene to the TCRα/δ locus. Interestingly, the FISH pattern in the BCL3 assay consisted of two co-localized signals and a single green signal, indicating that the chromosomal breakpoint is located in the genomic region telomeric from the BCL3 gene (Figures 2f and 2g). There was no evidence of a recurrent chromosomal breakpoint within the TCRα/δ genes among Lennert lymphoma cases, since none of the 19 remaining cases with this morphological subtype showed any indication for a breakpoint in this region.
The second case was a lymph node from a 50-year-old woman that showed partial effacement of the lymph node architecture. Increased numbers of follicular dendritic cells were present and visualized by CD21 and CD23 stains. The paracortex showed infiltration of small to medium-sized lymphocytes with clear cytoplasm and expression of the T-cell markers CD3, CD4, and CD5 (Figures 2h–2j). In a relapse sample that was obtained 5 years later, the lymph node had similar morphological features and the tumour cells showed an identical immunophenotype. PCR-based rearrangement analysis for the γ-chain of the TCR demonstrated identical tumour cell clones in the initial and relapse sample. Tumour cell proliferation was high (Ki-67 index: 80%) and no expression was observed for TdT. EBER in situ hybridization detected EBV sequences in a few activated, blastic B-cells; however, there was no indication for monoclonal proliferation of B-cells according to the PCR-based rearrangement analysis for the FR2A and FR3A regions of the immunoglobulin heavy chain variable region gene. In this case, all criteria for the diagnosis of angio-immunoblastic T-cell lymphoma (AILT) were fulfilled. Conventional cytogenetic analysis in the relapse sample yielded a complex karyotype (42 ∼ 47,XX, add(1)(p11), add(6)(q23)?dup(6)(q23qter), del(6) (q15), − 6, − 8, − 10, + 11, inv(14)(?q11q32), + mar1, + der(9)t(1;9)(p10;p22), + mar2?der(20)[cp21]). By FISH, 20% of the total cells in the primary biopsy and 50% of the cells in the relapse sample 5 years later showed evidence of a break in the TCRα/δ genes (Figure 2k). Interestingly, a metaphase preparation from the relapse sample showed this specific alteration in nearly all metaphases present. The centromeric probe of the TCRα/δ FISH assay remained on 14q11, whereas the telomeric probe was translocated to the region 14q32, a pattern indicating the presence of an inversion of chromosome 14. Subsequent FISH studies in this case identified no breakpoints affecting the BCL11b, MYC, HOX11A, BCL3, and the TCL gene cluster loci. However, hybridization with an IgH break-apart probe resulted in the detection of an additional single green signal, suggesting a chromosomal break located within the V-genes of the IgH locus (Figures 2l and 2m). Co-hybridization of the IgH and TCRα/δ probes showed fusion of the IgH signal to the centromeric TCRα/δ probe on 14q11. In conclusion, FISH studies suggest an inversion, inv(14)(q11q32), affecting the TCRα/δ and IgH genes loci in this case. Additional FICTION analysis confirmed that this aberration was confined to CD3-positive T-cells (Figure 2n) and was not present in CD79a-positive B-cells.
Numerical aberrations in mature T-cell lymphomas
Our FISH study also allowed the assessment of numerical aberrations in mature T-NHL. A gain and/or amplification of signals in all three TCR loci suggesting polyploidy was found in ALCL, PTCL NOS, and EATCL cases (Figure 3a). Trisomy or tetrasomy 7, as defined by the presence of gains in both the TCRβ and the TCRγ loci, but not in the TCRα/δ locus, was detected in PTCL NOS, EATCL, and ALK-negative ALCL (Figure 3b).
A predominance of 7q signals (at least two more signals for the TCRβ locus compared with the TCRα/δ and TCRγ loci) was evident in EATCL (14/51), ALK-negative ALCL (4/13), cutaneous ALCL (1/3), PTCL NOS (4/41), and ALK-positive ALCL (1/10). Some of the signal constellations observed may suggest the presence of an isochromosome 7q [i(7)(q10)] 18. In particular, in two of the 14 EATCL cases, the gain of the TCRβ locus coincided with a deletion of the TCRγ locus in 7p14, strongly suggesting the presence of an i(7)(q10) (Table 2). This is additionally supported by conventional CGH analysis 6 that demonstrated a gain of nearly the whole arm of 7q and a concomitant loss of 7p in these two cases (Table 2). Four of the remaining eight EATCL cases with CGH data available (Table 2) showed chromosomal gains of the total arm or major parts of 7q by CGH without clear evidence of an i(7)(q10).
Table 2. Comparison of FISH and CGH results in EATCL cases showing evidence of chromosome 7 alterations by FISH (CGH alterations affecting chromosome 7 are highlighted in bold)
We developed improved FISH assays for the detection of breakpoints within the TCR loci and performed a large-scale analysis in formalin-fixed and paraffin-embedded (FFPE) specimens of 245 mature T-NHLs in a tissue microarray (TMA) format. Translocations involving the TCR loci are well known to occur in precursor T-cell lymphoblastic leukaemias/lymphomas and T-cell prolymphocytic leukaemias, but recent studies suggest that, occasionally, mature T-NHLs may also harbour breakpoints in one of the TCR loci 12, 13. While previous studies were performed on cytogenetic suspensions in smaller series of cases, the applicability of our FISH assay to FFPE tissue allowed for an unbiased selection of a large number of mature T-NHLs. In our extended series of mature T-NHLs, only two cases harboured a translocation of the TCRα/δ locus, whereas no chromosomal translocation of the TCRβ and TCRγ loci was detected. The present study therefore establishes the frequency of breakpoints in one of the TCR loci to be less than 1% among mature T-NHLs other than T-PLL. This result may have been anticipated given that T-cells, in contrast to germinal centre B-cells, do not undergo a second genetic recombination at a mature developmental stage. However, the frequent chromosomal translocations t(11;14) and t(14;18) present in the B-NHL subtypes of mantle cell lymphoma and follicular lymphoma, respectively, are thought to occur at an early developmental stage (during VDJ recombination) as well 19, but, nevertheless, the tumour cells reach a mature phenotype, analogous to T-PLL. Why T-cell neoplasias with an early oncogenic hit in the TCR loci are only rarely capable of developing to a mature stage is currently unknown.
