| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Imaging, Diagnosis, Prognosis |
Authors' Affiliations: 1 Children's Cancer Research Institute; 2 St. Anna Kinderspital, Vienna, Austria; 3 M. Tettamanti Research Center, Pediatric Clinic, San Gerardo Hospital, University of Milan Bicocca, Monza, Italy; 4 Department of Immunology, Erasmus MC Rotterdam, Rotterdam, the Netherlands; 5 Department of Pediatrics, University of Padova, Padua, Italy; 6 Department of Pediatrics, Charite Medical Center, Berlin, Germany; 7 Department of Pediatrics, Hannover Medical School, Hannover, Germany; and 8 Department of Pediatrics, Justus Liebig University, Giessen, Germany
Requests for reprints: E. Renate Panzer-Grümayer, Children's Cancer Research Institute, Kinderspitalgasse 6, 1090 Vienna, Austria. Phone: 43-1-40170431; Fax: 43-1-4087230; E-mail: renate.panzer{at}ccri.at.
 |
Abstract |
|---|
 Top Abstract Patients, Materials, and MethodsResultsDiscussionReferences |
|---|
Patients and Methods: Immunoglobulin and T-cell receptor gene rearrangements were analyzed in 41 children with a TEL-AML1+ acute lymphoblastic leukemia and an early (up to 30 months after diagnosis; n = 12) or late (at 30 months or later; n = 29) disease recurrence by a standardized PCR approach.
Results: In 68% of the patients (group I), we identified differences in the immunogenotype patterns, whereas no changes were observed in the remaining 32% (group II). The divergence resulted more often from clonal selection than clonal evolution and consisted predominantly of losses (0-6, median 5) and/or gains (0-4, median 1) of rearrangements. The frequency and number of clonal immunoglobulin/T-cell receptor rearrangements in group I was higher at diagnosis (2-13, median 5) than at relapse (2-7, median 4), whereas it was the lowest in group II (1-5, median 3). Although group I children were younger at diagnosis, there was no correlation between particular immunogenotype patterns and remission duration.
Conclusion: These findings imply that the clonal heterogeneity in younger children most likely reflects an ongoing high recombinatorial activity in the preleukemic/leukemic cells, whereas the more uniform repertoire observed in older children mirrors end-stage rearrangement patterns of selected cell clones that evolved during the prolonged latency period.
25%, the translocation t(12;21)(p13;q22) and its molecular genetic counterpart, the TEL-AML1 gene fusion, is the most common specific genetic rearrangement in childhood B-cell precursor acute lymphoblastic leukemia (BCP-ALL). The initial view that this rearrangement concurs with an extraordinarily favorable prognosis has recently become a matter of discussion because in several therapy studies the incidence of TEL-AML1+ cases at diagnosis and relapse was found to be similar (reviewed in ref. 1).
The TEL-AML1 gene fusion is an early, or perhaps even the first, event in leukemia development. It commonly occurs already during fetal development but it is considered insufficient to cause clinically overt leukemia by itself (2). Further events are required but seem to be rare because the incidence of healthy newborns in whom TEL-AML1+ cells can be detected is 100 times that of the respective leukemias (3). The latency period is variable and has been reported to range up to10 years (4). Relapses occur predominantly late and are again responsive to chemotherapy (1). These data together with the observation that deletions of the nonrearranged TEL allele, one of the proposed essential secondary events, may differ between diagnosis and relapse, lead to the hypothesis that at least part of the relapses do not derive from the dominant leukemic clone at diagnosis but rather from a potentially therapy-resistant dormant "preleukemic" clone that is still characterized by a TEL-AML1 fusion gene (5). This notion is supported by several recent studies that focus on the analysis of TEL deletions in such cases (6–8).
