Detection of minimal residual disease (MRD) has proven to provide independent prognostic information for treatment stratification in several types of leukemias such as childhood acute lymphoblastic leukemia (ALL), chronic myeloid leukemia (CML) and acute promyelocytc leukemia. This report focuses on the accurate quantitative measurement of fusion gene (FG) transcripts as can be applied in 35–45% of ALL and acute myeloid leukemia, and in more than 90% of CML. A total of 26 European university laboratories from 10 countries have collaborated to establish a standardized protocol for TaqMan-based real-time quantitative PCR (RQ-PCR) analysis of the main leukemia-associated FGs within the Europe Against Cancer (EAC) program. Four phases were scheduled: (1) training, (2) optimization, (3) sensitivity testing and (4) patient sample testing. During our program, three quality control rounds on a large series of coded RNA samples were performed including a balanced randomized assay, which enabled final validation of the EAC primer and probe sets. The expression level of the nine major FG transcripts in a large series of stored diagnostic leukemia samples (n=278) was evaluated. After normalization, no statistically significant difference in expression level was observed between bone marrow and peripheral blood on paired samples at diagnosis. However, RQ-PCR revealed marked differences in FG expression between transcripts in leukemic samples at diagnosis that could account for differential assay sensitivity. The development of standardized protocols for RQ-PCR analysis of FG transcripts provides a milestone for molecular determination of MRD levels. This is likely to prove invaluable to the management of patients entered into multicenter therapeutic trials.
Current treatment protocols for acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML) and chronic myeloid leukemia (CML) are based on prognostic factors, which contribute to therapy stratification.1,2,3 Key prognostic factors identified in leukemia over the years include pretreatment characteristics such as age, WBC count, immunophenotypic profiles, specific chromosomal abnormalities,4,5 aberrant fusion genes (FGs), and mutations, such as FLT3 gene alterations in AML.6,7 Response to initial therapy provides a further well-known prognostic marker, in particular the presence or absence of blasts in the bone marrow after induction therapy in ALL and AML,8,9,10,11,12 or cytogenetic response in CML patients.13,14
However, it is important to stress that patient outcome cannot be reliably predicted on the basis of such classical parameters, thereby underlining the potential importance of minimal residual disease (MRD) testing. Over the last 10 years, technological developments have enabled the detection of leukemic cells beyond the threshold of cytomorphology or karyotyping.15,16 Three different techniques are currently used with a sensitivity of at least one leukemic cell in a background of 103 normal cells: (1) immunophenotyping; (2) polymerase chain reaction (PCR) using genomic DNA for detection of clonal rearrangements of immunoglobulin (Ig) and T-cell receptor (TCR) genes in ALL; and (3) reverse-transcriptase-PCR (RT-PCR) for detection of gene rearrangements, mainly FG transcripts resulting from chromosomal translocations. While the first two directly measure the tumor load, the latter method measures gene expression.
MRD information in leukemia patients has been clearly established as an independent prognostic factor in three pathological situations: (1) childhood ALL, in particular after induction therapy;17,18,19 (2) BCR-ABL detection in CML patients after allogeneic stem cell transplantation,20 enabling direction of donor leukocyte infusions; and (3) PML-RARA detection in acute promyelocytic leukemia (APL) patients after consolidation therapy,21 confering benefit for pre-emptive therapy at the point of molecular relapse in comparison to frank relapse.22 These data have led to the introduction of molecular monitoring in the stratification strategy in some current multicenter therapeutic trials. However, the clinical impact of MRD detection in other types of leukemias remains to be demonstrated in large series of patients.
In order to tackle the problem of lack of standardized diagnostic methodology, 8 years ago European laboratories conducted a collaborative program through a BIOMED-1 Concerted Action.23 This Concerted Action led to the development of standardized nested RT-PCR assays achieving sensitivities of at least 10−4 (RNA diluted into RNA) suitable for detection of MRD as well as diagnostic screening.24 However, ‘end point’ PCR analyses do not permit precise quantification of MRD levels. This limitation was underlined by the finding of low levels of AML1-ETO,25,26 CBFB-MYH111 or PML-RARA1,27 transcripts in patients in long-term clinical remission and by the detection of BCR-ABL mRNA at very low levels in healthy individuals.28,29 Quantitative PCR analysis has been achieved by competitive PCR.30 Expert laboratories showed that this technique enables accurate prediction of relapse suggesting that such analysis could be used for adapting treatment in BCR-ABL-positive CML31,32 or in AML patients with CBFB-MYH11 or AML1-ETO transcripts.33,34 However, this competitive PCR is labor intensive and time consuming which prohibits both standardization and large-scale multicenter analysis.
More recently, real-time quantitative PCR (RQ-PCR) has been introduced.35,36 There have been numerous manuscripts from individual laboratories demonstrating the reliability of this technology and its potential clinical value for MRD studies using FG transcripts as PCR targets such as BCR-ABL in CML37,38,39 or ALL,40 PML-RARA,41 AML1-ETO42,43,44 and CBFB-MYH1145,46 in AML and TEL-AML147,48,49 in ALL patients. However, standardized RQ-PCR procedures are warranted in order to apply this innovative technology for large-scale MRD studies within multicenter therapeutic trials.
In 1999, 25 university laboratories from 10 European countries designed a joint project of the health and consumer protection of the European Commission (SANCO) via the ‘Europe Against Cancer’ (EAC) program in order to develop standardization and quality control analysis for RQ-PCR, based on the ABI 7700 platform (Applied Biosystems, Foster City, USA) as it was the first such technology available. The major aim was to establish a standardized protocol allowing comparison of MRD data in order to assess the relative efficiency of each therapeutic strategy for leukemia bearing an appropriate molecular marker. The most frequently occurring FG transcripts in leukemia were selected, covering up to 30–40% of childhood and adult ALL and AML and more than 95% of CML patients.
A total of 25 laboratories were involved throughout the whole program. One laboratory (#28) joined the group for the latest phase. This EAC Concerted Action was supported by Applied Biosystems for the design and synthesis of most TaqMan probes and primers. For the extensive statistical analysis, the Concerted Action obtained support from the Department of Medical Information of the Université de la Méditerranée (Marseille, France).
The participating laboratories were divided into seven national networks (Austria–Germany, Belgium–The Netherlands, Denmark–Sweden, France, UK, Italy and Spain). For optimal efficiency, one control gene (CG) network (six laboratories) and nine FG networks (with 4–12 laboratories) were created, responsible for testing one particular target (Table 1). Each laboratory was involved in up to four gene networks.
Our program was divided into four main phases (I–IV) each completed by meetings to monitor progress made and to plan the experiments for the next phase. Common 96-well RQ-PCR reaction plate ‘set-up’ and presentation forms were designed in order to streamline the data analysis and the reporting of results. Work tasks were divided as follows (Table 2):
The aim of phases I and II was the initial selection of primers and probes in addition to training of the members since the majority only started to use this RQ-PCR methodology through involvement in this consortium.
During phase IIIa, different standard curves were compared (ie generated using RNA, cDNA or DNA plasmids) and the final validation of selected primer/probe sets within specific networks was made. In parallel, we performed our first quality control round (QC1).
During phase IIIb, testing of selected FG primer/probe sets was undertaken by all laboratories on centrally prepared coded quality control samples (QC2). This testing of the FG transcript targets was performed by randomly selected laboratories outside the original FG transcript networks.
During phase IVa, we performed the third quality control round (QC3) including undiluted cell line RNA samples.
During phase IVb, we determined the reference values of normalized FG transcript levels in leukemic cell lines and patient samples using the EAC primer/probe sets according to the EAC standardized protocol.
In addition, reference ranges were established for the CGs in normal peripheral blood (PB), bone marrow (BM) and PB stem cells in fresh samples (see accompanying manuscript by Beillard et al50).
Materials and methods
Currently available RQ-PCR technologies allow detection of fluorescence emission during the PCR reaction from one (TaqMan) or two (Light Cycler) internal oligonucleotide probes, or a fluorescent dye, the detected fluorescence being proportional to the amount of target present in the sample.51,52
We decided to run our EAC protocol using the ABI 7700 platform with TaqMan probes since this was the first robust RQ-PCR technology available permitting analysis of a large number of samples in a single run (96-well plate format). The 5′ nuclease assay (TaqMan technology) uses a single internal oligonucleotide probe bearing a 5′ reporter fluorophore (eg 6-carboxy-fluorescein) and 3′ quencher fluorophore (eg 6-carboxy-tetramethyl-rhodamine). During the extension phase, the TaqMan probe is hydrolyzed by the nuclease activity of the Taq polymerase, resulting in separation of the reporter and quencher fluorochromes and consequently in an increase in fluorescence (Figure 1). In the TaqMan technology, the number of PCR cycles necessary to detect a signal above the threshold is called the cycle threshold (Ct) and is directly proportional to the amount of target present at the beginning of the reaction (Figure 2a). The ΔRn corresponds to the increase in fluorescence intensity when the plateau phase is reached. Using standards or calibrators with a known number of molecules, one can establish a standard curve and determine the precise amount of target present in the test sample (Figure 2b). The theoretical slope of the standard curve is −3.32 for a PCR reaction with maximum efficiency. Using a known amount of DNA molecules for the standard curve, the intercept point on the Y-axis defines the number of cycles theoretically needed to detect one molecule.
Assays were designed to detect nine leukemia-associated FGs, including their more common breakpoint variants giving rise to 15 RNA targets: E2A-PBX1, MLL-AF4 (variants exon 9-exon 5, exon 10-exon 4 and exon 11-exon 5), TEL-AML1, BCR-ABL (M-bcr and m-bcr), SIL-TAL1, PML-RARA (bcr1, bcr2 and bcr3), CBFB-MYH11 (type A, D and E) and AML1-ETO.
In addition, 14 housekeeping genes were evaluated for their suitability to serve as CGs for sample to sample quality variations and gene expression quantification (see accompanying manuscript by Beillard et al50). Three CGs were ultimately selected (ABL (Abelson gene), B2M (beta-2-microglobulin) and GUS (beta-glucuronidase)); we analyzed the expression of the ABL gene during phases II–III and all three selected CGs (ABL, B2M and GUS) during phase IV.
Primer and probe design
Primers and probes were designed using Primer Express software (Applied Biosystems) based on their location on two separated exons and on the sequence of the amplicon generated by the primer sets described in the BIOMED-1 program.24 They are depicted in the schematic diagram of the exon/intron structure of the corresponding FG (see Sections 1–9). During the first two meetings, an initial selection of primers and probes was made; newly designed sets for each molecular target as well as already available ‘in house’ sets from experienced laboratories were evaluated. The set selection was based on: (1) the absence of nonspecific amplification artifacts; (2) a good efficiency with a slope close to −3.32 (100% theoretical efficiency); (3) a good sensitivity (at least 10−4 RNA dilution or 100 copies for plasmid dilution); and (4) the robustness of the reaction with a ΔRn value >1.0 at the plateau phase for the highest dilutions (Table 3). Such results had to be reached in at least 80% of the participating laboratories for a particular primer/probe set to be selected. Potential primer/probe sets were tested in parallel on serial dilutions of cell lines and plasmids, and the set with the best performance profile, particularly in terms of sensitivity, was selected. If none of the primer/probe sets satisfied the selection criteria, new sets were designed and evaluated. Overall, starting from 47 primer sets and 44 probes that were tested during the first phases, 12 primer sets and nine probes were finally selected for the 15 targets (see figures in individual sections). In each case, the sensitivity of the TaqMan-based RQ-PCR analysis appeared to be comparable to previously standardized nested RT-PCR analysis.24 On the BCR gene, Y (cytidine or thymidine) appears on the BCR primer at position 3188, according to the polymorphism recently described.53
RNA and cDNA from cell lines and leukemia samples at diagnosis
RNA and/or cDNA samples were prepared centrally by the FG network leaders (Table 1) and distributed on dry ice to members of their respective networks during phases I–IVa. FG transcript-positive cell line RNA was commonly used (see Table 1 for specific cell lines), except for rare FG transcripts (PML-RARA bcr2 and bcr3 and CBFB-MYH11 type D and E) for which patient RNA was provided by the network leader. These RNA samples were diluted in PB lymphocytes (PBL) RNA or FG transcript-negative cell line RNA. Cell lines were purchased from the DSMZ (Braunschweig, Germany), ATCC (Manassas, VA, USA) or directly provided by academical laboratories (TOM-1, ME-1 and PF382) and cultured according to the supplier's instructions. During phase III, network leader laboratories prepared equivalent dilution series of cDNA. During phase IVb, patient sample RNAs, positive for the relevant FG transcript and that had been stored for less than 18 months, were analyzed locally undiluted and/or diluted in a solution of 1 μg/μl E. coli 16S and 23S rRNA (Roche, Meylan, France) in duplicate. Overall, we analyzed 278 samples (BM and PB) mainly from different patients and 57 paired samples from a harvest of the same patient at the same time. Details are given in each FG section and corresponding tables. Although BM/PB data have some redundancy, analyses of BM and PB samples give an impression about the comparability of the results, particularly when paired BM/PB samples were used.
Plasmid DNA calibrators containing the target gene sequences
PCR products of the 15 different FG transcripts were generated from cell line or patient RNA by RT-PCR using BIOMED-1 A and B primers, as previously described.24 PCR products were cloned into the PCR II TOPO vector (Invitrogen, Groningen, The Netherlands). The selected plasmid clones were sequenced for confirmation of their insert (Genome Express, Grenoble, France). After subsequent bulk production, the plasmids were extracted using the QIAFILTER Plasmid MIDI kit (Qiagen, Courtaboeuf, France) and quantified spectrophotometrically. The copy number for 1 μg was estimated according to the molecular weight of the vector and the insert. Then 20 μg of plasmid was linearized with BamHI or HindIII restriction enzymes for 1 h at 37°C under agitation. The digested plasmid was serially diluted in a solution of Tris 10 mM, EDTA 1 mM pH 8, containing 20 ng/μl of E. coli 16S and 23S rRNA (Roche). Five successive dilutions (200 000, 20 000, 200, 20 and 2 copies/μl) were prepared. The corresponding standard curve generated a mean slope of −3.3 and an intercept of 39.8±1 Ct. A mean Ct value of 22.5±1 was obtained for the 20 000 copies/μl dilution. The plasmid dilutions were centrally prepared in Marseille (J Gabert's laboratory) during phases I–III. Thereafter, Ipsogen (Marseille, France) kindly provided plasmid dilutions for phase IV.
Standardized RT-PCR protocol
A common EAC protocol was established for all molecular targets (FG and CG transcripts) during phases I and II and then used by each laboratory throughout phases III and IV (Table 4).
This reaction was adapted from the BIOMED-1 protocol.24 Starting from 1 μg of total RNA, the main modifications involved alteration of the concentrations of random hexamers (25 μ M) and of reverse transcriptase (100 U) either murine Moloney leukaemia virus or Superscript (Invitrogen, Roche), which significantly enhanced the sensitivity of the assay (see accompanying manuscript by Beillard et al50).
All the RQ-PCR reactions were performed on a 7700 ABI platform (Applied Biosystems, Foster City, USA) using primers and TaqMan probes kindly provided by Applied Biosystems in conjunction with the TaqMan Universal Master Mix purchased from the same manufacturer. The number of amplification cycles was 50 (detailed protocol is described in Table 4).
Optimization of RQ-PCR assay (phase II)
In order to optimize the results and save costs, phase II involved assessment of the influence of primer and probe concentrations (900 and 300 nM for primers and 200 and 100 nM for probes) and the reaction volume (50 vs 25 μl) on the sensitivity of the assay.
After extensive testing within FG networks using serial dilutions of FG-positive cell line RNA in FG-negative RNA (10−1–10−5) and plasmid (106–10 copies), optimal results were obtained using a 25 μl volume, with 300 nM of primers and a 200 nM concentration of probe, except for AML1-ETO for which a 100 nM probe was chosen (see Section 9).
For comparative data analysis, a common threshold set at 0.1 was selected in order to be in the exponential phase. Such a threshold value typically lies above the so-called ‘creeping curves’ that were rarely observed in some negative controls; ‘creeping curves’ are defined as the amplification curve from a negative sample rising slowly during the PCR reaction. While the mechanism underlying the latter phenomenon is not entirely clear, examination of the ‘multicomponent’ view reveals that it is not indicative of specific amplification (see below). However, for PML-RARA, the threshold was fixed at 0.05 because of the relatively short exponential phase of the PCR amplification (low ΔRn). The baseline was calculated from cycles 3–15 except for high-expression CGs where cycles 3–10 were used (see accompanying manuscript by Beillard et al50). Additionally, we used dilutions of positive RNA (patient or cell line) into normal PBL or an FG-negative cell line RNA (10−1–10−6) during phases I–IIIa and identical dilution series using cDNA during phase IIIa.
Standard curve comparison (phase IIIa)
A comparison between RNA, cDNA and plasmid standard curves was performed in order to determine the influence of the RT step on the results (standard curve slope, sensitivity and reproducibility) for each laboratory. The FG network leaders sent to the different laboratories within their target network (4–12 members, see Table 1) centrally prepared serial dilutions of the FG transcript-positive cell line RNA and the corresponding cDNAs, in addition to serial dilutions of the corresponding FG control plasmid (200 000, 20 000, 200, 20 and 2 copies/μl).
Quality control (QC) rounds on coded samples
QC1 (phase IIIa) and QC3 (phase IVa) were performed by laboratories within their original FG networks (Table 1), whereas QC2 (phase IIIb) involved randomly chosen laboratories (see Balanced randomized assay section).
Control samples and definition of (false-) posivitity and (false-) negativity
The positive controls in all experiments and QC rounds concerned well-defined cell lines and patient samples (Tables 1 and 2). Two types of negative controls were used: (1) coded FG-negative RNA samples and (2) known negative controls for checking contamination of PCR products. These latter contamination controls concerned no-amplification controls (NAC), which contained E. coli RNA instead of human cDNA, and no-template controls (NTC), which contained water instead of human cDNA. Particularly, the NAC and NTC negative controls were regarded to be of utmost importance for identification of cross-contamination of PCR products, because this problem is frequently underestimated.
A positive well was defined as a sigmoidic amplification (log scale) with a Ct value below the Y-intercept Ct value of the plasmid standard curve+one Ct. Amplification on RNA samples of the FG was performed in triplicate and in duplicate for the CG expression. A false-negative sample was defined as a positive RNA sample with less than 50% of positive wells (0/2, 0/3 or 1/3). A false-positive result was defined as a negative sample, with at least 50% of positive wells (1/2, 2/3 or 3/3).
Sensitivity and reproducibility of the experiments
The criteria for sensitivity and reproducibility were defined during QC experiments via coded samples for each FG transcript (Table 3). An experiment was assumed to be reproducible for a particular dilution if more than 80% of the laboratories detected at least two positive wells with a Ct difference less than 1.5 Ct. The sensitivity of the experiment was defined as the last dilution showing at least 50% of positive wells in more than 80% of the laboratories whatever the Ct values were.
