Clonality profile in relapsed precursor-B-ALL children by GeneScan and sequencinganalyses. Consequences on minimal residual disease monitoring

G Germano1, L del Giudice1, S Palatron1, E Giarin1, G Cazzaniga2, A Biondi2 and G Basso1

1Laboratorio di Emato Oncologia, Dipartimento di Pediatria, Universita’ di Padova, Italy; and 2Centro Ricerca M Tettamanti,Clinica Pediatrica Universita’ di Milano Bicocca, Ospedale San Gerardo, Monza, Italy

Detection of minimal residual disease (MRD), using immuno-globulin (Ig) and T-cell receptor (TCR) gene rearrangements asclone-specific targets, represents the most recent developmentin diagnosis and treatment of acute lymphoblastic leukaemia(ALL). Nevertheless, risk of false-negative results, due tosecondary or ongoing rearrangements of Ig/TCR genes duringthe disease course, might hamper MRD detection. Therefore, togain extensive information on clonal stability, we performedPCR-GeneScan analysis of Ig/TCR gene rearrangements atdiagnosis and subsequent relapse in bone marrow samplesfrom 53 childhood precursor-B-ALL patients. In addition,sequencing analysis of junctional regions at diagnosis andrelapse provided a detailed insight in the stability and changesof Ig/TCR gene rearrangements during the disease course. Atleast one stable clonal Ig/TCR target was found in 94% ofpatients. In three patients complete differences in Ig/TCRrearrangements between diagnosis and relapse were observed,suggesting relapse with a new clone. At relapse, 71% ofdiagnostic clonal PCR targets was conserved. Since thecomparison of Ig/TCR gene rearrangements at diagnosis andrelapse in our precursor-B-ALL patients did not show signifi-cant difference in the stability of different clonal PCR targets(IGH, 70%; IGK, 71%; TCRD, 67%; TCRG, 75%), we concludethat there is no ‘preferential’ clone-specific target for MRDmonitoring.Leukemia (2003) 17, 1573–1582. doi:10.1038/sj.leu.2403008Keywords: relapse; precursor-B-ALL; Ig/TCR generearrangements; GeneScan analysis; MRD

Introduction

Despite substantial progress in treatment of newly diagnosedchildhood acute lymphoblastic leukaemia (ALL) with risk-tailored protocols, about 20–30% of children eventuallyrelapses.1,2 Recent clinical studies in ALL demonstrated theclinical relevance of minimal residual disease (MRD) detectionin the follow-up of the disease.3,4

The precise evaluation of MRD levels during and after thecourse of the disease is important to identify patients with higherrisk of recurrence that will therefore enter different risk-adaptedtherapeutic protocols,3,4 and it might help in giving a newdefinition of molecular remission in ALL.5,6 Quantitative MRDdata in the majority of haematological malignancies can beobtained by real-time quantitative PCR (RQ-PCR) using im-munoglobulin (Ig) and T-cell receptor (TCR) gene rearrange-ments7,8 as PCR targets. Ig and TCR rearrangements arecharacterized by a unique clone-specific junctional regions thatcan be considered as a leukaemia clone-specific fingerprint.9–11

At diagnosis these Ig and TCR rearrangements are routinelydetectable in about 90% of precursor-B-ALL and 95% of T-ALLpatients by PCR;9,12,13 a sensitivity of 10�4–10�6 can be reached

using these molecular targets in quantitative MRD experi-ments.7,8 Moreover, Ig/TCR-based clonality studies mightprovide an instrument to investigate the course of the diseasein children with ALL.14,15 However, clonal evolution of Ig andTCR gene rearrangements due to continuing and secondaryrecombination events throughout the disease course could limitthe effectiveness of V(D)J junctional sequences as targets indetecting MRD, since it may lead to false-negative results.16,17

The high-resolution PCR fragment analysis using GeneScantechnique (PCR-GSA) is an accurate procedure in determiningthe size of PCR products and in discriminating betweenpolyclonal and clonal DNA banding profiles.18–20 In this study,we carried out complete GeneScan and sequence analyses ofevery Ig and TCR gene rearrangement independently found atdiagnosis and relapse in a homogeneous series of 53 childhoodprecursor-B-ALL patients, in order to get detailed information onthe evolution and the stability of potentially useful MRD-PCRtargets.

Materials and methods

Patients and samples

Our study included 53 relapsed patients, presenting withprecursor-B-ALL between 1989 and 2002; this concerned 19females and 34 males with a mean age of 5.82 years (range0.58–14.33 years). Diagnosis of precursor-B-ALL was madeaccording to standard cytomorphology, cytochemistry andimmunophenotypic criteria.21,22 The selection of patients wasbased on the availability of cryopreserved bone marrow (BM)samples at initial diagnosis and at first relapse. Two patients,suffering from BM relapse more than once, were studied at twoand three subsequent leukaemia relapses, respectively. DNAwas extracted from BM mononuclear cells (MNCs) after Ficoll–Paque centrifugation (density 1.077 g/cm3; Pharmacia, Uppsala,Sweden), using either standard techniques (phenol–chloroformorganic extraction) or a commercial kit (Puregene DNA isolationkit, Gentra Systems, Minneapolis, MN, USA). The integrity ofeach DNA sample was checked amplifying albumin.

Comparative immunophenotyping analysis

Immunophenotyping was performed using a direct immuno-fluorescence technique on MNCs obtained at diagnosis andrelapse from BM. Monoclonal antibodies directly conjugatedwith fluorescein isothiocyanate, phycoerythrin and the phycoer-ythrin–cyanine 5 fluorochrome tandem conjugate were com-bined in each tube as previously reported.21,23 The analyseswere carried out in an Epics XL (Beckman Coulter, Miami, FL,USA) flow cytometer, and leukaemias were classified accordingto previously reported criteria.24 Immunophenotyping ofsamples at diagnosis allowed the identification of one (2%)Received 9 January 2003; accepted 2 April 2003

Correspondence: Dr G Germano, Laboratorio di Emato Oncologia,Dipartimento di Pediatria, Universita’ di Padova, via Giustiniani 3,35128 Padua, Italy; Fax: +39 049 8211 465

Leukemia (2003) 17, 1573–1582& 2003 Nature Publishing Group All rights reserved 0887-6924/03 $25.00

www.nature.com/leu

pre pre-B, 30 (57%) common and 22 (41%) pre-B leukaemias.Comparative immunophenotype analysis between diagnosisand first relapse showed intralineage shifts of antigen expressionin 28% (15/53) of cases, in particular, switches concerned 23%(7/30) of common and 36% (8/22) of pre-B leukaemias.

