PCR assays provide sensitive and quantitative assessments ... - Nature

6 downloads 0 Views 255KB Size Report
Non-Hodgkin's lymphoma (NHL) arises as a clonal transform- ation of normal B and T cell differentiation and is often charac- terized by a higher incidence of ...
Leukemia (1999) 13, 1833–1842  1999 Stockton Press All rights reserved 0887-6924/99 $15.00 http://www.stockton-press.co.uk/leu

Real-time t(11;14) and t(14;18) PCR assays provide sensitive and quantitative assessments of minimal residual disease (MRD) K Olsson1, CJ Gerard1,2, J Zehnder3, C Jones3, R Ramanathan1, C Reading1,4 and EG Hanania1,5 1

SyStemix Inc., A Novartis Company, Palo Alto, CA; and 3Department of Pathology, Stanford University, Stanford, CA, USA

Non-Hodgkin’s lymphoma (NHL) arises as a clonal transformation of normal B and T cell differentiation and is often characterized by a higher incidence of specific chromosomal translocations. We have developed real-time TaqMan PCR assays directed toward two of these tumor-associated DNA markers, the t(14;18)(q32;q21.3) at the major breakpoint region of the bcl2 gene and the t(11;14)(q13;q32) at the bcl-1 major translocation cluster. During analysis of serial dilutions of t(14;18)positive DNA, the t(14;18) real-time PCR was at least as sensitive as nested PCR and demonstrated enhanced quantitative potential. Moreover, in a blinded comparison of the t(14;18) real-time PCR and a clinically validated nested PCR protocol using 134 cell line and patient DNA samples, the real-time PCR detected the translocation in 30.0% more cases than nested PCR. Both the t(14;18) and t(11;14) real-time PCR assays were used to quantitate minimal residual disease (MRD) in an NHL clinical trial assessing the safety and efficacy of a tumor-purging protocol in autologous stem cell transplantation. The assays were also used to evaluate disease depletion in an ex vivo tumor spiking model in which normal peripheral blood was spiked with tumor cell lines and processed according to the clinical purging method. PCR data from both the clinical trial and the ex vivo model demonstrated a 4 to 6 log reduction in tumor cells during CD34+ and CD34+ Thy-1+ enrichment. Because the t(14;18) and t(11;14) real-time PCR assays are very sensitive, quantitative, rapid, and require no post-PCR manipulation, they may serve as practical alternatives to nested PCR. Keywords: real-time PCR; minimal residual disease (MRD); nonHodgkin’s lymphoma (NHL); TaqMan assay; quantitative assessment; bcl-1/bcl-2

Introduction Non-Hodgkin’s lymphomas (NHLs) represent a family of heterogeneous malignant solid tumors of lymphoid tissues that are often characterized by specific chromosomal translocations. Up to 85% of follicular and 20–30% of diffuse B cell lymphomas contain the t(14;18),1–8 and as many as 95% of mantle cell lymphomas may contain the t(11;14), as detected by overexpression of cyclin D1 mRNA.9–14 During these translocation events, the cyclin D1 gene on chromosome 11 or the bcl-2 gene on chromosome 18 fuses with the 5⬘ end of the immunoglobulin (Ig) heavy-chain joining region (JH) on chromosome 14.15,16 Thirty to 50% of t(11;14) occur in the bcl-1 major translocation cluster (MTC),2,5–7,15 while up to 70% of t(14;18) occur in an untranslated region of the bcl-2 gene known as the major breakpoint region (MBR).17–19 Universal polymerase chain reaction (PCR) assays can be designed to target the t(11;14) and t(14;18) molecular markers

for NHL. Primers hybridizing to consensus sequences in the JH region and the MTC or MBR may direct logarithmic PCR amplification only when translocations have occurred. Sensitivities as low as one cell in a 105–106 cell background can be achieved by performing a second, nested reaction with internal oligonucleotide pairs, or hybridizing PCR products with labelled probes.4,15,19–25 In spite of these advantages, nested PCR and blot analysis are time-consuming procedures that fail to be consistently quantitative. Using internal competitors or controls, ending reactions within the logarithmic amplification phase, or performing dilution experiments may improve template DNA quantitation, but these methods are often difficult to optimize, require large amounts of DNA, or rely exclusively on endpoint data collection.26–29 Since slight errors in amplification or shifts in PCR efficiency are logarithmically amplified during the cycling protocol, endpoint measurements are prone to error.30,31 The TaqMan real-time PCR assays have been shown to be an equally sensitive and more quantitative alternative to conventional single-round or nested PCR assays.31–33 The TaqMan assays include a nonextendable doubly labelled fluorescent probe that hybridizes internally to the primer pair. During each round of amplification, the 5⬘ nucleolytic activity of Taq polymerase cleaves the probe and results in increased reporter dye fluorescence that can be measured real-time. The cycle number at which samples’ fluorescence levels exceed a userdefined threshold may be correlated with starting copy number.34–37 This paper describes the evaluation and applications of two real-time TaqMan PCR assays for detecting t(14;18) and t(11;14) in NHL. A blinded comparison of the t(14;18) realtime PCR and a clinically validated nested PCR assay was performed using 134 cell line and patient DNA samples. Both the t(11;14) and t(14;18) real-time PCR assays were also used to detect minimal residual disease (MRD) in a phase I/II NHL clinical trial and for measuring the tumor-purging efficacy of an in-house cell separation method. Through optimization experiments and analysis of clinical samples, it was concluded that the real-time assays not only provide very sensitive, quantitative, and specific measures of MRD, but can also function as a rapid and suitable alternative to standard nested PCR. Materials and methods

Real-time t(11;14) and t(14;18) PCR Correspondence: EG Hanania, Oncosis, 6199 Cornerstone Court, Suite 111, San Diego, CA 92121, USA; Fax: (858) 550-1774 Current addresses: 2Targeted Genetics Corporation, 1100 Olive Way, Suite 100, Seattle, WA 98101 USA; 4Hollis-Eden Pharmaceuticals, Inc., 9333 Genesee Avenue, Suite 110, San Diego, CA 92121, USA; 5 Oncosis, 6199 Cornerstone Court, Suite 111, San Diego, CA 92121, USA Received 3 February 1999; accepted 16 June 1999

DNA isolation and quantitation: Genomic DNA was extracted from clinical blood samples using the QiaAmp Blood Kit (Qiagen, Chatsworth, CA, USA) and from normal peripheral blood and bone marrow samples using the QiaAmp Blood and Cell Culture DNA Maxi Kit (Qiagen). All DNA was eluted in 50–800 ␮l 10 mM Tris, pH 8.0

