| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Consultations in Molecular Diagnostics |
From the Department of Genetics,
*
Erasme Hospital; and Institute of Interdisciplinary Research,
Free University of Brussels, Brussels, Belgium
Abstract
Follicular lymphoma is characterized by the t(14;18)(q32;q21) translocation, which juxtaposes Ig heavy chain gene (IgH) sequences with the BclII gene. Several publications have highlighted the importance of molecular follow-up in follicular lymphoma, demonstrating that the detection of cells bearing the BclII-IgH rearrangement by real-time quantitative polymerase chain reaction (RQ-PCR) can anticipate a clinical relapse. In this context, we developed a BclII-IgH RQ-PCR. We began with SYBR Green I detection technology but observed that this system does not allow an accurate measurement of the tumor load when working with genomic DNA. While we were designing the assay using Taqman technology, Moppett et al (Moppett J, van der Velden VHJ, Wijkhuijs AJM, Hancock J, van Dongen JJM, Goulden N: Inhibition affecting RQ-PCR-based assessment of minimal residual disease in acute lymphoblastic leukemia: reversal by addition of bovine serum albumin. Leukemia 2003, 17:268–270) reported PCR inhibition problems in around 15% of blood and bone marrow samples, affecting the DNA quantification and thus the assessment of minimal residual disease. They demonstrated that this PCR inhibition could be partially resolved by adding nonacetylated bovine serum albumin. In our studies, we observed the same phenomenon in a single follicular lymphoma case and extended our study to other available cases. As a result, we suggest a new RQ-PCR procedure that is based on Taqman probe technology and that takes into account the PCR inhibition problems, making this assay more reliable in a routine molecular laboratory.
Follicular lymphoma (FL) is characterized by the juxtaposition of Ig heavy chain gene (IgH) sequences with BclII gene via the t(14;18)(q32;q21) translocation, which occurs in greater than 80% of cases.1, 2 The persistence of cells bearing the BclII-IgH rearrangement, which can be detected by real-time quantitative polymerase chain reaction (RQ-PCR), can anticipate a clinical relapse. As such, molecular follow-up is recommended in FL.3 Therefore, we sought to develop a BclII-IgH RQ-PCR using SYBR Green I detection technology. However, we observed that this system does not accurately measure tumor load when working with genomic DNA. Moppett et al4 have reported that about 15% of the blood and bone marrow samples examined by their lab exhibit PCR inhibition, which affects DNA quantification and assessment of minimal residual disease (MRD). They were partially able to resolve this PCR inhibition by addition of nonacetylated bovine serum albumin (BSA). We observed similar inhibition in a single FL case (outlined below) and extended our study to other available cases. As a result, we have developed a new procedure that improves the reliability of this assay in a routine molecular laboratory.
Case Reports
A 70-year-old male patient presented with an asymptomatic left inguinal mass of 10-cm length. Surgical examination disclosed several lymph nodes that were resected for pathological investigation. The histological findings were diagnostic of a diffuse large B-cell lymphoma arising from a follicular lymphoma. Bone marrow aspirate was normal on morphological grounds. Southern blot analysis was performed on fresh lymphomatous tissue, with the use of the S45/23 probe screening the MBR cluster (a generous gift from Dr. Bastard), demonstrating a BclII rearrangement. A BclII MBR-IgH PCR analysis was subsequently performed with specific BclII MBR and IgH primers (see below) to obtain a sensitive molecular marker for quantitative follow-up. The patient was then treated with six 3-week cycles of Rituximab and CHOP (adriamycin, cyclophosphamide, vincristine and prednisone) regimen. The evolution was favorable, and at present, the patient is alive and well.
Molecular Analysis
Five hundred nanograms of genomic DNA (corresponding to 83,333 diploid genomes) was extracted by a standard phenol/chloroform method from a lymphomatous lymph node. This DNA was subjected to 40 cycles of PCR amplification using the ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster City, CA). The PCR assay was performed using MBR and IgH primers and MBR probe according to Luthra et al5 (MBR, 5'-ttagagagttgctttacgtggcc-3'; IgH, 5'-actcacctgaggagacggtgac-3'; MBR probe, 5'-tttcaacacagacccacccagagcc-3'), at the final concentration of 300 and 200 nmol/L for primers and probe, respectively, and in a reaction mixture of 25 µl including the Taqman Universal PCR Master Mix (Applied Biosystems). The amplification protocol was as follows: initial denaturation of 10 minutes at 95°C followed by 40 cycles of 15-second denaturation at 95°C and 1-minute annealing/extension at 60°C.
The qualitative sensitivity we obtained in our follicular lymphoma cases was very similar to that described by Luthra et al,5 56% (18 cases of 32) versus 54% (13 cases of 24), respectively. A PCR assay was also designed targeting the ABL gene (see below), with the use of forward (5'-gctagtggtggttgcaggaga-3') and reverse (5'-actcccttagaaacatgcagttgt-3') primers and an ABL probe (5'-FAM-tggtaggtcaaaccgcaattcccagattt-TAMRA-3'). The PCR conditions as well as the amplification protocol were the same as for BclII MBR-IgH PCR (see above).
Theoretically, a difference of cycle threshold (Ct) values (
Ct) of 3.3 should be observed between two 10-fold dilutions for a PCR showing 100% efficiency. In practice, a Ct 
3 is acceptable.6
In our PCR design, the equation curve obtained by diluting the MBR-IgH cell line, DoHH2,7
(600 down to 0.06 ng diluted in normal DNA) in duplicates and in three independent assays was Y = –3.4934 x Log10(X) + 39.057 (where Y is the CtMBR-IgH and X is the number of cells; r2 = 0.99) (data not shown).
