In this article we will discuss about the novel DHPLC based strategy for identification and characterisation of (BI)-Clonal T-Cell receptor targets for the detection of minimal residual disease in childhood acute lymphoblastic leukemia.
Contents
Introduction:
Clonal antigen receptor rearrangements are widely used as patient-specific markers for detection of minimal residual disease (MRD) in acute lymphoblastic Leukemia (ALL). Since ALL blasts are derived from one single transformed lymphoid precursor cell, functional regions of immunoglobulin- or T-Cell Receptor (TCR) rearrangements are unique for each leukemia patient.
Southern Blot and/or PCR-based techniques allow the identification and characterization of unknown leukemic samples from the time point of diagnosis. Clonal, incomplete TCR- rearrangements are detectable in about 50% of B-cell precursor (BCP) ALL. Several studies demonstrated the applicability of these targets for detection of MRD.
Southern Blotting identifies only the type of rearrangement without further information about the junctional region. Therefore, only PCR based approaches generally allow for the subsequent analysis of the product by direct sequencing or by sequencing after cloning into a plasmid vector. Direct sequencing is often limited due to impurities from polyclonal background amplification or due to biclonal rearrangements on both alleles of the leukemic sample.
Heteroduplex induction followed by acrylamide gel electrophoresis has been described to circumvent these pitfalls. Here, PCR products have to be excised and re-amplified for successful downstream analysis. In cases were the resolution of the method used was insufficient, the samples had to be cloned and several clones had to be sequenced.
We recently described the use of partially denaturing high performance liquid chromatography (DHPLC) for identification of clonal TCR- rearrangements. This method permits the analysis of unknown products in a semi- automated way within 10 minutes after the initial PCR without further hands- on work and is highly reproducible using a target specific analysis protocol.
The aim of our recent study was to test feasibility of DHPLC analysis for identification and characterization of incomplete TCR- rearrangements in childhood precursor B-ALL. These rearrangements occur in about 50% of these leukemias and are widely used for MRD analysis. In a substantial number of cases both alleles are rearranged.
These biclonal leukemic samples were separated by DHPLC and fractionated for subsequent direct sequencing. The determined sequences were then tested for detection of MRD on the Light Cycler system. The method presented here allows us to combine the preparative advantages of DHPLC for characterization of targets with the analytical properties of real time PCR for quantification of MRD.
Materials and Methods:
i. Patients:
Bone marrow (BM) or peripheral blood (PB) from children with ALL at the time hi diagnosis was analyzed. Patients were treated according the COALL 92 or 97 protocol. Cell samples were collected after informed consent from the patients or their parents. The diagnosis was based on morphological and immunologic al examination of BM or PB. Cell samples contained between 40 to 95% malignant cells.
ii. PCR:
PCR analysis was performed with standard primers as described previously by Pongers. A germline V 2-D 3-gene rearrangement (without any nucleotide insertions or deletions around the junctional region) results in the amplification of a 150-bp PCR product. For subsequent analyses of initial leukemic DNA samples 50-µl reactions were performed containing 1 U Taq polymerase (Gibco) and 100 ng DNA as well as final concentrations of 2µM of each primer, 3mM MgCl2, and 200 nM of each dNTP.
The PCR buffer used was that recommended by supplier of the Taq polymerase. The PCR cycling protocol included a 5 min denaturation step at 95°C, followed by 35 cycles of 30 s at 95°C, 30 s at 55°C, 30 s at 72°C, and was followed by a final extension step of 2 minutes. PCR was performed in 200-µl, ultra-thin PCR tubes in a T-gradient thermocycler.
iii. Acrylamide Gel Electrophoresis:
An aliquot of the final PCR. product (1.5µl) was subjected to electrophoresis on an acrylamide gel followed by a rapid silver staining protocol. In brief, a 0.3 mm ultra-thin polyacrylamide was cast onto Gelbond (FMC) films. Samples were run in a discontinuous buffer system.