Balanced randomized assay (phase IIIb)
This assay was designed by the Department of Medical Information (Marseille, France). The aim was to evaluate to what extent RQ-PCR results were comparable between laboratories for MRD quantification of FG transcripts in a clinical setting according to the EAC protocol. The study focused on two main points: (1) comparison of the results between laboratories for a particular FG transcript which is crucial for multicenter studies and (2) the linearity of the RQ-PCR methodology in dilution experiments which is important to assess the potential tumor load during treatment. The balanced randomized assay appeared to be an appropriate statistical study to assess these two points without performing all the RQ-PCR analyses (n=15) in each participating laboratory (n=25).
The statisticians randomly assigned the 25 participating laboratories to nine networks. Each of the nine main FG transcripts were tested in 11 laboratories (except for CBFB-MYH11 network, n=12), including the involved FG transcript network leader, making a total of 100 RQ-PCR experiments. Each laboratory tested five coded samples for four different targets, making a total of 500 coded samples analyzed in this QC2 study (Table 2). FG transcript-positive control RNAs were diluted (10−1, 10−3 and 10−4 dilutions) in a negative RNA sample (HL60 or PBL).
A common plate design was used: the CG (ABL), the FG transcript and plasmid dilutions (ABL n=3, FG n=5) were amplified in triplicate, whereas four negative controls (NAC and NTC for ABL and FG targets) were run in duplicate. The raw data were collected and analyzed by each FG network leader. Common Excel worksheets were designed to collect the results within each FG network and were then forwarded to Marseille for subsequent statistical analysis (see below). The only exclusion criterion for coded RNA samples was an ABL Ct value outside the normal range [22–29.3] as defined by assays conducted within the CG network during phase IIIa (see accompanying manuscript by Beillard et al50).
The CG and the FG RQ-PCR data were collected and analyzed by each FG network leader. The mean Ct or mean ΔCt (mean Ct [FG]−mean Ct [CG]) values and the mean value of the log10 of the copy number (CN) for each gene were used. The normalized copy number (NCN) was defined as the CN of the FG per one copy of the CG transcript (mean value of log10[FG CN]−mean value of log10[CG CN]) except for normalization to B2M gene transcript level for which the results were expressed per 100 copies. Since CN did not show a normal distribution, the logarithmic value of the CN was used for statistical analysis. For an easier comprehension, results obtained in a logarithmic way were subsequently converted into decimal values. The level of significance was set at P<0.05. In tables and figures, when the P-value was between 0.05 and 0.1, even if not statistically significant, the numbers are noted. When the P-value was >0.1, the result appears as not significant (NS). The calculation of CV from the Ct's results in an underestimation of the real variation, since Ct's are based on log of the transcript numbers. The correlations between the expression level of different genes were measured by the Pearson correlation coefficient (r). The box-plots used for presenting data show the median value (dark line) within a box containing 50% of the samples (25–75th percentile). The statistical analysis was performed in Marseille, France, using the SPSS 10.1 Software (SPSS Inc., Chicago, USA).
Balanced randomized assay (phase IIIb)
Detection of significant differences in transcript quantification between laboratories: Four parameters per sample (Ct, ΔCt, CN and NCN) were tested using a global linear model. The laboratory number was set as a random effect and data were analyzed. When the results were not comparable between laboratories (P<0.05) for the selected parameter and transcript, a post hoc analysis (Tukey method) was used to evaluate the number of laboratories reproducible with others (P⩾0.05) defining a subgroup. For this, two criteria were chosen to estimate the reproducibility of the results: the number of subgroups (Sn) and the number of laboratories (Ln) reproducible with at least two other laboratories as defined by a nonsignificant difference between the mean (P⩾0.05) according to the Tukey method. When a particular laboratory was present in two or more subgroups, it was counted only once. The best parameter to compare results between laboratories was the one with the highest Ln value and the lowest Sn value. Ideally, only one subgroup (Sn=1) containing all the laboratories (Ln=11 or 12 per network or Ln=25 for the whole assay) was expected.
Evaluation of the linearity : The Pearson correlation coefficient was chosen to measure the linearity of the quantification of the FG transcripts using the three coded positive samples. For each FG transcript, the best parameter (ΔΔCt or NCN) to assess the results was the one with the highest correlation coefficient.
Reference values at diagnosis (phase IVb)
Leukemia samples were excluded from the analysis if the CG amplification was not within the normal range for fresh samples defined as follows: ABL Ct [21.8–29.4], B2M Ct [15.6–24.9] and GUS Ct [20.8–28.0] (see accompanying manuscript by Beillard et al50). The comparison between PB and BM was performed with nonparametric tests (Wilcoxon paired test for at least five paired samples or Mann–Whitney U-test for unpaired samples). At least five paired samples were expected per FG transcripts. The 95% range of expression for NCN refers to the range between the 3rd and the 97th percentile for the selected gene. Correlation coefficients between the FG CN and each CG CN are given before normalization by the CG. For cell line(s), the median value obtained on the same sample in eight different laboratories is shown.
All the figures, tables and raw data are available on-line at the following address: http://meidia.nord.univ-mrs.fr/EAC/publications.html. This web site will stay available for a long-term period, at least 5 years.
1. t(1;19)(q23;p13) with the E2A-PBX1 fusion gene transcript
JM Cayuela1, X Fund1, H Cavé2, G Brunie2, E Beillard3, C Glowackzower3, J Gabert3, VHJ van der Velden4, JM Wijkhuijs4, JJM van Dongen4, M Malec5, A Porwit-MacDonald5, F Watzinger6, R Baumgartinger6, T Lion6, O Spinelli7, A Rambaldi7, P Heimann8, H El Housni8 and F Sigaux1
1Hôpital Saint Louis, Paris, France; 2Hôpital Robert Debré, Paris, France; 3Institut Paoli Calmettes and present affiliation: Hôpital Universitaire Nord, Marseille, France; 4Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands; 5Karolinska Hospital, Stockholm, Sweden; 6Children Cancer Institut, Vienna, Austria; 7Riuniti Hospital, Bergamo, Italy; and 8Erasme Campus, Brussels, Belgium
The leukemogenic FG transcript E2A-PBX1 results from fusion of the E2A and PBX1 (formerly prl) genes, through the t(1;19)(q23;p13).54,55,56 The t(1;19)(q23;p13) is found in 3–5% of childhood and in 3% of adult precursor-B-ALL.57 In 95% of the cases, E2A-PBX1 transcripts are expressed.58 This expression is tightly associated with detection of cytoplasmic Ig μ chains.59,60 The remaining t(1;19) precursor-B-ALL are E2A-PBX1 negative and show no rearrangement of the E2A gene.61
The E2A gene, located on chromosome 19, encodes the helix–loop–helix Ig enhancer binding factors E12 and E47 and the PBX1 gene on chromosome 1 encodes a DNA-binding homeobox protein.55,56 The genomic organization of E2A is well defined and breakpoints occur almost exclusively in the 3.5 kb intron between exons 13 and 14.54,55,56 The genomic organization of PBX1 is not yet fully known and the breakpoints are dispersed over an intronic region of about 50 kb between exons 1 and 2. The majority of cases with E2A-PBX1 FG transcripts show a constant junction of E2A exon 13 to PBX1 exon 2 (Figure 3).62 A variant FG transcript has been described in about 5–10% of E2A-PBX1-positive patients. It is characterized by insertion of a stretch of 27 nucleotides at the E2A-PBX1 junction.61,63 This inserted sequence probably arises from an alternatively spliced exon of PBX1.
The FG encodes a chimeric transcriptional activator containing the N-terminal transcriptional activator domain of E2A joined to the C-terminal DNA-binding homeobox domain of PBX1.64 The transforming activity of E2A-PBX1 proteins has been demonstrated both in vitro and in vivo.65,66,67
Several studies using RT-PCR amplification of E2A-PBX1 FG transcripts to assess MRD have been reported.68,69,70,71,72 All of them were performed with a qualitative assay showing a detection threshold of up to 10−4/−5. In none of these reports does the presence or absence of E2A-PBX1 transcripts during follow-up predict treatment outcome. The largest series including 71 patients, published by Hunger et al,70 found no difference in event-free survival between PCR-positive and PCR-negative patients analyzed at the end of consolidation treatment. All these studies pinpoint to the limitations of qualitative assessment of MRD in monitoring t(1;19)-positive ALL patients, and underline the importance of quantitative approaches such as RQ-PCR methods.
1.2. EAC data
1.2.1. Primer design and optimization (phases I and II)
Among two primer and probe sets tested, one was chosen: ENF101, ENR161 and ENP141. Positions and nucleotide sequences are shown in Figure 3 and in Table 5, respectively. This set was selected on the basis of higher ΔRn values at the plateau phase with high dilutions of the E2A-PBX1-positive cell line 697 (also known as ACC42) RNA in HL60 RNA. During phase II and phase III experiments, this set of primers and probe allowed us to detect a 10−4 dilution of 697 RNA in HL60 RNA in seven out of seven laboratories; an example of typical amplification plots (10−1, 10−3 and 10−4 cell line RNA dilutions) is shown in Figure 4. We conclude that ENF101, ENR161 and ENP141 are suitable for quantification of all E2A-PBX1 fusion transcripts, including the variant form.
1.2.2. E2A-PBX1 expression in 697 cell line and diagnostic patient samples (phase IV)
One cell line, 697 and 27 diagnostic samples (14 BM and 13 PB samples) were analyzed. Blast percentages (defined morphologically) were available in all but two samples (one PB and one BM) and ranged from 74 to 100% (median=96%). Median values and 95% range for CGs and E2A-PBX1 Ct values, as well as normalized E2A-PBX1 copy number (corrected according to the blast percentage), are reported in Table 6. Ct values detected for E2A-PBX1 and CG in the 697 cell line and patient samples were comparable indicating that they are expressed at similar levels. The highest correlation coefficient was observed between E2A-PBX1 and ABL transcripts. Among PB and BM samples, 10 were paired samples, harvested in the same patient at presentation of the disease. No statistically significant difference could be observed between BM and PB, in terms of Ct or NCN in paired samples (Figure 5).
1.2.3. QC rounds (phases IIIa–IVa)
During the various QC rounds, 11 negative samples (five negative RNA, three NAC and three NTC) and five positive samples (10−3 (two samples) and 10−4 (three samples) dilutions of 697 cell line RNA in HL60 RNA) were tested in 8–10 labs (Table 7). E2A-PBX1 amplification of the negative samples accounts for 156 wells. Three wells (corresponding to three different samples) were found positive (2%), but none of these samples were considered positive according to the criteria defined in the Materials and methods section. E2A-PBX1 amplification of the positive samples accounts for 78 wells. Only one well was found negative (1%). According to the criteria defined in the Materials and methods section, all five positive samples were found positive, as expected from the sensitivity threshold defined in the previous phases.
2. t(4;11)(q21;q23) with the MLL-AF4 fusion gene transcript
E Beillard1, C Glowaczower1, D De Micheli2, E Gottardi2, VHJ van der Velden3, PG Hoogeveen3, G Cazzaniga4, V Rossi4, A Biondi4, JM Cayuela5, X Fund5, E Delabesse6, E Macintyre6, S Viehmann7, M Krahn1, R Dee8, E van der Schoot8, J Harbott7, JJM van Dongen3, G Saglio2 and J Gabert1
1Institut Paoli Calmettes and present affiliation: Hôpital Universitaire Nord, Marseille, France; 2University of Turin, Ospedale San Luigi Gonzaga, Orbassano-Torino, Italy; 3Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands; 4University of Milan Bicocca, San Gerardo Hospital, Monza, Italy; 5Hôpital Saint Louis, Paris, France; 6Hôpital Necker, Paris, France; 7Children's University Hospital, Giessen, Germany; and 8CLB, Amsterdam, The Netherlands
The t(4;11)(q21;q23) is the most frequent 11q23 translocation in precursor-B-ALL and involves MLL (HRX, Htrx, ALL1) and AF4 (FEL) genes.73 While MLL-AF4 positivity is observed in 5% of pediatric and adult ALL cases, this subgroup accounts for 40–60% of infant and therapy-induced ALL. The function of mammalian MLL is still largely unknown, but it seems to play a central role in segmentation during development.74 The recent mouse AF4 knockout suggested that this gene encodes a putative transcription factor that is involved in the ontogeny of the lymphoid lineage.75 Increased resistance of MLL-AF4-positive leukemia to stress-induced cell death shown in vitro has been suggested to contribute to their poor prognosis.76
At the molecular level, breakpoints in MLL and AF4 genes are spread within introns, between exons 8 and 12 (MLL) and exons 3 and 7 (AF4), some transcripts being more frequent in either adult or infant ALL.24 Uckun et al,77 using a nested RT-PCR technique with 10−4 sensitivity, reported low expression levels of MLL-AF4 transcripts in up to 13% of pediatric ALL at diagnosis, some of them being negative for the MLL-AF4 rearrangement by Southern analysis, and in around 25% of fresh normal BM or fetal liver samples. Based on these data, the authors suggested that RT-PCR assays for the MLL-AF4 FG transcripts were not suitable for MRD monitoring. However, these remarkable findings have not been confirmed so far by other groups or by other techniques.78,79
Nevertheless, in the same period, the first prospective MRD study, using nested RT-PCR, on 25 MLL-AF4-positive patients showed a significant correlation between PCR positivity, relapse and survival.80 The heterogeneity of the MLL-AF4 FG transcripts, their relatively low incidence in childhood and adult ALL and their poor prognosis with classical therapy explain the scarce number of reported MRD studies for this RT-PCR target.
2.2. EAC data
2.2.1. Primer design and optimization (phases I and II)
From the three probes and six primer sets tested, a common probe ENP242 and a common reverse primer ENR262, both located on AF4 exon 5, were adopted (Figure 6 and Table 8). Due to the large variety of MLL-AF4 FG transcripts, a common primer/probe set for all MLL-AF4 variants could not be designed. Two forward primers on the MLL gene were used in order to amplify at least 90% of non-infant and 60% of infant MLL-AF4 FG transcripts: ENF207 (exon 9) and ENF208 (exon 10).24 To our knowledge, no primer/probe sets for RQ-PCR have been published to date for MLL-AF4 or any other MLL FG transcripts.
Three cell lines were available for testing: RS4;11 (MLL exon 10-AF4 exon 4), MV4-11 (MLL exon 9-AF4 exon 5) and ALL-PO expressing two alternative transcripts (MLL exons 10 and 11-AF4 exon 5). Three plasmids were constructed: MLL-AF4 exon 9-exon 5, exon 10-exon 4 and exon 11-exon 5. An example of typical amplification plots (10−1, 10−3 and 10−4 cell line RNA dilutions) is shown in Figure 7; our sensitivity data on cell line material were comparable to those previously observed for detection of MLL-AF4 FG transcripts using nested PCR assay.24 The ‘creeping-curve’ phenomenon was a rare event within the negative samples and mostly laboratory related (see the Materials and methods section and Section 9).
2.2.2. MLL-AF4 expression in cell lines and diagnostic patient samples (phase IV)
Among the 22 samples included (14 BM and eight PB, including two paired samples), most of them (n=19) were recruited from the French therapeutic protocols LALA-FRALLE for ALL and tested in Marseille. The blast cell percentage in the samples was not known for all the samples; so results were not corrected according to this proportion. The normalized MLL-AF4 FG transcript expression (NCN) appeared to be similar between cell lines (n=3) and patients (Table 9). No differences in MLL-AF4 NCN were observed between PB and BM samples (Figure 8). The 95% range of expression was reduced when using either ABL or GUS as CG (Table 9) compared to the results obtained with B2M or without any CG (see web site). Furthermore, the correlation between MLL-AF4 and ABL transcript levels was the highest observed among the three CGs tested in diagnostic samples (Table 9).
2.2.3. QC rounds (phases IIIa–IVa)
Few false-negative samples (6%, 7/110) were observed for 10−3 and 10−4 dilutions. The amplification of MLL-AF4 cDNA within negative samples (FG-negative samples, NAC and NTC) also called false positivity was limited to 3% (5/176) and restricted to individual laboratories (Table 10). The Ct value in the false-positive wells was always more than 30 and most of the times higher than 37.
3. t(12;21)(p13;q22) with the TEL-AML1 fusion gene transcript
H Cavé1, G Brunie1, S Viehmann2, J Harbott2, G Cazzaniga3, V Rossi3, A Biondi3, M Malec4, A Porwit-MacDonald4, E Beillard5, C Glowaczower5, J Gabert5, F Watzinger6, R Baumgartinger6, T Lion6, X Fund7 and JM Cayuela7
1Hôpital Robert Debré, Paris, France; 2Children's University Hospital, Giessen, Germany; 3Ospedale San Gerardo, Monza, Italy; 4Karolinska Hospital, Stockholm, Sweden; 5Institut Paoli Calmettes and present affiliation: Hôpital Universitaire Nord, Marseille, France; 6St Anna Kinderspital, Vienna, Austria; and 7Hôpital Saint-Louis, Paris, France
The TEL(ETV6)-AML1(CBFA2,RUNX1) FG transcript results from the cryptic t(12;21)(p13;q22) and is found in about 25% of childhood precursor-B-ALL.81,82,83 Both TEL and AML1 encode nuclear transcription factors, which are critical for normal hematopoiesis. Their fusion leads to leukemogenesis by disrupting the normal function of TEL and/or creating a transcriptional repressor that impairs AML1 target gene expression.84 Two recurrent translocation breakpoints have been described. The major one breaks within TEL intron 5 and AML1 intron 1, generating TEL exon 5-AML1 exon 2 FG transcripts. The minor one is found in about 10% of TEL-AML1-positive ALL and breaks within AML1 intron 2, generating TEL exon 5-AML1 exon 3 FG transcripts which are 39 bp shorter (Figure 9). This latter FG transcript, together with transcripts lacking AML1 exon 3, is also detected in cases with AML1 exon 2 fusion, as a result of alternative splicing.85 Cases with breakpoints in TEL intron 4 have been described but remain exceptional.
The prognosis of TEL-AML1-positive ALL is still controversial. There is agreement that presence of the TEL-AML1 FG transcripts is associated with a high probability of 4-year event-free survival. However, some authors,86,87 but not others,88,89,90 reported relapse rates similar to those of ALL in general, with a majority of relapses occurring off-therapy. Whether these variations from one study to another are due to methodological bias or differences in efficacy of chemotherapy regimens is still unclear.
Few studies have reported MRD results for patients with TEL-AML1-positive ALL so far. Most of these studies relied on a qualitative or semiquantitative evaluation of the transcript level.85,90,91 The high frequency of TEL-AML1 transcript positivity in precursor-B-ALL prompted several groups to develop quantitative RT-PCR strategies targeted on this transcript to follow MRD.47,48,49,92 Preliminary data show that MRD is still detectable after induction therapy in 40–50% of patients, and that high MRD levels are found in some patients. However, despite the relatively high frequency of the TEL-AML1 fusion, only small series of patients have been analyzed so far (always less than 30 patients), and the rarity of relapses in addition to their possible late occurrence made it difficult to make any clinical correlations so far.
3.2. EAC data
3.2.1. Primer design and optimization (phases I and II)
The cell line used for testing was REH,93 which displays a TEL exon 5-AML1 exon 2 fusion. Two plasmid constructs were also used, one containing the ‘long transcript’ (TEL exon 5-AML1 exon 2) and the other containing the ‘short transcript’ (TEL exon 5-AML1 exon 3) (see the Materials and methods section).
