Large-scale clinical studies on detection of minimal residual disease (MRD) in acute lymphoblastic leukemia (ALL) have shown that quantification of MRD levels is needed for reliable MRD-based risk group classification. Recently, we have shown that 'real-time' quantitative PCR (RQ-PCR) can be applied for this purpose using patient-specific immunoglobulin (Ig) and T cell receptor (TCR) gene rearrangements as PCR targets with TaqMan probes at the position of the junctional region and two germline primers. Now, we tested an alternative approach on 35 immunoglobulin heavy chain (IGH) gene rearrangements, by designing three germline J H TaqMan probes to be used in combination with one of six corresponding germline J H primers and one allele specific oligonucleotide (ASO) primer complementary to the junctional region. In nine cases in which both approaches were compared, at least similar (n = 4) or slightly higher (n = 5) maximal sensitivities were obtained using an ASO primer. The ASO primer approach reached maximal sensitivities of at least 10 −4 in 33 out of 35 IGH rearrangements. The reproducible range for accurate quantification spanned four to five orders of magnitude in 31 out of 35 cases. In 13 out of 35 rearrangements the stringency of PCR conditions had to be increased to remove or diminish background signals; this only concerned the frequently occurring J H 4, J H 5 and J H 6 gene rearrangements. After optimization of the conditions (mainly by increasing the annealing temperature), only occasional aspecific amplification signals were observed at high threshold cycle (C T ) values above 42 cycles and at least six cycles above the C T value of the detection limit. Hence, these rare aspecific signals could be easily discriminated from specific signals. We conclude that the here presented set of three germline J H TaqMan probes and six corresponding germline J H primers can be used to develop patient-specific RQ-PCR assays, which allow accurate and sensitive MRD analysis in almost all IGH gene rearrangements. These results will facilitate standardized RQ-PCR analysis for MRD detection in large clinical studies. Leukemia (2000) 14, 1426-1435.

No abstract

Sensitive and quantitative detection of minimal residual disease (MRD) in bone marrow (BM) samples of children with acute lymphoblastic leukemia (ALL) is essential for evaluation of early treatment response. In this study, we evaluated whether the traumatic BM samplings can be replaced by peripheral blood (PB) samplings. MRD levels were analyzed in follow-up samples of 62 children with precursor-B-ALL (532 paired BM-PB samples) and 22 children with T-ALL (149 paired BM-PB samples) using real-time quantitative PCR (RQ-PCR) analysis of immunoglobulin and T cell receptor gene rearrangements with sensitivities of 10 −3 to 10 −5 (one ALL cell in 10 3 to 10 5 normal cells). In 14 of the 22 T-ALL patients, detectable MRD levels were found in 67 paired BM-PB samples: in 47 pairs MRD was detected both in BM and PB, whereas in the remaining pairs very low MRD levels were detected in BM (n = 11) or PB (n = 9) only. The MRD levels in the paired BM-PB samples were very comparable and strongly correlated (r s = 0.849). Comparable results were obtained earlier by immunophenotyping in 26 T-ALL patients (321 paired BM-PB samples), which also showed a strong correlation between MRD levels in paired BM and PB samples (r s = 0.822). In 39 of the 62 precursor-B-ALL patients, MRD was detected in 107 BM-PB pairs: in 48 pairs MRD was detected in both BM and PB, in 47 pairs MRD was solely detected in BM (at variable levels), and in 12 pairs only the PB sample was MRD-positive at very low levels (Յ10 −4 ). Furthermore, in the 48 double-positive pairs, MRD levels in BM and PB varied enormously with MRD levels in BM being up to 1000 times higher than in the corresponding PB samples. Consequently, BM samples cannot easily be replaced by PB sampling for MRD analysis in childhood precursor-B-ALL, in line with their BM origin. In T-ALL, which are of thymic origin, BM sampling might be replaced by PB sampling, because the dissemination of T-ALL cells to BM and PB appears to be comparable.

A total of 28 children and nine adults with relapsed T-ALL were analyzed for the configuration of their T-cell receptor (TCR) and TAL1 genes at diagnosis and relapse to evaluate their stability throughout the disease course. A total of 150 clonal TCR and TAL1 gene rearrangements were identified in the 37 patients at diagnosis. In 65% of cases all rearrangements and in 27% of cases most rearrangements found at diagnosis were preserved at relapse. Two children with unusually late T-ALL recurrences displayed completely different TCR gene rearrangement sequences between diagnosis and relapse. This indicates that a proportion of very late T-ALL recurrences might represent second T-ALL. Specifically, 88% of clonal rearrangements identified at diagnosis in truly relapsed T-ALL were preserved at relapse. This is significantly higher as compared to previously studied precursor-B-ALL (B70%). Thus, from biological point of view, immunogenotype of T-ALL is more stable as compared with precursor-B-ALL. The overall stability of TCR gene rearrangements was higher in adult T-ALL (97%) than in childhood T-ALL (86%). Based on the stability of TCR gene rearrangements, we propose a strategy for PCR target selection (TCRD þ TAL1-TCRB-TCRG), which probably allows reliable minimal residual disease detection in all T-ALL patients.