TO THE EDITOR
During the last decade, several studies have shown that quantification of minimal residual disease (MRD) provides prognostic information that can be used for treatment decisions in individual patients. In acute lymphoblastic leukemia (ALL), MRD information obtained at early time points during therapy allows the recognition of high-risk and low-risk patients, who may benefit from treatment intensification or treatment reduction, respectively. In acute promyelocytic leukemia (APL) and chronic myeloid leukemia (CML), MRD information at specific time-points enables effective early treatment intervention.1
Quantitative MRD data with sensitivities of at least 10−4 can be obtained by real-time quantitative PCR (RQ-PCR) analysis using fusion gene transcripts associated with chromosomal abnormalities as MRD-PCR targets.2 For this purpose, primer and probe sets have recently been developed within the Europe Against Cancer (EAC) program.3 In addition, three control gene transcripts (abelson (ABL), β2-microglobine (B2M) and glucuronidase (GUS)) were selected based on their suitability for MRD studies in leukemic patients.4
Shipment of freshly collected patient samples to the MRD-PCR laboratories may take considerable time. Especially in multicenter studies it may take 1–3 days before the patient sample can be processed. During this period, the quality of the RNA present in the sample may deteriorate;5 this can be due to induction of cell death with subsequent degradation of RNA, or to cellular changes resulting in altered stability or altered expression of transcripts. So far, only few studies have addressed the stability of fusion gene transcripts and control gene transcripts in patient samples.6,7
Differential stability over time of fusion gene transcripts and control gene transcripts may result in over- or underestimation of MRD levels. Therefore, careful assessment of the transcript stability is crucial for reliable MRD studies. We determined whether the stability of the control gene transcripts differed (1) between normal and leukemic cells, (2) between bone marrow (BM) and peripheral blood (PB) samples, and (3) between different leukemia (ALL, AML, and CML). Furthermore, we investigated the stability of fusion gene transcripts over time and evaluated whether this stability was comparable to the stability of control gene transcripts.
Three different sample protocols were used in this study: protocol 1 (whole BM and/or PB obtained at diagnosis or relapse; n=29), protocol 2 (mononuclear cells (MNC) obtained at diagnosis or relapse, diluted (20 ×) in MNC of a healthy control to mimic an MRD setting; n=9), and protocol 3 (whole BM and/or PB samples obtained during follow-up; n=26). In all three protocols, the cell sample was divided into five identical tubes. Tube 0 was processed immediately, whereas tube 1, 2, 3, and 4 were kept at room temperature and processed at days 1, 2, 3, and 4 after sample collection, respectively. On the day of processing, MNC were separated by Ficoll density centrifugation (protocol 1 and 3 only) and RNA was isolated from 2 × 106 MNC. cDNA and RQ-PCR reactions were performed according to the EAC protocol.3
As the control gene transcripts were the topic of our study, their absolute copy numbers could not be used to correct for the quantity and quality of the RNA. Therefore, ratios between copy numbers of the different control genes were used to determine whether they showed comparable degradation rates. Statistical analysis (mixed model with type of sample protocol (1/2/3), sample type (BM/PB), leukemia type (ALL/AML/CML), and time (0/1/2/3/4 days) as variables) showed that for none of the three ratios (ABL/B2M, ABL/GUS, and GUS/B2M) the type of protocol had a significant effect, indicating that the stability of the control gene transcripts is comparable between leukemic cells and normal cells, also during therapy. Therefore, for subsequent analyses, data from the three different sample protocols were combined.
Comparison of control gene transcript ratios between BM and PB samples showed significant differences (Figure 1). The ABL/B2M and GUS/B2M ratios differed significantly between PB and BM at days 1, 2, and 4, with higher ratios in BM (Figure 1a and b). In contrast, the ABL/GUS ratio did not differ between BM and PB (data not shown) and did not change significantly over time (Figure 1c). These effects were comparable between the three main leukemia types. Overall, these data indicate that (1) the stability of ABL, GUS, and B2M is comparable among ALL, AML, and CML cells; (2) ABL and GUS transcripts have a comparable degradation rate between BM and PB; and (3) B2M transcripts show a higher degradation rate than ABL and GUS transcripts in BM, but not in PB.
We subsequently analyzed the stability of fusion gene transcripts. It should be noted that at arrival of the sample in the laboratory it was not known whether the newly diagnosed leukemia harbored a fusion gene transcript (sample protocol 1 and 2) or whether MRD was present during follow-up (sample protocol 3). Consequently, the number of patients that could be analyzed for individual fusion gene transcript was generally low (one MLL-AF4; two TEL-AML1; three BCR-ABL p190; three AML1-ETO; two PML-RARA; one CBFB-MYH11; 18 BCR-ABL p210). Therefore, no statistical analysis was performed per type of fusion gene transcript and data were analyzed independent of the type of sample protocol and the sample type (except for BCR-ABL p210). As the fusion gene/control gene transcript ratio in follow-up samples or diluted diagnostic samples is obviously lower than the ratio in diagnostic samples, the ratio at day 0 was set at 1 and the ratio at the subsequent days was expressed relative to day 0.