Interestingly, one Lennert lymphoma (lympho- epithelioid cell variant) within the PTCL NOS category showed a signal constellation indicative of a t(14;19)(q11;q13) involving the TCRα/δ locus and BCL3. BCL3 has structural similarities to the IκB proteins that control NFκB activation 20 and promotes cell survival in B- and T-cells 21. BCL3 overexpression on the protein level was recently reported to be a regular feature of Hodgkin lymphoma, ALCL, and PTCL NOS 22. However, our data are in line with a previous report 12 indicating that BCL3 up-regulation is only rarely the result of a t(14;19) involving the TCRα/δ locus.
The second case in our series harbouring a chromosomal breakpoint in the TCRα/δ locus was an AILT carrying an inversion, inv(14)(q11q32). In contrast to common inversions involving the TCL or BCL11b gene loci in 14q32, eg in T-PLL 12, 23, 24, we could clearly demonstrate a breakpoint in the IgH locus in this case. Interestingly, a primary as well as a relapse sample taken 5 years later showed a FISH pattern indicating the inv(14)(q11q32). This finding points to an early event in tumourigenesis and a selection of tumour cells carrying the TCRα/δ/IgH rearrangement during tumour progression. Since AILT is most frequently associated with EBV-associated B-cell proliferation (as present in our case), the possibility had to be excluded that proliferating B-cells, rather than tumour T-cells, were hit by the inv(14)(q11q32). However, FICTION analysis clearly proved the presence of the inv(14)(q11q32) in CD3-positive T-cells, but not in CD79-positive B-cells.
The inv(14)(q11q32) involving the TCRα/δ and IgH loci appears to be a rare event. Denny et al25, 26 described this alteration in a childhood T-cell lymphoma and the corresponding T-cell line SUP-T1 and demonstrated an ‘in frame’ fusion transcript between the two loci suggesting translation into a functional protein. Whether the rearrangement of the TCRα/δ and IgH loci led to a fusion transcript in our AILT case was not investigated and the pathogenetic role of this event is unclear.
We also assessed the ploidy status of mature T-NHL within various entities. Cases with evidence of polyploidy among mature T-NHLs were restricted to the ALCL, PTCL NOS, and EATCL categories. Interestingly, both ALK-positive and ALK-negative ALCLs showed evidence of polyploidy in a high percentage of cases (approximately 60%), a feature that is shared with Hodgkin lymphoma 27. This genetic hallmark, however, may not be helpful in sharpening the diagnostic grey zone between ALK-negative ALCL and PTCL NOS cases, since the latter also show evidence of polyploidy in a significant subset of cases. Likewise, trisomy or tetrasomy 7 without evidence of polyploidy, in line with previously published data 7, occurred most frequently in PTCL NOS cases in the present study (almost 20%), but, again, this feature is also observed in occasional cases of ALK-negative ALCL and EATCL. Importantly, polyploidy or gains of chromosome 7 are not present in AILT, thus clearly separating this subset of T-NHLs from ALCL, PTCL NOS, and EATCL. Although the tumour cell content in AILT cases can be quite low, which may result in an underestimation of polyploidy using FISH assays, it is unlikely that this feature accounts for the absence of polyploidy in all the AILT cases investigated, since tumour cell-rich cases and areas were chosen for TMA preparation. In addition, recent investigations in AILT using an array CGH approach yielded chromosomal imbalances in 72% of cases 8.
Our study also provides indirect evidence for the presence of gains of the chromosomal arm 7q and, in particular, of isochromosome 7q in some cases of EATCL. Isochromosome 7 is currently viewed as a pathognomonic genetic alteration in hepatosplenic T-cell lymphoma and can therefore serve as a diagnostic tool for this entity 1, 18, 28, 29. Fourteen out of 51 EATCL cases in our series displayed a FISH pattern indicating a chromosomal gain of 7q and in ten cases, FISH results could be compared with conventional CGH data that were, in part, described earlier 6. The CGH data were in line with observed signal constellations in the FISH analysis and in six cases, CGH supported the notion that the whole 7q arm was affected by the chromosomal gain. In two of the ten cases, CGH demonstrated loss of the entire 7p arm, providing strong evidence for the presence of an isochromosome 7 in these cases. Although four EATCLs showed only a gain of the 7q arm, but no loss in 7p by FISH and CGH, this constellation may still be indicative of an isochromosome 7 in some cases, as pointed out in a previous FISH study in hepatosplenic T-cell lymphoma 18. Taken together, the presence of an isochromosome 7q may not be restricted to hepatosplenic T-cell lymphoma, but may also occur in a subset of EATCLs.
To summarize, we have shown that chromosomal breakpoints in the TCR loci are a rare, but recurrent event in mature T-NHL other than T-PLL. In addition, isochromosome 7q, a pathognomonic genetic feature of hepatosplenic T-cell lymphoma, may also characterize a subset of EATCLs.
We would like to thank Irina Eichelbrönner, Heike Brückner, Theodora Nedeva, and Sabine Roth for excellent technical assistance; R Siebert for providing the FISH probe set for the TCL gene cluster; and H. Stöcklein as well as E Hartmann for helpful discussions. AR, EL, and TR are supported by the Interdisciplinary Centre for Clinical Research (IZKF) of the University of Würzburg.