PCR amplification and GeneScan/heteroduplexclonality analyses

Each diagnostic and relapse BM sample was independentlyscreened for both Ig and TCR gene rearrangements, aspreviously described.9,25 Briefly, 28 primer pairs were used toamplify IGH (five VH–JH and seven DH–JH), IGK (four Vk–kdeand Irss–kde), TCRG (three Vg–Jg1.3/2.3 and three Vg–Jg1.1/2.1)and TCRD (complete Vd1–Jd1, Vd2–Jd1 and Dd2–Jd1, incom-plete Vd2–Dd3 and Dd2–Dd3) recombinations. GeneScananalysis was performed using fluorescent reverse primerslabelled at their 50 end with 6-carboxyfluorescein (FAM). Allreactions were carried out in a final volume of 50 ml with 100 ngof DNA sample, 200 nM of dNTPs, 12 pmol of each primer and2U of Taq polymerase. Primer sequences and PCR conditionswere previously published.7,9 Each PCR experiment included anappropriate positive clonal control, a polyclonal control (MNCDNA, MNC) and a nontemplate negative control.9

Aliquots of 1:25 diluted PCR products (1.5 ml) were mixedwith 13.5 ml formamide, and 1 ml of the internal size standard(Genescan-500 and Genescan-1000, Applied Biosystem, FosterCity, CA, USA) was included for precise determination of thelength of the amplicons. After denaturation for 10min at 941C,the products were separated on POP-4 polymer (AppliedBiosystem, Foster City, CA, USA) and analysed by 310GeneScan 3.1 software (Applied Biosystems Foster City, CA,USA). Scanning results of fluorescence-dye-labelled PCR pro-ducts by GSA were interpreted as follows: one or two strikingpeaks indicated a dominant leukaemic clone, while polyclon-ality was characterized by a number of peaks arranged in anormal distribution.20,26–28 A sensitivity test was also performed,diluting diagnostic or relapse samples in MNC DNA: clonalrearrangements of a characteristic size identified at presentationand relapse were detectable at a dilution of 1� 10�3 in morethan 50% of cases, and at a level of 5� 10�3 in 89% of samples.

Clonality identification was also performed by means oftraditional heteroduplex analysis as described:9,29–31 eachpositive PCR product was denatured at 951C for 5min andrapidly cooled to 41C to induce duplex formation, then loadedon 12% nondenaturing polyacrylamide gel in 1� Tris-borate-EDTA buffer (TBE) at 10mA and 41C, or 12.5% nondenaturingpolyacrylamide minigel (PhastGelt, Pharmacia Biotech, Up-psala, Sweden), visualized by ethidium bromide and silverstaining, respectively.31 Heteroduplex analysis enables discri-mination between PCR products derived from clonal andpolyclonal lymphoid cell populations, based on the presenceof homoduplexes (PCR products with identical junctionalregions) or a smear of heteroduplexes (derived from PCRproducts with heterogeneous junctional regions), respec-tively.13,29–31 In order to compare each diagnostic and relapsepatient sample, PCR products were mixed together (1:1) andsubjected to mixed heteroduplex analysis.13,32

Figure 1 gives some examples of the application of PCR-GSAand heteroduplex analysis in precursor-B-ALL patients.

Sequencing analysis

Each clonal PCR product independently found either atdiagnosis or at relapse was sequenced. Fluorescence sequencing

was performed using the BigDye Terminator Cycle SequencingReady Reaction Kit on ABI Prism 310 (Applied Biosystems,Foster City, CA, USA) according to the manufacturer’s guide-lines. Homoduplex PCR products were directly sequenced,whereas heteroduplex ones were either excised and eluted frompolyacrylamide gel or cloned into pCRt 2.1-TOPO vector(TOPO TA Cloning, Invitrogen, Carlsbad, CA, USA) in order toallow the identification of the different alleles. For each DNAsample, five to 10 randomly chosen clones were sequenced.Sequences of junctional regions (N regions) were determined bycomparison to germline sequences and to known humansequences in databases (www.mcr-cpe.cam.ac.uk/imt-doc andwww.ncbi.nlm.nih.gov/BLAST).

Statistical analysis

Statistical analysis using the w2 test on a 2� 2 table wasperformed to compare the frequencies of particular Ig/TCR generearrangements between different precursor-B-ALL patients’subgroups. SAS package (SAS Institute, Cary, NC, USA) wasused for computation, and P-values less than or equal to 0.05were considered statistically significant.

Results

GeneScan and heteroduplex PCR analyses

The clonality pattern at diagnosis and relapse of 53 precursor-B-ALL patients was independently studied by PCR-GSA, conven-tional PCR heteroduplex and sequence analyses. At diagnosis,all 53 (100%) patients presented at least one Ig/TCR generearrangement, with an average of four clonal PCR targets perpatient. At diagnosis, IGH, IGK, TCRD and TCRG rearrange-ments were found in 44 (83%), 22 (42%), 33 (62%) and 33(62%) precursor-B-ALL patients, respectively. A total number of218 clonal PCR targets were identified and analysed atdiagnosis. This concerned 86 IGH (70 complete VH–JH, 16incomplete DH–JH), 31 IGK (25 Vk–kde, 6 Irss–kde), 49incomplete TCRD and 52 TCRG gene rearrangements (Table 1).