At the end of the optimization phase, one primer/probe set (ENF301 on TEL exon 5, ENR361 on AML1 exon 3 and ENPr341 on TEL exon 5) was selected from the five sets that were initially tested (Figure 9 and Table 11). An example of typical amplification plots (10−1, 10−3 and 10−4 cell line RNA dilutions) is shown in Figure 10; the optimized RT and PCR conditions permitted us to detect a 10−4 dilution of the REH cell line in PBL RNA and 10 copies of plasmids in 100% of cases, which is consistent with the results obtained previously in the BIOMED-1 Concerted Action.24 It has to be noticed that molecular variants such as transcripts with a TEL-exon 4 fusion remain undetected using this set of primers.
3.2.2. TEL-AML1 expression in REH cell line and diagnostic patient samples (phase IV)
TEL-AML1 expression was studied in the REH cell line and in 57 TEL-AML1-positive precursor-B-ALL, 30 BM and 27 PB samples, including 23 paired samples (Table 12). Samples contained 18–100% leukemic blasts. A majority of these samples were obtained from patients of Hôpital Saint-Louis (Paris, France). NCN were calculated and adjusted for the percentage of blasts present in each sample.
TEL-AML1 expression in the REH cell line was within the range detected in primary leukemia samples.
No significant difference for ABL NCN was observed in TEL-AML1 expression when comparing PB and BM samples collected at diagnosis in the same patients (Figure 11).
3.2.3. QC rounds (phases IIIa–IVa)
As expected from the sensitivity threshold defined in previous phases, 10−3 and 10−4 RNA dilutions of the REH cell line were always found to be positive (Table 13) according to the criteria defined in the Meterials and methods section. A 5 × 10−5 dilution was found to be positive in 7/7 laboratories. During the various QC rounds, 11 negative samples (six negative RNA and five NAC or NTC samples) were tested in 7–11 labs, corresponding to a total of 279 amplification wells (Table 13). Nine of these wells (3%) were found falsely positive, corresponding to three false-positive samples out of 93 (3%) according to the criteria mentioned above. These false-positive wells were observed in six different labs. No case of false positivity with three positive wells was observed, and Ct values were always higher than 39. All false-positive wells corresponded to negative RNA, while NAC and NTC were always negative. This observation suggests that these false-positive results could be due to contaminations achieved prior to amplification, such as the RT step.
4. t(9;22)(q34;q11) with the BCR-ABL m-bcr fusion gene transcript
F Pane1, M Intrieri1,2, F Salvatore1, N Cross3, J Kaeda3, G Barbany4, G Cazzaniga5, V Rossi5, A Biondi5, H Cavé6, G Brunie6, P Heimann7, JM Delroisse7, N Pallisgaard8, P Hokland8, LS Mikkelsen8, G Martinelli9, S Buonamici9, S Viehmann10 and J Harbott10
1University of Naples, Naples, Italy; 2University of Molise, Isernia, Italy; 3Hammersmith Hospital, London, UK; 4University Hospital, Uppsala, Sweden; 5University of Milan Bicocca, San Gerardo Hospital, Monza, Italy; 6Hôpital Robert Debré, Paris, France; 7Erasme Campus, Brussels, Belgium; 8Aarhus University Hospital, Aarhus, Denmark; 9Institute of Hematology and Medical Oncology ‘Seragnoli’, University of Bologna, Bologna, Italy; and 10Children's University Hospital, Giessen, Germany
The BCR-ABL FG is associated with the formation of the Philadelphia translocation (Ph) and is one of the most common genetic abnormalities detected in leukemias.94 In ALL, Ph is detected in 25–30% of adult and 2–5% of childhood cases.95 Less frequently, it is associated with AML.96 In the ALL subset, this genetic lesion is known to confer a very poor prognosis,95,97 and, consequently, its detection is important in planning aggressive therapies, including allogeneic bone marrow transplant. In addition, the Ph chromosome is found in more than 95% of CML cases and is the hallmark of this disease.98 At the molecular level, the Ph chromosome or t(9;22) results in the juxtaposition of the 5′ part of the BCR gene (chromosome 22) to the 3′ part of the ABL gene (chromosome 9). In the vast majority of patients, the breakpoints in the BCR gene are clustered within three well-defined regions: (i) a 55 kb sequence of the first intron, called the minor breakpoint cluster region (m-bcr),99 (ii) a 5.8 kb region spanning exons 12–16, called the major breakpoint cluster region (M-bcr),99 and finally (iii) intron 19, called μ-bcr.100,101 Analysis of μ-bcr breakpoints will not be discussed further due to their extreme rarety. In the case of m-bcr breakpoints, the first exon of the BCR gene (e1) is juxtaposed to the second exon of the ABL gene (a2). The resultant fusion transcript (e1-a2) encodes a 190 kDa chimeric protein (p190).99 This type of BCR-ABL FG is found in 65% of adults and 80% of children with Ph-positive ALL.102 Only in sporadic cases is the p190 encoding BCR-ABL gene found in CML.103
Despite huge efforts, the molecular mechanisms by which the hybrid BCR-ABL protein gains transforming capability are still not fully understood. However, the BCR-ABL protein shows an increased and deregulated tyrosine kinase activity98 and it seems to deregulate the normal cytokine-dependent signal transduction pathways leading to the inhibition of apoptosis and by growth factor independent growth.104
All the studies about the clinical value of MRD in Ph+ ALL patients indicate that BCR-ABL-positive cells cannot be eradicated even by intensive chemotherapy.105,106,107 In a series of 36 Ph+ ALL patients treated by SCT, Radich et al108 reported that RT-PCR assessment of MRD was, by multivariate analysis, the best prognostic indicator for continuous complete remission (CR). Recently, Yokota et al40 reported the follow-up by RQ-PCR analysis of 13 m-bcr-positive ALL patients. All these data suggest that, in Ph+ ALL patients, quantitative monitoring of residual leukemic cells could prove more valuable than their qualitative detection to assist in clinical decision-making.
4.2. EAC data
4.2.1. Primer design and optimization (phases I and II)
The efficiency of four different primer/probe sets, designed using the ‘Primer Express™’ software, was tested in 1:10 serial dilution experiments of RNA from the TOM-1 cell line (e1-a2 junction or m-bcr) into RNA from HL60 cells. All primer/probe sets were free from nonspecific amplification artifacts, but two sets were superior in terms of sensitivity and ΔRn value at plateau phase. Both sets had comparable amplification efficiencies and reached 10−5 sensitivity and ΔRn >3.0 at the 10−1 cell line dilution. After extensive testing, the set that included ENF402 (located in BCR exon 1), ENR561 and probe ENP541 (both located in ABL exon 2) was selected. Both the reverse primer and probe are common to the set used for the RQ-PCR detection of BCR-ABL (M-bcr) FG transcripts (Figure 12 and Table 14). An example of typical amplification plots (10−1, 10−3 and 10−4 cell line RNA dilutions) is shown in Figure 13.
4.2.2. BCR-ABL m-bcr FG expression in TOM-1 cell line and diagnostic patient samples (phase IV)
To establish reference intervals of m-bcr transcripts, we determined FG expression in 17 BM samples and seven PB samples from sequential ALL patients at diagnosis as well as centrally prepared and distributed RNA from the TOM-1 cell line (Table 15). For the FG and the three CG transcripts assayed, separate series of plasmid dilutions were amplified in each experiment to calculate transcript copy numbers. Although samples with ABL Cts >29.3 were excluded from the analysis because they were not suitable for amplification, the Ct values of all transcripts were generally lower in cell lines than in patient material (both BM and PB), most probably due to low quality of some of the patient RNAs. However, after normalization to CG expression, m-bcr transcript levels were comparable in patients and the TOM-1 cell line, and very similar in four paired BM and PB samples (Figure 14).
The primers and probe used to amplify ABL mRNA are also able to amplify BCR-ABL FG mRNA, and hence the assay of ABL gene expression used to normalize data may be affected by the levels of m-bcr transcript in the samples.
4.2.3. Quality control rounds (phases IIIa–IVa)
Only 3/129 wells gave false-negative results (see the Materials and methods section for details) in the three QC rounds, and in all cases the false negativity was obtained for 10−4 TOM-1 dilutions. Furthermore, false-positive results were detected in 5.3% of PCR tests (13/246, Table 16).
5. t(9;22)(q34;q11) with the BCR-ABL M-bcr fusion gene transcript
G Barbany1, C Chillón2, M Silva2, D De Micheli3, E Gottardi3, P Evans4, F Watzinger5, R Baumgartinger5, N Pallisgaard6, F Pané7, M Intrieri7,8, V Montefusco9, S Bounamici9, E Delabesse10, V Asnafi10, JM Delroisse11, J Kaeda12, G Balazenko13, N Cross12, P Heimann11, E Macintyre10, G Martinelli9, P Hokland6, T Lion5, G Saglio3 and M González2
1University Hospital, Uppsala, Sweden; 2University Hospital, Salamanca, Spain; 3University of Turin, Ospedale Luigi Gonzaga, Turin, Italy; 4Institute of Pathology, HMDS, Leeds, UK; 5Children's Cancer Research Institute, St Anna Children Hospital, Vienna, Austria; 6University Hospital, Aarhus, Denmark; 7University Hospital, Naples, Italy; 8University of Molise, Isernia, Italy; 9Institute of Hematology and Medical Oncology ‘Seragnoli’, University of Bologna, Italy; 10Hôpital Necker, Paris, France; 11Erasme Campus, Brussels, Belgium; 12Hammersmith Hospital, London, UK; and 13National Center of Haematology and Transfusiology, Sofia, Bulgaria
Most cases of CML are associated with the presence of t(9;22) resulting in a small derivative chromosome 22 known as the Ph.109 As a consequence, the ABL proto-oncogene on chromosome 9 is fused to the BCR gene on chromosome 22.110 In CML patients and approximately 35% of Ph-positive adult ALL patients, the breakpoint on chromosome 22 is located between exons 12 and 16 of the BCR gene, in the so-called major breakpoint cluster region (M-bcr).99 The breakpoint on chromosome 9 is located in most cases between exons 1 and 2 in the ABL gene. The transcription product of this BCR-ABL FG is an 8.5-kb aberrant fusion RNA with two junction variants b2a2 and/or b3a2 that gives rise to the BCR-ABL chimeric protein (p210), a tyrosine kinase with deregulated activity.111 Rare cases with b2a3 and b3a3 BCR-ABL transcripts can be observed.
Because of its high sensitivity, qualitative RT-PCR has been extensively used to monitor residual disease in CML, yielding partially contradictory results. Sequential analysis of patients who received allogeneic BM transplantation (BMT) showed that repeated PCR positivity correlated with an increased risk of relapse.112,113 On the contrary, other studies did not find any correlation between PCR positivity and subsequent relapse, and showed that long-term survivors of allogeneic BMT could be PCR positive even years after transplant without ever relapsing.114,115,116,117
A competitive RT-PCR method to quantify the level of BCR-ABL FG transcripts was developed by several groups118,119 in an effort to improve the predictive value of BCR-ABL mRNA detection. The sequential analysis of patients who had undergone BMT showed that monitoring of BCR-ABL FG transcripts levels can be useful to predict an impending relapse while the patient is still in hematological and cytogenetic remission.31,120 Based on quantitative RT-PCR studies, two groups have proposed to define a ‘molecular relapse’ parameter.31,121 In addition, quantification of BCR-ABL FG transcripts has also proven useful to monitor response to α-interferon-122 and imatinib-treated patients.123,124
Despite many encouraging reports, monitoring of BCR-ABL FG transcripts with competitive RT-PCR has had limited clinical impact, partly due to the fact that this approach is difficult to standardize. Since the advent of real-time PCR, several groups have published reports that describe the feasibility of monitoring CML patients with this technique.37,38,39,125,126,127 However, most RQ-PCR studies included too few patients or patients from different therapeutic protocols that together with methodological differences make it difficult to evaluate the clinical impact and to define general guidelines to monitor CML patients with RQ-PCR. The availability of a standardized protocol for RQ-PCR will facilitate data comparison among different centers, making it possible to define a threshold where a patient is likely to relapse and ultimately to assess the impact of an early therapeutic intervention based on the kinetics of BCR-ABL FG transcripts.
5.2. EAC data
5.2.1. Primer design and optimization (phases I and II)
Two alternative forward BCR primers, one located on BCR exon 13 (exon b2) and the second on BCR exon 14 (exon b3), and a reverse ABL primer and probe, both on the second exon of the ABL gene, were designed (Figure 12). This set, together with five comparable sets already available within the network, was evaluated for sensitivity as well as for robustness of the PCR reaction (see the Materials and methods section). Based on these parameters, the set designed by the EAC group was chosen. Since dilution series of the K-562 cell line was amplified with equal efficiency with either of the two forward primers, one common forward primer ENF501, located on BCR exon 13 (exon b2), was selected to amplify both fusion variants BCR-ABL b3-a2 and b2-a2 (Figure 12, Table 17). An example of typical amplification plots (10−1, 10−3 and 10−4 cell line RNA dilutions) is shown in Figure 15.
5.2.2. BCR-ABL M-bcr expression in K-562 cell line and diagnostic patient samples (phase IV)
18.104.22.168. For CML, in K-562 and patients at diagnosis
The expression of BCR-ABL M-bcr transcripts was quantified in 29 CML patients in order to establish the range of FG expression levels in diagnostic samples (Table 18 and Figure 16a). These samples included 15 BM and 14 PB samples. The expression of the BCR-ABL FG relative to the CG was very similar between BM and PB. No large variations in BCR-ABL M-bcr expression levels were found between individual patients. Expression of BCR-ABL mRNA in the K-562 cell line was found to be higher than CML diagnostic samples by two logs relative to B2M and one log relative to GUS gene (see Table 18).
22.214.171.124. In BCR-ABL M-bcr-positive ALL
In addition to diagnostic samples from CML patients in chronic phase, BCR-ABL M-bcr expression was quantified in diagnostic BCR-ABL M-bcr-positive ALL samples (Table 19 and Figure 16b). Expression of BCR-ABL relative to B2M and GUS was found to be higher in ALL samples compared to CML samples (Tables 18 and 19). We observed a statistical difference in the FG level of expression between M-bcr ALL and CML patients either with B2M or GUS as CG (P=0.001), while no difference was observed between m-bcr and M-bcr in ALL (Figure 17).
The primer set designed to amplify the ABL CG is located on exon 2 and also amplifies the BCR-ABL FG. For this reason, the use of ABL as a CG could introduce a bias for quantifying BCR-ABL in CML and Ph+ ALL samples when a large proportion of the cells express BCR-ABL. Using the ratio (BCR-ABL/ABL) would theoritically lead to an underestimation of the tumor load in these samples since the maximum ratio is one. However, this bias had a minor impact on relative quantification of FG transcripts at diagnosis (see Section 11.1.3). We found values up to 3.0 in BM and up to 4.4 in PB samples of CML patients at diagnosis (Table 18), although all median BCR-ABL/ABL ratios were below 1, except for PB CML patients at diagnosis (Tables 15 and 18). These unexpected results obtained with plasmid calibrators were confirmed without calibrators (ΔCt method, see web site) and clearly illustrate the limits of accuracy of gene transcript quantification by RQ-PCR. Similar results were observed in a oligocenter context.125,128
5.2.3. Quality control rounds (phases IIIa–IVa)
The percentage of false negatives was 4.9% (14/285) for the first and second quality control rounds (Table 20). The laboratories that showed the false negatives had a consistent reduction in sensitivity for all the targets in a particular phase, which indicated that the cause for the lower sensitivity was a lower RT efficiency rather than a PCR-related problem.
In the third quality control round (phase IVa), a rate of false positivity (5.5%, 6/110) similar to the previous phases was observed despite a particularly high frequency (18%, 4/22) of false-positive results within FG-negative samples (Table 20). This observation is possibly explained by the inclusion of the undiluted K-562 RNA among the coded samples instead of the 10−1 dilution and thereby increasing the risk of accidentally contaminating neighboring wells when pipeting the cDNA onto the PCR plate. It should be noted that the majority of the false-positive samples (NAC/NTC) were concentrated in individual laboratories while the rest of the laboratories (one laboratory per QC round) showed only occasionally single false-positive wells or no false positives.
6. Intrachromosomal microdeletion on 1p32 with the SIL-TAL1 fusion gene transcript
VHJ van der Velden1, PG Hoogeveen1, N Boeckx1, MJ Willemse1, E Delabesse2, V Asnafi2, E MacIntyre2, N Pallisgaard3, P Hokland3, LS Mikkelsen3, JM Cayuela4, O Spinelli5 and JJM van Dongen1
1Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands; 2Hôpital Necker, Paris, France; 3University Hospital, Aarhus, Denmark; 4Hôpital St Louis, Paris, France; and 5Riuniti Hospital Bergamo, Italy
The microdeletion on 1p32 is the most frequent chromosome aberration found in childhood T-ALL.129,130,131,132,133,134,135 The microdeletion involves the TAL1 gene (T-cell acute leukemia 1 gene, also known as stem cell leukemia (SCL) or T-cell leukemia gene 5 (TCL5)) and the SIL gene (SCL interrupting locus), which is located approximately 90 kb upstream.133 As a result, the TAL1 coding sequences are placed under the control of the SIL promotor, which is expressed in T cells, and consequently the TAL1 gene becomes ectopically expressed in the involved T-ALL.
The TAL1 gene, in particular exons 4–6, encode a 42 kDa protein, which is a basic helix–loop–helix (bHLH) transcription factor. The TAL1 protein can heterodimerize with other bHLH transcription factors, including members of the E2A family, and is an essential factor for the development of all hematopoietic lineages.136,137,138 The SIL gene is a member of the immediate-early gene family, but its function in hematopoietic cells is not yet well defined.129,139
Although both the SIL and TAL1 genes contain several conserved deletion breakpoints, most cases (⩾95%) involve the sildb1 breakpoint in combination with the taldb1 or taldb2 breakpoint.131,132,140 By alternative splicing, three different SIL-TAL1 transcripts can be formed, of which the type II transcript is the most predominant one.134 There is no apparent relationship between the occurrence of SIL-TAL1 transcripts and prognosis or outcome.130
SIL-TAL1 transcripts are exclusively found in T-ALL, in which they are present in 5–25% of the patients.129,132,134,141 The frequency is related to the immunophenotype of the T-ALL (the presence of the SIL-TAL1 FG is restricted to CD3− and TCRαβ+ T-ALL) and the occurrence of TCRD gene deletions.132,142 The SIL-TAL1 FG transcripts seem to be more frequent in children compared to adults.134
The detection of TAL1 deletions at the DNA level has already been described.143,144,145 A recent report described a TaqMan-based RQ-PCR method for the detection of TAL1 deletions at the DNA level in T-ALL patients.145 In that report, the forward primer and probe were positioned in SIL exon 1b (and part of the following intron) and the reverse primer was located in TAL1 exon 1b. Using the CEM cell line, a sensitivity of 10−5 could be obtained, which is equivalent to a single leukemic genome. To our knowledge, no RQ-PCR primers/probe sets for SIL-TAL1 FG transcripts have been published so far.
6.2. EAC data
6.2.1. Primer design and optimization (phases I and II)
Initially, three TaqMan probes (located in SIL exon 1a, TAL1 exon 4 and TAL1 exon 5), two forward primers (both in SIL exon 1a) and five reverse primers (two in TAL1 exon 3, two in TAL1 exon 4 and one in TAL1 exon 6) were tested for specificity and efficiency.