Immunoglobulin (IG) and T-cell receptor (TCR) gene rearrangements occur during normal lymphoid development and are widely used as clone-specific markers for clonally expanded lymphoid cells (9). Although these rearrangements are not involved in the leukemogenic process, they nevertheless may provide essential information about the affected cell type and the time of their manifestation during the fetal or postnatal life (10, 11). Specific patterns and types of rearrangements have been attributed to particular, genetically defined leukemia subtypes (12). In line with this observation, we also reported distinct age-related changes of the IG/TCR rearrangements in TEL-AML1+ cases (13). Furthermore, clonotypic IG/TCR rearrangements are also used as specific markers for the surveillance of minimal residual disease. This method has recently been implemented in clinical trials for treatment stratification of childhood leukemia (14–17).
Earlier studies of BCP-ALL (18, 19) as well as an initial study of three relapsed TEL-AML1+ cases suggested that the remission duration could be one of the crucial factors that influence the likelihood that clonal patterns change between diagnosis and relapse (5). However, this notion could not be substantiated in a follow-up study with a larger number of TEL-AML1+ cases (6). Moreover, this latter study also revealed that the differences of the immunogenotype patterns at diagnosis and relapse did not concord with the respective TEL deletion patterns in 6 of 12 cases. To explore these issues further, we determined the frequency of IG/TCR gene rearrangements in a large number of TEL-AML1+ cases at diagnosis and relapse and correlated the respective patterns with clinical variables to gain some insight into their potential clinical or biological relevance.
|
Patients, Materials, and Methods |
|---|
|
TopAbstractPatients, Materials, and MethodsResultsDiscussionReferences |
|---|
Identification of PCR targets. DNA extraction, quality control of DNA, and screening PCRs for the subsequent analyses were conducted according to standardized protocols at diagnosis and relapse (21). Screening PCRs were done in all samples and included those for the incomplete and complete IG heavy chain (IGH) genes, the IG light chain 
-deleting element (IGK-Kde), as well as the incomplete TCR 
(TCRD) and the TCR 
(TCRG) rearrangements (9). In some of the diagnostic samples, additional screening PCRs for the most common V2-J
(22) and TCR ß (TCRB) rearrangements (9) were done. The clonal nature of the amplified products was assessed by heteroduplex analysis. In case multiple bands were present, sequences were obtained from cloned PCR products. The involved segments of IGH, IGK, TCRD, V2-J, TCRG, and TCRB gene rearrangements zzzwere identified through comparison with published germ line sequences using the sequence alignment software (V-QUEST) from IMGT (http://imgt.cines.fr:8104/). VDJH gene segments were additionally analyzed in the VBASE directory of human IG genes (http://www.vbasemrc-cpe.cam.ac.uk/).
Definition of monoclonal and oligoclonal B-cell precursor acute lymphoblastic leukemia. A sample was assumed to be monoclonal if no more than one IG/TCR rearrangement per allele was present (i.e., the number of rearrangements detected by PCR had to be equal to or less than two per IG/TCR locus). Oligoclonality, on the other hand, was defined by the occurrence of more than two rearrangements per gene or, in the case of IGH rearrangements, also by the detection of related sequences.
Statistical analysis. The different variables, such as the number of IG/TCR gene rearrangements, the age at initial diagnosis, the age at relapse, the remission duration, and its correlation with the patterns of IG/TCR rearrangements, were analyzed with the Wilcoxon two-sample test and, when children were grouped according to age, with the Fisher's exact test. A difference in the frequencies of IG/TCR rearrangements between different occasions was analyzed by McNemar's test. A correlation between age and number of rearrangements at diagnosis was assessed by Kendall's rank correlation.
|
Results |
|---|
|
TopAbstractPatients, Materials, and MethodsResultsDiscussionReferences |
|---|
Comparison of immunoglobulin/T-cell receptor rearrangement patterns between diagnosis and relapse. Depending on the altered or conserved relapse rearrangement pattern, we defined two groups: group I is composed of 28 cases with a complete or partial change and group II is composed of 13 cases with identical rearrangements (Fig. 1). Thus, the immunogenotype had changed in approximately two thirds of the patients (68%).