Stability of Ig and TCR gene rearrangements

In all, 23 patients (43%) maintained at relapse all therearrangements identified at diagnosis, while 50 (94%) and 40(75%) cases presented, respectively, at least one and twoconserved PCR clonal targets. Only three patients (6%) had fullydifferent Ig/TCR gene rearrangement patterns at the two phasesof the disease.

Concerning the clonal rearrangements, 154 out of 218 (71%)rearrangements amplified at diagnosis were conserved atsubsequent relapse: in particular, 60 out of 86 IGH (70%), 22out of 31 IGK (71%), 33 out of 49 TCRD (67%) and 39 out of52 TCRG (75%) gene rearrangements were maintained (moredetails are shown in Table 1).

The stability profile of Ig and TCR rearrangements, as a resultof the comparison between diagnosis and relapse of eachpatient, is schematized in Figure 2.

Analysis of Ig gene rearrangements

The stability of IGH gene rearrangements was 70%, with 60 outof 86 clonal PCR targets maintained identical at relapse: in

Clonality profiles at diagnosis and relapseG Germano et al

1574

Leukemia

#3C

D

#3C

R

MIX

#28 D

#28 R

MIX

#61 D

#61 R

MIX

M

VH3JH

#3C D

#3C R

#28 D

#28 R

#61 D

#61 R

320 bp

404 bp

a b

321 335

321 335

354

354

327

333

Figure 1 Comparison between heteroduplex and GeneScan PCR analyses. Polyacrylamide silver-stained Minigels 12.5% (a, c, e) and GeneScan electropherograms from ABI 310 Genetic Analyzer(b, d, f) are represented. The size of each product (in base pairs) is displayed on the x-axis with its relative fluorescence intensity on the y-axis. The first lane in the gels and the red plot in theelectropherograms represent the molecular weight marker and the size standard, respectively. In each case, the same clonality pattern was obtained both by mixed heteroduplex and GeneScananalysis. (a, b) In patient #3C, diagnosis and relapse samples showed two identical PCR amplified rearranged alleles (VH3–JH) also in mixed heteroduplex analysis, as confirmed by GeneScan peaks.Patient #28 presented identical clonal VH3–JH PCR products at diagnosis and relapse, while patient #61 showed differences between the two phases of the disease, as shown both by mixedheteroduplex and GeneScan analysis. (c, d) The analysis of IGK rearrangements showed that diagnosis and relapse of patient #1G (VkIII–kde) had different clonal PCR products, while both in patient#5D and in #7E completely identical VkIII PCR products were found at the two phases of the disease. The formation of additional heteroduplex bands in the mix sample of patient #1G confirmed thediversity of the rearrangements at diagnosis and relapse. (e, f) Patient #91 showed two fully identical Vd2–Dd3 amplified targets at diagnosis and relapse; identical clonal Vd2–Dd3 PCR products werefound in patient #1E, while different clonal ones were identified in patient #1G at diagnosis and relapse, as revealed by the presence of additional heteroduplex bands. M: molecular weight marker; D:diagnosis; R: relapse; MIX: sample obtained by mixing of diagnosis and relapse PCR products.

Clon

alityprofiles

atdiagn

osisan

drelapse

GGerm

anoetal

1575

Leukemia

#1

G D

#1

G R

MIX

#5D

D

#5D

R

MIX

#7E

D

#7E

R

MIX

M

VkIIIkde

404 bp

501 bp

c d

#1G D

#1G R

#5D D

#5D R

#7E D

#7E R

419

424

431

431

422

422

Figure 1 (continued)

Clon

alityprofiles

atdiagn

osisan

drelapse

GGerm

anoetal

1576

Leukemia

#91 D

#91 R

MIX

#1E

D

#1E

R

MIX

#1

G D

#1

G R

MIX

M

Vδ2Dδ3

404 bp

501 bp

e f

#9 1 D

#9 1 R

#1E D

#1E R

#1G D

#1G R

498 506

498 506

499

499

481

477

Figure 1 (continued)

Clon

alityprofiles

atdiagn

osisan

drelapse

GGerm

anoetal

1577

Leukemia

particular, complete IGH clonal targets accounted for 71% ofIGH stability, while incomplete ones for 63% (conservation of50/70 and 10/16 targets, respectively).

With regard to the 44 patients characterized by IGH clonalrearrangements at diagnosis, 34 (83%) out of 41 with VH–JHclonal targets and eight (80%) out of 10 with DH–JH clonal

targets at diagnosis relapsed with at least one stable rearrange-ment. Moreover, 27 patients (66%) conserved all their completeIGH rearrangements, while six cases (60%) maintained all theirincomplete IGH rearrangements (Table 1).

The clonality pattern of 18 relapsed patients showeddifferences (loss, change, gain of rearrangements or a combina-tion of these events) compared to the original IGH clonal profileidentified at diagnosis; 10 cases showed loss of IGH clonaltargets, seven patients gained at relapse new rearrangements notpreviously detected at diagnosis, six patients showed somechanges in rearrangements already present at diagnosis.

A different VH3 allele added to a preserved N–JH4 joiningwas observed in patient #73; other two patients (#7F, #69)presented two VH3 PCR products at diagnosis and relapsed withone different VH3 rearrangement, which conserved the same N–DH–N–JH join, suggesting a typical VH replacement. In otherthree patients (#1A, #25, #61) sequence analysis of IGHrearrangements revealed possibly new secondary rearrange-ments.

IGK gene rearrangements had a stability of 71%, with 22 outof 31 conserved clonal PCR targets at relapse (18/25, 72% Vk–kde and 4/6, 67% Irss–kde).

Vk–kde gene rearrangements were present in 19/53 (36%)patients, and at least one of these targets was preserved in 89%of the cases while 14 (74%) patients presented all diagnosticVk–kde rearrangements at relapse. Irss gene rearrangement,characterizing 11% of the patients at diagnosis, was preserved atrelapse in 67% of cases (Table 1).