Real-time PCR provides sensitive and quantitative assessment of MRD K Olsson et al

1834

(BioWhittaker, Walkersville, MD, USA). DNA was quantitated by comparing fluorescence emitted by standard or unknown DNA after binding to picoGreen dye (picoGreen dsDNA Quantitation Kit; Molecular Probes, Eugene, OR, USA). Unknown samples were diluted 1:100 and 1:1000 in 1× TE, and a series of standards were prepared by serially diluting lambda DNA (2 ng/␮l) 1:5 into 1× TE buffer. A 100 ␮l volume of sample DNA was added to 100 ␮l of 1× picoGreen dye in clear 96-well flat-bottomed Nunc-Immuno Plates (Nalge Nunc International, Rochester, NY, USA), and plates were read using the CytoFluor II Microplate Fluorescence Reader (Biosearch, Subsidiary of Millipore Corporation, Bedford, MA, USA). Each plate contained at least eight no-template controls. Regression analysis was performed on fluorescence data of standards to assess the DNA concentration of unknown samples. Preparation of standards: DNA derived both from peripheral blood lymphocytes (PBLs) of normal donors and cell lines harboring the bcl-1 (MTC) and bcl-2 (MBR) translocations was diluted to a concentration of 0.03 ␮g/␮l with 10 mM Tris, pH 8.0. Standard A contained 100% cell line DNA, and Standard B contained 20% cell line DNA in a PBL background. Standard B was serially diluted 1:10 into PBL DNA to prepare Standards C–G. Standard H was a PBL negative control. Cell line DNA was isolated from MO2058 cells (obtained from Timothy Meeker, MD; University of Kentucky, Lexington, KY, USA)38 for the t(11;14) assay and from SUDHL-6 cells (Cell Therapeutics, Seattle, WA, USA) for the t(14;18) assay. Protocols for t(11;14) and t(14;18) quantitative PCR: The t(11;14) quantitative assay amplified 0.3 ␮g of DNA (equivalent to 50 000 cells) in duplicate using 0.65 ␮M MTC primer (5⬘-TGG ATA AAG GCG AGG AGC ATA A-3⬘), 0.5 ␮M JH primer (5⬘-ACC TGA GGA GAC GGT GAC C-3⬘), 0.2 ␮M MTC probe (5⬘-ACT GCA TAT TCG GTT AGA CTG TGA TTA GCT TT-3⬘), 5% dimethyl sulfoxide (Sigma Chemical, St Louis, MO, USA), and standard master mix. The t(14;18) quantitative assay amplified 0.15–0.30 ␮g of DNA in two to four replicates using 0.5 ␮M MBR primer (5⬘-GTG TTG TCC CTT TGA CCT TGT TTC-3⬘), 0.5 ␮M JH primer (5⬘-GAC CTG AGG AGA CGG TGA CC-3⬘), 0.2 ␮M MBR probe (5⬘-TCT GTG TTG AAA CAG GCC ACG TAA AGC A-3⬘), and standard master mix. Single dGTP nucleotides were added to the 5⬘ end of the MBR and JH primers to facilitate TA cloning. All oligonucleotide primers and probes were synthesized by the Oligo Factory (PerkinElmer Applied Biosystems Division, Foster City, CA, USA). The standard master mix for both assays was composed of reagents obtained from Perkin-Elmer Corporation (Norwalk, CT, USA), and included 10% TaqMan 10× buffer, 2.5 mM MgCl2, 0.2 ␮M of dATP, dCTP, and dGTP, 0.4 ␮M dUTP, 0.5 units AmpErase Uracil N-Glycosylase (UNG), and 1.5 units TaqGold DNA polymerase. All reactions were brought to a 50 ␮l volume using sterile H2O (Baxter Healthcare Corporation, Deerfield, IL, USA). Reactions were placed in MicroAmp Optical Tubes and covered with MicroAmp Optical Caps (Perkin-Elmer Applied Biosystems Division). PCR was completed in an ABI PRISM 7700 Thermal Cycler (Perkin-Elmer Applied Biosystems Division), and included a 2-min 50°C UNG activation, a 10-min 95°C TaqGold activation and pre-denaturation, and 42–45 cycles, each consisting of a 15-s 95°C denaturation and a 1-min 60–62°C

annealing step. Data were normalized to the quencher dye TAMRA and analyzed using the Sequence Detection software (Perkin-Elmer Applied Biosystems Division). One cell equivalent (CE) was assumed to contain 6 pg DNA.31 Percent tumor in unknowns was calculated by performing weighted averages of all data collected for a given sample. PCR results were confirmed through electrophoresis with a 3% NuSieve/1% SeaKem GTG Agarose gel (FMC BioProducts, Rockland, ME, USA) in 1× TBE buffer (BioWhittaker, Walkersville, MD, USA) according to the protocol previously described.31

Clinically validated t(14;18) nested PCR DNA isolation and quantitation: DNA from frozen tissue was extracted using the salting-out protocol from the Puregene DNA Isolation Kit (Gentra Systems, Minneapolis, MN, USA), and DNA from formalin-fixed and paraffin-embedded specimens was extracted using a previously described method.39 All DNA was quantitated using OD 260 measurements. Validated nested t(14;18) PCR protocol: The t(14;18) nested PCR was adapted from Poteat et al40 and validated for use in the Department of Pathology at Stanford University Medical Center. The reaction mix for the first round of PCR consisted of 1 ␮g DNA, 2.5 U AmpliTaq Gold polymerase (Perkin-Elmer Corporation), 100 ng of both external JH (5⬘ACC TGA GGA GAC GGT GAC C-3⬘) and MBR (5⬘-CAG CCT TGA AAC ATT GAT GG-3⬘) primers, 200 ␮M of each dNTP, 1.5 mM MgCl2, and 1× Perkin-Elmer PCR Buffer II in a 50 ␮l volume. Primers directed to Intron 10 of the factor V gene were also included in the first-round mix to confirm the amplifiability of all DNA samples. The touch-down PCR was performed in a thermal cycler PE 9600 (Perkin-Elmer Applied Biosystems Division) and consisted of TaqGold activation for 10 min at 95°C and 11 touch-down cycles of a 30-s denaturation at 95°C and a 30-s annealing step at 65°C (which decreased by 1°C per subsequent cycle). This was followed by 19 standard cycles of a 30-s denaturation at 95°C and a 30-s annealing step at 55°C. A final extension was performed at 72°C for 10 min. The second-round reaction mix contained no endogenous control, and included 200 ng of both the internal JH (5⬘-ACC AGG GTC CCT TGG CCC CA-3⬘) and MBR (5⬘-TCT ATG GTG GTT TGA CCT TTA G-3⬘) primers and 5 ␮l of a 1/100 dilution of the first-round product. Otherwise, the second-round mix and cycling protocol were identical to those of the first round. Analysis of PCR amplification was performed through electrophoresis with a 2% ethidiumstained agarose gel, and confirmed through standard Southern blot analysis when necessary. Positive, negative, and internal controls were analyzed concurrently with patient samples.