|
|
Eight leukemias had oligoclonal IGH rearrangements at diagnosis (Table 2). The two or even three related clones in seven of them had evolved exclusively from VH replacements (Fig. 1). At relapse, one or two of these related rearrangements were lost in all but one patient (patient 22), who retained two related clones. The original clone (i.e., the clone with the most upstream VH segment) was only preserved in two patients, whereas in the other four cases the clone with the most downstream VH segment (i.e., the clone with the more "advanced" rearrangement) was preserved. This finding supports the notion that the relapse clone emerged from an oligoclonal population. Only one case (patient 23) was also oligoclonal at relapse. A new IGH rearrangement had developed, which was related to the persistent IGH rearrangement. These data show that subclones with related IGH rearrangements due to VH replacement are not only frequently detected at diagnosis but also, albeit to a lower extent, at relapse (Table 1).
|
Immunoglobulin light chain -locus. Twenty cases in group I had an IGK-Kde rearrangement either at initial diagnosis and/or at relapse compared with 10 cases in group II (Fig. 1). Up to five rearrangements per leukemia were identified at either occasion in group I, whereas in group II only one and two rearrangements were present in seven and three patients, respectively. In group I, 16 of the 18 cases with an IGK-Kde rearrangement at diagnosis had at least one rearrangement conserved at relapse, including five patients with additional losses and/or gains of rearrangements (Fig. 1). In two cases (patients 11 and 21) oligoclonal populations with three and five IGK-Kde rearrangements were present at diagnosis. The latter patient retained three of the five rearrangements at relapse, whereas another patient (patient 18) kept one IGK rearrangement, but also gained two additional ones (Fig. 1; Table 2).
T-cell receptor -locus. Twenty patients from group I had at least one TCRD rearrangement at diagnosis and/or at relapse, whereas only five patients in group II had one (n = 4) or two (n = 1) rearrangements (Fig. 1). In group I, only one rearrangement per leukemia was preserved at relapse (n = 12). No further changes took place in five of them; however, in the others, losses (n = 6) predominated over losses and gains (n = 1). Eight patients did not have a conserved rearrangement at relapse. Table 1 and Fig. 1 provide detailed information about the respective changes. The original oligoclonality with three and four rearrangements in two patients was lost at relapse (Table 2). The ongoing rearrangements from TCRD to one of the J segments of the TCRA locus were not analyzed systematically but data from individual cases indicate that, as expected, heterogeneity of these leukemias increases further, so that even some clonally appearing cases eventually became oligoclonal (patients 29 and 31; Fig. 1).
T-cell receptor -locus. With 27 cases in group I and 10 in group II, the overall frequency of TCRG rearrangements at diagnosis and/or relapse was high in both groups (Fig. 1). Ten of the 26 patients with a TCRG rearrangement at diagnosis in group I did not have a stable rearrangement because of isolated losses of rearrangements (n = 1) or, more frequently, of losses combined with gains (n = 9). Eight of the 16 patients with at least one stable rearrangement had additional losses or gains of rearrangements.
Because TCRG rearrangements are end-stage recombinations, a loss of rearrangements at this locus indicates that another subclone must have evolved to constitute the relapse clone. The deletion of a preexisting rearrangement by upstream V segments and downstream J segments, however, cannot be formally excluded in this study because the sequences were too short to distinguish between J1.1. and J2.1. or J1.3. and J2.3. It seems unlikely, however, that V-J replacements of rearrangements are common because the relapse V segments were more downstream than the one in the initial clone-specific rearrangement. Only one leukemia at diagnosis and two at relapse appeared oligoclonal. In two additional cases, a third very weak TCRG rearrangement was present at diagnosis but did not evolve to a relapse clone.
Oligoclonal immunogenotype pattern. At diagnosis, 11 cases (27%, 95% confidence interval 14-42%) appeared oligoclonal in up to three immunoreceptor loci. The highest frequency was found in the IGH gene (20%, 95% confidence interval, 5-35%; Table 2). In keeping with the high continuous recombinatorial potential of these leukemias, oligoclonal rearrangements were also identified at relapse in six patients.
