Detection of clonal T cell receptor γ (TCRG) gene rearrangements by PCR is widely used in both the diagnostic assessment of lymphoproliferative disorders and the follow-up of acute lymphoblastic leukaemia (ALL), when residual positivity in excess of 10−3 at morphological complete remission is increasingly recognised to be an independent marker of poor prognosis. This is largely based on specific detection of V–J rearrangements from childhood cases. We describe rapid, multifluorescent Vγ and Jγ PCR typing of multiplex amplified diagnostic samples, as applied to 46 T-ALL. These strategies allow selected analysis of appropriate cases, immediate identification of Vγ and Jγ segments in over 95% of alleles, improved resolution and precision sizing and a sensitivity of detection at the 10−2–10−3 level. We demonstrate preferential V–J combinations but no difference in V–J usage between children and adults, nor between SIL-TALI-negative and -positive cases. A combination of fluorescent multiplex and Vγ–Jγ-specific monoplex follow-up, as described here, will allow detection of both significant clonal evolution and of the diagnostic clone at a level of prognostic significance, by techniques which can readily be applied to large-scale prospective studies for which real-time analysis is required.
Detection of clonal T cell receptor γ (TCRG) gene rearrangements is widely used in both the diagnostic assessment of lymphoproliferative disorders and the detection of minimal residual disease (MRD) in acute lymphoblastic leukaemia (ALL). While initial analyses were performed by Southern blotting (SB), these are increasingly done by the polymerase chain reaction (PCR) from DNA. PCR detection of clonal TCR or immunoglobulin (Ig) gene rearrangements exploits the fact that polyclonal lymphoid populations generate products which demonstrate a Gaussian distribution as a result of the variable length and nucleotide sequence of the heterogeneous V–(D)–J junctional sequences, whereas a clonal population, with identical V–(D)–J sequences, will generate discrete band(s), depending on the number of rearranged alleles.1 Detection of PCR products is classically performed by non-denaturing polyacrylamide gel (PAGE) analysis, with a sensitivity of detection of approximately 5%.
TCRG represents a preferential target for PCR analyses since it is rearranged at an early stage of T lymphoid development in both TCRαβ and TCRγδ lineage precursors and in the majority of B lineage ALLs. It contains a limited number of Vγ and Jγ segments, thus allowing amplification of all major Vγ–Jγ combinations with a limited number of specific primers. The human TCRG locus contains 14 Vγ segments, only 10 of which have been shown to undergo rearrangement (Figure 1a). The expressed Vγ repertoire includes only six Vγ genes (Vγ2, Vγ3, Vγ4, Vγ5, Vγ8 and Vγ9) but rearrangement also occurs with the Vγψ7, Vγψ10, Vγψ11.2 Vγψ10 and Vγψ11 have recently been shown to be pseudogenes, due to absence of splicing of their leader introns.3 Rearranging Vγ segments can be subdivided into those belonging to the VγI family (Vγ2, Vγ3, Vγ4, Vγ5, Vγψ7 and Vγ8) and the single member Vγ9, Vγψ10, Vγψ11 families. Within Vγfl, the highest levels of homology (95%) exist between Vγ2 and Vγ4 and Vγ3 and Vγ5, reflecting probable gene duplication.4 Of the five Jγ segments (JγP1 (Jγ1.1), JγP [Jγ1.2], Jγ1 [Jγ1.3], JγP2 [Jγ2.1], Jγ2 [Jγ2.3]), Jγ1 and Jγ2 are highly homologous, as are JγP1 and JγP2.5 It is therefore possible to identify all possible Vγ–Jγ combinations using four Vγ and three Jγ primers (Figure 1b and c).
Whilst the restricted germline-encoded repertoire facilitates PCR amplification, the limited junctional diversity of TCRG rearrangements complicates distinction of clonal from polyclonal PCR products. Junctional diversity results from the incorporation of diversity (D) segments and N and P (palindromic) nucleotides and the removal of germline-encoded nucleotides at the extremities of the V, D and J segments. The TCRG locus does not contain D segments and demonstrates relatively limited nucleotide additions, at least in ALL, where the average number of added nucleotides is six.6 TCRG V–J junctional length therefore varies by approximately 20 bp, whereas TCRD and IGH junctional length varies by approximately 60 bp.78 It is therefore extremely important to analyse TCRG PCR products using high resolution electrophoretic techniques. A further possible source of false positivity results from the presence of TCRγδ expressing T lymphocytes demonstrating ‘canonical’ TCRG rearrangements which do not demonstrate N nucleotide additions. The most commonly recognised human canonical TCRG rearrangement involves the Vγ9–JγP segments and occurs in approximately 1% of PBL.910
Several strategies for the detection of minor clonal TCRG populations have been described. The most sensitive and specific involve sequencing of the clonal Vγ–Jγ junction and synthesis of an allele-specific oligonucleotide (ASO) which is then used as a hybridisation probe.61112 The sensitivity of such strategies is approximately 10−5. Their clinical use in ALL follow-up has demonstrated that a significant proportion of patients remain PCR positive during apparent complete remission but that an increased risk of relapse is most marked for patients who demonstrate levels of residual positivity in excess of 10−2–10−3.1112 While these strategies represent optimal specificity and sensitivity they are labour intensive and do not permit detection of clonal evolution. It is therefore possible that simpler strategies which provide a level of detection intermediate to the 5% level of non-denaturing PAGE and ASO methods may be more easily applicable to large-scale clinical follow-up.