Comparison of t(14;18) detection in 134 samples by real-time PCR and clinically validated nested PCR One hundred and thirty-four DNA samples were obtained from Stanford University’s Department of Pathology and analyzed in a blinded comparison of real-time PCR and the clinically validated nested PCR protocols described above. Samples consisted of two cell lines positive for the t(14;18), OCI and SU-DHL-4, as well as histopathologically characterized clinical samples from patients with NHL, mantle cell lymphoma, Burkitt’s lymphoma, Hodgkin’s disease, Sezary

Real-time PCR provides sensitive and quantitative assessment of MRD K Olsson et al

syndrome, myeloma, T-LGL, renal disease and thrombosis. All samples were coded. Nested PCR data were generated by the Stanford Department of Pathology during routine sample analysis. These clinical samples were quantitated with the picoGreen assay after they were appropriately diluted for use in real-time PCR. At least 1.2 ␮g DNA (when available) was then analyzed in one to three independent experiments.

Use of t(11;14) and t(14;18) real-time PCR assays in NHL MRD analysis Real-time t(11;14) and t(14;18) PCR assays were used to monitor bone marrow and peripheral blood samples for six evaluable patients (two with detectable t(11;14) and four with detectable t(14;18)) in a phase I/II clinical trial designed to test the safety and efficacy of a tumor purging process for peripheral blood stem cell transplantation.41 The study consisted of 21 patients and included 12 diagnosed with follicular lymphoma and seven diagnosed with mantle cell lymphoma. When available, 2 to 4 × 105 cells were aliquoted for MRD analysis at time of apheresis and after both CD34+ and CD34+ Thy-1+ enrichment. Bone marrow samples were also collected for PCR analysis at accrual and day 100 post-transplant.

of the t(11;14) and t(14;18) assays, and final primer/probe combinations were selected based on assay sensitivity, specificity, and strength of signal in dilution experiments with MO2058 and SU-DHL-6 standards. The final t(11;14) assay consisted of a MTC primer and probe located on the sense strand of the cyclin D1 sequence, upstream of commonly observed MTC breakpoints (Figure 1a). The t(14;18) primer and probe were located in the sense and antisense strand (respectively) of the bcl-2 gene; breakpoints were generally detected downstream of the probe’s hybridization region on the sense strand (Figure 1b). Once primers and probes were selected, the t(11;14) and t(14;18) real-time assays were optimized by measuring effects of adjuncts and varied magnesium and primer concentrations on detection of serially diluted standards. After optimization, both assays were capable of detecting less than or equal to 1 cell equivalent (CE) in a 2.5 to 5.0 × 104 CE background (0.15–0.30 ␮g total DNA). Standard curves consistently reflected an R2 value of greater than or equal to 0.99, and the quantitative range for both assays extended for greater than four logs. Figures 2 and 3 show amplification plots and standard curves obtained for typical experiments using the t(11;14) and t(14;18) PCR assays.

Comparison of t(14;18) nested and real-time PCR in dilution experiment Evaluation of tumor purging processes with t(11;14) and t(14;18) real-time PCR An ex vivo tumor spiking model was designed to evaluate the purging potential of clinical cell processing stages. Ten clinical scale tissues (each containing greater than 1.6 × 1010 mobilized peripheral-blood cells) from normal donors were spiked with different combinations of tumor cell lines representing breast cancer, NHL, and multiple myeloma. Both the SUDHL-6 and MO2058 cells, which carry the t(14;18) and t(11;14) translocations respectively, were spiked into four tissues. The cell processing schema, which was identical to the clinical protocol used in the phase I/II trial, included CD34+ enrichment with the Isolex 300 Magnetic Cell Separator (Nexell Therapeutics, Irvine, CA, USA) followed by CD34+ Thy-1+ cell selection with high-speed sorters (SyStemix, A Novartis Company, Palo Alto, CA, USA). Samples collected at the different separation phases were analyzed using t(11;14) and t(14;18) real-time PCR in at least three independent experiments. Results

Design of t(11;14) and t(14;18) quantitative PCR assays Primers and probes for the t(11;14) and t(14;18) real-time PCR assays were designed using OLIGO Primer Analysis Software, version 5.0 for Macintosh (Informagen, Newington, NH, USA). When possible, dGTP/dCTP content of oligos did not exceed 40–70%, melting temperatures (Tms) of primer pairs were comparable and approximately 10°C greater than the Tm of the corresponding probe (calculated using the nearestneighbor method), and the probes’ 5⬘ end did not contain dGTPs. A dGTP was added to the 5⬘ end of both MBR and JH primers to facilitate TA cloning, and had no observable effect on the sensitivity and specificity of the PCR (data not shown). Several primers and probes were tested during development

A modified t(14;18) nested PCR protocol described in Figure 4 was used to evaluate directly the detection limits of nested and real-time PCR. The first round of nested PCR resulted in a 2.0 × 10−2 detection limit (corresponding to 1000 CEs in 50 000), and a second, nested reaction increased sensitivity by an additional two logs. Typical t(14;18) real-time assays could consistently detect template concentrations of 2.0 to 4.0 × 10−4, and achieved an additional one-log sensitivity for at least 50% of replicates (corresponding to single-copy detection). Following nested PCR, relative starting copy numbers of serial dilutions could not be accurately resolved through gel densitometry due to signal saturation (data not shown). By contrast, regression analysis of real-time data for the serially diluted standards yielded R2 values greater than or equal to 0.99. It was therefore concluded that real-time t(14;18) PCR was more quantitative and at least as sensitive as a traditional nested PCR approach.