A single forward primer (ENF601; located in SIL exon 1a), reverse primer (ENR664; located in TAL1 exon 3) and probe (ENP641; located in SIL exon 1a) were selected (Figure 18 and Table 21). An example of typical amplification plots (10−1, 10−3 and 10−4 cell line RNA dilutions) is shown in Figure 19.
The selected primer/probe set will detect virtually all SIL-TAL1 transcripts (the most common type II as well as type III), but will not detect TAL1 translocations or ‘aberrant’ expression of the TAL1 gene without apparent rearrangements.
6.2.2. SIL-TAL1 expression in cell lines and diagnostic patient samples (phase IV)
Undiluted RNA of six different SIL-TAL1-positive cell lines was tested in duplicate for CG expression (ABL, B2M and GUS) and in triplicate for SIL-TAL1 FG expression (Table 22). In three laboratories (Erasmus MC in Rotterdam, Hôpital Necker and Hôpital St Louis in Paris), a total of 16 SIL-TAL1-positive patients at diagnosis were also included (10 BM and 10 PB samples, including five pairs).
Ct values for the CG transcripts were significantly lower in the cell lines as compared to the patient samples (Table 22), which could be due to the fact that stored patient samples were used or alternatively that their expression is higher in cell lines than in primary patient samples. SIL-TAL1 transcript expression (Ct, CN and NCN) was comparable between the six cell lines tested, but in patients a slightly larger variation in SIL-TAL1 FG transcript expression was found (Table 22). Nevertheless, the normalized SIL-TAL1 FG transcript expression did not differ between cell lines and patients. Furthermore, comparison between BM and PB samples (including five pairs) showed that SIL-TAL1 transcript expression was similar in both compartments (Figure 20 ).
6.2.3. Quality control rounds with blind samples (phases IIIa–IVa)
False positivity was observed in two out of 46 FG-negative samples (4.3%) and in none out of 162 NAC/NTC wells (0%) resulting in a total false positivity of 1.0% (2/208). False negativity was only observed in the first QC round (phase IIIa), but was absent in the next two QC rounds (Table 23). Therefore, false negativity does not seem to be a problem, although this may be dependent on the level of SIL-TAL1 FG transcripts in the leukemic cell sample.
7. t(15;17)(q22;q21) with the PML-RARA fusion gene transcript
G Cazzaniga1, V Rossi1, R Flora2, C Chillon3, E Beillard4, R Dee5, M Malec6, MJ Mozziconacci7, C Glawaczower4, Y Toiron7, M Lafage-Pochitaloff7, J Krauter8, D Diverio9, R Garcia-Sanz3, E van der Schoot5, A Porwitt McDonald6, G Heil8, F Lo Coco9, J Gabert4, D Grimwade2 and A Biondi1
1Ospedale San Gerardo, Monza, Italy; 2Guy's, King's & St Thomas’ School of Medicine, London, UK; 3Hospital Clínico Universitario, Salamanca, Spain; 4Institut Paoli Calmettes and present affiliation: Hôpital Universitaire Nord, Marseille, France; 5CLB, Amsterdam, The Netherlands; 6Karolinska Hospital, Stockholm, Sweden; 7Institut Paoli-Calmettes, Marseille, France; 8Medical School, Hannover, Germany; and 9Università la Sapienza, Roma, Italy
The PML-RARA FG transcripts, which are the molecular result of the t(15;17)(q22;q21) translocation, are associated with the majority of APL cases, a distinct AML subset with M3 cytomorphology.146
APL accounts for 10–15% of de novo AML in younger adults in Southern Europe (reviewed in Biondi et al147). Among pediatric patients, the incidence of APL is usually considered to be lower and accounting for 3–9%, although published data from Italian cooperative studies indicate that APL occurs in Italian children with the same incidence as observed in adults. Moreover, several small series from different countries in Central and South America have noted a higher-than expected frequency of pediatric APL.147
The two genes fused in the t(15;17) are PML, located on chromosome 15,148,149,150,151 and the retinoic acid receptorα (RARA) gene on chromosome 17.152,153 Other genes have been shown to be fused to RARA in rare instances of morphological APL cases negative for the t(15;17), such as PLZF on chromosome 11q23, NPM on 5q35, NUMA on 11q13 and STAT5B on 17q21.154
The chimeric PML-RARA protein is a transcriptional repressor. In the absence of ligand (retinoic acid, RA), it binds DNA together with co-repressors such as SMRT (silencing mediator for RAR and TR) and N-CoR (nuclear receptor co-repressor) and renders chromatin inaccessible to transcriptional activators or basal transcription machinery.154,155
RARA breakpoints always occur in intron 2 which is 17 kb in length (Figure 21).149 By contrast, three regions of the PML locus are involved in the t(15;17) translocation breakpoints: intron 6 (bcr1; 55% of cases), exon 6 (bcr2; 5%) and intron 3 (bcr3; 40%) (Figure 21). As a consequence, there are three possible PML-RARA isoforms, referred to as long (L, or bcr1), variant (V, or bcr2) and short (S, or bcr3). It should be noted that the size of the PCR products varies in bcr2-positive cases, because of the variable breakpoint positions in exon 6 of the PML gene and inclusion of a variable number of RARA intron 2-derived nucleotides in the FG transcript.156 Chimeric PML-RARA and RARA-PML transcripts are formed as a consequence of the reciprocal translocation between the PML and RARA loci. However, the observation that RARA-PML FG transcripts are present in most but not all APL cases, has favored the use of PML-RARA FG transcripts as PCR target for detection of APL cells at diagnosis and during monitoring. Standardized conditions for RT-PCR analysis of PML-RARA FG transcripts have been developed by the BIOMED-1 Concerted Action.24 Primer sets have been designed that allow the detection of the various PML-RARA FG transcripts, generated by the existence of different PML breakpoint regions as well as the presence of alternative splicing between central exons of PML.
In the last decade, the availability of differentiation therapy with all-trans retinoic acid (ATRA) has produced a remarkable improvement in the outcome of patients with APL (reviewed in Grimwade155). The challenge is how to identify the relatively small subgroup of patients at particular risk of relapse who cannot be reliably distinguished on the basis of pretreatment characteristics and who could potentially benefit from more intensive treatment in first remission. Overall, there is general agreement that a positive PML-RARA test after consolidation is a strong predictor of subsequent hematological relapse, whereas repeatedly negative results are associated with long-term survival in the majority of patients.21,157 The Italian GIMEMA group (Gruppo Italiano Malattie Ematologiche Maligne Adulto) reported that recurrence of PCR positivity, detected by 3-monthly BM surveillance performed after completion of therapy, was highly predictive of relapse.21 Using such a strategy, approximately 70% of relapses were successfully predicted.21 A different perspective in the application of MRD to identify APL patients at higher risk of relapse has been used by the MRC ATRA trial,158 where the kinetics of achieving a molecular remission was evaluated. Finally, the benefit of early treatment at the time of molecular relapse has still to be proven, but preliminary evidence supports such a strategy.22
Among the different methods (conventional karyotyping, FISH and PML immunostaining with specific antibodies), RT-PCR detection of the PML-RARA FG transcripts appears to be the only approach suitable for MRD detection.24 Moreover, quantitative PCR could provide information on the correlation between different levels of disease at early phases of therapy and clinical outcome. However, there have been relatively few studies reporting the use of RQ-PCR in APL patients.41,159,160,161
Although the molecular diagnosis and monitoring of APL patients represents one of the most relevant examples of the impact of molecular genetics in clinical hematology, further investigations are still needed.
7.2. EAC data
7.2.1. Primer design and optimization (phases I and II)
One probe and two reverse primers on RARA gene exon 3 in combination with seven forward primers on the PML gene were evaluated. Five forward primers were designed in PML exon 6, two and three specific primers for bcr1 and bcr2 breakpoints respectively, while two primers for bcr3 were designed in PML exon 3. Based on published data on the localization of bcr2 PML breakpoints,162,163,164 the respective forward primers on PML exon 6 were designed in order to cover at least 80% of bcr2 cases. The only cell line available for testing was NB-4,165 which has a bcr1 PML breakpoint; for the evaluation of bcr2 and bcr3 primer/probe sets, diagnostic patient BM RNA was used. Plasmid constructs for all three PML-RARA breakpoint variants were made (see the Materials and methods section).
After extensive testing on cell line and patient RNA and plasmid dilutions, three specific primer/probe sets were selected, based on the maximum sensitivity: the probe RARA ENP942, the common reverse primer RARA ENR962 and three PML forward primers ENF903 (for bcr1), ENF906 (for bcr2) and ENF905 (for bcr3), respectively (Figure 21). Sequences are listed in Table 24. Although ENF906 can potentially be used to amplify bcr1 as well as bcr2 cases, direct comparison of ENF906 and ENF903 for amplification of PML-RARA in serial dilutions of the NB-4 cell line revealed the latter primer to provide a more sensitive assay for amplification of bcr1 cases. An example of typical amplification plots (10−1, 10−3 and 10−4 NB-4 RNA (bcr1) in PBL RNA) is shown in Figure 22.
7.2.2. PML-RARA expression in cell lines and diagnostic patient samples (phase IV)
PML-RARA expression was studied in the NB-4 cell line and in 16 positive AML-M3 bcr1 patients (Table 25 and Figure 23). Patient samples consisted of 14 BM and nine PB samples, including six paired BM/PB samples, obtained at diagnosis. Samples contained 10–100% of leukemia blasts. NCN were calculated and adjusted for the percentage of blasts present in each sample. PML-RARA expression in the NB-4 cell line was within the range of FG expression detected in primary leukemia samples.
PML-RARA expression was studied in six bcr3-positive patients at diagnosis, consisting of six BM and four PB samples (including four paired BM/PB samples). Although the number was very limited, no significant difference was observed in PML-RARA bcr3 expression when comparing PB and BM on paired samples except for B2M normalized results (see web site and Figure 23).
7.2.3. QC rounds (phases IIIa–IVa)
During the various QC rounds, 145 negative samples were tested in 7–11 labs during the three phases (Table 26). Five out of 100 NAC/NTC samples (5%) and five out of 45 FG-negative samples (11%) were falsely positive for bcr1 amplification (Table 26). Overall, the frequency of false positivity was 6.9% (10/145). The so-called false positivity was limited to individual laboratories and the Ct value in the false-positive well was always more than 30 and most of the time higher than 35.
By contrast, according to the criteria mentioned above, no false-negative samples (n=96) for 10−3 and 10−4 dilutions were observed, either for bcr1, bcr2 or bcr3. None of the 42 wells tested independently for bcr2 and bcr3 at 10−4 dilution falsely resulted as negative. Only eight out of 180 wells (4.4%) tested for bcr1 at 10−4 dilution falsely resulted as negative.
8. Inv(16) (p13q22) with the CBFB-MYH11 fusion gene transcript
E Gottardi1, D De Micheli1, F Pane2, M Intrieri3, F Salvatore2, C Preudhomme4, N Grardel-Duflos5, G Martinelli5, S Buonamici5, J Krauter6, G Heil6, P Vandekerckhove7, A Navarrete7, JLE Aerts7 and G Saglio1
1University of Turin, Ospedale San Luigi Gonzaga, Orbassano-Torino, Italy; 2University of Naples, Naples, Italy; 3University of Molise, Isernia, Italy; 4Hôpital Calmette, Lille, France; 5Institute of Haematology and Medical Oncology ‘Seragnoli’, University of Bologna, Bologna, Italy; 6Medizinische Hochschule, Hannover, Germany; and 7University Hospital, Leuven, Belgium
Pericentric inversion of chromosome 16, inv(16)(p13q22), is found in about 8–9% of newly diagnosed AML cases.166 The inv(16)-positive AMLs are included with those with t(8;21) translocation in a group generally referred to as ‘core binding factor’ (CBF) leukemias, as both are characterized by rearrangements of genes that code for components of the heterodimeric transcription factor CBF, which plays an essential role in hematopoiesis.167 Inv(16) or the rarer t(16;16)(p13;q22) leads to fusion of the CBFB chain gene with the smooth muscle myosin heavy chain gene MYH11.168 The resulting FG mRNA can be detected by RT-PCR and represents a suitable molecular marker for both diagnostic and monitoring studies.169,170,171 So far, 10 different CBFB-MYH11 FG transcripts have been reported. The nomenclature used is derived and updated from the review by Liu et al.167 More than 85% of positive patients have the type A transcript; type D and E transcripts each represent nearly 5%, whereas all other types occur in sporadic cases. CBFB-MYH11-positive AML are usually considered to have a favorable prognosis, with more than 50% of patients obtaining long-term CR.4 Such favorable results with conventional chemotherapy led some authors to consider that allo-BMT is not indicated to consolidate first CR in these patients, even when a suitable donor is available.46,172,173 Nevertheless, the relapse rate is still high indicating that reliable methods to detect MRD during hematologic CR are needed in order to better adapt the intensity of postremission therapy to specific cohorts of patients. So far, the use of qualitative RT-PCR-based methods employed to detect CBFB-MYH11 FG transcripts did not allow consistent discrimination of prognostic subgroups of patients in CR.169,171,174,175 In fact, the use of standard nested RT-PCR has produced conflicting MRD results: while in most reports the vast majority of patients in prolonged CR were found to be PCR-negative, a few long-term survivors never converted to RT-PCR negativity. Moreover, 10–20% of PCR-negative patients eventually relapsed, suggesting that the achievement of PCR negativity is not synonymous with cure. Some of the difficulties in interpreting the above results may derive from lack of standardization of methodologies involved. Quantitative RT-PCR studies using competitive PCR or RQ-PCR enabled monitoring of the decrease in CBFB-MYH11 FG transcripts during early phases of induction and consolidation therapies.33,45,46,176,177 However, due to the low number of patients so far examined, it was not possible to define a kinetic or a cutoff level for predicting relapse.178
8.2. EAC data
8.2.1. Primer design and optimization (phases I and II)
During phase I, we tested six primer/probe sets: three for the A, two for D and one for the E form. As CBFB-MYH11 transcripts type A, D and E represent approximately 95% of all cases, in order to amplify these transcripts we decided to use a common forward primer located on CBFB exon 5 (ENF803) and a common probe located on CBFB exon 5 (ENPr843). Three different reverse primers located respectively on MYH11 exon 12 for type A (ENR862), MYH11 exon 8 for type D (ENR863) and MYH11 exon 7 for type E (ENR865) (Figure 24 and Table 27) were chosen. An example of typical amplification plots (10−1, 10−3 and 10−4 RNA dilutions) is shown in Figure 25 for all three types of CBFB-MYH11 transcripts.
8.2.2. CBFB-MYH11 expression in the ME-1 cell line and diagnostic patient samples (phase IV)
Pure RNA of the ME-1 cell line (type A) was tested in eight different laboratories (Table 28). In addition, diagnostic BM or PB samples of 24 type A patients, three type D and four type E were analyzed (Figure 26). The values of FG transcripts at diagnosis were similar in all three categories of patients, but show a range of variation from patient to patient of one log.
8.2.3. QC rounds (phases IIIa–IVa)
No false-negative results for the 10−3 dilution were observed, whereas at the 10−4 dilution a maximum of 12% of false-negative results were observed (Table 29). False positivity was absent during all phases in NAC and NTC wells (0%, n=104), whereas a single case (two wells out of 16) of false positivity was observed during phase IVa in a coded FG-negative sample and was due to contamination (Table 29).
9. t(8;21)(q22;q22) with the AML1-ETO fusion gene transcript
JLE Aerts1, D Grimwade2, P Vandekerckhove1, A Navarrete1, R Flora2, J Krauter3, G Heil3, S Viehmann4, C Preudhomme5, N Grardel-Duflos5, F Pane6, M Intrieri7, F Salvatore6, NCP Cross8, J Kaeda8, M Malec9, A Porwitt McDonald9, Y Toiron10, M-J Mozziconacci10 and M Lafage-Pochitaloff10
1University Hospitals Gasthuisberg, Leuven, Belgium; 2Guy's, King's & St Thomas’ School of Medicine, London, UK; 3Medical School, Hannover, Germany; 4Children's University Hospital, Giessen, Germany; 5Hôpital Calmette – CHU, Lille, France; 6University Hospital, Naples, Italy; 7Università del Molise, Isernia, Italy; 8 Hammersmith Hospital, London, UK; 9Karolinska Hospital, Stockholm, Sweden; and 10Institut Paoli Calmettes, Marseille, France
The AML1(CBFA2, RUNX1)-ETO (MTG8) gene fusion results from the t(8;21)(q22;q22), which is the commonest chromosomal rearrangements associated with AML, being detected in approximately 8% of AML cases in children and young adults.4 The AML1 gene encodes the α2 subunit of the heterodimeric transcription factor CBF, which is critical for hemopoietic development and whose β subunit is disrupted by the inv(16)/t(16;16) which leads to the CBFB-MYH11 FG (see Section 8).179
As shown in Figure 27, AML1 breakpoints are located within intron 5, while ETO breakpoints occur upstream of exon 2. This gives rise to a single type of AML1-ETO FG transcript in which AML1 exon 5 is fused to ETO exon 2,24 thereby simplifying molecular screening strategies and MRD monitoring as compared to FG with multiple breakpoint regions, such as t(4;11) with MLL-AF4, t(15;17) with PML-RARA and inv(16) with CBFB-MYH11.
AML1-ETO is an important PCR target for MRD detection in view of the generally favorable outcome of patients with the t(8;21), such that routine use of BMT in first CR has been shown to confer no overall survival benefit.172,173 Therefore, it is of paramount importance to identify the relatively small subgroup of patients at high risk of relapse who could benefit from additional therapy. However, the role of MRD detection in AML1-ETO-positive AML has been somewhat controversial in view of the detection of FG transcripts in patients in long-term remission following chemotherapy, autologous BMT/PBSCT and even alloBMT.25,180,181 The detection of residual transcripts in patients who are cured of their disease has been seen as providing evidence that AML1-ETO alone is insufficient to mediate AML, and has recently been shown to relate to a fraction of stem cells, monocytes and B cells present in remission marrow.182 Hence, the relatively frequent reports of PCR positivity in patients considered to be cured of t(8;21)-positive AML is likely to reflect the higher levels of sensitivity commonly achieved for AML1-ETO RT-PCR assays (typically 1 in 105 to 1 in 106), as compared to those for other AML FG targets (typically 1 in 104–105). Despite the fact that a recent study has shown that conventional qualitative RT-PCR has the potential to provide an independent prognostic factor in AML1-ETO-positive AML,183 there has been some concern regarding the suitability of sensitive ‘end point’ assays for MRD detection as a means of determining treatment approach in this subgroup of patients.