Analysis of fluorescent PCR products on automated DNA fragment analysers represents a simple, reliable, non-toxic form of high resolution analysis which is widely used for the analysis of expressed TCRβ repertoires following reverse transcriptase PCR (RT-PCR) with family specific Vβ and Cβ primers and internal, linear fluorescent Jβ run-off reactions.13 It has also been applied to analysis of IGH genomic repertoires, using consensus primers and monofluorescence.8 In this study, we have developed a two-stage multiplex PCR strategy for the rapid detection of TCRG clonality and subsequent multifluorescent identification of Vγ and Jγ segment utilisation. It allows identification of virtually all recognised Vγ–Jγ combinations in two PCR reactions with a maximum of three fluorescent run-off reactions. We have tested the applicability of this approach on a panel of T-ALLs, both with regard to informativity and sensitivity. These strategies will allow improved identification of Vγ and Jγ segment utilisation and detection of minor clonal populations at a level which may have clinical relevance.
Materials and methods
Diagnostic samples from T-ALL blood or bone marrow were obtained with patient consent and separated by density gradient separation using Ficoll (Pharmacia, Sweden) prior to immunophenotype and genotype analysis, either immediately or after cryopreservation in 10% DMSO. Immunophenotyping for surface expression was performed by indirect staining and flow cytometric analysis, using classical techniques and a panel of T and B lymphoid and myeloid antigens.
Ethidium bromide-stained polyacrylamide gel electrophoresis (EB PAGE):
DNA was isolated by phenol–chloroform extraction after digestion with proteinase K. Detection of a clonal TCRG population was undertaken by PCR amplification of TCRG genes from DNA, modified from Ref. 6 to allow multiplex detection of all major Vγ–Jγ combinations in two PCR reactions. Both reaction mixes included three Jγ primers (JγP, JγP1/2, Jγ1/2) capable of recognising all five known Jγ segments.25 Vγ primers included a consensus VγfI primer (Vγ2, Vγ3, Vγ4, Vγ5, Vγψ7 and Vγ8) and specific primers for the Vγ9, Vγψ10 and Vγψ11 segments. All primer sequences are shown in Table 1 and their positions in Figure 1b and c. The mixes were designed to allow partial determination of Vγ usage on the basis of PCR product size. The first mix (VγfI/Vγψ10) detects rearrangements involving the VγI family (PCR product size 230–280 base pairs (bp)) and Vγψ10 TCRG (160–190 bp) gene segments. The second mix (Vγ9/Vγψ11) detects rearrangements involving the Vγ9 (230–280 bp) and Vγψ11 (140–170 bp) segments.
PCR were performed on an automated thermocycler (Biometra, Göttingen, Germany) using 1 μg DNA template, 1 U Taq polymerase (Boerhinger, Mannheim, Germany) and 0.5 μM each primer, 0.2 mM dNTPs, 2.5 mM MgCl2 in a total volume of 50 μl. DNA was denatured at 92°C for 5 min, followed by 35 cycles consisting of denaturation for 30 s, primer annealing for 1 min at 57°C and extension for 1 min 72°C, prolonged to 10 min for the last cycle. For ethidium bromide 29:1 PAGE (EB-PAGE) analysis (Eurobio, Les Ulis, France), 15 μl of PCR product were size-separated on 8% 20-cm non-denaturing gels, run in 0.5× TBE (Tris, Borate, EDTA) and visualized under UV illumination. In order to maximise heteroduplex formation, PCR samples were heated at 99°C for 3 min, at 65°C for 3 min, then cooled to 4°C prior to electrophoresis. A clonal population of T cells corresponded to a discrete band visible within the size range predicted for each set of primers.
High resolution analysis and partial determination of Vγ and Jγ segment usage of PCR products was undertaken by rendering them fluorescent prior to analysis on an automated DNA fragment analyser. Fluorescent primers were provided by PE biosystems. For cases demonstrating approximately 250 bp fragments with the VγI/Vγψ10 reaction mix, PCR products were incubated with a mixture of internal TCRG Vγ primers specific for the Vγ2 and Vγ4 (Vγ2/4 INT-6FAM), Vγ3 and Vγ5 (Vγ3/5 INT-HEX) and Vγ8 (Vγ8 INT-TET) segments labelled with three different fluorophores, as indicated, in a limited cycle, linear amplification of VγfI PCR products, adapted from fluorescent IGH run-off (FluRO) strategies8 (Table 1 and Figure 1b). Identification of Jγ usage for Vγfl/Vγψ10 and Vγ9/Vγψ11 PCR products was undertaken by FluRO analysis with a mix of three internal Jγ primers, Jγ1/2 INT-6FAM, JγP INT-HEX and JγP1/2 INT-TET (Table 1 and Figure 1c). FluRO reactions contained 2.1 mM MgCl2 0.15 mM dNTP, 0.25 μM of each fluorescent primer (0.5 μM JγP-HEX), 0.25 U Taq polymerase and 2 μl PCR product in a total volume of 12 μl and were amplified for seven cycles comprising 1 min denaturation at 95°C, 1 min hybridization at 64°C and 1 min elongation at 72°C. TCRG FluRO products (2 μl) were denatured by heating to 95°C for 2 min in the presence of formamide followed by rapid cooling to 4°C prior to analysis on an ABI373 DNA fragment analyser (PE Biosystems, Roissy, France), using 10% denaturing PAGE, 1 × TBE buffer at 1000 V for 4 h. Results were analysed using Genescan 672 software.
Direct fluorescence analysis was performed by inclusion of Jγ1/2-TET, JγP-HEX and JγP1/2–6FAM (with the same sequence as the external Jγ primers used in EB PAGE; Table 1 and Figure 1c) at a final concentration of 0.25 μM in the initial PCR (35 cycles) and analysis of 1 μl of PCR product on an ABI310 DNA fragment analyser using POP4 polymer (PE Biosystems).
Specific Vγ PCR:
Verification of TCRG VγfI segment utilisation in selected cases was undertaken by specific monoplex PCR using the Vγ specific primers shown in Table 1 and Figure 1b and the reaction conditions indicated above. Specific PCR reactions were analysed by 8% non-denaturing PAGE. For sensitivity testing DNA samples were diluted into PBL polyclonal DNA or a 10% PBL/90% U937 mix prior to amplification.