Regarding our patient, an amplification plot for MBR-IgH rearrangement was demonstrated with a Ct value of 26.13 (threshold, 0.1). To rule out a putative PCR inhibition phenomenon as described by Moppett et al,4 a subsequent PCR reaction was performed using the same DNA quantity (500 ng), as well as 10- and 100-fold dilutions, with and without the addition of 0.04% nonacetylated BSA (Biolabs).
Results and Discussion
In our patient, we observed in two independent assays a strong BclII MBR-IgH PCR inhibition with a DNA quantity of 500 ng as demonstrated by a Ct between 50 and 500 ng (Ct50–500) of 1.82 (data not shown). This inhibition was partially resolved with addition of 0.04% BSA (Ct50BSA-500BSA = 2.41) and, furthermore, disappeared when the DNA quantity used was limited to 50 ng (Ct5–50 = 3.1).
These results prompted us to verify the possible presence of inhibitors in DNA extracted from 21 snap-frozen benign lymphoid hyperplasia cases. The ABL gene was used as an internal reference to assess the amount of amplifiable target DNA, hence the number of detectable cells in a given sample. A large spectrum of Ct values for the ABL gene was obtained when identical amounts (500 ng based on absorbance at 260 nm measurement) of target DNA extracted from our 21 samples were used. The 99% confidence interval of the CtABL was 19.7 to 24.5 (data not shown). Because an identical amount of DNA was loaded for each sample, this broad range of Ct values should be due to PCR inhibitors. In keeping with this hypothesis, the inclusion of nonacetylated BSA (Biolabs) in the reaction mixture greatly reduced the 99% confidence interval of the CtABL (20.8 to 22.8, data not shown). However, as shown in our reported case, adding nonacetylated BSA is not always sufficient to prevent completely the inhibition.
A second source of potential error in quantification results from the fact that PCR efficacy is known to depend on the size and/or the sequence of the template DNA. In molecular oncology laboratories, RQ-PCR is usually mostly used to quantify fusion transcripts displaying the same sequence and size in all patients. In BclII-IgH rearrangements, a Ct value could vary from sample to sample despite an identical starting DNA amount because both the size and sequence of the BclII-IgH rearrangement may vary among different patients. For this reason, we believe it would be better to report the results as a range of tumor cells relative to the whole detectable cell population present in the PCR reaction instead of an inaccurate single value. A set of three dilution curves derived from the DoHH2 cell line was then tested in duplicate (from 600 down to 0.06 ng, diluted in normal DNA), to define a standard curve as well as an average SD of the CtMBR-IgH values (data not shown). The number of lymphoma cells was then estimated for each sample, based on the 99% confidence interval calculated from the experimental CtMBR-IgH value ± (2.5 x the average SD) (legend of Figure 1
).
|
This strategy allowed us to establish and propose an algorithm (described in Figure 1
) that can be routinely applied to give reliable quantitative assessments of the MRD. To disclose potential inhibition phenomena, two DNA quantities per sample were used: 50 and 500 ng. Both DNA amounts were chosen to ensure sufficient sensitivity for the MRD detection as well as to control the problems of inhibition: 500 ng is the minimum required to obtain a sensitivity of one tumor cell on 105 total cells (loading 500 ng corresponds to 83,333 diploid genomes), whereas a 10-fold dilution should significantly reduce the concentration of potential inhibitors.
A parallel study was also performed using the cheaper SYBR Green I detection technology. However, we observed an amazing atypical amplification curve generated with 500 ng of genomic DNA from healthy patients even in the absence of primers (Figure 2A)
. The corresponding melting curve showed a flattened profile (Figure 2B)
, probably due to accumulation of PCR products of different sizes and sequences, generated by hybridization between repetitive sequences scattered throughout the genome or between small degraded DNA fragments (explaining why no primers were needed to generate such an amplification profile). This spurious amplification curve persisted when decreasing the amount of target DNA down to 100 ng (data not shown). We did not test less than 100 ng (corresponding to 16,666 diploid genomes) because this DNA amount is the minimum requirement for achieving a quantitative sensitivity of at least one tumor cell among 103 total cells. This feature was not observed with the pure DoHH2 cell line but was detected with low dilution (one tumor cell among 102 total cells) of this cell line DNA in normal DNA (Figure 2C)
. This artifact did not hinder the detection of one tumor cell among a total of 1000 cells because the melting curve obtained from this dilution could be superimposed on the one observed with the diagnostic sample (Figure 2D)
. However, the estimated Ct value did not fit the starting amount of tumor cells present in a tube (Figure 2C)
, rendering this approach unusable.
|
Footnotes
Address reprint requests to Barbara Dessars, Department of Genetics, Erasme Hospital, 808 Lennikstreet, 1070 Brussels, Belgium. E-mail: Bdessars{at}ulb.ac.be
Accepted for publication August 16, 2005.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and 10–2 diluted (•) DoHH2 cell line DNA in normal DNA. Each sample was tested in duplicate. The horizontal dotted line represents the threshold. Although the amplification plot obtained with the pure cell line shows a classical profile, the curve generated with the 10–2 diluted cell line displays a "bumpy profile," most likely due to genomic DNA autoamplification (see arrow). In the absence of this artifact, the