Comparison of t(14;18) detection in 134 samples by real-time PCR and clinically validated nested PCR protocol Concordance in binary data acquired through real-time t(14;18) PCR or a validated nested PCR protocol was observed for 122 of 134 samples analyzed (91.0%). The translocation was detectable with real-time PCR in nine of the 12 discrepant cases, and was detectable with nested PCR in three of the 12 discrepant cases. Both nested and real-time PCR detected translocations in samples for one renal patient and one patient with T cell lymphoma; Southern blot hybridization reactions confirmed the presence of t(14;18) breakpoints in both samples. Twenty-nine samples possessed detectable t(14;18) in one or both PCR assays. Of these, 26 samples were positive with real-time PCR, while only 20 were positive with nested PCR. All samples included in the study were concluded to be amplifiable in the validated nested PCR using the factor V endogenous control.

1835

Real-time PCR provides sensitive and quantitative assessment of MRD K Olsson et al

1836

Figure 1 (a) Partial sequence of the PRAD-1 gene showing the position of the MTC primer (`) and MTC probe (䊉–䊉) used in the real-time PCR assay. The forward arrow (→) indicates the amplification direction. The black inverted triangles (왔) show the position of the common breakpoints reported in the literature. The basepair numbers on the right are arbitrary and are used to show the relative distances among the MTC primer, probe and common breakpoints on the sense strand. (b) Partial sequence of the bcl-2 gene showing the position of the MBR primer and probe used in the real-time PCR assay. Also shown are the EXT and INT primers corresponding to the external and internal primers used for the nested PCR. The forward and backward arrows (→ ←) indicate the amplification direction. The MBR TaqMan probe on the antisense strand is indicated by the gray double-circled arrow (䊉–䊉). The inverted triangles (왔) show the position of the common breakpoints reported in the literature. The basepair numbers on the right are arbitrary and are used to show the relative distances among the various MBR primers, TaqMan probe and common breakpoints on the sense strand.

MRD analysis for phase I/II NHL clinical trial The t(11;14) and t(14;18) real-time PCR assays were used to analyze baseline and follow-up samples from six evaluable NHL patients. Bone marrow samples were collected for all patients prior to autologous peripheral blood stem cell (PBSC) transplantation and at day 100. The t(11;14) translocation was detected in baseline bone marrow samples for two patients at 0.0002% and 0.1260%, and the t(14;18) translocation was detected in baseline bone marrow samples for four patients at levels ranging from 0.0008% to 0.0602% (Table 1). One t(11;14)-positive patient showed an increase in tumor levels

at day 100 post-transplant, but all other patients showed a decrease in malignant cell numbers during follow-up. The t(11;14) and t(14;18) PCR assays were also used to analyze samples acquired during the CD34+ and CD34+ Thy-1+ enrichment process. Absolute fold reduction in tumor load after CD34+ enrichment ranged from greater than or equal to 2.78 logs to as high as 5.07 logs. Absolute fold reduction in tumor burden after sorting for CD34+ Thy-1+ was always greater than or equal to 3.26 logs (Table 1). Amplifiability of DNA was assessed through a real-time ␤-actin assay using a probe and primers from the TaqMan ␤-actin Detection Reagents Kit (Perkin-Elmer Applied Biosystems Division). All

Real-time PCR provides sensitive and quantitative assessment of MRD K Olsson et al

1837

Figure 2 Amplification plots, standard curve and gel analysis from a single t(11;14) real-time PCR assay. (a) Amplification plot graphing the change in fluorescence (⌬Rn) vs cycle number for two replicates of each of eight standards labeled A to H. The threshold line (indicated by the arrow) was set within all samples’ range of exponential amplification, and the threshold cycle (Ct) was defined as the cycle number at which a given sample’s ⌬Rn crossed the threshold. No logarithmic signal increase was observed for standards G or H. (b) Standard curve calculated for the real-time t(11;14) assay using serial dilutions of MO2058 DNA. Threshold cycle numbers (Ct) for standards A to E were graphed against the log of their starting quantities and the plot was fitted with a least-squares regression equation. (c and d) Post-PCR analysis of the standards A to H. Samples were loaded into an agarose gel and electrophoresis was performed at 200 volts. Gel bands were visualized by staining with ethidium bromide. The two gels represent the duplicate sets of standards used. Bands were seen for standards A to F, which is in agreement with the TaqMan analysis. Lane L is the 100-basepair DNA marker.

samples which were not consistently PCR positive were concluded to contain tumor levels at or below the assays’ sensitivity limits (notated by ‘⭐’ in Table 1).

Evaluation of tumor purging processes in the ex vivo spike model Mean tumor depletion for all spiking experiments was comparable using either the t(11;14) and t(14;18) assays (4.99 ± 0.075 for t(14;18) and 5.35 ± 0.22 for t(11;14)). Table 2 shows quantitative PCR data obtained for one representative experiment. The SU-DHL-6 cell concentration was initially 0.124%, but was reduced by greater than or equal to 4.86 logs during processing (⭓2.44 logs through CD34+ enrichment and ⭓2.42 logs through CD34+ Thy-1+ sorting). Similarly, MO2058 was depleted by greater than or equal to 5.13 logs from its initial concentration of 0.230%.

Discussion The t(11;14) and t(14;18) real-time PCR assays provide rapid and convenient alternatives to traditional PCR approaches. Detection of probe cleavage during PCR amplification is at least as sensitive as nested PCR, and real-time data collection enables improved quantitation of templates’ starting copy number. Use of a control for carry-over contamination (UNG/dUTP) and elimination of post-PCR processing steps greatly reduce the risk of false positives, and inclusion of a template-specific probe increases reaction stringency, limiting real-time detection of spurious and nonspecific PCR products. Moreover, the universal t(11;14) and t(14;18) PCRs require only small amounts of DNA (less than or equal to 300 ng per reaction), eliminate the need for developing patient-specific assays, and can be used in conjunction with endogenous control primers and probes in either multiplexed or separatetube reactions.

Real-time PCR provides sensitive and quantitative assessment of MRD K Olsson et al

1838

Figure 3 Amplification plots, standard curve and gel analysis from a single t(14;18) real-time PCR assay. (a) Amplification plot graphing the change in fluorescence (⌬Rn) vs cycle number for two replicates of each of 8 standards A to H. The threshold line (indicated by the arrow) was set within all samples’ range of exponential amplification, and the threshold cycle (Ct) was defined as the cycle number at which a given sample’s ⌬Rn crossed the threshold. Only one of the replicates for F was positive, while neither replicate for G or H exhibited logarithmic signal increase. (b) Standard curve calculated for the real-time t(14;18) assay using serial dilutions of SU-DHL-6 DNA. Standards’ threshold cycle numbers (Ct) were graphed against the log of their starting quantities, and the plot was fitted with a least-squares regression equation. The equation was then used to estimate the initial target copy number for unknown samples. Only standards A to E were used to construct this curve. (c and d) Post-PCR analysis of the standards A to H. Gel bands were visualized by staining with ethidium bromide and capturing the digital image using Stratagene’s Eagle Eye II Still Video System. The two gels represent the duplicate sets of standards used. As predicted by the real-time PCR data, a band was seen for only one of the replicates for F (Figure 3a). Lane L is the 100-basepair DNA marker.