Cloning and sequencing of TCRG PCR products
PCR products were purified from agarose gels and cloned in pGemT-easy vector (Promega, Madison, WI, USA), as indicated by the manufacturer. Nucleotide sequence was determined by the dideoxynucleotide method using the ABI373 sequencing apparatus with the M13 dye-primer sequencing kit, as indicated by the manufacturer (PE Biosystems).
SIL-TAL1 fusion transcripts were detected by RT-PCR from RNA, as previously described.14
For routine diagnostic detection of clonal populations, we developed two multiplex PCR mixes (VγfI/Vγψ10 and Vγ9/Vγψ11) which allow detection of all Vγ–Jγ rearrangements. All DNA samples were routinely tested with both mixes and analysed on ethidium bromide-stained PAGE (EB PAGE). They were considered positive if one or the other demonstrated at least one discrete band. PCR products with VγfI/Vγψ10 generated approximately 250 bp bands from VγfI–Jγ rearrangements and approximately 170 bp bands from Vγψ10–Jγ rearrangements. Vγ9 PCR products measured approximately 250 bp and Vγψ11 products approximately 150 bp. It was not possible to determine Jγ utilisation on the basis of size. Each PCR experiment included control reactions without DNA, a polyclonal PBL negative control and the PEER T cell line (Vγ8–Jγ2 and Vγ9–Jγ2) clonal positive control at log dilutions into PBL DNA (Figure 2a). Repeated analysis showed the 100% positive control to be strongly positive, the 10% control reproducibly weakly positive, and the 1% control negative or at the limit of detection, thus limiting the sensitivity of detection to approximately 5%.
TCRG fluorescent run-off analysis (FluRO)
Vγ utilisation: Vγ segment identification for VγfI members was performed by fluorescent analysis of appropriately-sized positive VγfI/Vγψ10 PCR products after limited cycle linear run-off amplification (FluRO) with a mixture of internal Vγ2/Vγ4, Vγ3/Vγ5 and Vγ8 fluorescent internal primers, generating fluorescent PCR products of approximately 140 bp (Figure 1b). No attempt was made to distinguish Vγ2 from Vγ4 and Vγ3 from Vγ5 by FluRO analysis, although this was determined by specific PCR in appropriate cases. Two overlapping Gaussian peaks were observed for Vγ2/4, reflecting the fact that the Vγ4 segment is 21 bp longer than Vγ2 (Figures 1b and 3). Jγ utilisation: Jγ utilisation was determined by the incorporation of one of a mix of three internal Jγ-labelled primers on VγfI/Vγψ10 and Vγ9/Vγψ11 PCR, when positive by EB-PAGE. Two Jγ1/2 and JγP1/2 peaks were demonstrated in VγfI/Vγψ10 PBL samples (Figure 3a), the larger corresponding to VγfI members and the smaller to rearrangements involving Vγψ10. Incorporation of JγP into VγfI and Vγψ10 rearrangements was rare. FluRO analysis of Vγ9/Vγψ11 PCR products demonstrated predominantly Vγ9–Jγ1/2 rearrangements and a small number of Vγ9–JγP which varied from virtually absent to the expected restricted number of peaks with a 3 bp size distribution. Vγ9–JγP1/2 and Vγψ11 rearrangements were rare.
The applicability of FluRO TCR Vγ and Jγ determination was tested on 46 T-ALLs, all of which demonstrated cytoplasmic CD3 (cCD3) positivity. EB PAGE analysis (Figure 2) demonstrated at least one clonal band in all but one case. Eight (17%) demonstrated a single band in the VγfI/Vγψ10 PCR and Vγ9/Vγψ11 negativity, compatible with a monoallelic Vγ–Jγ rearrangement. Vγ9/Vγψ11 PCR demonstrated a single band in seven cases (two Vγ9, five Vγψ11), all of which also showed a monoallelic VγfI/Vγψ10 rearrangement, and a bi-allelic Vγ9 and Vγψ11 rearrangement in one case (UPN 219).
Vγ and Jγ FluRO analysis was performed on all VγfI/Vγψ10-positive cases and Jγ on all Vγ9/Vγψ11-positive cases. The Vγ and Jγ primers were specific when used in a multiplex run-off, although appropriate dilution of FluRO products was necessary to avoid ‘cross-fluorescence’ due to over-loading. Vγ and Jγ assignment was clear in the majority of cases, as detailed in Table 2 and as shown for selected cases in Figure 3b. Three cases classified as monoallelic on ethidium bromide analysis were shown to have undergone bi-allelic VγfI–Jγ1/2 rearrangement by FluRO, with the peaks separated by less than 10 bp (Figure 3b, UPN1126). Retrospectively, the presence of heteroduplex bands in two of these cases, with biallelic Vγ4–Jγ1/2 (UPN 1126) and Vγ5–Jγ1/2 and Vγ2–Jγ1/2 (UPN 94) rearrangements, should have been suggestive of bi-allelic rearrangements. In general, heteroduplex formation was observed in the majority of cases having undergone the same VγfI member–Jγ rearrangement on both alleles but was also seen between certain VγfI members. This was neither constant, nor was it obviously restricted to specific pairs of Vγ segments. Fluorescent intensity differed between the two alleles, particularly, but not exclusively, when Vγ or Jγ utilisation differed (see UPN 1026 in Figure 3b). Fluorescent intensity cannot therefore be used in these conditions to accurately assess the proportion of cells with a given rearrangement.
FluRO analysis allowed Jγ determination in 81/82 alleles (99%) from the 46 patients. The single unidentified Jγ allele (UPN 1613) was shown to correspond to a Vγ3/5–Jγ1/2 rearrangement by Vγ FluRO and Jγ-specific PCR, respectively. Sequence analysis of this rearrangement confirmed a Vγ3–Jγ1/2 with no Jγ mismatch to explain the failure of the internal Jγ1/2 primer to recognise the VγfI–Jγ1/2 PCR product (data not shown and see below).