Frequency of immunoglobulin/T-cell receptor rearrangements. We further analyzed the frequency of rearrangements for the individual IG/TCR loci at diagnosis and, in those from group I, also at relapse (Table 3). The highest frequency of IGH and TCRD rearrangements was found in the initial leukemias of group I. At relapse, IGH rearrangements were slightly less common, but the TCRD ones were significantly less common (P = 0.004, McNemar's test), whereas the incidence of IGK and TCRG remained constantly high at both occasions. In cases with a conserved pattern (group II), however, frequency of rearrangements of all immunoreceptor loci was similar to that of the relapses in group I. This difference in the frequencies of rearrangements at the individual gene loci is similar to the one reported previously for the age-related IG/TCR rearrangement patterns of TEL-AML1+ BCP-ALL at diagnosis (13).
|
The median age of the 28 children in group I was 3.15 years (range 1.7-13.4) and only four of them (14%) were 6 years or older at initial diagnosis. In contrast, the median age of the 13 patients in group II was 4.9 years (range 1.8-12.0) and six of them (46%) were 6 years or older (Table 4). The ungrouped age difference showed a trend toward a higher age in group II with the conserved pattern (P = 0.09, Wilcoxon two-sample test). This difference became statistically significant when the patients were grouped according to age (either younger or equal to and older than 6 years; P = 0.048, Fisher's exact test). The remission duration (group I: range 11-60 months, median 34; group II: range 6-56 months, median 32) did not differ significantly between the two groups (P = 0.7, Wilcoxon two-sample test). There was no association between age and number of rearrangements at diagnosis (P = 0.17, Kendall's rank correlation).
|
|
Discussion |
|---|
|
TopAbstractPatients, Materials, and MethodsResultsDiscussionReferences |
|---|
Previous comparisons of the diagnosis and relapse IG/TCR gene rearrangement patterns in BCP-ALL were either confined to specific gene loci or did not take potential peculiarities of particular genetic or age groups into consideration (20, 23, 24). Consequently, some authors noted a higher likelihood for clonal changes in patients with longer remission durations (18, 25), whereas others did not (26). In our cohort of TEL-AML1+ cases, the remission duration was neither associated with particular immunogenotype patterns nor did it depend on whether the relapse was the consequence of clonal evolution or selection. In support of subclone selection are the data from an earlier study demonstrating that in all six cases with a changed immunogenotype, the relapse clone was detectable at diagnosis irrespective of the remission duration (6). Such clones may withstand the initial treatment and may, therefore, also persist at a low level for several months, as was previously shown in two patients with a late recurrence of the disease (8). However, the biology behind the therapy resistance of these relapse clones also seems to differ from the ones that evolve into early relapses because the treatment response of such late relapses is usually as favorable as it is for the initial disease (27).
The incidence of oligoclonal IGH rearrangements of the cases reported herein was 20% and was, therefore, not significantly different from the 12% of the 214 previously analyzed TEL-AML1+ patients. Further, in this study, we could not establish that oligoclonality has a prognostic meaning or represents a more aggressive disease form as has been reported for BCP-ALL (18, 25, 28). It merely seems to be an expression of the high and persistent recombinatorial potential of these preleukemic/leukemic cells, which is also evidenced by their pronounced recombination-activating gene activity (29) and high number of IG/TCR gene rearrangements (13).