Over the last few years, quantitative RT-PCR methods have been investigated to determine whether they can more reliably identify the relatively small subgroup of patients destined to relapse.184,185,186 Competitive RT-PCR assays have revealed variation in AML1-ETO expression relative to ABL between cases at diagnosis (10-fold in BM, 32-fold in PB) and suggest that AML1-ETO and ABL mRNAs have comparable stability.186 Furthermore, these studies revealed varying kinetics of FG transcript reduction following chemotherapy. Patients with low or undetectable levels of AML1-ETO transcripts were associated with maintenance of CCR, while high or rising transcript numbers predicted relapse.186 These promising preliminary data suggest that RQ-PCR is likely to be valuable for MRD monitoring in this subset of AML, with the added advantages that the latter technique is less labor intensive, more reproducible and amenable to standardization lending itself to use in large-scale clinical trials. Preliminary studies of RQ-PCR42,43,44,187,188 have essentially confirmed the results obtained via competitive RT-PCR, revealing a 3.5- to 20-fold variation in AML1-ETO FG expression levels in diagnostic BM that was not related to blast percentage and which needs to be taken into account when assessing response to therapy. Furthermore, variability in kinetics of response to chemotherapy was noted, and interestingly AML1-ETO transcripts were also detected in patients in long-term remission from AML. However, the predictive value of RQ-PCR remains to be established in a large number of patients subject to a consistent treatment approach.
9.2. EAC data
9.2.1. Primer design and optimization (phases I and II)
Initially, the primer and probe sequences published by Marcucci et al42 were tested together with two ‘in house’ sets. While the former primer/probe set was superior in terms of sensitivity, we observed significant recurrent background signals in the negative control samples. We therefore decided to design two new probes compatible with this primer set leading to the selection of ENP747 positioned on the breakpoint, in conjunction with the forward primer ENF701 positioned on AML1 exon 5 and the reverse primer ENR761 on ETO exon 2 (Figure 27 and Table 30).24 The cell line available for testing was KASUMI-1.189 The AML1-ETO RQ-PCR assay was found to be particularly robust, associated with a relatively high ΔRn (see Figure 28a). Reduction in the probe concentration from 200 to 100 nM was therefore evaluated. The decrease in probe concentration did not affect assay sensitivity, and the ‘baseline creeping’ artifact was only very rarely observed (Figure 28b).
9.2.2. AML1-ETO expression in KASUMI cell line and diagnostic patient samples (phase IV)
Undiluted RNA of the KASUMI-1 cell line was tested in five different laboratories (Table 31). Analysis for archived diagnostic RNA from 22 patient samples was undertaken by four different laboratories. All differences between BM and PB per CG were not significant on paired samples (n=10, Wilcoxon test). After applying the exclusion criteria, 12 PB and 10 BM samples were evaluated (Table 31).
AML1-ETO FG transcript expression (Ct and CN) was higher in the KASUMI-1 cell line than in the patient samples (median difference of approximately two Ct values). Among the patient samples, no difference was seen in the expression of the AML1-ETO FG transcript. With regard to the CGs, the expression of ABL was comparable between the cell line and patient samples. Significant variation was seen in the expression of B2M, both between patient samples and cell line and between BM and PB samples. For GUS, intermediate results were observed (Table 31 and Figure 29).
9.2.3. QC rounds (phases IIIa–IVa)
No false-negative results (out of a total of 112 analyzed samples) for 10−3 and 10−4 dilutions were observed, which is in line with the good sensitivity of the RQ-PCR assay (Table 32). Overall, the frequency of false positivity was 9.7% (15/154). Six out of 108 NAC/NTC samples (5.6%) and nine out of 46 FG-negative samples (20%) were falsely positive for AML1-ETO amplification (Table 32). In coded FG-negative samples, false positivity was in most cases restricted to individual laboratories.
10. Overall results
10.1. Standard curve comparison (phase IIIa)
No significant differences were observed in Ct values for the lowest RNA, cDNA (10−3 and 10−4) and plasmid (106, 105 and 103 copies) dilutions, even between centrally and locally prepared cDNA samples. For the highest dilutions, the Ct values were more reproducible between laboratories for centrally distributed cDNA (10−3 and 10−4 dilutions) and plasmid (10 and 100 copies) than for RNA dilutions. The respective CV were below 5% for cDNA and plasmids and 11% for RNA at the highest dilution. In two target-networks (AML1-ETO and CBFB-MYH11 type A), this observation even resulted in significantly fewer positive results for the samples for which RT was performed locally compared to the centrally prepared cDNA samples. The slopes established with cDNA or plasmid dilutions were close to the theoretical slope −3.32 (100% efficiency) for the vast majority of the participating laboratories. In contrast, slopes from RNA dilutions were indicative of lower reaction efficiency probably due to RNA degradation during transportation. Finally, the sensitivity levels of RNA dilutions for all targets were comparable to standardized nested PCRs designed in the BIOMED-1 program24 and were even better for bcr2 and bcr3 variants of PML-RARA. A total of 10 plasmid copies could generally be detected by all laboratories (see web site, phase IIIa and b).
10.2. Balanced randomized assay (phase IIIb)
This assay was set up in order to detect differences in transcript quantification between laboratories. The laboratories were randomly chosen to amplify four different FG transcripts, generally outside their original FG networks (see the Materials and methods section). This methodology is also of importance as a QC round to detect if false negativity and positivity are proportionally identical to other phases for which only laboratories focusing on a particular FG were performing the experiments.
10.2.1. ABL amplification
10.2.1.1. On plasmid dilutions
We observed very similar results for the three ABL plasmid dilutions among the 25 laboratories. We found, for the ABL 105 copies plasmid dilution, 22.30±0.38 (Ct±s.d., n=296) and a corresponding CV of 1.7%. We found a significant laboratory effect (set as a random effect) on ABL Ct measurement for the three ABL plasmid dilution (P<0.001, n=296). But the difference between opposites laboratories was no more than 1.2 Ct. Such a difference might well not be relevant in a clinical point of view. According to the criteria defined in the Materials and methods section, all laboratories had reproducible results for Ct value of each ABL plasmid dilution (see web site).
10.2.1.2. On coded RNA samples
In the same way, we focused within each network on a laboratory effect for quantification of a given gene transcript. We used ABL gene amplification as a model to investigate this. No significant difference was found within networks between highly diluted samples (10−3 and 10−4) and negative samples when ABL expression was compared using Ct or CN values (results of PML-RARA network given as an example in Figure 30a), except for the CBFB-MYH11 network (P=0.03). On the contrary, we found highly significant differences (P<0.001) between laboratories within each network for the same samples (results of PML-RARA network given as an example in Figure 30b). Thus Ct and CN values were dependent on the testing laboratory and not on the sample. These data strongly suggest that variations occur during the pre-PCR steps (transportation and RT efficiency).
10.2.2. Fusion gene transcript amplification
We focused on the best parameter (Ct, CN, ΔCt or NCN) to express RQ-PCR results in a serial dilution model. We calculated the corresponding correlation coefficients on three diluted samples measured in 11 different laboratories (PML-RARA network as an example in Figure 31). The correlation coefficients were higher for the NCN method in five FGs, slightly higher with the ΔCt method in two FGs, whereas both methods were equivalent for the remaining two FGs (Table 33). Only for quantification of E2A-PBX1 transcripts, ABL appeared not to be helpful as CG to normalize the RQ-PCR results. Overall, use of ABL gene transcripts for normalization was found to greatly improve the linearity of the results (PML-RARA network as an example in Figure 32) and avoided overlap in FG Ct and CN values observed between different dilutions (see web site).
Finally, we checked the results of this model for MRD quantification. The correlation curve between Ct value (Y-axis) and CN value (log10, X-axis) for wells related to coded FG-positive samples (n=824, covering all nine FG targets) showed a mean slope and an intercept of −3.35 and 39.7, respectively. These good results indicated that even with nine different plasmid sets (one per transcript), the quantification of the FG transcripts in the coded RNA samples was similar whatever the plasmid set was.
10.3. Fusion gene expression levels in diagnostic leukemia samples
We compared (Wilcoxon test) the 55 paired BM and PB samples taken at the time of diagnosis, mainly from E2A PBX1- and TEL AML1-positive patients. Two additional pairs in MLL-AF4 network were not analyzed due to insufficient number. The statistical analysis revealed that sample source had no significant impact on the expression level at diagnosis of the relevant FG expressed as NCN, except for PML-RARA using B2M or GUS as the CG (see web site). These analyses would suggest that either source of diagnostic material could act as a suitable reference for MRD studies. Overall, median NCN of each FG transcript was variable, ranging from 8.5 copies (E2A-PBX1) down to 0.1 copy (SIL-TAL1) per copy of ABL gene transcript (Figure 33a). Such result was also observed with B2M and GUS CG. Relative expression levels of the different FG were generally consistent irrespective of the CG used for normalization (Figure 33b,c), although some heterogeneity was observed in a few cases (see PML-RARA, AML1-ETO and SIL-TAL1 sections). In most cases, NCN of the FG transcript in patient samples were comparable to those observed in the corresponding cell line controls (see each section). These results were corrected according to the blast cell proportion in the sample when the information was available.
10.4. RNA dilution series in E. coli RNA
FG-positive cell line or patient RNA was diluted in E. coli RNA, using 10-fold dilution steps (10−1–10−6, see the Materials and methods section). We observed a concordance in the limiting dilution experiments between the last positive dilution of the cell line or patient RNA and the sensitivity predicted on the basis of FG and CG levels in the corresponding undiluted RNA (see Section 11.3). These data indicated that the quantification of the CG and the FG with CN in the pure sample was correct. A linear regression analysis on the same dilutions showed that the FG/CG ratio did not change significantly (less than a factor two) over 3–5 logarithmic dilutions depending on the FG expression level in the pure sample. These results indicated equal RQ-PCR amplification efficiencies for the FG and the CG on a large dilution range.
As molecular analysis plays an increasing role in therapeutic stratification in clinical trials, standardization and external QC programs become mandatory. This large-scale collaborative study of leading laboratories in the field of RQ-PCR-based detection of FG transcripts is the first ever published on this scale in the molecular oncology field. It is an excellent example on how European collaboration can stimulate progress. Furthermore, this initiative has recently been complemented by a collaboration involving 47 expert European laboratories aiming to achieve standardization in the detection of Ig and TCR gene rearrangements as DNA targets for molecular diagnosis in lymphoproliferative disorders (BIOMED-2 Concerted Action).
11.1. Overall achievements
The EAC program revealed that the PCR component of the RQ-PCR assay was very robust as demonstrated by use of the plasmid calibrators. The ABL plasmid amplification during phase IIIb suggested that the threshold of interlaboratory reproducibility for gene expression quantification is a factor two (one Ct). Such precision should be sufficient for clinical application. EAC standardization was obtained leading to the development of a common RQ-PCR protocol followed by the 26 member laboratories. Such common protocols (Table 4) and the use of robust calibrators by each participant enabled reliable comparison of results. We have selected EAC primer and probe sets (see tables in each particular section) allowing the detection by RQ-PCR of 15 different FG transcripts (nine FGS, with six breakpoint variants), corresponding to the main leukemia-associated translocations. The selected primer and probe sets using the common protocol showed a sensitivity of 10 plasmid molecules or 10−4 RNA cell line dilution for the majority of the targets. Three reference genes (ABL, B2M and GUS) have also been selected out of 17 tested based on their high expression stability in PB and BM of normal and patient samples (see accompanying manuscript by Beillard et al50). This program was designed for MRD studies and the EAC primer and probe sets have not been designed for the detection of each FG variant at initial diagnosis. Consequently, the EAC sets could lead to false-negative results if used for molecular screening at diagnosis, failing to detect approximately 25% cases with underlying MLL-AF4 FG transcripts as well as some rare variants of BCR-ABL, CBFB-MYH11 and PML-RARA. Therefore for screening at initial diagnosis, the classical BIOMED-1 primers might be used.24
11.1.1. QC data: false positivity and false negativity
The QC analysis was performed on a large series of coded samples. False negativity was 1% for the 10−3 dilution (316 tested samples) and 3.8% for the 10−4 dilution (366 tested samples) (Table 34).
False positivity was observed in 6.0% (34/546) of the FG-negative RNA samples tested, always with low expression (Ct>30) and mainly restricted to few laboratories and FG networks (Table 35 ), suggesting a cross-contamination during the experimental set-up (Table 34). The false positivity was lower in NAC/NTC samples (3.2%) than in FG-negative RNA samples (6.0%), and could be related to the additional RT step associated with an increasing risk of contamination. False positivity was clearly a potential problem for the reliable detection of low levels of FG transcripts. One likely explanation is the large number of samples tested per laboratory in each experiment, with the inherent risk of cDNA carry-over between adjacent wells of the reaction plate. In any case, our data emphasize the need for very strict manipulation rules, even when using a closed system such as in RQ-PCR. Furthermore, negative controls (NTC, NAC) are mandatory in each test but patient sample analysis should also ideally include ‘RT negative’ controls, whereby patient RNA is subjected to all stages of the RQ-PCR protocol with the omission of the RT enzyme in the cDNA synthesis step.
11.1.2. QC data: reproducibility of RQ-PCR results
The reproducibility of the RQ-PCR experiments was assessed by the balanced randomized assay (phase IIIb). Despite all our efforts to reduce variability, we still noticed significant differences between results for some laboratories. We observed that these discrepancies were related to the pre-PCR steps and are likely to have resulted from RNA degradation during storage and shipment, and/or intra- and interlaboratory variations in RT efficiency. The use of a CG to normalize results (Table 33) greatly improved the data reproducibility between laboratories suggesting that variations in RNA quality/quantity and RT efficiency can at least in part be compensated by normalization.
Even though normalization can reduce the effect of RNA degradation or inefficiency of the RT or PCR steps, it should be kept in mind that RNA degradation and RT efficiency can dramatically reduce the sensitivity of MRD detection.
During phase IIIb, the QC round was performed by a number of laboratories with no previous experience of PCR amplification of the involved FG transcripts. We did not notice significant changes in false positivity or false negativity proportions compared to other phases. These results suggest that the RQ-PCR protocol can be easily used for MRD detection by any other laboratory with the same sensitivity.
11.1.3. Expression level of fusion gene transcripts at diagnosis
We have for the first time compared the level of expression of the main FG transcripts in cell lines and a large series of leukemic samples at diagnosis (n=278) correlating FG expression with that of three reference genes (ABL, B2M, GUS). RQ-PCR revealed that marked differences in FG expression between transcripts in leukemic samples at diagnosis (Figure 33) could account for differential assay sensitivity. It would probably contribute to ‘false-negative’ results described in the literature for FG that are less expressed (eg PML-RARA or SIL-TAL1) and to the detection of FG transcripts in long-term remission in situations in which the FG is more highly expressed (eg AML1-ETO). Such information is likely to prove invaluable for clinical laboratories involved in MRD monitoring and also for basic research focusing on gene expression profiles (using DNA microarrays for example). However, we restricted the analyses to recently stored samples; hence it is currently unclear to what extent RNA degradation could influence the quantification. Over-time stability experiments are in progress to assess whether the degradation kinetics of FG and CG transcripts are similar (V van der Velden et al, in preparation). Furthermore, we compared FG expression in a large series of paired and unpaired diagnostic BM and PB samples. We observed that after correction for blast cell percentage and normalization by a CG, the relative expression of the FG transcript did not differ significantly between different sample sources. This suggests that either BM or PB at diagnosis can be used to estimate the NCN of the FG transcripts in individual cases of leukemia, serving as a reference for subsequent MRD assessments during follow-up.
11.1.4. Choice of the control gene
In the accompanying manuscript50 we show that ABL gene expression did not differ significantly between normal and leukemic samples at diagnosis. Moreover, of the three extensively tested CGs, ABL gene expression had the highest correlation with the FG transcripts in diagnostic samples. In our study, in a model of MRD detection (phase IIIb), normalization of FG expression to that of ABL as the CG improved the reproducibility of FG transcript results obtained in comparison to raw (not normalized) Ct or CN values. Therefore, we propose to use ABL as the CG of first choice for normalization in diagnostic and follow-up samples. As second choice, we recommend to use either B2M or GUS depending on their relative correlation with the respectice FG transcript expression and the variability of the NCN at diagnosis. In the context of BCR-ABL quantification, the use of ABL should give only values up to 1.0 from a theoretical point. Our data, in accordance with the literature, show that practice is different. The values higher than 1.0 are most probably caused by the use of separate PCR methods and separate standards for BCR-ABL and ABL. In practice, identification of isolated samples with a low expression level of CG suggests RNA degradation or presence of inhibitors in such samples. On the other hand, observation of reduced or absent CG amplification for all samples tested is indicative of a reagent problem during the RT or PCR reactions. Finally, the CG transcript CN for each patient RNA sample allows normalization of efficiency of the pre-PCR steps.
11.2. Expression of RQ-PCR data
MRD monitoring by RQ-PCR analysis is becoming a tool for decision-making in multicenter therapeutic trials. For this reason, it is of capital importance for RQ-PCR results to be expressed in a uniform way. In the literature, no study addresses the issue of multicenter RQ-PCR analysis on a large scale (26 laboratories in our EAC network). Most publications show data from a single laboratory, but, to our knowledge, just one study involves the standardization between three laboratories for one target.159 The majority of publications used a copy number ratio between the FG and the CG with a standard curve, this ratio being expressed as a decimal value or a percentage. To obtain the standard curve, laboratories used either cell line cDNA37,39,128,190 or plasmid DNA.38,42,125,191 So far, few authors used the ΔΔCt method49,192 and, to our knowledge, a standard curve of diagnostic cDNA has not been used so far.
In our EAC network, we discussed four possibilities: (1) cell line RNA dilutions, (2) percentage of positive cell number relative to the diagnostic sample, (3) copy number ratios and (4) the ΔΔCt method.
(1) The cell line RNA dilutions appeared to be very sensitive to degradation during transportation as shown in our study. The variability of expression for one cell line can be subject to large variations. Such potential variation depends on the source of the cell line, the timing of cell culture when RNA extraction has been done, and finally the RT efficiency. For multicenter studies, the option to overcome such difficulties could be to centrally prepare and distribute the cDNA.
(2) Results expressed as frequency of positive cells would have the huge advantage to allow direct comparison with other MRD techniques. Furthermore, it would enable more reliable determination of kinetics of FG transcript reduction within individual patients, given the variability in FG transcript expression levels between patients as observed in this study. However, such an approach is dependent on availability of diagnostic material against which relative levels of MRD can be judged. Since the precise level of FG expression and its variations during treatment at the single cell level are not entirely clear, we ultimately decided to express results in terms of ratios between the target (FG transcript) and the reference (CG transcript) in our experiments.
(3) The ratio (NCN) was expressed as FG copies per copy of ABL or GUS gene transcript and per 100 copies of B2M gene transcript due to its high expression level. This ratio should be independent of the starting RNA quantity. For this purpose, we used plasmid standard curves, which offer the possibility to quantify directly the copy number of the transcripts. Our data show that plasmids are suitable calibrators for inter- or intralaboratory normalization of RQ-PCR analysis. Plasmid DNA is probably a good option providing stability and robustness, which is unlikely to be achieved when using large-scale production of cDNA as a potential QC material.
The potential drawbacks of this method are as follows. (i) The risk of contamination, although we used plasmid dilutions containing FG copies within the same range as patient samples. Usual rigorous precautions for PCR analysis are always required for limiting this risk. (ii) The use of plasmid calibrators reduces the number of wells available for patient samples and slightly increases the cost. (iii) The calibrators introduce additional steps/calculations potentially increasing the variability. (iv) DNA plasmids do not directly assess the RT efficiency, but the CG expression level in patient samples clearly represents a control of the pre-PCR steps.