Subtyping of VγfI members allowed detection of all Vγ8 and Vγ3/5 rearrangements. Eleven alleles from nine patients (UPN 54, 94, 1026, 1029, 1126, 1314, 1456, 1613 and 1813) demonstrated a VγfI–Jγ rearrangement using a consensus VγfI primer which was detected by the Jγ1/2 INT–6FAM fluorescent primer, but not by the three Vγ fluorescent primers. Specific PCR analyses demonstrated that this was due to a Vγψ7 rearrangement for two alleles (UPN 1456 and 1813) and failure of the Vγ2/4 INT–6FAM primer to recognise eight Vγ4 rearrangements and one Vγ2 rearrangement. This was likely to be due to the fact that the C nucleotides at positions 2 and, more importantly, 15, of Vγ2/4 INT–6FAM were complementary to Vγ2 but not to Vγ4, when they were represented by a G. Vγ2/4 INT–6FAM was therefore modified by adding one 5′ nucleotide and introducing a C/G degeneracy (Vγ2/4 INT2–6FAM) at position 15, as shown in Table 1. The Vγ2 and 7/8 Vγ4 and rearrangements were detected with Vγ2/4 INT2–6FAM, although in one Vγ4 (UPN 1314), the fluorescent intensity was at the limit of detection. Sequence analysis of this case confirmed a Vγ4–Jγ1/2, with a PCR product size compatible with that observed on Vγ2/4 INT2–6FAM FluRO analysis. No mutation was observed in the sequences recognised by either the Vγ2/4 INT-6FAM or Vγ2/4 INT2–6FAM FluRO primers (data not shown). This case demonstrated a Vγ3/5–Jγ1/2 rearrangement on the other allele, confirmed by sequence analysis.
The only remaining T-ALL was the single case (UPN 1613) in which a Jγ1/2 rearrangement was not recognised by FluRO. This case therefore demonstrated a Vγ3–Jγ1/2 of 247 bp which was recognised by the Vγ3/5 INT-HEX but not by the Jγ1/2INT–6FAM and a Vγ4–Jγ2 of 248 bp recognised by Jγ1/2 INT–6FAM but not by Vγ2/4 INT2–6FAM. Sequence analysis of both alleles failed to identify any mismatches within the target sequences for the FluRO primers (data not shown). We therefore amplified both alleles using Vγ3S and Vγ4S specific primers and either Jγ1/2 or Jγ1/2-TET, which have identical sequences and differe only by the presence of the fluorophore. Primer concentration was identical in all reactions. As shown in Figure 2b, both alleles were amplified with unlabelled Jγ1/2, using either consensus or specific Vγ primers. The Vγ4–Jγ1/2 allele was also amplified, albeit very weakly, with Jγ1/2-TET, whereas this primer failed to amplify the Vγ3–Jγ1/2 allele. We therefore conclude that the presence of the fluorophore in both Jγ1/2 INT–6FAM and Jγ1/2-TET led to failure of amplification in this case.
Analysis of cases detected by the initial Vγ2/4 INT–6FAM fluorescent primer demonstrated that these were predominantly Vγ2, with only 2/10 tested involving Vγ4. Retesting with Vγ2/4 INT2–6FAM demonstrated that both Vγ4–Jγ1/2 and one Vγ2–Jγ1/2 rearrangements were poorly identified with the latter, although they were not completely negative. We therefore now routinely use the Vγ2/4 INT2–6FAM primer but retest with Vγ2/4 INT–6FAM and with a Vγψ7-specific PCR if the Vγ FluRO is negative in the presence of a VγfI-sized fragment and Jγ FluRO positivity.
In total, Vγ identification by evaluation of PCR product size on EB PAGE (Vγ9–11) and Vγ FluRO was successful in typing of 79/82 (96%) Vγ alleles, the three undetected alleles corresponding to two Vγψ7–Jγ1/2 and one Vγ4–Jγ1/2.
Vγ and Jγ utilisation
The majority of rearrangements involved the VγfI (66/82; 80%) and Jγ1/2 (68/82; 83%) segments (Table 2). The overall incidence of Vγ2/4 and Vγ3/5 rearrangement was similar and both were rearranged to Jγ1/2 in all but four alleles (45/49; 92%). Vγ8 rearrangements represented 20% of alleles and a higher proportion (5/16; 31%) involved JγP1/2. Rearrangements involving the Vγψ7 pseudogene were rare and the Vγψ10 and Vγψ11 pseudogenes were found in four and six alleles, respectively. Overall, pseudogenes were involved in 15% of rearrangements. Fifty percent of Vγψ11 rearrangements involved JγP1/2. In total, JγP1/2 rearrangements were found in 16% of alleles and involved Jγ-proximal Vγ segments in the majority (Vγ8–11 in 10/13 alleles). They were over-represented (4/8) in cases with monoallelic TCRG rearrangements. A single, monoallelic, JγP rearrangement, confirmed by specific Vγ3/5–JγP PCR, was observed.
One T-ALL (UPN 1769) demonstrated a normal-sized Vγ8–Jγ1/2 rearrangement and an approximately 70 bp VγfI–Jγ1/2 fragment on EB PAGE, which was shown to be a Vγ2–Jγ1/2 on specific amplification with Vγ2S (Figure 2b). Based on the size, this is likely to correspond to a Vγ2–Jγ1/2 rearrangement which has undergone internal deletion.15
Vγ2–5 utilisation was determined by specific PCR for all cases with available DNA with Vγ2/4 or Vγ3/5 by FluRO. Of 24 Vγ2/4 cases, 11 had undergone Vγ2 rearrangement and 13 Vγ4. Of the 17 Vγ3/5 cases tested, 11 had undergone Vγ3 rearrangement and six Vγ5.