Gels contained 7.1% glycerol and 33 mM Tris-sulfate buffer, pH 9.0, and were run with a buffer strip containing 140 mM Tris-borate, pH 9.0, onto a cooled (8°C) Pharmacia Multiphore System. The whole procedure takes about 1 h and allows the staining of up to 50 samples. Up to 2µl of the 50µl PCR product were used for separation.
iv. DHPLC Analysis:
Eight µl of each PCR product were analyzed by DHPLC on a WAVE System from Transgenomic. The shortest and best suitable gradient for analysis of the targets was determined by running universal gradients at different temperatures, starting with 50°C, for several products (Data not shown). An optimized gradient was then used for analysis of the ALL samples (Table 7-1).
v. Fragment Collection and Sequencing:
Fragment collection was performed in a two-step procedure. In an initial DHPLC run the time window for the elution of the peaks of interest was determined. During a second injection the samples of interest were collected during the previously defined time interval. Usually, 80 to 100µl of eluent containing the peak of interest was collected. Collection volume is determined by the product of DHPLC buffer flow rate (0.9 ml/min) and collection time.
In detail, the fragment collection software was programmed as follows: Using the first run as a reference for specific peak collection, the detection type was set to “whole window modus.” The number of collected fractions then depends on the vial volume and corresponds to the time window in the chromatogram over which the fragments are collected. Collected fractions were analyzed in a “result chart” in the WAVEMAKER software.
A collection table displays each collected vial relative to the corresponding segment of the peak of interest in the elution diagram. Positive fractions were dried under vacuum until complete removal of the eluent. The latter DNA served as template for direct sequencing with the target specific primers used for amplification.
vi. Real Time PCR:
Real Time PCR using the Light-Cycler System was employed to perform quantitative PCR analyses. A germline D 3 TaqMan probe (5′- 6-FAM-ATATCCTCACCCTGGGTCCCATGCCXT-P; X = TAMRA) was used in combination with patient specific ASO primers and a D 3 specific PCR primer (5′- CTGCTTGCTGTGTTTGTCTCCT). Quantification of residual disease was performed in 20 µl reaction volumes using 1 U Platinum Taq Polymerase (Invitrogen) and 2 µl 10x buffer, 200 µM of each dNTP, 100 nM TaqMan probe, 500 nM antisense primer, and 500 nM ASO primer and 5% DMSO.
Amplification conditions for quantification were as follows:
8 min initial denaturation followed by 50 cycles with 5 s denaturation and 23 s annealing/extension at the appropriate temperature between 60 and 68°C. Single point fluorescence measurements were performed during the extension step. Serial ten-fold dilutions of initial leukemic cell DNA in DNA from healthy controls was used for the determination of sensitivity limits and for optimization of the RQ-PCR reaction conditions.
Results:
i. DHPLC Analysis:
DHPLC analysis was performed for the most common V 2-D 3 rearrangement. In order to determine the optimal time window for analysis a 20-min acetonitrile (ACN) gradient from 40 to 72% buffer B, with an increase of 2% buffer B per minute was chosen. The positive control elutes under these conditions between 45 and 60% buffer B. Based on these results the gradient conditions described in Table 1 were chosen as general analysis conditions.
Higher column oven temperatures do not lead to a better quality of separation. The length of the gradient was selected within a relatively broad time window, which allows the detection of clonal or biclonal samples with different junctional regions. These junctional regions sometimes differ by more than 30 nucleotides (due to deletions and insertions). Characteristic elution profiles are shown in Figure 7- 1.
For characterization of the specific junctional regions clonal samples were sequenced directly. Biclonal rearrangements elute as a sharp heteroduplex peak one to two minutes before the two homoduplexes. Biclonal samples lead to negative sequencing reactions due to overlapping junctional regions.
In order to circumvent cloning or gel excision of these bands we used a novel fragment collection principle based on time-dependent, single-peak collection of DHPLC separated products. Oligoclonal rearrangements were not further analyzed due to their relative instability during therapy and the risk of false negative MRD results.
ii. Fragment Collection:
Biclonal samples were subjected to a second DHPLC run. After a first run, elution profiles were analyzed on the WAVE system and results were evaluated with the WAVEMAKER software. A typical biclonal fragment reveals 3 peaks. The heteroduplex fraction elutes from the column at a lower acetonitrile concentration than the corresponding homoduplexes.
The resulting elution profile serves as a matrix for subsequent fragment collections. The first homoduplex peak of interest was marked with a time window for initialization of the collection software during the next run. In our application, we started fragment collection from the beginning of the first clonal/homoduplex peak to the end of the second clonal/homoduplex peak with an individual fraction volume of 80µl.