In 11 patients, we found differences (loss, change, gain ofrearrangements or a combination of these events) between theclonality pattern at diagnosis and relapse. Loss of IGK generearrangements regarded six patients, whereas six patientsacquired new clonal PCR targets at relapse. Modifications wereidentified in only one patient (#1G): he presented VkI, VkIII andIrss clonal PCR targets at diagnosis. At relapse he lost VkI–kdegene rearrangement and presented new VkIII and Irss both withdifferent junctional regions.

Analysis of TCR gene rearrangements

The stability of TCRD rearrangements was 67%, with 33 out of49 diagnostic clonal targets found identical at relapse of thedisease. TCRD clonal targets at diagnosis were represented onlyby incomplete rearrangements (Vd2–Dd3, Dd2–Dd3); at leastone stable clonal target was found at relapse in 76% of thepatients presenting TCRD rearrangements at diagnosis, while 19(56%) patients maintained all their TCRD rearrangements(Table 1).

In all, 18 patients presented differences between diagnosticand relapse TCRD gene rearrangements; nine cases showed lossof TCRD clonal targets, eight patients gained at relapse newrearrangements not previously detected at diagnosis. In threecases (#3D, #1G, #55) sequence comparison revealed differ-ences in the junctional regions of Vd2–Dd3 target, indicating

Pa

tie

nt

VkI

VkII

VkIII

VkIV

Irss

VH

1V

H2

VH

3V

H4

VH

6D

H1

DH

2D

H3

DH

4D

H6

DH

7

Vδ

δ 2

Dδ

δ 3

Dδ

δ 2

Dδ

δ 3

Vγ γ I

Jγ1

.3/2

.3V

γ γ IIJ

γ γ 1.3

/2.3

Vγ γ IV

J γ γ1.3

/2.3

Vγ γ IJ

γ γ 1.1

/2.1

Vγ γ I

IJ γ γ

1.1

/2.1

Vγ γ I

VJ

γ γ 1.1

/2.1

Re

lap

se

tim

ing

#3B 6#25 6#59 9#5E 11#57 11#83 11#67 13#89

§ 15#5C 16#23 17#7D 18#7F 20#49 21#63 21#91 21#3C 24#1G 24#3G 26#1F 26#55 26#65 28#71 29#69 30#3E 30#35 31#1A 32#51 33#7A 34#1B 34#3D 36#9D

§ 36#7E 36#5G 36#53 37#5A 39#1E 39#77 39#7B 40#9C 40#15G 40#93 40#5D 41#81 41#75 44#1C 46#73 47#28 48#9A 49#9E 54#13G 55#61 61#7G 62#7C 67

Ig TCR

§In patient #89 and #9D diagnosis and first relapse are compared. Figure 2 Schematic view of the Ig/TCR clonal stability in 53

precursor-B-ALL at diagnosis and at relapse. Green squares representstable clonal PCR targets; red squares, fully different Ig/TCR generearrangements; yellow squares, partially preserved gene rearrange-ments; black squares, absence of rearrangement. VH5 and completeTCRD (Vd1–Jd1, Vd2–Jd1, Dd2–Jd1) rearrangements were not found inany patient. Patients are listed on the first vertical column and codedby #, while rearrangements are reported on the horizontal upper line.The last column indicates the time of relapse expressed in months afterdiagnosis.

Clonality profiles at diagnosis and relapseG Germano et al

1578

Leukemia

possible secondary rearrangements. Other two patients (#3C,#23) relapsed with only one out of the two original Vd2–Dd3rearrangements found at diagnosis. In case #81, relapse wascharacterized by two Vd2–Dd3 targets, one of which wasidentical to the clonal Vd2–Dd3 target detected at diagnosis.GeneScan analysis allowed also the interpretation of TCRD generearrangement pattern in patient #51, which was characterizedby a series of dominant peaks for Vd2–Dd3 gene rearrangementat diagnosis and only two dominant peaks at relapse: mostlikely, the two peaks at relapse were already present within thepeak profile of diagnostic sample.The stability of TCRG rearrangements was 75% (39/52 stable

rearrangements). In all, 27 patients (82%) showing TCRG generearrangements at diagnosis had at least one stable target atrelapse, while 17 (52%) maintained all diagnostic TCRG targets(Table 1).On the other hand, some diagnostic TCRG trait fragments

differed at relapse in 21 patients. In 11 cases TCRG clonaltargets were lost, and in 16 patients new rearrangementsundetected at diagnosis were gained at relapse. None of thepatients showed modifications in the sequence of junctionalregions of rearrangements found both at diagnosis and atrelapse.

New rearrangements at relapse

The independent analysis of diagnostic and relapse samplesallowed the identification of new clonal PCR targets in 24patients at relapse (45%): this concerned 13 IGH, 8 IGK, 14TCRD and 23 TCRG gene rearrangements. To verify whetherthese clonal targets were already present at diagnosis, patient-specific forward primers were designed complementary to thesequences of relapse junctional regions and tested in diagnosticsamples by RQ-PCR.7,8 All IGH and IGK gene rearrangementswere tested for presence at diagnosis. Only two IGK gene

rearrangements (both VkII) were detected in the diagnosticsamples of two patients (cases #56 and #76), with a positivityless than 10�4 and 10�5, respectively, indicating that they werecharacteristic of subclones already present at diagnosis. Inter-estingly, none of the tested new IGH rearrangements was foundin diagnostic samples, with a sensitivity of 1� 10�5.

Discussion

At present, relapse remains the most important obstacle toovercome in childhood leukaemia.1,2,33 The investigation ofclonality profiles by means of Ig/TCR gene rearrangements is avery powerful tool for understanding whether relapse representsthe reoccurrence of the original leukaemic clone/s or asecondary malignancy.17,34–42 The existence of at least oneconserved rearranged clonal target confirms that relapse isprobably related to the diagnostic clone.