TCR Vγ–Jγ as a function of age and SIL-TAL1 expression (Table 2)
Twenty-eight cases (61%) were aged less than 15 years. The proportion of childhood cases undergoing JγP1/2 rearrangement was higher than in adults, but no other significant differences were observed, including in Vγ2 vs Vγ4 and Vγ3 vs Vγ5 usage.
The SIL-TAL1 fusion transcript is found in approximately 20% of childhood and 10% of adult T-ALL.14 The present series included 12 SIL-TAL1-positive cases, the majority of which were children. No differences in TCRG usage in this group were observed on comparison with the overall group or with the paediatric subgroup (Table 2). It is noteworthy that, while monoallelic TCRG rearrangements were seen in 7/31 (23%) SIL-TAL1-negative cases, all SIL-TAL1-positive cases had undergone bi-allelic rearrangement (Table 2).
Sensitivity (Figure 4)
In order to estimate the sensitivity of detection of the fluorescent strategies, we initially performed dilutions of clonal DNA from the Peer T lymphoid line and one T-ALL (UPN 1029) into polyclonal PBL DNA. Sensitivity of detection of the appropriately sized bands after multiplex PCR and EB PAGE varied from 2.5 to 10%. FluRO analysis with the internal Jγ primers used above led to a minor, two-to four-fold, improvement in sensitivity for two of the four alleles tested. In contrast, Vγ and fluorescent Jγ-specific PCR with direct analysis led to a 10- to 200-fold improvement in sensitivity for three of the four alleles. Direct fluorescent analysis was therefore chosen for further evaluation.
Since competition from polyclonal TCRG rearrangements significantly reduces sensitivity, both due to competition for reagents and loss of the clonal peak within the Gaussian population, dilution into PBL DNA is not representative of bone marrow assessment, when the proportion of polyclonal rearrangements is lower, particularly at early stages following induction chemotherapy. We therefore tested six TCRG rearrangements from three T-ALLs by multiplex and Vγ–Jγ-specific direct fluorescent PCR following dilution into a mixture of 10% PBL in U937 myeloid cell line DNA, to simulate early remission bone marrow conditions. The Vγ-specific (VγS) and labelled external Jγ primers from Table 1 were used. PCR products were analysed both by EB PAGE and fluorescence analysis. As shown in Figure 4, multiplex amplification allowed detection at the 2.5–5 × 10−2 level (mean 3.3 × 10−2) on EB PAGE and predominantly at the 5 × 10−3 level by fluorescence analysis (mean 8.3 × 10−3). Vγ–Jγ-specific PCR was more sensitive but more variable, allowing detection by EB PAGE at the 5 × 10−3 to 5 × 10−2 level (mean 2.3 × 10−2) and detection by fluorescent analysis at the 5 × 10−4 to 5 × 10−3 level (mean 10−3). Variability between alleles for a given T-ALL was up to 1 log and depended predominantly on the position of the clonal peak with respect to the polyclonal peaks. In each of the three cases tested, the peak within the polyclonal size range was detected at 5 × 10−3 and the peak outside this range at 5 × 10−4.
These data demonstrate that improved sensitivity can be obtained with fluorescent and/or Vγ–Jγ-specific strategies and that both are necessary to obtain reproducible detection at the 5 × 10−3 (0.5%) level, demonstrated to be of prognostic significance for early molecular assessment in ALL at achievement of complete remission.1112 It should, however, be emphasised that the Vγ–Jγ specific fluorescent PCR analyses were not optimised to maximise specificity, being used under multiplex PCR conditions. It is therefore possible that their potential sensitivity is higher.
The aim of this study was to use T-ALLs to evaluate the use of a multifluorescent PCR strategy for efficient identification of TCR Vγ and Jγ utilisation in ALL, with a view to improving simplified molecular follow-up. We chose to develop a two-stage strategy based on two multiplex reactions analysed by heteroduplex non-denaturing EB PAGE which are used for all diagnostic indications for TCRG clonality analysis, thus limiting more extensive (and expensive) fluorescent analysis to selected cases. The former have now been used in our laboratory for the analysis of over 1000 diagnostic DNA samples. They provide a reproducible sensitivity of 1–10%, as demonstrated here, and rarely given non-specific bands which could be interpreted as minor clonal populations, at least in younger individuals. Partial Vγ segment identification (VγfI, Vγ9, Vγψ10, Vγψ11) is possible on the basis of PCR product size. Caution is obviously required with regard to canonical Vγ9–JγP rearrangements, but this is rarely a problem once their existence and profile is recognised. A single multiplex TCRG reaction incorporating eight Vγ and three Jγ primers has been described.16 Individual Vγ primers for the VγfI members were used, potentially explaining the high proportion of cases (53%) which demonstrated three Vγ–Jγ primer combinations when the multiplex primer mix was split. In our experience, it is difficult to obtain specific amplification of VγfI members in multiplex conditions, particularly Vγ2 vs Vγ4 and Vγ3 vs Vγ5. A VγfI consensus primer is probably preferable.
To our knowledge, this paper represents the first report of multi-fluorescent TCRG analysis. Monofluorescent analysis of TCRG rearrangements with high resolution analysis has been described but was not adapted for Vγ or Jγ typing.17 By performing up to three multifluorescent run-off reactions (1 Vγ and 2 Jγ) on multiplex PCR, it is possible to identify all Vγ and Jγ segments other than Vγψ12, which is rearranged very infrequently2 and is rarely used in diagnostic PCR strategies. The sensitivity of these strategies was 99% (81/82) for Jγ identification and 96% (79/82) for Vγ. The three undetected Vγ alleles included two Vγψ7 and one Vγ4. The former are to be expected, since Vγψ7 can be detected with the VγfI consensus primer but not with the VγFluRO mix. Their negativity demonstrates the specificity of the Vγ2/4, Vγ3/5 and Vγ8 fluorescent primers. Vγψ7 is used rarely (2/82 alleles in the present series) and can be screened for by specific PCR if desired.