When developing the t(11;14) and t(14;18) real-time assays, PCR sensitivity and specificity were enhanced by selecting optimal primer and probe combinations. In the t(11;14) PCR, the most sensitive MTC primer and probe were located together on the sense strand. Although this could potentially enable probe cleavage during amplification of the untranslocated allele, no fluorescence increase was detected in negative control samples. The probe selected for the t(14;18) realtime assay was located on the antisense strand, and therefore could only be cleaved by extension from the consensus JH primer following a translocation event. While placing the probe closer to potential breakpoints (Figure 1) could increase the probability of a false negative result, it reduced amplicon size and thus likely improved amplification efficiency and sensitivity. In order to compensate for observed instability of the internal reference dye (ROX) during the cycling protocol, data for both assays were normalized using the quencher dye TAMRA. This improved detection of low-copy-number samples and frequently improved R2 values of the regression equations (data not shown). Both the t(11;14) and t(14;18) real-time assays were optim-

ized until they consistently achieved a 2.0 to 4.0 × 10−4 detection limit and reflected an additional 10-fold sensitivity in at least 50% of replicates (corresponding roughly to the Poisson probability of observing single-copy events). Since the t(14;18) nested PCR assay described in Figure 4 reflected a similar limit of detection (2.0 × 10−4), it was concluded that the optimized real-time PCR assays were at least as sensitive as nested PCR. These observations support the conclusions of Do¨lken et al,33 who found that a t(14;18) real-time assay’s detection limit was at least comparable to that of nested PCR, and Luthra et al,32 who observed that single-round conventional PCR reflected a two-log lower detection limit than a t(14;18) real-time assay (1 × 10−3 vs 1 × 10−5, respectively). Although hybridizing PCR products with internal radiolabelled probes might further increase nested PCR sensitivity, it is a labor-intensive procedure that would not be practical in a diagnostic laboratory environment. It has been documented that PCR assays relying on endpoint data collection are usually semiquantitative at best.31 Through the dilution experiment used to compare real-time and nested t(14;18) PCR, it was observed that plateaus in pro-

Real-time PCR provides sensitive and quantitative assessment of MRD K Olsson et al

Figure 4 Evaluation of the limit of detection (LOD) for t(14;18) nested PCR using SU-DHL-6 standards A to H. The assay was adapted from Gribben et al43 and used external JH (5⬘-ACC TGA GGA GAC GGT GAC-3⬘), external MBR (5⬘-CAG CCT TGA AAC ATT GAT GG3⬘), internal JH (5⬘-ACC AGG GTC CCT TGG CCC CA-3⬘) and internal MBR (5⬘-TAT GGT GGT TTG ACC TTT AG-3⬘) primers. Reaction mixes consisted of 200 ng of each primer, 200 ␮M of each dNTP, 2.25 mM MgCl2, and I⫻ Perkin-Elmer PCR Buffer II in a 50 ␮l volume. The cycling protocol for each round included 25–30 cycles of a 1min denaturation (94°C), a 1-min annealing step (55 to 58°C), and a 1-min extension (72°C). Final 10-min extensions (72°C) were also performed. Panels (a) and (b) show the amplification from the external and internal MBR primers respectively. The LOD for the first-round PCR is 2.0 × 10−2 as indicated by the faint bands for B and C (Figure 4a). The second nested round increases the LOD to 2.0 × 10−4 as shown by strong bands through dilution E (Figure 4b). Although the sensitivity of the PCR is improved by the nested reaction and the LOD is close to that of the real-time PCR (Figure 3a), no quantitation can be performed due to strong saturated signal throughout the dilution curve, A–E. The first lane L on the gel is a 100-basepair DNA marker. Duplicate samples of sample H were loaded on each of the two gels.

duct accumulation restricted accurate quantitation of starting copy number following nested PCR. Real-time PCR, by contrast, was very effective in quantitating template concentrations at or above 5 to 10 CEs (苲0.02% of total CEs) per reaction. Samples containing detectable tumor DNA below this level reflected greater data variability due to increased significance of sampling errors and stochastic effects.5,42 In order to minimize errors in quantitation and increase assay sensitivity, reactions containing less than 0.02% tumor levels were excluded from regression analysis, and at least 200 000 total CEs were analyzed in two or more real-time PCR assays when DNA was available. Moreover, since amplification efficiencies of standard and unknown samples often varied due to differences in amplicon size, real-time PCR was only used to compare tumor levels in samples from a single patient