Although Southern blot analysis is considered to be the goldstandard for detection of clonal Ig and TCR gene rearrange-ments, previous studies12,13,25,34 demonstrated a high concor-dance (490%) between Southern blot and heteroduplex PCRanalysis. In this Ig/TCR clonality study, we used an exclusivelyPCR-based strategy, which is the method routinely employed forMRD target detection within International BFM Study Group.3,9

A molecular relation between diagnosis and relapse wasestablished in 50 of the 53 childhood precursor-B-ALL patients.In 43% of patients all clonal PCR targets identified at diagnosiswere detected also at relapse. In only three patients (cases #1A,#1C and #61, relapsed after 32, 46 and 61 months afterdiagnosis) the Ig/TCR gene rearrangements profile at diagnosisand relapse were completely different (Figure 2), with unrelatedjunctional region sequences. Eventually, these three cases ofleukaemia relapse may turn out to represent new malignancies.

In our extensive molecular study, we did not find a correlationbetween remission duration and clonality profile of the patients,

Table 1 Stability of Ig/TCR gene rearrangements in relapsed precursor-B-ALL patients

MRD-PCR targets Precursor-B-ALL patients

Stable atrelapse

Incidence of MRD-PCRtargets at diagnosis

At least one MRD-PCRtarget stable at relapse

All MRD-PCR targetsstable at relapse

VH–JH 71% 77% 83% 66%(50/70) (41/53) (34/41) (27/41)

DH–JH 63% 19% 80% 60%(10/16) (10/53) (8/10) (6/10)

IGH 70% 83% 86% 68%(60/86) (44/53) (38/44) (30/44)

Vk–kde 72% 36% 89% 74%(18/25) (19/53) (17/19) (14/19)

Irss–kde 67% 11% 67% 67%(4/6) (6/53) (4/6) (4/6)

IGK 71% 42% 86% 68%(22/31) (22/53) (19/22) (15/22)

TCRDa 67% 62% 76% 56%(33/49) (33/53) (25/33) (19/33)

TCRG 75% 62% 82% 52%(39/52) (33/53) (27/33) (17/33)

aOnly incomplete Vd2–Dd3, Dd2–Dd3.

Clonality profiles at diagnosis and relapseG Germano et al

1579

Leukemia

and this is in line with previous similar reports.17 Case #9Dpresented three consecutive relapses and the Ig/TCR clonalityprofile was determined for each relapse. Interestingly, atdiagnosis he showed three clonal PCR targets (VH3–JH5, Vg3–Jg1.3/2.3, Vg8–Jg1.1/2.1), all preserved both at first and secondrelapse; a new Vd2–Dd3 gene rearrangement appeared at firstrelapse, being conserved at second one. At six months afterallogeneic BM transplantation he relapsed for the third time andall clonal PCR targets previously identified were lost. Theanalysis of Ig/TCR gene rearrangements at third relapse showedonly a new Vg9–Jg1.3/2.3 clonal PCR target. Consequently, itseems that the third relapse, being completely different from theprevious phases of the disease, might represent a newleukaemia. Further studies such as comparative Southern blotanalysis and/or identification of other gene rearrangements(TCRB, TCRA, IGL, fusion genes) should be performed in orderto demonstrate real secondary malignancies.

GeneScan analysis employed in this study for clonalityassessment in ALL demonstrated to be effective: the sensitivityof 5� 10�3 is highly satisfactory at diagnosis or at relapse; inaddition, the automatic size determination was shown to be avaluable tool for interpreting clonal profiles, since resolution ofPCR products within a narrow size range is inefficient byconventional polyacrylamide gel electrophoresis; in fact, exactassessment with discrimination of one base pair is only possibleusing GeneScan analysis.28,43 We achieved a very goodagreement in the results of the two techniques (Figure 1), andGeneScan analysis has the advantage that allows a directcomparison of Ig/TCR gene rearrangements, skipping the mixingof PCR products followed by heteroduplex analysis.32 Therefore,GeneScan analysis could represent a valid alternative for rapiddiscrimination between identical or different clonal PCRproducts at diagnosis and at relapse. Recent studies3,4 demon-strated that early response to treatment, measured by MRDdetection exploiting Ig/TCR gene rearrangements, is one of themost relevant and reliable prognostic factors in ALL; therefore,the choice of suitable and stable clonal PCR targets is a crucialissue in MRD monitoring.

Recently, Szczepanski et al17 reported a particularly highstability for IGK rearrangements, suggesting the preferentialchoice of these clonal PCR targets for MRD strategy.44

The results of our detailed PCR-GSA and sequencing analysisof Ig/TCR gene rearrangement profile add a contribution forselecting PCR targets for MRD detection in childhood precursor-B-ALL. Ig/TCR clonal targets for MRD detection were identifiedin all patients (100%), with 71% of these targets being preservedat relapse. The stability of Ig/TCR gene rearrangements wasapproximately 70% (IGH 70%, IGK 71%, TCRD 67%, TCRG75%) with no significant difference among the different clonalPCR targets (Table 1). Consequently, this suggests that there is no‘preferential’ MRD-PCR target and that all targets may beequally useful for MRD monitoring.

We therefore propose that the selection of MRD-PCR targetsshould be based mainly on the sensitivity tests in RQ-PCRexperiments: each target with a minimum sensitivity of 10�43

should represent the first choice for evaluation of MRD inprecursor-B-ALL patients. Although the use of all available andsensitive PCR clonal targets would be the best strategy for MRDdetection in ALL patients, this might be unrealistic especially forlarge routine diagnostic laboratories; therefore, the widely usedcombination of two MRD-PCR targets could be acceptable toreduce false-negative results.3,4,11,17 It has been reported thatMRD detection using TCRG gene rearrangements in precursor-B-ALL patients often results in low sensitivities due tononspecific amplification of rearrangements in normal T-

lymphocytes,45 and this feature generally limits the applicabilityof this target for MRD detection. Consequently, TCRG targetsshould be used when samples do not present any nonspecificamplification in reproducible RQ-PCR experiments. Further-more, in approximately 10% of our precursor-B-ALL patients weobserved the occurrence of ongoing and secondary generearrangements involving Ig/TCR loci. As regards IGH generearrangements, patient-specific primers should be designedcomplementary to DH–N–JH junctional regions in order tominimize false-negative results due to clonal evolution: in twoof our cases (#7F, #69) primers designed on DH–JH stems mighthave allowed the detection of relapse.