No attempt was made here to distinguish Vγ2 from Vγ4 and Vγ3 from Vγ5 by fluorescent run-off, in view of the >95% homology. Despite this, the initial internal Vγ2/4 INT-6FAM recognised Vγ2 rearrangements preferentially, due to a G–C mismatch with Vγ4 at position 15 of a 24-mer. We encountered an additional problem which appears to be due to a failure of amplification of occasional DNAs in the presence of labelled primers. FluRO of both alleles in 1/46 (2%) T-ALLs failed, when parallel amplification using the same unlabelled primers was successful. The fact that this was observed with two independent multiplex PCR, with both Vγ and Jγ segments on different alleles and with two different fluorophores suggest that it is DNA related. We postulate that DNA from certain patients may be particularly susceptible to inhibition (as in UPN 1613 and, to a lesser extent UPN 1314). It is noteworthy that these two patients had both undergone Vγ3–Jγ1/2 and Vγ4–Jγ1/2 rearrangements. This combination was only seen in one other patient. Recognition of this phenomenon can reduce misinterpretation of results.
Comparable assessment by monoplex PCR would require a total of 21 reactions per patient, although a sequential approach based on relative frequency of Vγ and Jγ utilisation would obviously reduce this number in the majority of cases. A possible alternative to the run-off strategies performed here would be to perform direct fluorescent analysis with similar Vγ and labelled Jγ primer mixes. This would reduce the number of analyses to two initial multiplex PCR with three differently labelled Jγ primers, with identification of Vγfl, Vγ9, Vγψ10 and Vγψ11 on the basis of size and VγfI subtyping as a third PCR with fluorescent Vγ primers, if desired. This approach would be justified for initial characterisation of material from patients destined for future molecular follow-up, such as ALL. It remains to be determined whether initial fluorescent analysis of all diagnostic samples will represent an improvement over EB PAGE.
Several laboratories limit rearrangement analysis to the most frequent combinations, commonly to VγfI–Jγ1/2, VγfI–JγP1/2 and Vγ9–Jγ1/2. This would allow detection of 70/81 (85%) of the alleles detected here. The number of non-informative cases would rise from one to four and those with a mono-allelic rearrangement from eight to 14. While this does not have a major impact on the number of patients being followed by TCRG PCR, it does limit the number of available molecular targets. It is also conceivable that Vγψ10 and Vγψ11 usage will be more frequent in other T cell malignancies.
It should be emphasised that the multifluorescent run-off strategies described here are designed for the characterisation of major clonal rearrangements and not for improved detection of minor populations, particularly with regard to Vγ analysis. The more the TCRG repertoire is subdivided, the more difficult it becomes to separate clonal from oligoclonal samples, as demonstrated for the under-represented Vγ–Jγ combinations, since the Gaussian distribution becomes more irregular.
High resolution fragment separation is extremely important in TCRG analysis since the junctional diversity is relatively limited and the different peaks in a polyclonal population are much more closely packed than for IGH rearrangements. This results from the fact that IGH PCR products demonstrate a 3 bp size distribution, since in-frame rearrangements are selected in virtually all B cells during lymphoid development, whereas TCRG rearrangements only undergo selection of in-frame rearrangements in a minority of circulating T lymphocytes. TCRG PCR products therefore demonstrate a 1 bp size distribution. The increased resolution provided by FluRO analysis allowed identification of unsuspected bi-allelic rearrangements in 3/46 T-ALLs in the present series and increased precision in size identification of clonal populations, useful for follow-up. We did not attempt to exploit precision sizing for the identification, for example, of in-frame rearrangements. This should be possible, if appropriate size markers and electrophoretic conditions are used. Other techniques for obtaining high resolution of TCRG PCR products include denaturing-gradient gel electrophoresis,1819 temperature-gradient gel electrophoresis20 or single-stranded conformational polymorphism.21
Preferential pairing of certain TCRG segments was observed: Vγ2–5 paired almost exclusively with Jγ1/2 (91%) whereas Vγ8 and Vγψ11 paired frequently with JγP1/2. The latter segment is known to rearrange preferentially to Jγ-proximal Vγ segments.2223 JγP1/2 rearrangements were moreover frequent in T-ALLs demonstrating monoallelic rearrangement, suggesting that this segment may decrease the incidence of rearrangement on the second allele, or that these rearrangements occur preferentially in a population of cells destined to shut down TCRG rearrangement capacity at an early stage. As previously described,24 JγP rearrangements were rare, only occurring in one, surface CD3-negative, T-ALL.
Recent studies using ASO-specific probes suggest that detection at complete remission of minor clonal populations at the 10−3 level will identify approximately 10–15% of childhood ALL at a high risk of relapse and that 10−2 may suffice at initial achievement of complete remission (approximately day 30).1112 Our current analyses demonstrate that this level of detection should be possible using optimised direct fluorescent PCR, but that this is likely to be reproducibly achieved only by using Vγ–Jγ-specific strategies. The rapid Vγ and Jγ typing strategies described here should facilitate identification of appropriate follow-up primers. It is, however, noteworthy that multiplex fluorescence analysis, which led to a five- to 10-fold increase in sensitivity compared to EB PAGE analysis, gave a 5 × 10−3 level of sensitivity in 5/6 alleles tested. It should be emphasised that the levels of sensitivity attained here are only possible if the exact size of the diagnostic peaks are known. This approach has the advantage of being universal and of allowing detection of novel or minor rearrangements not detected in the diagnostic sample. Clonal evolution between diagnosis and relapse occurs in approximately 10% of cases on at least one Ig/TCR rearrangement, with the incidence varying for the target gene. Such evolution should be less frequent in early remission samples. However, Evans et al25 recently identified novel IGH rearrangements by monofluorescent genescan analysis in five of nine childhood ALLs demonstrating residual molecular positivity at week 20, including four which subsequently relapsed. We therefore conclude, that while Vγ–Jγ-specific strategies will provide the required sensitivity for detection of high risk patients, the reduced capacity to detect novel rearrangements is a limitation of this, and ASO, approaches. This can obviously be reduced by the use of multiple targets. We propose a combination of follow-up analysis by fluorescent multiplex (sensitivity 10−2–5 × 10−3) for the detection of significant clonal evolution and optimised Vγ–Jγ-specific monoplex for detection of the diagnostic rearrangement(s) at the 5 × 10−4 to 5 × 10−3 level. The simplified approaches described here are obviously not capable of detecting patients in extremely good prognostic groups, as previously identified,12 since a sensitivity of approximately 10−5 is required.