source. Should direct comparisons of DNA samples from different patients or tissues become necessary, the assays could be modified to include multiplexed endogenous control reactions using probes with distinct reporter dyes. A blinded comparison of nested and real-time t(14;18) PCR was performed using 134 cell line and clinical DNA samples. The t(14;18) target was chosen over t(11;14) due to the increased frequency of t(14;18) translocations in NHL patient populations. The nested PCR assay was validated by the Department of Pathology at Stanford University Medical Center for use as the standard molecular method for detecting t(14;18). Our study showed a 91.0% correlation between nested and real-time PCR. Nine of the 12 discordant samples were positive in realtime PCR but negative in nested PCR. One of these was obtained from a patient who has since relapsed, suggesting that the real-time data has positive predictive value. Three samples were positive in nested PCR and negative in real-time PCR. Gel electrophoresis performed after real-time PCR revealed potentially specific 200–500 basepair amplification products. The amplicons’ specificities must be confirmed through sequence analysis, but these preliminary data suggest that the samples’ breakpoints were located within the TaqMan probe hybridization region and thus disrupted probe binding and cleavage but not PCR amplification. The real-time and nested t(14;18) PCR assays used in our comparison study demonstrated comparable sensitivities (detecting approximately 1 in 104–105 total CEs) in independent validation experiments with serially diluted standards (data not shown). Thus, the 30.0% increase in t(14;18)-positive samples detected in the real-time vs nested PCR assays was attributed primarily to the effects of breakpoint location and oligonucleotide placement (Figure 1). While the MBR probe hybridized 32 basepairs downstream of the internal nested MBR primer, the internal nested JH primer extended 17 basepairs downstream of the real-time assay’s JH primer. Since fewer samples were positive in nested PCR, it was concluded that breakpoints were more likely to be located within the nested internal JH primer sequence. As shown in Figure 4, failure to complete a second-round PCR would reduce nested assay sensitivity by up to two logs, and thus limit detection of t(14;18) in clinical samples. Among the 17 samples that contained detectable t(14;18) in both assays, two were from patients with renal disease and T cell lymphoma. Detection of the t(14;18) could be attributed to a false initial diagnosis or undetected B cell lymphoma. Gribben et al43 failed to detect the translocation in bone marrow samples of 12 normal donors using a nested PCR protocol. However, several studies have shown that approximately 43.0 to 66.7% of normal individuals possess detectable levels of the t(14;18) translocation.42,44,45 Moreover, Limpens et al45 observed that frequency of normal donors with detectable t(14;18) greatly exceeds the number of individuals with follicular lymphoma, suggesting that presence of t(14;18) may not reflect a subclinical malignancy. Until the relationship between malignancy and detectable t(14;18) is firmly established, the real-time PCR targeting the MBR should not be used as the only means of diagnosing NHL, but can provide a sensitive and specific means of assessing MRD once a patient is confirmed positive. The t(11;14) and t(14;18) real-time PCR assays were used in the SyStemix Phase I/II NHL clinical trial to evaluate baseline and follow-up patient samples as well as assess the purging potential of cell-enrichment processes. At day 100 posttransplant, malignancy was undetectable in bone marrow samples from two previously t(14;18)-positive patients and

1839

Real-time PCR provides sensitive and quantitative assessment of MRD K Olsson et al

1840

Table 1

Evaluation of tumor reduction in an NHL clinical trial setting

Patient

CD34+

Apheresis No. of tumor cells

CD34+/Thy-1+

No. of tumor Absolute fold cells reduction (log10) 7.44 × 106 0.82 × 106 10.5 × 106 117 × 106 12.1 × 106 45.1 × 106

105a 111 119 121a 125 126

NA ⭐1360 2620 NA 1360 384

No. of tumor Absolute fold cells reduction (log10) ⭐360 ⭐451 ⭐771 ⭐261 ⭐457 ⭐830

NA ⭓2.78 3.6 NA 3.95 5.07

⭓4.32 ⭓3.26 ⭓4.14 ⭓5.65 ⭓5.68 ⭓5.61

Baseline BM percent tumor cells

Day 100 BM percent tumor cells

0.0002 0.0008 0.0602 0.1260 0.0119 0.0010

0.0005 ⭐0.0005 ⭐0.0005 NA 0.0003 0.0003

The t(11;14) and t(14;18) real-time PCR assays were used to analyze samples acquired during CD34+ and CD34+/Thy-1+ enrichment for six evaluable patients who were confirmed to harbor the t(11;14) (denoted bya) or the t(14;18) translocations. Absolute fold reduction in tumor load after CD34+ ranged from ⭓2.78 logs to 5.07 logs. Absolute fold reduction in tumor burden following CD34+/Thy-1+ enrichment ranged from ⭓3.26 logs to ⭓5.68 logs, reflecting an approximate 0.5 log enhancement in tumor elimination. The same assays were also used to analyze baseline bone marrow samples collected prior to transplantation and at day 100 post-transplant. Tumor cells found in the baseline bone marrow samples ranged from 0.0002 to 0.126% as compared to ⭐0.0005–0.0003% detected in the day 100 samples. Since no tumor cells were detected at day 100 for patients 111 and 119, it was assumed that actual tumor levels were ⭐0.0005% (representing the t(14;18) assay’s lower limits of detection).

Table 2

Implementation of bcl-2 and bcl-1 real-time assays in an ex vivo tumor spike model

Assay ⇒ Samples ⇓

bcl-2/MBR real-time PCR Total cells

No. tumor cells

% tumor cells

Absolute depletion Fold

Pre-Isolex Spiked pre Isolex Post Isolex CD34+ Post Isolex CD34− Presort post stain Post sort CD34+/Thy-1+ Post sort CD34+/Thy-1− Overall process

1.74 × 1010 1.61 × 1010

1.99 × 107

0.124

1.64 × 108

7.22 × 104

0.044

1.87 × 1010

2.20 × 107

1.14 × 108

bcl-1/MTC real-time PCR No. tumor cells

% tumor cells

Log

Fold 3.69 × 107

0.230

9.02 × 104

0.55

0.118

3.31 × 107

0.177

4.10 × 104

0.036

3.42 × 104

0.030

2.75 × 107

2.75 × 102

0.001

2.75 × 102

0.001

6.92 × 107

7.61 × 103

0.011

5.53 × 103

0.008

⭓275.80

⭓262.40

⭓72370.91

⭓2.44

⭓2.42

⭓4.86

Absolute depletion Log

⭓409.26

⭓2.61

⭓328.00

⭓2.52

⭓134236.36

⭓5.13

Detailed analysis of a single representative ex vivo tumor spike experiment. 1.74 × 1010 PBLs (normal donor apheresis) were spiked with tumor cell lines including SU-DHL-6 and MO2058 cells. More than 2.4 logs of absolute tumor depletion were achieved by each stage of the cell processing protocol, resulting in more than 4.8 log tumor depletion through the entire process. In the table, each value represents the mean of at least three PCR assays with samples run in duplicate.

had decreased from baseline levels in three additional patients. One t(11;14)-positive patient showed an increase in bone marrow tumor levels during follow-up, but additional DNA samples would be required to accurately assess the significance of this increase. When calculating absolute tumor depletion of the different processing steps, samples with undetectable t(11;14) or t(14;18) were concluded to possess tumor levels at the assays’ sensitivity limits. Therefore, absolute fold reduction values for one CD34+-selected sample and all CD34+ Thy-1+-enriched samples were probable underestimates. Accurate assessment of tumor purging efficiency in the NHL clinical trial was limited largely by the number of cells available for analysis. An ex vivo tumor spiking model was therefore developed in order to evaluate at least 1.0 to 1.5 × 105