Detailed sequencing analysis of Ig/TCR gene rearrangementsat diagnosis and relapse is the most reliable way to understandhow rearrangements persist or change/evolve during the courseof the disease. In order to assess if new rearrangements observedat relapse represented minor clones already present albeit atvery low levels at diagnosis, we performed a very sensitivequantitative PCR approach, but the overall stability of IGH andIGK gene rearrangements was not affected by the results. Thenew rearrangements found at relapse had probably developedduring the course of the disease. Our findings suggest that cellsfrom the original clone/s can remain dormant for long periods oftime. The mechanisms responsible for this dormancy are notknown, but probably they involve either an ability of the clonalcells to remain out of cell cycle, or the inability of host immunesurveillance to eradicate minimal residual leukaemia, or acombination of the two.40 Another hypothesis is that some cellsescape the activity of chemotherapy, remaining protected inparticular sites and resulting therefore undetectable in bonemarrow.46 It has also been reported that chemotherapy couldeliminate an overt leukaemic subclone that dominates atdiagnosis, but may fail to eradicate a ‘preleukaemic’ clone thatpersist after cessation of therapy: the proliferative stress duringthe regenerative rebound may convert a ‘preleukaemic’ cell intoa new leukaemic clone.47

In conclusion, our results show that none of the clonal Ig andTCR gene rearrangements we tested stands out for a strikingstability. Clonal evolution phenomena and ongoing secondaryrearrangement processes may determine changes in clonalitypatterns. The use of two sensitive Ig/TCR targets in clinicalsetting is a good compromise between the applicability of themethod in large routine analyses and the stability of the clonaltargets; in particular, when the informative time points used forclinical choices are within the first 3 months of therapy. Itremains to be demonstrated whether the kinetic of clearance ofthe clones at early time points is different for targets which laterwill demonstrate their instability.

Acknowledgements

GG and LdG were equally responsible for designing the study andwriting the paper; SP, EG, GC contributed to analysis and torevision of the results; GB and AB critically revised the manuscriptand gave the final approval for its submission. GG and LdGgratefully acknowledge Dr Lucia De Zen, Dr Geertruy Te Kronniefor helpful comments and suggestions and Dr Sandra Trevisan forstatistical analysis. This study was supported by CNR, Murst ex60%, Fondazione Citta’ della Speranza, Fondazione Tettamantiand Associazione Italiana Ricerca sul Cancro (AIRC).

References

1 Pui CH, Relling MV, Campana D, Evans WE. Childhood acutelymphoblastic leukemia. Rev Clin Exp Hematol 2002; 6: 161–180.

Clonality profiles at diagnosis and relapseG Germano et al

1580

Leukemia

2 Conter V, Arico M, Valsecchi MG, Basso G, Biondi A, Madon Eet al. Long-term results of the Italian Association of PediatricHematology and Oncology (AIEOP) acute lymphoblastic leukemiastudies, 1982–1995. Leukemia 2000; 14: 2196–2204.

3 van Dongen JJ, Seriu T, Panzer-Grumayer ER, Biondi A, Pongers-Willemse MJ, Corral L et al. Prognostic value of minimal residualdisease in acute lymphoblastic leukaemia in childhood. Lancet1998; 352: 1731–1738.

4 Cave H, van der Werff ten Bosch J, Suciu S, Guidal C, WaterkeynC, Otten J et al. Clinical significance of minimal residual disease inchildhood acute lymphoblastic leukemia. European Organizationfor Research and Treatment of Cancer, Childhood LeukemiaCooperative Group. N Engl J Med 1998; 339: 591–598.

5 Pui CH, Campana D. New definition of remission in childhoodacute lymphoblastic leukemia. Leukemia 2000; 14: 783–785.

6 Arico’ M, Germano G, del Giudice L, Ziino O, Locatelli F, BassoG. Late relapse of childhood acute lymphoblastic leukemia andPCR-monitoring of minimal residual disease: how much time canelapse between ‘molecular’ and clinical relapse? Haematologica2002; 87: ELT19. Available from URL: http://www.haematologi-ca.it/e-letters/2002_04/19.htm

7 Pongers-Willemse MJ, Verhagen OJ, Tibbe GJ, Wijkhuijs AJ, deHaas V, Roovers E et al. Real-time quantitative PCR for thedetection of minimal residual disease in acute lymphoblasticleukemia using junctional region specific TaqMan probes.Leukemia 1998; 12: 2006–2014.

8 Verhagen OJ, Willemse MJ, Breunis WB, Wijkhuijs AJ, Jacobs DC,Joosten SA et al. Application of germline IGH probes in real-timequantitative PCR for the detection of minimal residual disease inacute lymphoblastic leukemia. Leukemia 2000; 14: 1426–1435.

9 Pongers-Willemse MJ, Seriu T, Stolz F, d’Aniello E, Gameiro P, PisaP et al. Primers and protocols for standardized detection ofminimal residual disease in acute lymphoblastic leukemia usingimmunoglobulin and T cell receptor gene rearrangements andTAL1 deletions as PCR targets: report of the BIOMED-1 CON-CERTED ACTION: investigation of minimal residual disease inacute leukemia. Leukemia 1999; 13: 110–118.

10 Davis MM, Bjorkman PJ. T-cell antigen receptor genes and T-cellrecognition. Nature 1988; 334: 395–402.

11 van Dongen JJ, Szczepanski T, de Bruijn MA, van den Beemd MW,de Bruin-Versteeg S, Wijkhuijs JM et al. Detection of minimalresidual disease in acute leukemia patients. Cytokines Mol Ther1996; 2: 121–133.

12 Szczepanski T, Beishuizen A, Pongers-Willemse MJ, Hahlen K,Van Wering ER, Wijkhuijs AJ et al. Cross-lineage T cell receptorgene rearrangements occur in more than ninety percent ofchildhood precursor-B acute lymphoblastic leukemias: alternativePCR targets for detection of minimal residual disease. Leukemia1999; 13: 196–205.