It is to be expected that minor residual leukaemic populations using rare Vγ–Jγ combinations will be detected with greater sensitivity, due to less marked competition from polyclonal rearrangements, as observed for the Vγψ11 rearrangement here. The exclusion of this type of rearrangement in the selective TCR Vγ–Jγ combination strategies mentioned above, may have a practical impact on the proportion of patients showing residual molecular positivity. Other variables affecting sensitivity include the position of the clonal peak relative to the polyclonal peaks, as demonstrated here. While these considerations should be borne in mind when analysing different follow-up series, they will probably represent a secondary effect when large patient series are examined.
Our data also demonstrate that TCRG utilisation is not significantly different in children and adults, suggesting that fluorescent strategies should be applicable with equivalent informativity and success in both clinical situations. This is in keeping with recent data from Ref. 24. Furthermore, TCRG rearrangement patterns were not significantly different in SIL-TAL1-positive and -negative cases. This fusion transcript, or its tald genomic equivalent, represent alternative markers for molecular follow-up. Our data demonstrate that all SIL-TAL1 patients have bi-allelic TCRG rearrangements, suggesting that the tald genomic target will rarely be required because of lack of alternative markers. It is, of course, possible that tald may represent other comparative advantages to TCRG.
Fluorescent IGH strategies have been shown to increase sensitivity by approximately 1 log (Ref. 8) and to allow identification of patients at risk of relapse in B lineage childhood ALL.25 The present study extends application of this type of technology to both T and B lineage ALL and provides a level of sensitivity which has recently been shown to have clinical relevance with techniques which are attainable in a routine molecular diagnostic laboratory with large-scale patient throughput. Their use should allow improved follow-up of patients with ALL or other disorders characterised by TCRG clonality.
d'Auriol L, Macintyre E, Galibert F, Sigaux F . In vitro amplification of T-cell gamma gene rearrangements: a new tool for the assessment of minimal residual disease in acute lymphoblastic leukemias Leukemia 1989 3: 155–158
Chen Z, Font MP, Loiseau P, Bories JC, Degos L, Lefranc MP, Sigaux F . The human T-cell V gamma gene locus: cloning of new segments and study of V gamma rearrangements in neoplastic T and B cells Blood 1988 72: 776–783
Zhang XM, Tonnelle C, Lefranc MP, Huck S . T-cell receptor gamma cDNA in human fetal liver and thymus: variable regions of gamma chains are restricted to V gamma I or V9, due to the absence of splicing of the V10 and V11 leader intron Eur J Immunol 1994 24: 571–578
Font MP, Chen Z, Bories JC, Duparc N, Loiseau P, Degos L, Cann H, Cohen D, Dausset J, Sigaux F . The V gamma locus of the human T-cell receptor gamma gene. Repertoire polymorphism of the first variable gene segment subgroup J Exp Med 1988 168: 1383–1394
Huck S, Lefranc MP . Rearrangements to the JP1, JP and JP2 segments in the human T-cell rearranging gamma gene (TRG gamma) locus FEBS Lett 1987 224: 291–296
Macintyre EA, d'Auriol L, Duparc N, Leverger G, Galibert F, Sigaux F . Use of oligonucleotide probes directed against T-cell antigen receptor gamma delta variable–(diversity)–joining junctional sequences as a general method for detecting minimal residual disease in acute lymphoblastic leukemias J Clin Invest 1990 86: 2125–2135
Macintyre E, d'Auriol L, Amesland F, Loiseau P, Chen Z, Boumsell L, Galibert F, Sigaux F . Analysis of junctional diversity in the preferential V delta 1–J delta 1 rearrangement of fresh T-acute lymphoblastic leukemia cells by in vitro gene amplification and direct sequencing Blood 1989 74: 2053–2061
Landman-Parker J, Aubin J, Delabesse E, Tabone MD, Adam M, Millien C, Leboeuf D, Buzyn-Veil A, Dollfus C, Leverger G, Macintyre EA . Simplified strategies for minimal residual disease detection in B-cell precursor acute lymphoblastic leukaemia Br J Haematol 1996 95: 281–290
Delfau MH, Hance AJ, Lecossier D, Vilmer E, Grandchamp B . Restricted diversity of V gamma 9-JP rearrangements in unstimulated human gamma/delta T lymphocytes Eur J Immunol 1992 22: 2437–2443
Porcelli S, Brenner MB, Band H . Biology of the human gamma delta T-cell receptor Immunol Rev 1991 120: 137–183
Cave H, van der Werff ten Bosch J, Suciu S, Guidal C, Waterkeyn C, Otten J, Bakkus M, Thielemans K, Grandchamp B, Vilmer E . Clinical significance of minimal residual disease in childhood acute lymphoblastic leukemia. European Organization for Research and Treatment of Cancer–Childhood Leukemia Cooperative Group New Engl J Med 1998 339: 591–598
van Dongen JJ, Seriu T, Panzer-Grumayer ER, Biondi A, Pongers-Willemse MJ, Corral L, Stolz F, Schrappe M, Masera G, Kamps WA, Gadner H, van Wering ER, Ludwig WD, Basso G, de Bruijn MA, Cazzaniga G, Hettinger K, van der Does-van den Berg A, Hop WC, Riehm H, Bartram CR . Prognostic value of minimal residual disease in acute lymphoblastic leukaemia in childhood Lancet 1998 352: 1731–1738