Although the mean stability of most fusion gene transcripts seemed to be comparable to the control genes (as reflected by a constant ratio around 1), large differences in stability could be observed between patients expressing the same fusion gene transcript. For example, the maximum and minimum fusion gene/control gene transcript ratio (ie the assumed MRD level) in individual samples could differ up to 25-fold for AML1-ETO (at day 2), more than 100-fold for BCR-ABL p210 normalized to B2M (days 2, 3, and 4), and up to 40-fold for BCR-ABL p210 normalized to ABL or GUS (days 2, 3, and 4). An example is shown in Figure 2 for the ratio between BCR-ABL p210 and ABL in PB samples of CML patients processed at five subsequent days. Apparently, the stability of the fusion gene transcripts can vary substantially between different patients. Consequently, MRD data obtained from samples that are not processed immediately after sampling may not accurately reflect actual fusion gene transcript levels.
In order to obtain data on the absolute degradation rate of gene transcripts over time, we evaluated the stability of the three control gene transcripts (ABL, B2M, and GUS) in PB samples obtained from healthy subjects, using an in vitro transcribed non-human RNA as an exogenous internal positive control in order to correct for variations in MNC number, RNA isolation, and cDNA synthesis. Whole-PB samples (n=12) were divided into four tubes. From tube 1, MNC were immediately separated by Ficoll density centrifugation, counted, and lyzed (first step RNA isolation). Per 1 × 106 cells, 5 μl of exogenous control RNA (1000-fold diluted; kindly provided by Wanli Bi, Applied Biosystems, Foster City, CA, USA) was added, followed by completion of the RNA isolation, cDNA synthesis, and RQ-PCR analysis. Tubes 2–4 were processed identically at days 1–3. The exogenous control normalized Ct values of all three control gene transcripts increased with time and generally an increase of 1 Ct was observed on consecutive days (data not shown). These results indicate that the absolute degradation rate is approximately 50% per day (ie an increase of the Ct value with one cycle) in PB MNC.
Although our data are based on a limited and heterogeneous set of samples, they clearly indicate that transcripts are rapidly degraded ex vivo and that the rate of degradation can differ between different types of transcripts, between PB and BM, and between patients. As such differential degradation will results in an over- or underestimation of MRD levels, samples should preferably be processed on the day of sampling; this processing should include at least the Ficoll density centrifugation-based separation of MNC and the cell lysis step of the RNA extraction. Even better, because changes in transcript levels can already occur within the initial 4 h,5 samples should be collected in tubes with immediate stabilization of intracellular RNA, thereby preventing any degradation of control gene and/or fusion gene transcripts. Such reagents for stabilization of RNA have recently successfully been applied in a multi-center study (Mueller et al, Blood 2003; 102: 64a abstract).
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.
van der Velden VH, Hochhaus A, Cazzaniga G, Szczepanski T, Gabert J, van Dongen JJ . Detection of minimal residual disease in hematologic malignancies by real-time quantitative PCR: principles, approaches, and laboratory aspects. Leukemia 2003; 17: 1013–1034.
Gabert J, Beillard E, van der Velden VHJ, Bi W, Grimwade D, Pallisgaard N et al. Standardization and quality control studies of ‘real-time’ quantitative reverse transcriptase polymerase chain reaction (RQ-PCR) of fusion gene transcripts for minimal residual disease detection in leukemia – a Europe Against Cancer Program. Leukemia 2003; 17: 2318–2357.
Beillard E, Pallisgaard N, van der Velden VHJ, Bi W, 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; 17: 2474–2486.
Tanner MA, Berk LS, Felten DL, Blidy AD, Bit SL, Ruff DW . Substantial changes in gene expression level due to the storage temperature and storage duration of human whole blood. Clin Lab Haematol 2002; 24: 337–341.
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.
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.
About this article
Cite this article
van der Velden, V., Boeckx, N., Gonzalez, M. et al. Differential stability of control gene and fusion gene transcripts over time may hamper accurate quantification of minimal residual disease – a study within the Europe Against Cancer Program. Leukemia 18, 884–886 (2004). https://doi.org/10.1038/sj.leu.2403309
Standardisation and consensus guidelines for minimal residual disease assessment in Philadelphia-positive acute lymphoblastic leukemia (Ph + ALL) by real-time quantitative reverse transcriptase PCR of e1a2 BCR-ABL1
Development and evaluation of armored RNA-based standards for quantification ofBCR-ABL1p210/p190fusion gene transcripts
Journal of Clinical Laboratory Analysis (2018)
Is Next-Generation Sequencing the way to go for Residual Disease Monitoring in Acute Lymphoblastic Leukemia?
Molecular Diagnosis & Therapy (2017)
International Journal of Laboratory Hematology (2017)
Rationale and efficacy of proteasome inhibitor combined with arsenic trioxide in the treatment of acute promyelocytic leukemia