(4) The ΔΔCt method (Applied Biosystems User's bulletin #2) does not have these disadvantages but has its own limitations. The method relies on the relative efficiencies of the FG and CG assays being comparable and consistent from plate to plate; therefore, it is critical that positive RNA or cDNA standards are routinely included, to enable deterioration in assay performance to be detected by a rise in Ct value as encountered once in our study for the analysis of 70 CML samples, including 17 paired samples. This method can be very efficient in expert laboratories and can be used to determine the relative level of MRD in comparison to the diagnostic sample. There are concerns that this approach may not lend itself to assessment of interexperimental variations in the intra- or interlaboratory setting. This may create difficulties in comparing RQ-PCR data between different groups, particularly when different machines are used (there are at least eight providers today).
11.3. Proposal for assessing the sensitivity level of RQ-PCR experiments based on EAC data
When one encounters an absence of FG transcript amplification in a patient sample during follow-up, it is necessary to assess the detection limit for the particular assay to determine the reliability and clinical relevance of the result obtained. To address this issue, we propose two formulae to calculate the sensitivity level of a given experiment. While the ΔΔCt method is explained in the accompanying manuscript,50 we report here the use of copy number values.
The formula is based on the results of E. coli dilution experiments and the correlation between FG CN and CG CN with a slope close to 1 in diagnostic samples (see each particular section). In this formula, the sensitivity is directly related to the NCN of the FG at diagnosis and the CG CN of the sample. Ideally, the calculation should be based on the patient's diagnostic NCN, after correction for blast percentage. If not available, EAC data can be used. In this model, 10 copies of the FG plasmid should be amplified for any particular fusion transcript. If only 100 copies can be amplified, the sensitivity should be reduced by one log10:
In this formula, SENS is the sensitivity (log10) of the experiment for the diagnostic sample and should be expressed as 10SENS. NCN is either the ratio of the patient sample at diagnosis or if not available the corresponding median NCN from the EAC data.
The formula is valid for all CG at diagnosis, but one should be aware of the bias toward underestimation for BCR-ABL/ABL ratio for samples containing a high level of leukemic cells. However, only the ABL gene did not show any significant difference between BM and PB and between normal samples and leukemic samples at diagnosis (see accompanying manuscript).50 Thus this formula can be used only with ABL as CG without any correction for assessing the sensitivity level of the experiment during the follow-up.
11.3.3. Three examples through real cases
126.96.36.199. At diagnosis
Patient A presents a pediatric T-ALL at diagnosis. The search of SIL-TAL1 FG transcript in its PB sample remains negative. The quantification of ABL gene transcript using RQ-PCR with our EAC protocol is 46 000 copies. Thus the estimated sensitivity of the experiment based on EAC data for the median SIL-TAL1 FG transcript expression in PB (0.09, Table 22) at diagnosis is
188.8.131.52. At relapse
Patient B presents a late relapse of a TEL-AML1-positive precursor-B-ALL. The quantification of TEL-AML1 FG transcript in its PB sample is 419 000 copies. The quantification of ABL gene transcript in the same sample is 17 000 copies. The TEL-AML1/ABL ratio for this patient is 25 (419 000/17 000). Thus the estimated sensitivity based on TEL-AML1 FG expression in this patient is
184.108.40.206. During follow-up
The same patient B is followed 3 months later by RQ-PCR for the detection of TEL-AML1 FG transcript in PB and BM samples. TEL-AML1 FG transcript is not detected by RQ-PCR in both samples. The results of ABL gene transcript quantification on the PB and BM samples are respectively 7700 and 13 700 copies. Based on the observation that TEL-AML1 NCN does not differ significantly between PB and BM (Figure 11), the ratio in this patient at relapse (419 000/17 000) is used to calculate the sensitivity of this experiment:
Compared to classical RT-PCR FG transcript follow-up, the sensitivity in this case relies on patient and not on cell line samples. The sensitivity threshold calculated with this methodology is clearly more accurate than the one based on classical RT-PCR on cell line dilutions.24
11.4. EAC protocol in clinical laboratories involved in current therapeutic trials for leukemias
Current therapeutic protocols for leukemias tend to be prognosis adapted, one consequence being that large number of institutions are required to enroll patients to ensure that sufficient sample sizes are reached in order to address reliably which randomized treatment approaches afford a superior outcome. Analysis of MRD appears to be a key prognostic indicator allowing treatment optimization. Today, there is no standardized technique for fusion transcript detection during MRD follow-up. Our EAC protocol could provide the basis for an international reference of MRD studies using RQ-PCR analysis of FG transcripts. Indeed, up to 80 laboratories from 30 countries have already applied, upon a confidentiality agreement, for details of the methods employed in this study. Now, we are in a position to assess the clinical relevance of MRD analysis for the main FG transcripts in leukemia within therapeutic protocols. One direct application will be therapeutic trials with the innovative tyrosine kinase inhibitor (Imatinib) for BCR-ABL-positive ALL in Europe where the efficacy of each protocol will be followed at the molecular level using the EAC protocol. We would like to suggest that in the future, therapeutic strategies will be adapted according to such standardized MRD evaluations. It would be a great improvement if such biological data assist in therapeutic decision-making, such as in unrelated allogeneic transplant, infusion of donor lymphocytes or basically the choice of the most efficient therapeutic strategy.
In this large multicenter study, we aimed to standardize the only quantitative step of a sample RQ-PCR analysis: the PCR step. Assays were performed by the CG group on the RT step (see accompanying manuscript50). This standardization effort is clearly a first step since notably all the previous steps from harvesting to RNA extraction were not studied. We believe that such standardization protocol on the PCR step should allow data comparison by measuring the quality of the sample.50 Clearly in order to improve the quality of samples for such analysis in a routine laboratory, a large effort of optimization and standardization remains to be done.
Our EAC protocol has been set up using an ABI 7700 platform, but the TaqMan technology (5′ nuclease assay) employed in this study can be in principle applied to various machines available today. The next step will be an inter-RQ-PCR machine standardization using robust calibrators. Furthermore, this standardization and QC program in leukemia could be a model for other biological markers in onco-hematology and more broadly in the oncology field. Standardization and quality control programs for novel technologies such as semiautomated MRD detection today, but also high-throughput chip DNA technologies tomorrow, are mandatory in ensuring that advances achieved through innovative genomic methodologies yield maximal benefit in improving the outcome of patients with leukemia.
Lowenberg B, Downing JR, Burnett A . Acute myeloid leukemia. N Engl J Med 1999; 341: 1051–1062.
Verma A, Stock W . Management of adult acute lymphoblastic leukemia: moving toward a risk-adapted approach. Curr Opin Oncol 2001; 13: 14–20.
Appelbaum FR . Perspectives on the future of chronic myeloid leukemia treatment. Semin Hematol 2001; 38: 35–42.
Grimwade D, Walker H, Oliver F, Wheatley K, Harrison C, Harrison G et al. The importance of diagnostic cytogenetics on outcome in AML: analysis of 1,612 patients entered into the MRC AML 10 trial. The Medical Research Council Adult and Children's Leukaemia Working Parties. Blood 1998; 92: 2322–2333.
Groupe FdCH. Cytogenetic abnormalities in adult acute lymphoblastic leukemia: correlations with hematologic finding and outcome. A collaborative study of the Groupe Français de Cytogénétique Hématologique. Blood 1996; 87: 3135–3142.
Nakao M, Yokota S, Iwai T, Kaneko H, Horiike S, Kashima K et al. Internal tandem duplication of the flt3 gene found in acute myeloid leukemia. Leukemia 1996; 10: 1911–1918.
Kottaridis PD, Gale RE, Frew ME, Harrison G, Langabeer SE, Belton AA et al. The presence of a FLT3 internal tandem duplication in patients with acute myeloid leukemia (AML) adds important prognostic information to cytogenetic risk group and response to the first cycle of chemotherapy: analysis of 854 patients from the United Kingdom Medical Research Council AML 10 and 12 trials. Blood 2001; 98: 1752–1759.
Abshire TC, Buchanan GR, Jackson JF, Shuster JJ, Brock B, Head D et al. Morphologic, immunologic and cytogenetic studies in children with acute lymphoblastic leukemia at diagnosis and relapse: a Pediatric Oncology Group study. Leukemia 1992; 6: 357–362.
Hoelzer D, Thiel E, Loffler H, Buchner T, Ganser A, Heil G et al. Prognostic factors in a multicenter study for treatment of acute lymphoblastic leukemia in adults. Blood 1988; 71: 123–131.
Liu Yin JA, Wheatley K, Rees JK, Burnett AK . Comparison of ‘sequential’ versus ‘standard’ chemotherapy as re-induction treatment, with or without cyclosporine, in refractory/relapsed acute myeloid leukaemia (AML): results of the UK Medical Research Council AML-R trial. Br J Haematol 2001; 113: 713–726.
Estey EH . New drugs for therapy of AML. Leukemia 2002; 16: 306–309.
Wheatley K, Burnett AK, Goldstone AH, Gray RG, Hann IM, Harrison CJ et al. A simple, robust, validated and highly predictive index for the determination of risk-directed therapy in acute myeloid leukaemia derived from the MRC AML 10 trial. United Kingdom Medical Research Council's Adult and Childhood Leukaemia Working Parties. Br J Haematol 1999; 107: 69–79.
Kantarjian H, Smith T, OBrien S, Beran M, Pierce S, Talpaz M . Prolonged survival in chronic myelogenous leukemia following cytogenetic response to a interferon therapy. Ann Intern Med 1995; 122: 254–261.
Italian Cooperative Study Group on Chronic Myeloid Leukemia and Italian Group for Bone Marrow Transplantation. Monitoring treatment and survival in chronic myeloid leukemia. J Clin Oncol 1999; 17: 1858–1868.
Campana D, Pui C-H . Detection of minimal residual disease in acute leukemia: methodologic advances and clinical significance. Blood 1995; 85: 1416–1434.
Szczepanski T, Orfao A, van der Velden VH, San Miguel JF, van Dongen JJ . Minimal residual disease in leukaemia patients. Lancet Oncol 2001; 2: 409–417.
Cave H, van der Werff ten Bosch J, Suciu S, Guidal C, Waterkeyn C, Otten J et al. Clinical significance of minimal residual disease in childhood acute lymphoblastic leukemia. European Organization for Research and Treatment of Cancer – Childhood Leukemia Cooperative Group [see comments]. N Engl J Med 1998; 339: 591–598.
van Dongen JJ, Seriu T, Panzer-Grumayer ER, Biondi A, Pongers-Willemse MJ, Corral L et al. Prognostic value of minimal residual disease in acute lymphoblastic leukaemia in childhood. Lancet 1998; 352: 1731–1738.
Coustan-Smith E, Behm FG, Sanchez J, Boyett JM, Hancock ML, Raimondi SC et al. Immunological detection of minimal residual disease in children with acute lymphoblastic leukaemia. Lancet 1998; 351: 550–554.
Radich JP, Gehly G, Gooley T, Bryant E, Clift RA, Collins S et al. PCR detection of the BCR-ABL fusion transcript after allogeneic bone marrow transplantation for chronic myeloid leukemia: results and implications in 346 patients. Blood 1995; 85: 2632–2638.
Diverio D, Rossi V, Avvisati G, De Santis S, Pistilli A, Pane F et al. Early detection of relapse by prospective reverse transcriptase-polymerase chain reaction analysis of the PML/RARalpha fusion gene in patients with acute promyelocytic leukemia enrolled in the GIMEMA-AIEOP multicenter ‘AIDA’ trial. GIMEMA-AIEOP Multicenter ‘AIDA’ Trial. Blood 1998; 92: 784–789.
Lo Coco F, Diverio D, Avvisati G, Petti MC, Meloni G, Pogliani EM et al. Therapy of molecular relapse in acute promyelocytic leukemia. Blood 1999; 94: 2225–2229.
San Miguel J, Van Dongen J, Bartram C, Parreira A, Wormann B, Biondi A et al. In: I. press (ed). Investigation of Minimal Residual Disease (MRD) in Acute Leukemia (AL): International Standardization and Evaluation Cancer research supported under BIOMED1. Amsterdam: Baig SS, 1998, pp 300–306.
van Dongen JJ, Macintyre EA, Gabert JA, Delabesse E, Rossi V, Saglio G et al. Standardized RT-PCR analysis of fusion gene transcripts from chromosome aberrations in acute leukemia for detection of minimal residual disease. Report of the BIOMED-1 Concerted Action: investigation of minimal residual disease in acute leukemia. Leukemia 1999; 13: 1901–1928.
Jurlander J, Caligiuri MA, Ruutu T, Baer MR, Strout MP, Oberkircher AR et al. Persistence of the AML1/ETO fusion transcript in patients treated with allogeneic bone marrow transplantation for t(8;21) leukemia. Blood 1996; 88: 2183–2191.
Miyamoto T, Nagafuji K, Akashi K, Harada M, Kyo T, Akashi T et al. Persistence of multipotent progenitors expressing AML1/ETO transcripts in long-term remission patients with t(8;21) acute myelogenous leukemia. Blood 1996; 87: 4789–4796.
Tobal K, Saunders MJ, Grey MR, Yin JA . Persistence of RAR alpha-PML fusion mRNA detected by reverse transcriptase polymerase chain reaction in patients in long-term remission of acute promyelocytic leukaemia. Br J Haematol 1995; 90: 615–618.
Biernaux C, Loos M, Sels A, Huez G, Stryckmans P . Detection of major bcr-abl gene expression at a very low level in blood cells of some healthy individuals. Blood 1995; 86: 3118–3122.
Bose S, Deininger M, Gora-Tybor J, Goldman JM, Melo JV . The presence of typical and atypical BCR-ABL fusion genes in leukocytes of normal individuals: biologic significance and implications for the assessment of minimal residual disease. Blood 1998; 92: 3362–3367.
Zimmermann K, Mannhalter JW . Technical aspects of quantitative competitive PCR. Biotechniques 1996; 21: 268–272, 274–269.
Lion T, Henn T, Gaiger A, Kalhs P, Gadner H . Early detection of relapse after bone marrow transplantation in patients with chronic myelogenous leukaemia. Lancet 1993; 341: 275–276.
Hochhaus A, Reiter A, Saussele S, Reichert A, Emig M, Kaeda J et al. Molecular heterogeneity in complete cytogenetic responders after interferon-alpha therapy for chronic myelogenous leukemia: low levels of minimal residual disease are associated with continuing remission. German CML Study Group and the UK MRC CML Study Group. Blood 2000; 95: 62–66.
Laczika K, Novak M, Hilgarth B, Mitterbauer M, Mitterbauer G, Scheidel-Petrovic A et al. Competitive CBFbeta/MYH11 reverse-transcriptase polymerase chain reaction for quantitative assessment of minimal residual disease during postremission therapy in acute myeloid leukemia with inversion(16): a pilot study. J Clin Oncol 1998; 16: 1519–1525.
Tobal K, Liu Yin JA . Molecular monitoring of minimal residual disease in acute myeloblastic leukemia with t(8;21) by RT-PCR. Leuk Lymphoma 1998; 31: 115–120.
Gabert J . Detection of recurrent translocations using real time PCR; assessment of the technique for diagnosis and detection of minimal residual disease. Haematologica 1999; 84: 107–109.
van der Velden V, Hochhaus A, Cazzanigia G, Szczepanski T, Gabert J, van Dongen J . Detection of minimal residual disease in hematopoietic malignancies by real-time quantitative PCR: principles, approaches, and laboratory aspects. Leukemia 2003; 17: 1013–1034.
Mensink E, van de Locht A, Schattenberg A, Linders E, Schaap N, Geurts van Kessel A et al. Quantitation of minimal residual disease in Philadelphia chromosome positive chronic myeloid leukaemia patients using real-time quantitative RT-PCR. Br J Haematol 1998; 102: 768–774.
Preudhomme C, Revillion F, Merlat A, Hornez L, Roumier C, Duflos-Grardel N et al. Detection of BCR-ABL transcripts in chronic myeloid leukemia (CML) using a ‘real time’ quantitative RT-PCR assay. Leukemia 1999; 13: 957–964.
Radich JP, Gooley T, Bryant E, Chauncey T, Clift R, Beppu L et al. The significance of bcr-abl molecular detection in chronic myeloid leukemia patients ‘late,’ 18 months or more after transplantation. Blood 2001; 98: 1701–1707.
Yokota H, Tsuno NH, Tanaka Y, Fukui T, Kitamura K, Hirai H et al. Quantification of minimal residual disease in patients with e1a2 BCR-ABL-positive acute lymphoblastic leukemia using a real-time RT-PCR assay. Leukemia 2002; 16: 1167–1175.
Cassinat B, Zassadowski F, Balitrand N, Barbey C, Rain JD, Fenaux P et al. Quantitation of minimal residual disease in acute promyelocytic leukemia patients with t(15;17) translocation using real-time RT-PCR. Leukemia 2000; 14: 324–328.
Marcucci G, Livak KJ, Bi W, Strout MP, Bloomfield CD, Caligiuri MA . Detection of minimal residual disease in patients with AML1/ETO-associated acute myeloid leukemia using a novel quantitative reverse transcription polymerase chain reaction assay. Leukemia 1998; 12: 1482–1489.
Krauter J, Wattjes MP, Nagel S, Heidenreich O, Krug U, Kafert S et al. Real-time RT-PCR for the detection and quantification of AML1/MTG8 fusion transcripts in t(8;21)-positive AML patients. Br J Haematol 1999; 107: 80–85.
Sugimoto T, Das H, Imoto S, Murayama T, Gomyo H, Chakraborty S et al. Quantitation of minimal residual disease in t(8;21)-positive acute myelogenous leukemia patients using real-time quantitative RT-PCR. Am J Hematol 2000; 64: 101–106.
Marcucci G, Caligiuri MA, Dohner H, Archer KJ, Schlenk RF, Dohner K et al. Quantification of CBFbeta/MYH11 fusion transcript by real time RT-PCR in patients with INV(16) acute myeloid leukemia. Leukemia 2001; 15: 1072–1080.
Buonamici S, Ottaviani E, Testoni N, Montefusco V, Visani G, Bonifazi F et al. Real-time quantitation of minimal residual disease in inv(16)-positive acute myeloid leukemia may indicate risk for clinical relapse and may identify patients in a curable state. Blood 2002; 99: 443–449.
Drunat S, Olivi M, Brunie G, Grandchamp B, Vilmer E, Bieche I et al. Quantification of TEL-AML1 transcript for minimal residual disease assessment in childhood acute lymphoblastic leukaemia. Br J Haematol 2001; 114: 281–289.
Seeger K, Kreuzer KA, Lass U, Taube T, Buchwald D, Eckert C et al. Molecular quantification of response to therapy and remission status in TEL-AML1-positive childhood ALL by real-time reverse transcription polymerase chain reaction. Cancer Res 2001; 61: 2517–2522.
Pallisgaard N, Clausen N, Schroder H, Hokland P . Rapid and sensitive minimal residual disease detection in acute leukemia by quantitative real-time RT-PCR exemplified by t(12;21) TEL-AML1 fusion transcript. Genes Chromosomes Cancer 1999; 26: 355–365.
Beillard E, Pallisgaard N, Bi W, van der Velden VHJ, Dee R, van der Schoot CE et al. Evaluation of candidate control genes for diagnosis and residual disease detection in leukemic patients using ‘real-time’ quantitative reverse-transcriptase polymerase chain reaction (RQ-PCR) – A Europe Against Cancer Program. Leukemia 2003, (in press).