13 Langerak AW, Szczepanski T, van der Burg M, Wolvers-Tettero IL,van Dongen JJ. Heteroduplex PCR analysis of rearranged T cellreceptor genes for clonality assessment in suspect T cell prolifera-tions. Leukemia 1997; 11: 2192–2199.

14 Beishuizen A, Verhoeven MA, Hahlen K, van Wering ER, vanDongen JJ. Differences in immunoglobulin heavy chain generearrangement patterns between bone marrow and blood samplesin childhood precursor B-acute lymphoblastic leukemia atdiagnosis. Leukemia 1993; 7: 60–63.

15 Breit TM, Wolvers-Tettero IL, Beishuizen A, Verhoeven MA, vanWering ER, van Dongen JJ. Southern blot patterns, frequencies, andjunctional diversity of T-cell receptor-delta gene rearrangements inacute lymphoblastic leukemia. Blood 1993; 82: 3063–3074.

16 Langlands K, Craig JI, Anthony RS, Parker AC. Clonal selection inacute lymphoblastic leukaemia demonstrated by polymerase chainreaction analysis of immunoglobulin heavy chain and T-cellreceptor delta chain rearrangements. Leukemia 1993; 7: 1066–1070.

17 Szczepanski T, Willemse MJ, Brinkhof B, van Wering ER, van derBurg M, van Dongen JJ. Comparative analysis of Ig and TCR generearrangements at diagnosis and at relapse of childhood precursor-B-ALL provides improved strategies for selection of stable PCRtargets for monitoring of minimal residual disease. Blood 2002; 99:2315–2323.

18 Simon M, Kind P, Kaudewitz P, Krokowski M, Graf A, Prinz J et al.Automated high-resolution polymerase chain reaction fragment

analysis: a method for detecting T-cell receptor gamma-chain generearrangements in lymphoproliferative diseases. Am J Pathol 1998;152: 29–33.

19 Dippel E, Assaf C, Hummel M, Schrag HJ, Stein H, Goerdt S et al.Clonal T-cell receptor gamma-chain gene rearrangement by PCR-based GeneScan analysis in advanced cutaneous T-cell lympho-ma: a critical evaluation. J Pathol 1999; 188: 146–154.

20 Guidal C, Vilmer E, Grandchamp B, Cave H. A competitive PCR-based method using TCRD, TCRG and IGH rearrangements forrapid detection of patients with high levels of minimal residualdisease in acute lymphoblastic leukemia. Leukemia 2002; 16:762–764.

21 Basso G, Buldini B, De Zen L, Orfao A. New methodologicapproaches for immunophenotyping acute leukemias. Haemato-logica 2001; 86: 675–692.

22 van der Does-van den Berg A, Bartram CR, Basso G, Benoit YC,Biondi A, Debatin KM et al. Minimal requirements for thediagnosis, classification, and evaluation of the treatment ofchildhood acute lymphoblastic leukemia (ALL) in the ‘BFM Family’Cooperative Group. Med Pediatr Oncol 1992; 20: 497–505.

23 De Zen L, Orfao A, Cazzaniga G, Masiero L, Cocito MG,Spinelli M et al. Quantitative multiparametric immunophenotyp-ing in acute lymphoblastic leukemia: correlation with specificgenotype. I. ETV6/AML1 ALLs identification. Leukemia 2000; 14:1225–1231.

24 Putti MC, Rondelli R, Cocito MG, Arico M, Sainati L, Conter V et al.Expression of myeloid markers lacks prognostic impact in childrentreated for acute lymphoblastic leukemia: Italian experience inAIEOP-ALL 88-91 studies. Blood 1998; 92: 795–801.

25 Szczepanski T, Langerak AW, Wolvers-Tettero IL, OssenkoppeleGJ, Verhoef G, Stul M et al. Immunoglobulin and T cell receptorgene rearrangement patterns in acute lymphoblastic leukemia areless mature in adults than in children: implications for selection ofPCR targets for detection of minimal residual disease. Leukemia1998; 12: 1081–1088.

26 Klemke CD, Dippel E, Dembinski A, Ponitz N, Assaf C, HummelM et al. Clonal T cell receptor gamma-chain gene rearrangementby PCR-based GeneScan analysis in the skin and blood of patientswith parapsoriasis and early-stage mycosis fungoides. J Pathol2002; 197: 348–354.

27 Evans PA, Short MA, Owen RG, Jack AS, Forsyth PD, Shiach CRet al. Residual disease detection using fluorescent polymerasechain reaction at 20 weeks of therapy predicts clinical outcome inchildhood acute lymphoblastic leukemia. J Clin Oncol 1998; 16:3616–3627.

28 Delabesse E, Burtin ML, Millien C, Madonik A, Arnulf B, BeldjordK et al. Rapid, multifluorescent TCRG Vgamma and Jgammatyping: application to T cell acute lymphoblastic leukemia and tothe detection of minor clonal populations. Leukemia 2000; 14:1143–1152.

29 Martinelli G, Farabegoli P, Testoni N, Terragna C, Vittone A,Raspadori D et al. Detection of clonality by heteroduplex analysisof amplified junctional region of T-cell receptor gamma in patientswith T-cell acute lymphoblastic leukemias. Haematologica 1997;82: 161–165.

30 Bottaro M, Berti E, Biondi A, Migone N, Crosti L. Heteroduplexanalysis of T-cell receptor gamma gene rearrangements fordiagnosis and monitoring of cutaneous T-cell lymphomas. Blood1994; 83: 3271–3278.

31 Germano G, Songia S, Biondi A, Basso G. Rapid detection ofclonality in patients with acute lymphoblastic leukemia. Haema-tologica 2001; 86: 382–385.

32 Szczepanski T, Willemse MJ, Kamps WA, van Wering ER,Langerak AW, van Dongen JJ. Molecular discriminationbetween relapsed and secondary acute lymphoblastic leukemia:proposal for an easy strategy. Med Pediatr Oncol 2001; 36:352–358.