CEs of DNA from each processing stage in both the t(11;14) and t(14;18) assays. Predicted purging potential for the combined CD34+ and CD34+Thy-1+ enrichment processes was approximately equivalent using either t(11;14) or t(14;18) realtime PCR (Table 2). Slight discrepancies in PCR tumor purging data could be attributed to variability in cell sampling, the generic nature of the PCR assays, or effects of target (MBR vs MTC) differences on assay performance. Quantitating the purging potential of CD34+ and CD34+Thy-1+ enrichment methods and patients’ molecular tumor levels is very important for assessing the potential benefit of treatment programs, and establishing patient responses to therapy. We have presented two real-time PCR assays targeting the t(11;14) and t(14;18) in B cell lymphomas. By evaluating clinical samples using these universal PCR assays, one

Real-time PCR provides sensitive and quantitative assessment of MRD K Olsson et al

may obtain more quantitative and rapid assessments of minimal residual disease, and ultimately, more accurately predict long-term prognosis. References 1 Aisenberg AC, Wilkes BM, Jacobson JO. The bcl-2 gene is rearranged in many diffuse B-cell lymphomas. Blood 1988; 71: 969–972. 2 Crescenzi M, Seto M, Herzig GP, Weiss PD, Griffith RC, Korsmeyer SJ. Thermostable DNA polymerase chain amplification of t(14;18) chromosome breakpoints and detection of minimal residual disease. Proc Natl Acad Sci USA 1988; 85: 4869–4873. 3 Graninger WB, Seto M, Boutain B, Goldman P, Korsmeyer SJ. Expression of Bcl-2 and Bcl-2-Ig fusion transcripts in normal and neoplastic cells. J Clin Invest 1987; 80: 1512–1515. 4 Gribben JG, Nadler LM. Detection of minimal residual disease in patients with lymphomas using the polymerase chain reaction. Important Adv Oncol 1994; 117–129. 5 Lambrechts AC, de Ruiter PE, Dorssers LC, vant Veer MB. Detection of residual disease in translocation (14;18) positive non-Hodgkin’s lymphoma, using the polymerase chain reaction: a comparison with conventional staging methods. Leukemia 1992; 6: 29–34. 6 Korsmeyer SJ. Chromosomal translocations in lymphoid malignancies reveal novel proto-oncogenes. Annu Rev Immunol 1992; 10: 785–807. 7 Ngan BY, Nourse J, Cleary ML. Detection of chromosomal translocation t(14;18) within the minor cluster region of bcl-2 by polymerase chain reaction and direct genomic sequencing of the enzymatically amplified DNA in follicular lymphomas. Blood 1989; 73: 1759–1762. 8 Weiss LM, Warnke RA, Sklar J, Cleary ML. Molecular analysis of the t(14;18) chromosomal translocation in malignant lymphomas. New Engl J Med 1987; 317: 1185–1189. 9 Aguilera NS, Bijwaard KE, Duncan B, Krafft AE, Chu WS, Abbondanzo SL, Lichy JH, Taubenberger JK. Differential expression of cyclin D1 in mantle cell lymphoma and other non-Hodgkin’s lymphomas. Am J Pathol 1998; 153: 1969–1976. 10 Avet-Loiseau H, Garand R, Gaillard F, Daviet A, Mellerin MP, Robillard N, Bouyge I, Arcot S, Batzer M, Talmant P, Harousseau JL, Milpied N, Bataille R. Detection of t(11;14) using interphase molecular cytogenetics in mantle cell lymphoma and atypical chronic lymphocytic leukemia. Genes Chromosomes Cancer 1998; 23: 175–182. 11 Fan H, Gulley ML, Gascoyne RD, Horsman DE, Adomat SA, Cho CG. Molecular methods for detecting t(11;14) translocations in mantle-cell lymphomas. Diagn Mol Pathol 1998; 7: 209–214. 12 Nakamura S, Yatabe Y, Kuroda H, Ueda R, Takahashi T, Suchi T, Seto M. Immunostaining of PRAD1/cyclin D1 protein as a marker for the diagnosis of mantle cell lymphoma. Leukemia 1997; 11: 536–537. 13 Ott MM, Bartkova J, Bartek J, Durr A, Fischer L, Ott G, MullerHermelink HK, Kreipe H. Cyclin D1 expression in mantle cell lymphoma is accompanied by downregulation of cyclin D3 and is not related to the proliferative activity. Blood 1997; 90: 3154–3159. 14 Pott C, Tiemann M, Linke B, Ott MM, von Hofen M, Bolz I, Hiddemann W, Parwaresch R, Kneba M. Structure of Bcl-1 and IgHCDR3 rearrangements as clonal markers in mantle cell lymphomas. Leukemia 1998; 12: 1630–1637. 15 Griesser H. Applied molecular genetics in the diagnosis of malignant non-Hodgkin’s lymphoma. Diagn Mol Pathol 1993; 2: 177–191. 16 Tsujimoto Y, Gorham J, Cossman J, Jaffe E, Croce CM. The t(14;18) chromosome translocations involved in B-cell neoplasms result from mistakes in VDJ joining. Science 1985; 229: 1390–1393. 17 Chibbar R, Leung K, McCormick S, Ritzkalla K, Strickler J, Staggs R, Poppema S, Brunning RD, McGlennen RC. bcl-1 gene rearrangements in mantle cell lymphoma: a comprehensive analysis of 118 cases, including B-5-fixed tissue, by polymerase chain reaction and Southern transfer analysis. Mod Pathol 1998; 11: 1089–1097. 18 Luthra R, Hai S, Pugh WC. Polymerase chain reaction detection of the t(11;14) translocation involving the bcl-1 major translocation cluster in mantle cell lymphoma. Diagn Mol Pathol 1995; 4: 4–7.