Livak K, Flood S, Marmaro J, Giusti W, Deetz K . Oligonucleotides with fluorescent dyes at apposite ends provides a quenched probe system useful for detecting PCR product and nucleic acid hybridization. PCR Meth Appl 1995; 4: 357–362.
van der Velden VHJ, Szczepanski T, van Dongen JJM . Polymerase chain reaction, real-time quantitative. In: Brenner S, Miller JH (eds). Encyclopedia of Genetics. London: Academic Press, 2001, pp 1503–1506.
Saussele S, Weisser A, Muller MC, Emig M, La Rosee P, Paschka P et al. Frequent polymorphism in BCR exon b2 identified in BCR-ABL positive and negative individuals using fluorescent hybridization probes. Leukemia 2000; 14: 2006–2010.
Mellentin JD, Nourse J, Hunger SP, Smith SD, Cleary ML . Molecular analysis of the t(1;19) breakpoint cluster region in pre-B cell acute lymphoblastic leukemias. Genes Chromosomes Cancer 1990; 2: 239–247.
Nourse J, Mellentin JD, Galili N, Wilkinson J, Stanbridge E, Smith SD et al. Chromosomal translocation t(1;19) results in synthesis of a homeobox fusion mRNA that codes for a potential chimeric transcription factor. Cell 1990; 60: 535–545.
Kamps MP, Murre C, Sun XH, Baltimore D . A new homeobox gene contributes the DNA binding domain of the t(1;19) translocation protein in pre-B ALL. Cell 1990; 60: 547–555.
Crist WM, Carroll AJ, Shuster JJ, Behm FG, Whitehead M, Vietti TJ et al. Poor prognosis of children with pre-B acute lymphoblastic leukemia is associated with the t(1;19)(q23;p13): a Pediatric Oncology Group study. Blood 1990; 76: 117–122.
Hunger SP, Sun T, Boswell AF, Carroll AJ, McGavran L . Hyperdiploidy and E2A-PBX1 fusion in an adult with t(1;19)+ acute lymphoblastic leukemia: case report and review of the literature. Genes Chromosomes Cancer 1997; 20: 392–398.
Borowitz MJ, Hunger SP, Carroll AJ, Shuster JJ, Pullen DJ, Steuber CP et al. Predictability of the t(1;19)(q23;p13) from surface antigen phenotype: implications for screening cases of childhood acute lymphoblastic leukemia for molecular analysis: a Pediatric Oncology Group study. Blood 1993; 82: 1086–1091.
Privitera E, Kamps MP, Hayashi Y, Inaba T, Shapiro LH, Raimondi SC et al. Different molecular consequences of the 1;19 chromosomal translocation in childhood B-cell precursor acute lymphoblastic leukemia. Blood 1992; 79: 1781–1788.
Privitera E, Luciano A, Ronchetti D, Arico M, Santostasi T, Basso G et al. Molecular variants of the 1;19 chromosomal translocation in pediatric acute lymphoblastic leukemia (ALL). Leukemia 1994; 8: 554–559.
Hunger SP, Galili N, Carroll AJ, Crist WM, Link MP, Cleary ML . The t(1;19)(q23;p13) results in consistent fusion of E2A and PBX1 coding sequences in acute lymphoblastic leukemias. Blood 1991; 77: 687–693.
Izraeli S, Kovar H, Gadner H, Lion T . Unexpected heterogeneity in E2A/PBX1 fusion messenger RNA detected by the polymerase chain reaction in pediatric patients with acute lymphoblastic leukemia. Blood 1992; 80: 1413–1417.
Van Dijk MA, Voorhoeve PM, Murre C . Pbx1 is converted into a transcriptional activator upon acquiring the N-terminal region of E2A in pre-B-cell acute lymphoblastoid leukemia. Proc Natl Acad Sci USA 1993; 90: 6061–6065.
Monica K, LeBrun DP, Dedera DA, Brown R, Cleary ML . Transformation properties of the E2a-Pbx1 chimeric oncoprotein: fusion with E2a is essential, but the Pbx1 homeodomain is dispensable. Mol Cell Biol 1994; 14: 8304–8314.
Dedera DA, Waller EK, LeBrun DP, Sen-Majumdar A, Stevens ME, Barsh GS et al. Chimeric homeobox gene E2A-PBX1 induces proliferation, apoptosis, and malignant lymphomas in transgenic mice. Cell 1993; 74: 833–843.
Kamps MP, Baltimore D . E2A-Pbx1, the t(1;19) translocation protein of human pre-B-cell acute lymphocytic leukemia, causes acute myeloid leukemia in mice. Mol Cell Biol 1993; 13: 351–357.
Lanza C, Gottardi E, Gaidano C, Vivenza C, Parziale A, Perfetto F et al. Persistence of E2A/PBX1 transcripts in t(1;19) childhood acute lymphoblastic leukemia: correlation with chamotherapy intensity and clinical outcome. Leukemia Res 1996; 20: 441–443.
Izraeli S, Janssen JW, Haas OA, Harbott J, Brok SF, Walther JU et al. Detection and clinical relevance of genetic abnormalities in pediatric acute lymphoblastic leukemia: a comparison between cytogenetic and polymerase chain reaction analyses. Leukemia 1993; 7: 671–678.
Hunger SP, Fall MZ, Camitta BM, Carroll AJ, Link MP, Lauer SJ et al. E2A-PBX1 chimeric transcript status at end of consolidation is not predictive of treatment outcome in childhood acute lymphoblastic leukemias with a t(1;19)(q23;p13): a Pediatric Oncology Group study. Blood 1998; 91: 1021–1028.
Privitera E, Rivolta A, Ronchetti D, Mosna G, Giudici G, Biondi A . Reverse transcriptase/polymerase chain reaction follow up and minimal residual disease detection in t(1;19)-positive acute lymphoblastic leukaemia. Br J Haematol 1996; 92: 653–658.
Devaraj PE, Foroni L, Janossy G, Hoffbrand AV, Secker-Walker LM . Expression of the E2A-PBX1 fusion transcripts in t(1;19)(q23;p13) and der(19)t(1;19) at diagnosis and in remission of acute lymphoblastic leukemia with different B lineage immunophenotypes. Leukemia 1995; 9: 821–825.
Rowley JD . Molecular genetics in acute leukemia. Leukemia 2000; 14: 513–517.
Yu BD, Hess JL, Horning SE, Brown GAJ, Korsmeyer SJ . Altered Hox expression and segmental identity in MLL-mutant mice. Nature 1995; 378: 505–508.
Isnard P, Core N, Naquet P, Djabali M . Altered lymphoid development in mice deficient for the mAF4 proto-oncogene. Blood 2000; 96: 705–710.
Kersey JH, Wang D, Oberto M . Resistance of t(4;11) (MLL-AF4 fusion gene) leukemias to stress-induced cell death: possible mechanism for extensive extramedullary accumulation of cells and poor prognosis. Leukemia 1998; 12: 1561–1564.
Uckun FM, Herman-Hatten K, Crotty ML, Sensel MG, Sather HN, Tuel-Ahlgren L et al. Clinical significance of MLL-AF4 fusion transcript expression in the absence of a cytogenetically detectable t(4;11)(q21;q23) chromosomal translocation. Blood 1998; 92: 810–821.
Trka J, Zuna J, Hrusak O, Michalova K, Muzikova K, Kalinova M et al. No evidence for MLL/AF4 expression in normal cord blood samples. Blood 1999; 93: 1106–1107, discussion 1108–1110.
Kim-Rouille MH, MacGregor A, Wiedemann LM, Greaves MF, Navarrete C . MLL-AF4 gene fusions in normal newborns. Blood 1999; 93: 1107–1108.
Cimino G, Elia L, Rapanotti MC, Sprovieri T, Mancini M, Cuneo A et al. A prospective study of residual-disease monitoring of the ALL1/AF4 transcript in patients with t(4;11) acute lymphoblastic leukemia. Blood 2000; 95: 96–101.
Romana SP, Poirel H, Leconiat M, Flexor MA, Mauchauffe M, Jonveaux P et al. High frequency of t(12;21) in childhood B-lineage acute lymphoblastic leukemia. Blood 1995; 86: 4263–4269.
Raynaud S, Mauvieux L, Cayuela J, Bastard C, Bilhou-Nabera C, Debuire B et al. TEL/AML1 fusion gene is a rare event in adult acute lymphoblastic leukemia. Leukemia 1996; 10: 1529–1530.
Borkhardt A, Cazzaniga G, Viehmann S, Valsecchi MG, Ludwig WD, Burci L et al. Incidence and clinical relevance of TEL/AML1 fusion genes in children with acute lymphoblastic leukemia enrolled in the German and Italian multicenter therapy trials. Associazione Italiana Ematologia Oncologia Pediatrica and the Berlin–Frankfurt–Munster Study Group. Blood 1997; 90: 571–577.
Heibert SW, Lutterbach B, Durst K, Wang L, Linggi B, Wu S et al. Mechanisms of transcriptional repression by the t(8;21)-, t(12;21)-, and inv(16)-encoded fusion proteins. Cancer Chemother Pharmacol 2001; 48 (Suppl 1): S31–S34.
Satake N, Kobayashi H, Tsunematsu Y, Kawasaki H, Horikoshi Y, Koizumi S et al. Minimal residual disease with TEL-AML1 fusion transcript in childhood acute lymphoblastic leukaemia with t(12;21). Br J Haematol 1997; 97: 607–611.
Harbott J, Viehmann S, Borkhardt A, Henze G, Lampert F . Incidence of TEL/AML1 fusion gene analyzed consecutively in children with acute lymphoblastic leukemia in relapse. Blood 1997; 90: 4933–4937.
Seeger K, Adams HP, Buchwald D, Beyermann B, Kremens B, Niemeyer C et al. TEL-AML1 fusion transcript in relapsed childhood acute lymphoblastic leukaemia. The Berlin–Frankfurt–Munster Study Group. Blood 1998; 91: 1716–1722.
Loh ML, Silverman LB, Young ML, Neuberg D, Golub TR, Sallan SE et al. Incidence of TEL/AML1 fusion in children with relapsed acute lymphoblastic leukemia. Blood 1998; 92: 4792–4797.
Maloney K, McGavran L, Murphy J, Odom L, Stork L, Wei Q et al. TEL-AML1 fusion identifies a subset of children with standard risk acute lymphoblastic leukemia who have an excellent prognosis when treated with therapy that includes a single delayed intensification. Leukemia 1999; 13: 1708–1712.
Zuna J, Hrusak O, Kalinova M, Muzikova K, Stary J, Trka J . TEL/AML1 positivity in childhood ALL: average or better prognosis? Czech Paediatric Haematology Working Group. Leukemia 1999; 13: 22–24.
Cayuela JM, Baruchel A, Orange C, Madani A, Auclerc MF, Daniel MT et al. TEL-AML1 fusion RNA as a new target to detect minimal residual disease in pediatric B-cell precursor acute lymphoblastic leukemia. Blood 1996; 88: 302–308.
Ballerini P, Landman Parker J, Laurendeau I, Olivi M, Vidaud M, Adam M et al. Quantitative analysis of TEL/AML1 fusion transcripts by real-time RT-PCR assay in childhood acute lymphoblastic leukemia. Leukemia 2000; 14: 1526–1528.
Rosenfeld C, Goutner A, Choquet C, Venuat AM, Kayibanda B, Pico JL et al. Phenotypic characterisation of a unique non-T, non-B acute lymphoblastic leukaemia cell line. Nature 1977; 267: 841–843.
Groffen J, Stephenson JR, Heisterkamp N, de Klein A, Bartram CR, Grosveld G . Philadelphia chromosomal breakpoints are clustered within a limited region, bcr, on chromosome 22. Cell 1984; 36: 93–99.
Hoelzer D . Treatment of acute lymphoblastic leukemia. Semin Hematol 1994; 31: 1–15.
Cuneo A, Ferrant A, Michaux JL, Demuynck H, Boogaerts M, Louwagie A et al. Philadelphia chromosome-positive acute myeloid leukemia: cytoimmunologic and cytogenetic features. Haematologica 1996; 81: 423–427.
Gleissner B, Gokbuget N, Bartram CR, Janssen B, Rieder H, Janssen JW et al. Leading prognostic relevance of the BCR-ABL translocation in adult acute B-lineage lymphoblastic leukemia: a prospective study of the German Multicenter Trial Group and confirmed polymerase chain reaction analysis. Blood 2002; 99: 1536–1543.
Faderl S, Talpaz M, Estrov Z, O’Brien S, Kurzrock R, Kantarjian HM . The biology of chronic myeloid leukemia. N Engl J Med 1999; 341: 164–172.
Melo JV . The diversity of BCR-ABL fusion proteins and their relationship to leukemia phenotype. Blood 1996; 88: 2375–2384.
Saglio G, Guerrasio A, Rosso C, Zaccaria A, Tassinari A, Serra A et al. New type of Bcr/Abl junction in Philadelphia chromosome-positive chronic myelogenous leukemia. Blood 1990; 76: 1819–1824.
Pane F, Frigeri F, Sindona M, Luciano L, Ferrara F, Cimino R et al. Neutrophilic-chronic myeloid leukemia: a distinct disease with a specific molecular marker (BCR/ABL with C3/A2 junction). Blood 1996; 88: 2410–2414.
Secker-Walker LM, Craig JM, Hawkins JM, Hoffbrand AV . Philadelphia positive acute lymphoblastic leukemia in adults: age distribution, BCR breakpoint and prognostic significance. Leukemia 1991; 5: 196–199.
Melo JV, Myint H, Galton DAG, Goldman JM . P190BCR/ABL chronic myeloid leukemia: the missing link with chronic myelomonocytic leukemia? Leukemia 1994; 8: 208–211.
Sirard C, Laneuville P, Dick JE . Expression of bcr-abl abrogates factor-dependent growth of human hematopoietic M07E cells by an autocrine mechanism. Blood 1994; 83: 1575–1585.
Radich JP, Kopecky KJ, Boldt DH, Head D, Slovak ML, Babu R et al. Detection of BCR-ABL fusion genes in adult acute lymphoblastic leukemia by the polymerase chain reaction. Leukemia 1994; 8: 1688–1695.
Brisco MJ, Sykes PJ, Hughes E, Dolman G, Neoh SH, Peng LM et al. Monitoring minimal residual disease in peripheral blood in B-lineage acute lymphoblastic leukaemia. Br J Haematol 1997; 99: 314–319.
Annino L, Ferrari A, Lamanda M, Mandelli F, Lo Coco F, Pane F . PCR-detectable transcripts in long-term remission of P190(BCR/ABL)-positive acute lymphoblastic leukemia. Blood 2000; 95: 4018–4019.
Radich J, Gehly G, Lee A, Avery R, Bryant E, Edmands S et al. Detection of bcr-abl transcripts in Philadelphia chromosome-positive acute lymphoblastic leukemia after marrow transplantation. Blood 1997; 89: 2602–2609.
Nowell PC, Hungerford DA . A minute chromosome in human chronic granulocytic leukemia. Science 1960; 132: 1497.
de Klein A, van Kessel AG, Grosveld G, Bartram CR, Hagemeijer A, Bootsma D et al. A cellular oncogene is translocated to the Philadelphia chromosome in chronic myelocytic leukaemia. Nature 1982; 300: 765–767.
Lugo TG, Pendergast AM, Muller AJ, Witte ON . Tyrosine kinase activity and transformation potency of bcr-abl oncogene products. Science 1990; 247: 1079–1082.
Hughes TP, O’Shea P, Morgan G, Martiat P, Goldman JM . Persistence of BCR/ABL transcripts after BMT for CML detected by PCR reflects a high risk of relapse. Bone Marrow Transplant 1991; 7 (Suppl 2): 23.
Radich JP, Gehly G, Gooley T, Bryant E, Clift RA, Collins S et al. Polymerase chain reaction detection of the BCR-ABL fusion transcript after allogeneic marrow transplantation for chronic myeloid leukemia: results and implications in 346 patients. Blood 1995; 85: 2632–2638.
Gabert J, Thuret I, Carcassonne Y, Maraninchi D, Mannoni P . Residual bcr/abl transcripts in chronic myeloid leukaemia. Lancet 1990; 335: 417–418.
Pignon JM, Henni T, Amselem S, Vidaud M, Duquesnoy P, Vernant JP et al. Frequent detection of minimal residual disease by use of the polymerase chain reaction in long-term survivors after bone marrow transplantation for chronic myeloid leukemia. Leukemia 1990; 4: 83–86.
Miyamura K, Tahara T, Tanimoto M, Morishita Y, Kawashima K, Morishima Y et al. Long persistent bcr-abl positive transcript detected by polymerase chain reaction after marrow transplant for chronic myelogenous leukemia without clinical relapse: a study of 64 patients. Blood 1993; 81: 1089–1093.
van Rhee F, Lin F, Cross NC, Reid CD, Lakhani AK, Szydlo RM et al. Detection of residual leukaemia more than 10 years after allogeneic bone marrow transplantation for chronic myelogenous leukaemia. Bone Marrow Transplant 1994; 14: 609–612.
Thompson JD, Brodsky I, Yunis JJ . Molecular quantification of residual disease in chronic myelogenous leukemia after bone marrow transplantation. Blood 1992; 79: 1629–1635.
Lion T, Izraeli S, Henn T, Gaiger A, Mor W, Gadner H . Monitoring of residual disease in chronic myelogenous leukemia by quantitative polymerase chain reaction. Leukemia 1992; 6: 495–499.
Cross NC, Feng L, Chase A, Bungey J, Hughes TP, Goldman JM . Competitive polymerase chain reaction to estimate the number of BCR-ABL transcripts in chronic myeloid leukemia patients after bone marrow transplantation. Blood 1993; 82: 1929–1936.
Cross NC . Minimal residual disease in chronic myeloid leukaemia. Hematol Cell Ther 1998; 40: 224–228.
Hochhaus A, Lin F, Reiter A, Skladny H, Mason PJ, van Rhee F et al. Quantification of residual disease in chronic myelogenous leukemia patients on interferon-alpha therapy by competitive polymerase chain reaction. Blood 1996; 87: 1549–1555.
Merx K, Muller MC, Kreil S, Lahaye T, Paschka P, Schoch C et al. Early reduction of BCR-ABL mRNA transcript levels predicts cytogenetic response in chronic phase CML patients treated with imatinib after failure of interferon alpha. Leukemia 2002; 16: 1579–1583.
Stentoft J, Pallisgaard N, Kjeldsen E, Holm MS, Nielsen JL, Hokland P . Kinetics of BCR-ABL fusion transcript levels in chronic myeloid leukemia patients treated with STI571 measured by quantitative real-time polymerase chain reaction. Eur J Haematol 2001; 67: 302–308.
Emig M, Saussele S, Wittor H, Weisser A, Reiter A, Willer A et al. Accurate and rapid analysis of residual disease in patients with CML using specific fluorescent hybridization probes for real time quantitative RT-PCR. Leukemia 1999; 13: 1825–1832.
Saffroy R, Lemoine A, Brezillon P, Frenoy N, Delmas B, Goldschmidt E et al. Real-time quantitation of bcr-abl transcripts in haematological malignancies. Eur J Haematol 2000; 65: 258–266.