33 Conter V, Arico M, Valsecchi MG, Rizzari C, Testi A, Miniero Ret al. Intensive BFM chemotherapy for childhood ALL: interimanalysis of the AIEOP-ALL 91 study. Associazione ItalianaEmatologia Oncologia Pediatrica. Haematologica 1998; 83:791–799.

34 Steward CG, Goulden NJ, Katz F, Baines D, Martin PG, LanglandsK et al. A polymerase chain reaction study of the stability of Igheavy-chain and T-cell receptor delta gene rearrangements

Clonality profiles at diagnosis and relapseG Germano et al

1581

Leukemia

between presentation and relapse of childhood B-lineage acutelymphoblastic leukemia. Blood 1994; 83: 1355–1362.

35 Beishuizen A, Verhoeven MA, van Wering ER, Hahlen K,Hooijkaas H, van Dongen JJ. Analysis of Ig and T-cell receptorgenes in 40 childhood acute lymphoblastic leukemias at diagnosisand subsequent relapse: implications for the detection of minimalresidual disease by polymerase chain reaction analysis. Blood1994; 83: 2238–2247.

36 Taylor JJ, Rowe D, Kylefjord H, Chessells J, Katz F, Proctor SJ et al.Characterisation of non-concordance in the T-cell receptor gammachain genes at presentation and clinical relapse in acutelymphoblastic leukemia. Leukemia 1994; 8: 60–66.

37 Baruchel A, Cayuela JM, MacIntyre E, Berger R, Sigaux F.Assessment of clonal evolution at Ig/TCR loci in acute lympho-blastic leukaemia by single-strand conformation polymorphismstudies and highly resolutive PCR derived methods: implication fora general strategy of minimal residual disease detection. Br JHaematol 1995; 90: 85–93.

38 Rosenquist R, Thunberg U, Li AH, Forestier E, Lonnerholm G,Lindh J et al. Clonal evolution as judged by immunoglobulin heavychain gene rearrangements in relapsing precursor-B acute lym-phoblastic leukemia. Eur J Haematol 1999; 63: 171–179.

39 Li AH, Rosenquist R, Forestier E, Lindh J, Roos G. Detailedclonality analysis of relapsing precursor B acute lymphoblasticleukemia: implications for minimal residual disease detection.Leukemia Res 2001; 25: 1033–1045.

40 Vora A, Frost L, Goodeve A, Wilson G, Ireland RM, Lilleyman Jet al. Late relapsing childhood lymphoblastic leukemia. Blood1998; 92: 2334–2337.

41 Lo Nigro L, Cazzaniga G, Di Cataldo A, Pannunzio A, D’Aniello E,Masera G et al. Clonal stability in children with acute lympho-blastic leukemia (ALL) who relapsed five or more years afterdiagnosis. Leukemia 1999; 13: 190–195.

42 Balduzzi A, Corral L, Gaipa G, Germano G, Giudici G, Cantu’-Rajnoldi A et al. Molecular assessment of clonality in a pre-B ALLunusually relapsing as a mature-B ALL. Leukemia 1999; 13: 2114–2116.

43 Wickham CL, Lynas C, Ellard S. Detection of clonal T cellpopulations by high resolution PCR using fluorescently labellednucleotides; evaluation using conventional LIS-SSCP. Mol Pathol2000; 53: 150–154.

44 van der Velden VH, Willemse MJ, van der Schoot CE, Hahlen K,van Wering ER, van Dongen JJ. Immunoglobulin kappa deleting

element rearrangements in precursor-B acute lymphoblasticleukemia are stable targets for detection of minimal residualdisease by real-time quantitative PCR. Leukemia 2002; 16: 928–936.

45 van der Velden VH, Wijkhuijs JM, Jacobs DC, van Wering ER, vanDongen JJ. T cell receptor gamma gene rearrangements as targetsfor detection of minimal residual disease in acute lymphoblasticleukemia by real-time quantitative PCR analysis. Leukemia 2002;16: 1372–1380.

46 Valetto A, Anselmi G, Scuderi F, Lanciotti M, Chiesa V, Dini G.Use of molecular techniques to confirm true re-emergence of theoriginal clone and to track minimal residual disease in a case oflate extramedullary relapse of childhood acute lymphoblasticleukemia. Haematologica 2000; 85: 1102–1103.

47 Ford AM, Fasching K, Panzer-Grumayer ER, Koenig M, Haas OA,Greaves MF. Origins of ‘late’ relapse in childhood acutelymphoblastic leukemia with TEL-AML1 fusion genes. Blood2001; 98: 558–564.

Appendix

The following institutions enrolled patients in the study: Bari,Clinica Pediatrica I (Dr N Santoro); Bari, Clinica Pediatrica II(Professor N Rigillo); Bologna, Clinica Pediatrica (Professor GPaolucci, Dr A Pession); Brescia, Clinica Pediatrica (Professor LNotarangelo, Dr F Porto); Catanzaro, Div. di Ematologia(Professor S Magro); Firenze, Ospedale Meyer, Dipartimentodi Pediatria, U. O. Oncoematologia Pediatrica (Professor GBernini); Genova, Ist ‘G. Gaslini’ (Dr C Micalizzi); Modena,Clinica Pediatrica (Dr M Cellini); Monza, Clinica Pediatrica(Professor G Masera); Napoli, Ospedale Pausilipon (Professor VPoggi), Napoli, II Universita’, Dipartimento di Pediatria, ServizioAutonomo di Oncologia Pediatrica (Professor MT Di Tullio);Padova, Clinica Pediatrica II (Professor L Zanesco); Parma,Clinica Pediatrica (Dr G Izzi); Pescara, Divisione di Ematologia(Dr A Di Marzio); Pisa, Clinica Pediatrica III (Dr C Favre);Torino, Clinica Pediatrica (Professor E Madon, Dr E Barison).

Clonality profiles at diagnosis and relapseG Germano et al

1582

Leukemia