The collected fractions were dried under vacuum and sequenced directly after reconstitution. Typical result chromatograms and the corresponding sequences from collected fractions are shown in Figure 7-2. Table 7-2 summarizes the results from 10 collected biclonal samples. Sequence variations in the junctional regions are listed and compared to germ line sequences.
iii. Real Time PCR:
For RQ-PCR a 10-fold dilution series of patient DNA (time point of diagnosis) in DNA derived from 5 healthy unrelated donors was used. The described RQ-PCR approach allowed unequivocal amplification and quantification of both alleles for detection of MRD in each patient (Figure 7-3). Sensitivities of 10-3 to 10-5 were measurable in all samples tested. Thus, both targets from a biclonal rearrangement can be used for a quantitative analysis at different time points during therapy.
Discussion on Novel and Innovative Applications of DHPLC:
The purpose of our study was to evaluate the feasibility of DHPLC for the detection and discrimination of several forms of incomplete TCR- rearrangements (clonal, biclonal and oligoclonal) in childhood precursor B-ALL.
Furthermore, we tested the application of a tandem fragment collector arrangement for purification of biclonal rearrangements for subsequent sequencing reactions. These biclonal PCR products were detectable in about 30% of our V 2-D 3 positive samples.
All biclonal samples gave a characteristic elution profile and were unequivocally distinguishable from clonal or oligoclonal samples (Figure 7-1), which was in complete agreement with acrylamide gel electrophoresis and heteroduplex analysis results. Whereas clonal rearrangements were sequenced directly, overlapping junctional regions within a biclonal product prohibits this quite simple sequencing approach.
We recently described the feasibility of DHPLC for discrimination of polyclonal from clonal/biclonal TCR- rearrangements. We applied and extended this methodology to the identification of incomplete TCR- rearrangements. V 2-D 3 rearrangements are the most common rearrangements (except IgH) in B-Cell Precursor ALL (BCP-ALL) and therefore widely used for clone specific detection of minimal residual disease.
Direct sequencing can identify single rearranged fragments. PCR fragments with double rearranged alleles render sequencing inconclusive just at the beginning of the region of interest, i.e., the junctional region.
Compared to the heteroduplex analysis of several other investigators, DHPLC has an automated setup with no gel preparation or post PCR handling, as we performed no heteroduplex induction in our samplest. Additionally, the short and reproducible DHPLC running conditions allow for the analysis of 96 samples overnight.
For a detailed molecular analysis of biclonal products, i.e., sequencing, we performed 2 DHPLC runs for each PCR product. The first run serves as a positioning step for the second collection run. Five to ten fractions were collected on a microtiter plate, 2 to 3 best matching fractions from each peak were prepared for direct sequencing. This procedure allows for a complete analysis of biclonal samples without further PCR or cloning steps.
A detailed molecular analysis of incomplete TCR- PCR products seems to be absolutely necessary for these rearrangements. In contrast to Southern blotting analysis, PCR also amplifies minor clonal rearrangements. The resulting small sub-clones are relatively frequent in precursor B-ALL. Steenbergen points out that in precursor B-ALL’s this may be a common feature.
Sczcepanski et al. clarified this in their comparative study of T- and B-cell receptor rearrangements between initial and relapse diagnosis. Because of the relative instability of oligoclonal targets, incomplete TCR- rearrangements should only be used if they are clonal or biclonal. Especially in oligoclonal samples the response to chemotherapy may vary within different (sub)-clones leading to false negative measurement of minimal residual disease during therapy.
In order to circumvent or minimize these problems we established a novel analytical and preparative procedure for identification and characterization of biclonal rearrangements prior to MRD analysis. RQ-PCR with allele specific oligonucleotides (ASO – primer) corresponding to each allele was performed to test the specificity of the determined sequences.
This allows one to use both allelic variants for the exact examination of minimal residual disease during therapy. Figure 7-4 describes a possible procedure for the identification, characterization and application of V 2 3 targets for MRD analysis.
Conclusion:
We present a novel DHPLC-based method for identification of (bi) clonal V 2-D 3 rearrangements in childhood ALL. Biclonal rearrangements with length differences in the junctional region down to 4 base pairs, compared to germ line sequences, were resolved on the DHPLC column. This high peak resolution allows us to purify biclonal samples for direct sequencing using an automated fragment collector.
Our results show:
(1) DHPLC analysis for clonality assessment is a rapid and reliable tool for identification of MRD targets in newly diagnosed leukemias under standardized conditions.
(2) Fragment collection simplifies the identification of single alleles within biclonal rearrangements.
(3) Subsequent real time PCR on the Light Cycler now enables us to measure MRD quantitatively. The described strategy of DHPLC followed by RQ-PCR will standardize the molecular detection of MRD in childhood ALL.