Barbany G, Hagberg A, Olsson-Stromberg U, Simonsson B, Syvanen AC, Landegren U . Manifold-assisted reverse transcription-PCR with real-time detection for measurement of the BCR-ABL fusion transcript in chronic myeloid leukemia patients. Clin Chem 2000; 46: 913–920.
Schoch C, Schnittger S, Bursch S, Gerstner D, Hochhaus A, Berger U et al. Comparison of chromosome banding analysis, interphase- and hypermetaphase-FISH, qualitative and quantitative PCR for diagnosis and for follow-up in chronic myeloid leukemia: a study on 350 cases. Leukemia 2002; 16: 53–59.
Aplan PD, Lombardi DP, Reaman GH, Sather HN, Hammond GD, Kirsch IR . Involvement of the putative hematopoietic transcription factor SCL in T-cell acute lymphoblastic leukemia. Blood 1992; 79: 1327–1333.
Bash RO, Crist WM, Shuster JJ, Link MP, Amylon M, Pullen J et al. Clinical features and outcome of T-cell acute lymphoblastic leukemia in childhood with respect to alterations at the TAL1 locus: a Pediatric Oncology Group study. Blood 1993; 81: 2110–2117.
Bernard O, Lecointe N, Jonveaux P, Souyri M, Mauchauffe M, Berger R et al. Two site-specific deletions and t(1;14) translocation restricted to human T-cell acute leukemias disrupt the 5′ part of the tal-1 gene. Oncogene 1991; 6: 1477–1488.
Breit TM, Mol EJ, Wolvers-Tettero IL, Ludwig WD, van Wering ER, van Dongen JJ . Site-specific deletions involving the tal-1 and sil genes are restricted to cells of the T cell receptor alpha/beta lineage: T cell receptor delta gene deletion mechanism affects multiple genes. J Exp Med 1993; 177: 965–977.
Brown L, Cheng JT, Chen Q, Siciliano MJ, Crist W, Buchanan G et al. Site-specific recombination of the tal-1 gene is a common occurrence in human T cell leukemia. EMBO J 1990; 9: 3343–3351.
Delabesse E, Bernard M, Landman-Parker J, Davi F, Leboeuf D, Varet B et al. Simultaneous SIL-TAL1 RT-PCR detection of all tal(d) deletions and identification of novel tal(d) variants. Br J Haematol 1997; 99: 901–907.
Delabesse E, Bernard M, Meyer V, Smit L, Pulford K, Cayuela JM et al. TAL1 expression does not occur in the majority of T-ALL blasts. Br J Haematol 1998; 102: 449–457.
Begley CG, Green AR . The SCL gene: from case report to critical hematopoietic regulator. Blood 1999; 93: 2760–2770.
Begley CG, Aplan PD, Denning SM, Haynes BF, Waldmann TA, Kirsch IR . The gene SCL is expressed during early hematopoiesis and encodes a differentiation-related DNA-binding motif. Proc Natl Acad Sci USA 1989; 86: 10128–10132.
Aplan PD, Nakahara K, Orkin SH, Kirsch IR . The SCL gene product: a positive regulator of erythroid differentiation. EMBO J 1992; 11: 4073–4081.
Aplan PD, Lombardi DP, Kirsch IR . Structural characterization of SIL, a gene frequently disrupted in T-cell acute lymphoblastic leukemia. Mol Cell Biol 1991; 11: 5462–5469.
Chen Q, Cheng JT, Tasi LH, Schneider N, Buchanan G, Carroll A et al. The tal gene undergoes chromosome translocation in T cell leukemia and potentially encodes a helix–loop–helix protein. EMBO J 1990; 9: 415–424.
Huang W, Kuang SQ, Huang QH, Dong S, Zhang T, Gu LJ et al. RT/PCR detection of SIL-TAL-1 fusion mRNA in Chinese T-cell acute lymphoblastic leukemia (T-ALL). Cancer Genet Cytogenet 1995; 81: 76–82.
Macintyre EA, Smit L, Ritz J, Kirsch IR, Strominger JL . Disruption of the SCL locus in T-lymphoid malignancies correlates with commitment to the T-cell receptor alpha beta lineage. Blood 1992; 80: 1511–1520.
Breit TM, Beishuizen A, Ludwig WD, Mol EJ, Adriaansen HJ, van Wering ER et al. tal-1 deletions in T-cell acute lymphoblastic leukemia as PCR target for detection of minimal residual disease. Leukemia 1993; 7: 2004–2011.
Pongers-Willemse MJ, Seriu T, Stolz F, d’Aniello E, Gameiro P, Pisa P et al. Primers and protocols for standardized detection of minimal residual disease in acute lymphoblastic leukemia using immunoglobulin and T cell receptor gene rearrangements and TAL1 deletions as PCR targets: report of the BIOMED-1 CONCERTED ACTION: investigation of minimal residual disease in acute leukemia. Leukemia 1999; 13: 110–118.
Chen X, Pan Q, Stow P, Behm FG, Goorha R, Pui CH et al. Quantification of minimal residual disease in T-lineage acute lymphoblastic leukemia with the TAL-1 deletion using a standardized real-time PCR assay. Leukemia 2001; 15: 166–170.
Rowley JD, Golomb HM, Dougherty C . 15/17 translocation, a consistent chromosomal change in acute promyelocytic leukaemia. Lancet 1977; 1: 549–550.
Biondi A, Rambaldi A . Acute promyelocytic leukemia. In: Henderson ES, Lister TA, Greaves MS (eds), Leukemia, 7th edn, Chap. 23. Philadelphia: WB Saunders, 2002, pp 529–543.
Longo L, Pandolfi PP, Biondi A, Rambaldi A, Mencarelli A, Lo Coco F et al. Rearrangements and aberrant expression of the retinoic acid receptor alpha gene in acute promyelocytic leukemias. J Exp Med 1990; 172: 1571–1575.
Borrow J, Goddard AD, Sheer D, Solomon E . Molecular analysis of acute promyelocytic leukemia breakpoint cluster region on chromosome 17. Science 1990; 249: 1577–1580.
de The H, Chomienne C, Lanotte M, Degos L, Dejean A . The t(15;17) translocation of acute promyelocytic leukaemia fuses the retinoic acid receptor alpha gene to a novel transcribed locus. Nature 1990; 347: 558–561.
Lemons RS, Eilender D, Waldmann RA, Rebentisch M, Frej AK, Ledbetter DH et al. Cloning and characterization of the t(15;17) translocation breakpoint region in acute promyelocytic leukemia. Genes Chromosomes Cancer 1990; 2: 79–87.
Petkovich M, Brand NJ, Krust A, Chambon P . A human retinoic acid receptor which belongs to the family of nuclear receptors. Nature 1987; 330: 444–450.
Giguere V, Ong ES, Segui P, Evans RM . Identification of a receptor for the morphogen retinoic acid. Nature 1987; 330: 624–629.
Zelent A, Guidez F, Melnick A, Waxman S, Licht JD . Translocations of the RARalpha gene in acute promyelocytic leukemia. Oncogene 2001; 20: 7186–7203.
Grimwade D . The pathogenesis of acute promyelocytic leukaemia: evaluation of the role of molecular diagnosis and monitoring in the management of the disease. Br J Haematol 1999; 106: 591–613.
Reiter A, Sauβele S, Grimwade D, Wiesmels JL, Segal M, Lafage-Pochitaloff M et al. Genomic anatomy of the reciprocal translocation t(15;17) in acute promyelocytic leukemia. Gene Chromosome Cancer 2003; 36: 175–188.
Lo Coco F, Diverio D, Pandolfi PP, Biondi A, Rossi V, Avvisati G et al. Molecular evaluation of residual disease as a predictor of relapse in acute promyelocytic leukaemia. Lancet 1992; 340: 1437–1438.
Burnett AK, Grimwade D, Solomon E, Wheatley K, Goldstone AH . Presenting white blood cell count and kinetics of molecular remission predict prognosis in acute promyelocytic leukemia treated with all-trans retinoic acid: result of the Randomized MRC Trial. Blood 1999; 93: 4131–4143.
Slack JL, Bi W, Livak KJ, Beaubier N, Yu M, Clark M et al. Pre-clinical validation of a novel, highly sensitive assay to detect PML-RARalpha mRNA using real-time reverse-transcription polymerase chain reaction. J Mol Diagn 2001; 3: 141–149.
Visani G, Buonamici S, Malagola M, Isidori A, Piccaluga PP, Martinelli G et al. Pulsed ATRA as single therapy restores long-term remission in PML-RARalpha-positive acute promyelocytic leukemia patients: real time quantification of minimal residual disease. A pilot study. Leukemia 2001; 15: 1696–1700.
Kwong YL, Au WY, Chim CS, Pang A, Suen C, Liang R . Arsenic trioxide- and idarubicin-induced remissions in relapsed acute promyelocytic leukaemia: clinicopathological and molecular features of a pilot study. Am J Hematol 2001; 66: 274–279.
Gallagher RE, Li YP, Rao S, Paietta E, Andersen J, Etkind P et al. Characterization of acute promyelocytic leukemia cases with PML-RAR alpha break/fusion sites in PML exon 6: identification of a subgroup with decreased in vitro responsiveness to all-trans retinoic acid. Blood 1995; 86: 1540–1547.
Grimwade D, Howe K, Langabeer S, Davies L, Oliver F, Walker H et al. Establishing the presence of the t(15;17) in suspected acute promyelocytic leukaemia: cytogenetic, molecular and PML immunofluorescence assessment of patients entered into the M.R.C. ATRA trial. M.R.C. Adult Leukaemia Working Party. Br J Haematol 1996; 94: 557–573.
Slack JL, Willman CL, Andersen JW, Li YP, Viswanatha DS, Bloomfield CD et al. Molecular analysis and clinical outcome of adult APL patients with the type V PML-RARalpha isoform: results from intergroup protocol 0129. Blood 2000; 95: 398–403.
Lanotte M, Martin-Thouvenin V, Najman S, Balerini P, Valensi F, Berger R . NB4, a maturation inducible cell line with t(15;17) marker isolated from a human acute promyelocytic leukemia (M3). Blood 1991; 77: 1080–1086.
Rowe D, Cotterill SJ, Ross FM, Bunyan DJ, Vickers SJ, Bryon J et al. Cytogenetically cryptic AML1-ETO and CBF beta-MYH11 gene rearrangements: incidence in 412 cases of acute myeloid leukaemia. Br J Haematol 2000; 111: 1051–1056.
Liu P, Hajra A, Wijmenga C, Collins F . Molecular pathogenesis of the chromosome16 inversion in the M4Eo subtype of acute myeloid leukemia. Blood 1995; 85: 2289–2302.
Claxton DF, Liu P, Hsu HB, Marlton P, Hester J, Collins F et al. Detection of fusion transcripts generated by the inversion 16 chromosome in acute myeloid leukemia. Blood 1994; 83: 1750–1756.
Hébert J, Cayuela JM, Daniel MT, Berger R, Sigaux F . Detection of minimal residual disease in acute myelomonocytic leukemia with abnormal marrow eosinophils by nested polymerase chain reaction with allele specific amplification. Blood 1994; 84: 2291–2296.
Poirel H, Radford-Weiss I, Rack K, Troussard X, Veil A, Valensi F et al. Detection of the chromosome 16 CBFβ-MHY11 fusion transcript in myelomonocytic leukemias. Blood 1995; 85: 1313–1322.
Costello R, Sainty D, Lecine P, Cusenier A, Mozziconacci MJ, Arnoulet C et al. Detection of CBFbeta/MYH11 fusion transcripts in acute myeloid leukemia: heterogeneity of cytological and molecular characteristics. Leukemia 1997; 11: 644–650.
Burnett AK, Goldstone AH, Stevens RM, Hann IM, Rees JK, Gray RG et al. Randomised comparison of addition of autologous bone-marrow transplantation to intensive chemotherapy for acute myeloid leukaemia in first remission: results of MRC AML 10 trial. UK Medical Research Council Adult and Children's Leukaemia Working Parties. Lancet 1998; 351: 700–708.
Burnett AK, Wheatley K, Goldstone AH, Stevens RF, Hann IM, Rees JHK et al. The value of allogeneic bone marrow transplant in patients with acute myeloid leukaemia at differing risk of relapse: results of the MRC AML10 trial. Br J Heamatol 2002; 118: 385–400.
Marcucci G, Caligiuri MA, Bloomfield CD . Defining the ‘absence’ of the CBFbeta/MYH11 fusion transcript in patients with acute myeloid leukemia and inversion of chromosome 16 to predict long-term complete remission: a call for definitions. Blood 1997; 90: 5022–5024.
Costello R, Sainty D, Blaise D, Gastaut JA, Poirel H, Buzyn-Veil A et al. Prognosis value of residual disease monitoring by polymerase chain reaction in patients with CBF beta/MYH11-positive acute myeloblastic leukemia [letter]. Blood 1997; 89: 2222–2223.
Evans PA, Short MA, Jack AS, Norfolk DR, Child JA, Shiach CR et al. Detection and quantitation of the CBFbeta/MYH11 transcripts associated with the inv(16) in presentation and follow-up samples from patients with AML. Leukemia 1997; 11: 364–369.
Krauter J, Hoellge W, Wattjes MP, Nagel S, Heidenreich O, Bunjes D et al. Detection and quantification of CBFB/MYH11 fusion transcripts in patients with inv(16)-positive acute myeloblastic leukemia by real-time RT-PCR. Genes Chromosomes Cancer 2001; 30: 342–348.
Guerrasio A, Pilatrino C, De Micheli D, Cilloni D, Serra A, Gottardi E et al. Assessment of minimal residual disease (MRD) in CBFbeta/MYH11-positive acute myeloid leukemias by qualitative and quantitative RT-PCR amplification of fusion transcripts. Leukemia 2002; 16: 1176–1181.
Friedman AD . Leukemogenesis by CBF oncoproteins. Leukemia 1999; 13: 1932–1942.
Nucifora G, Larson RA, Rowley JD . Persistence of the t(8;21) translocation in patients with AML-M2 in long-term remission patients. Blood 1993; 82: 712–715.
Kusec R, Laczika K, Knobl P, Friedl J, Greinix H, Kahls P et al. AML1/ETO fusion mRNA can be detected in remission blood samples of all patients with t(8;21) acute myeloid leukemia after chemotherapy or autologous bone marrow transplantation. Leukemia 1994; 8: 735–739.
Miyamoto T, Weissman IL, Akashi K . AML1/ETO-expressing nonleukemic stem cells in acute myelogenous leukemia with 8;21 chromosomal translocation. Proc Natl Acad Sci USA 2000; 97: 7521–7526.
Morschhauser F, Cayuela JM, Martini S, Baruchel A, Rousselot P, Socie G et al. Evaluation of minimal residual disease using reverse-transcription polymerase chain reaction in t(8;21) acute myeloid leukemia: a multicenter study of 51 patients. J Clin Oncol 2000; 18: 788–794.
Muto A, Mori S, Matsushita H, Awaya N, Ueno H, Takayama N et al. Serial quantification of minimal residual disease of t(8;21) acute myelogenous leukaemia with RT-competitive PCR assay. Br J Haematol 1996; 95: 85–94.
Tobal K, Yin JA . Monitoring of minimal residual disease by quantitative reverse transcriptase-polymerase chain reaction for AML1-MTG8 transcripts in AML-M2 with t(8; 21). Blood 1996; 88: 3704–3709.
Tobal K, Newton J, Macheta M, Chang J, Morgenstern G, Evans PA et al. Molecular quantitation of minimal residual disease in acute myeloid leukemia with t(8;21) can identify patients in durable remission and predict clinical relapse. Blood 2000; 95: 815–819.
Kondo M, Kudo K, Kimura H, Inaba J, Kato K, Kojima S et al. Real-time quantitative reverse transcription-polymerase chain reaction for the detection of AML1-MTG8 fusion transcripts in t(8;21)-positive acute myelogenous leukemia. Leuk Res 2000; 24: 951–956.
Fujimaki S, Funato T, Harigae H, Imaizumi M, Suzuki H, Kaneko Y et al. A quantitative reverse transcriptase polymerase chain reaction method for the detection of leukaemic cells with t(8;21) in peripheral blood. Eur J Haematol 2000; 64: 252–258.
Asou H, Tashiro S, Hamamoto K, Otsuji A, Kita K, Kamada N . Establishment of a human acute myeloid leukemia cell line (Kasumi-1) with 8;21 chromosome translocation. Blood 1991; 77: 2031–2036.
Hochhaus A, Weisser A, La Rosee P, Emig M, Muller MC, Saussele S et al. Detection and quantification of residual disease in chronic myelogenous leukemia. Leukemia 2000; 14: 998–1005.
Wattjes MP, Krauter J, Nagel S, Heidenreich O, Ganser A, Heil G . Comparison of nested competitive RT-PCR and real-time RT-PCR for the detection and quantification of AML1/MTG8 fusion transcripts in t(8;21) positive acute myelogenous leukemia. Leukemia 2000; 14: 329–335.
Livak KJ, Schmittgen TD . Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 2001; 25: 402–408.
This work has been supported by the SANCO European Commission (no. SI2.129294 (99CVF2-016) and Applied Biosystems (Foster City, CA, USA). Additional support was given by national grants: ARC no. 5484, Ligue Contre le Cancer, Dutch Cancer Society/Koningin Wilhelmina Fonds (Grant SNWLK 2000-2268), Leukaemia Research Fund of Great Britain and Special Trustees of Guy's Hospital (DG); Fondos de Investigación de la Seguridad Social (00/1079), Beca del Instituto de Salut Carlos III (99/4230) and Fondos Feder (AC G7); Associazione Italiana per la Ricerca sul Cancro (AIRC), MURST and Fondazione M Tettamanti (GC), CNR PF Biotecnologie (Rome), MIUR (Rome), AIL (Rome), Regione Campania and Swedish Cancer Society.
We like to thank W Mayser for his logistic support during meetings, K Livak for useful discussions and C Marcou for secretarial assistance.
About this article
Cite this article
Gabert, J., Beillard, E., van der Velden, V. et al. Standardization and quality control studies of ‘real-time’ quantitative reverse transcriptase polymerase chain reaction of fusion gene transcripts for residual disease detection in leukemia – A Europe Against Cancer Program. Leukemia 17, 2318–2357 (2003). https://doi.org/10.1038/sj.leu.2403135
- real-time quantitative PCR
- fusion gene transcript
Influence of major BCR-ABL1 transcript subtype on outcome in patients with chronic myeloid leukemia in chronic phase treated frontline with nilotinib
Expression differences of miR-142-5p between treatment-naïve chronic myeloid leukemia patients responding and non-responding to imatinib therapy suggest a link to oncogenic ABL2, SRI, cKIT and MCL1 signaling pathways critical for development of therapy resistance
Experimental Hematology & Oncology (2020)
Prognostic value of MRD monitoring based on BCR-ABL1 copy numbers in Philadelphia chromosome positive acute lymphoblastic leukemia
Leukemia & Lymphoma (2020)
Annals of Laboratory Medicine (2020)
Measurable residual disease after the first consolidation predicts the outcomes of patients with acute promyelocytic leukemia treated with all-trans retinoic acid and chemotherapy
International Journal of Hematology (2020)