19 Pinyol M, Campo E, Nadal A, Terol MJ, Jares P, Nayach I, Fernandez PL, Piris MA, Montserrat E, Cardesa A. Detection of the bcl1 rearrangement at the major translocation cluster in frozen and paraffin-embedded tissues of mantle cell lymphomas by polymerase chain reaction. Am J Clin Pathol 1996; 105: 532–537. 20 Andersen NS, Donovan JW, Borus JW, Poor CM, Neuberg D, Aster JC, Nadler LM, Freedman AS, Gribben JG. Failure of immunologic purging in mantle cell lymphoma assessed by polymerase chain reaction detection of minimal residual disease. Blood 1997; 90: 4212–4221. 21 Corradini P, Astolfi M, Cherasco C, Ladetto M, Voena C, Caracciolo D, Pileri A, Tarella C. Molecular monitoring of minimal residual disease in follicular and mantle cell non-Hodgkin’s lymphomas treated with high-dose chemotherapy and peripheral blood progenitor cell autografting. Blood 1997; 89: 724–731. 22 Gascoyne RD, Adomat SA, Krajewski S, Krajewska M, Horsman DE, Tolcher AW, O’Reilly SE, Hoskins P, Coldman AJ, Reed JC, Connors JM. Prognostic significance of Bcl-2 protein expression and Bcl-2 gene rearrangement in diffuse aggressive non-Hodgkin’s lymphoma. Blood 1997; 90: 244–251. 23 Hardingham JE, Kotasek D, Sage RE, Dobrovic A, Gooley T, Dale BM. Molecular detection of residual lymphoma cells in peripheral blood stem cell harvests and following autologous transplantation. Bone Marrow Transplant 1993; 11: 15–20. 24 Negrin RS, Kiem HP, Schmidt-Wolf IG, Blume KG, Cleary ML. Use of the polymerase chain reaction to monitor the effectiveness of ex vivo tumor cell purging. Blood 1991; 77: 654–660. 25 Scholten C, Hilgarth B, Hilgarth M, Jaeger U, Greinix HT, Mannhalter C. Predictive value of clone-specific CDR3 PCR in high-grade non-Hodgkin’s lymphoma: long-term persistence of a clone-specific rearrangement in a patient with secondary highgrade NHL in remission after high-dose therapy. Br J Haematol 1997; 97: 246–247. 26 Cave H, Van Der Werff Ten Bosch J, Suciu S, Guidal C, Waterkeyn C, Otten J, Bakkus M, Thielemans K, Grandchamp B, Vilmer E. Clinical significance of minimal residual disease in childhood acute lymphoblastic leukemia. New Engl J Med 1998; 339: 591–598. 27 Cross NC. Quantitative PCR techniques and applications. Br J Haematol 1995; 89: 693–697. 28 Heid CA, Stevens J, Livak KJ, Williams PM. Real time quantitative PCR. Genome Res 1996; 6: 986–994. 29 Lin F, Cross NC. Molecular methods for monitoring malignant disease after bone marrow transplantation. Curr Opin Hematol 1995; 2: 460–467. 30 Billadeau D, Blackstadt M, Greipp P, Kyle RA, Oken MM, Kay N, Van Ness B. Analysis of B-lymphoid malignancies using allelespecific polymerase chain reaction: a technique for sequential quantitation of residual disease. Blood 1991; 78: 3021–3024. 31 Gerard CJ, Olsson K, Ramanathan R, Reading C, Hanania EG. Improved quantitation of minimal residual disease in multiple myeloma using real-time polymerase chain reaction and plasmidDNA complementarity determining region III standards. Cancer Res 1998; 58: 3957–3964. 32 Luthra R, McBride JA, Cabanillas F, Sarris A. Novel 5⬘ exonuclease-based real-time PCR assay for the detection of t(14;18)(q32;q21) in patients with follicular lymphoma. Am J Pathol 1998; 153: 63–68. 33 Do¨lken L, Schu¨ler F, Do¨lken G. Quantitative detection of t(14;18)positive cells by real-time quantitative PCR using fluorogenic probes. Biotechniques 1998; 6: 1058–1064. 34 Gibson UE, Heid CA, Williams PM. A novel method for real time quantitative RT-PCR. Genome Res 1996; 6: 995–1001. 35 Holland PM, Abramson RD, Watson R, Gelfand DH. Detection of specific polymerase chain reaction product by utilizing the 5⬘→3⬘ exonuclease activity of Thermus aquaticus DNA polymerase. Proc Natl Acad Sci USA 1991; 88: 7276–7280. 36 Lee MS, Chang KS, Cabanillas F, Freireich EJ, Trujillo JM, Stass SA. Detection of minimal residual cells carrying the t(14;18) by DNA sequence amplification. Science 1987; 237: 175–178. 37 Livak KJ, Flood SJ, Marmaro J, Giusti W, Deetz K. Oligonucleotides with fluorescent dyes at opposite ends provide a quenched probe system useful for detecting PCR product and nucleic acid hybridization. PCR Meth Appl 1995; 4: 357–362. 38 Meeker TC, Sellers W, Harvey R, Withers D, Carey K, Xiao H,

1841

Real-time PCR provides sensitive and quantitative assessment of MRD K Olsson et al

1842

39 40 41

42

Block AM, Dadey B, Han T. Cloning of the t(11;14)(q13;q32) translocation breakpoints from two human leukemia cell lines. Leukemia 1991; 5: 733–737. Frank TS, Svoboda-Newman SM, Hsi ED. Comparison of methods for extracting DNA from formalin-fixed paraffin sections for nonisotopic PCR. Diagn Mol Pathol 1996; 5: 220–224. Poteat HT, Sklar J. A simplified polymerase chain reaction assay for detection of chromosomal translocations in hematologic malignancies. Diagn Mol Pathol 1997; 6: 3–9. Vose JM, Bierman PJ, Atkinson K, Lynch JC, Juttner C, Hanania E, Reading C, Ramanathan R, Armitage JO. Transplantation following high-dose chemo and CD34+ Thy-1+/Lin- peripheral blood stem cells (PBSC) in patients (PTS) with non-Hodgkin’s lymphoma (NHL). Blood 1998; 12: 726a (Abstr). Ji W, Qu GZ, Ye P, Zhang XY, Halabi S, Ehrlich M. Frequent detection of bcl-2/JH translocations in human blood and organ

samples by a quantitative polymerase chain reaction assay. Cancer Res 1995; 55: 2876–2882. 43 Gribben JG, Freedman AS, Woo SD, Blake K, Shu RS, Freeman G, Lontine JA, Pinkus JS, Nadler LM. All advanced stage nonHodgkin’s lymphomas with a polymerase chain reaction amplifiable breakpoint of bcl-2 have residual cells containing the bcl-2 rearrangement at evaluation and after treatment. Blood 1991; 78: 3275–3280. 44 Liu Y, Hernandez AM, Shibata D, Cortopassi GA. BCL2 translocation frequency rises with age in humans. Proc Natl Acad Sci USA 1994; 91: 8910–8914. 45 Limpens J, Stad R, Vos C, de Vlaam C, de Jong D, van Ommen GJ, Schuring E, Kluin PM. Lymphoma-associated translocation t(14;18) in blood B cells of normal individuals. Blood 1995; 85: 2528–2536.