Pannetier C, Cochet M, Darche S, Casrouge A, Zoller M, Kourilsky P . The sizes of the CDR3 hypervariable regions of the murine T-cell receptor beta chains vary as a function of the recombined germ-line segments Proc Natl Acad Sci USA 1993 90: 4319–4323
Delabesse E, Bernard M, Landman-Parker J, Davi F, Leboeuf D, Varet B, Valensi F, Macintyre EA . Simultaneous SIL-TALI RT-PCR detection of all tal(d) deletions and identification of novel tal(d) variants Br J Haematol 1997 99: 901–907
Castellanos A, Martin-Seisdedos C, Toribio M, San Miguel J, Gonzalez-Sarmiento R . TCR-gamma gene rearrangement with interstitial deletion within the TRGV2 gene segment is not detected in normal T lymphocytes Leukemia 1998 12: 251–253
Fodinger M, Buchmayer H, Schwarzinger I, Simonitsch I, Winkler K, Jager U, Knobler R, Mannhalter C . Multiplex PCR for rapid detection of T-cell receptor-gamma chain gene rearrangements in patients with lymphoproliferative diseases Br J Haematol 1996 94: 136–139
Simon S, Kind P, Kaudewitz P, Krokowski M, Graf A, Prinz J, Puchta U, Medeiros L, Sandler C . Automated high-resolution polymerase reaction fragment analysis. A method for detecting T-Cell Receptor gamma-chain gene rearrangements in lymphoproliferative diseases Am J Pathol 1998 152: 29–33
Wood GS, Tung RM, Haeffner AC, Crooks CF, Liao S, Orozco R, Veelken H, Kadin ME, Koh H, Heald P, Uluer AZ . Detection of clonal T-cell receptor gamma gene rearrangements in early mycosis fungoides/Sezary syndrome by polymerase chain reaction and denaturing gradient gel electrophoresis (PCR/DGGE) J Invest Dermatol 1994 103: 34–41
Greiner TC, Raffeld M, Lutz C, Dick F, Jaffe ES . Analysis of T-cell receptor-gamma gene rearrangements by denaturing gradient gel electrophoresis of GC-clamped polymerase chain reaction products. Correlation with tumor-specific sequences Am J Pathol 1995 146: 46–55
Menke MA, Tiemann M, Vogelsang D, Boie C, Parwaresch R . Temperature gradient gel electrophoresis for analysis of a polymerase chain reaction-based diagnostic clonality assay in the early stages of cutaneous T-cell lymphomas Electrophoresis 1995 16: 733–738
Baruchel A, Cayuela JM, MacIntyre E, Berger R, Sigaux F . Assessment of clonal evolution at Ig/TCR loci in acute lymphoblastic leukaemia by single-strand conformation polymorphism studies and highly resolutive PCR derived methods: implication for a general strategy of minimal residual disease detection Br J Haematol 1995 90: 85–93
Krangel MS, Yssel H, Brocklehurst C, Spits H . A distinct wave of human T-cell receptor gamma/delta lymphocytes in the early fetal thymus: evidence for controlled gene rearrangement and cytokine production J Exp Med 1990 172: 847–859
MacVay L, Carding S, Bottomly K, Hayday A . Regulated expression and structure of T-cell receptor gamma-delta transcripts in human thymic ontogeny EMBO J 1991 10: 83–91
Szczepanski T, Langerak A, Wolvers-Tetteroo I, Ossenkoppele G, Verhoef G, Stul M, Petersen E, de Bruijn M, van't Veer M, van Dongen J . Immunoglobulin and T-cell receptor gene rearrangement patterns in acute lymphoblastic leukemia are less mature than in children: implications for selection of PCR targets for selection of minimal residual disease Leukemia 1998 12: 1081–1088
Evans PA, Short MA, Owen RG, Jack AS, Forsyth PD, Shiach CR, Kinsey S, Morgan GJ . Residual disease detection using fluorescent polymerase chain reaction at 20 weeks of therapy predicts clinical outcome in childhood acute lymphoblastic leukemia J Clin Oncol 1998 16: 3616–3627
We thank P Cartier, J Landman-Parker, M-H Estienne, L Croisille, C Bayle, X Troussard and F Picard for providing T-ALL samples. This work was supported by the Fondation de France, the League Contre le Cancer and the Direction de la Recherche Clinique de l'Assistance Publique - Hopitaux de Paris.
About this article
Cite this article
Delabesse, E., Burtin, M., Millien, C. et al. Rapid, multifluorescent TCRG Vγ and Jγ typing: application to T cell acute lymphoblastic leukemia and to the detection of minor clonal populations. Leukemia 14, 1143–1152 (2000) doi:10.1038/sj.leu.2401750
- TCR gamma
- PCR diagnosis
- fluorescent PCR
Evaluation of next‐generation sequencing‐based clonality analysis of T‐cell receptor gamma gene rearrangements based on a new interpretation algorithm
International Journal of Laboratory Hematology (2019)
A Single-Tube, EuroClonality-Inspired, TRG Clonality Multiplex PCR Aids Management of Patients with Enteropathic Diseases, including from Formaldehyde-Fixed, Paraffin-Embedded Tissues
The Journal of Molecular Diagnostics (2019)
Anaplastic large cell lymphoma arises in thymocytes and requires transient TCR expression for thymic egress
Nature Communications (2016)
Accurate detection of the tumor clone in peripheral T-cell lymphoma biopsies by flow cytometric analysis of TCR-Vβ repertoire
Modern Pathology (2012)
EMC - Hématologie (2012)