Nevertheless, it is intriguing to note that the clonal heterogeneity between diagnosis and relapse changes with the age of the patients at diagnosis. Because the vast majority of our patients were consecutively enrolled in the various BFM-based national therapy studies, we can rule out a selection bias. Moreover, the restriction to TEL-AML1+ cases also avoids biases that might have been due to more general age- and genetic subtype–related variations in the IG/TCR gene recombination process (12, 30). The observed age-dependent differences can, therefore, only be attributed to two factors, namely the differential accessibility of the respective IG/TCR loci, which is influenced by the developmental stage of the B precursor cell at the time of transformation and the continuation of the recombination process in these cells during the entire latency period. On the one hand, this process would lead to a continuous formation of new rearrangements at the accessible gene loci (IGK, TCRD, TCRG) and, on the other hand, it would also modify preexisting rearrangements (IGH, TCRD; ref. 23). In contrast to loci that can undergo further recombination, potential end-stage recombinations, such as IGK and TCRG rearrangements (31), are not expected to change with age. Whereas IGH loci may continue to rearrange their genes resulting in V replacements of VDJ rearrangements or in gene deletions, TCRD may further recombine with one of the J segments (22, 32). The age-associated decreasing frequency of IGH and TCRD rearrangements at diagnosis and the increase in biallelic end stage IGK and TCRG rearrangements (13), together with the lack of an apparent association between immunogenotype changes at relapse and the length of remissions, support the idea that the vast majority of these rearrangements are already formed in the preleukemic phase of the disease.
The current notion is that the TEL-AML1 gene fusion is one of the first steps in transformation. It is predominantly generated already during fetal development but only additional mutations are supposed to finally set off the development to a clinically apparent leukemia (2). The target cell, in which this gene fusion is formed, is currently unknown. However, recent experiments in mice indicate that the expression of the TEL-AML1 fusion gene inhibits already the earliest stages of B-cell development (33). Our data concord with these observations because the vast majority of leukemias has clonally appearing IGH rearrangements. Furthermore, we can deduce from the type of changes at relapse, namely VH replacements, that the respective target cell is indeed a B-cell precursor at the pro-B to pre-B transition. In this developmental stage, incomplete to complete VDJH rearrangements take place. After the completion of a VDJH rearrangement, subclone formation results most frequently or even exclusively in this type of clonal evolution (34). Notably, our data fit also the concept that TEL-AML1 does not transform hematopoietic "stem cells" (35) because relapses with completely unrelated IG/TCR rearrangements were hardly ever detected. Such a pattern would, in analogy with the BCR-ABL+ and the MLL-AF4+ acute leukemias, be expected if a stem cell is targeted (12, 36, 37).
How do our data now fit into this proposed scheme? If a change of the immunogenotype only depended on the age at initial diagnosis, then also the duration of the first remission should matter, yet we did not find any evidence for that. The important elements in the development of these leukemias and relapses are once again the two Darwinian principles of clonal evolution and selection (38, 39) whereby the various IG/TCR gene rearrangements serve only as very valuable clonal markers that probably do not have any further biological function or prognostic meaning.
In our model, schematically represented in Fig. 2, a precursor B cell with particular VDJH and IGK gene rearrangements expands after the TEL-AML1 gene fusion has taken place. Somatic recombination can then generate a multitude of clones with different TCRD and TCRG gene rearrangements. Those cells, which experience advantageous mutations, continue to expand, rearrange further, and modify their IG/TCR joints accordingly. Ongoing mutation and selection processes during the latency period produce many clones, which are defined, among others, by their individual types of IG/TCR gene rearrangements. One cell of these clones eventually acquires the critical mutation and progresses into a clinically overt leukemia. In addition, a large number of much smaller preleukemic subclones, whose common denominator is the TEL-AML1 fusion gene, may continue to linger around. The subclones, which escape the therapeutic interventions, will again provide the reservoir for the development of second leukemias in form of relapses.
|
Finally, we want to briefly refer to the potential consequences of our data for minimal residual disease analyses. As we have pointed out above, the presence of a PCR-detectable oligoclonality does not predispose to a higher relapse rate. Nevertheless, the higher number of rearrangements in younger children, together with the particular propensity of the occurrence of rearrangements in the relapse, which differs from those in the initial clone, requires specific attention. In our study, the frequency and stability of IG/TCR rearrangements (Table 2), with the exception of a higher instability of TCRD rearrangements, concur with those reported by Szscepansky et al. (23) on unselected childhood BCP-ALL. Therefore, our findings reinforce the currently used recommendation for the hierarchical order of applying the individual immunoreceptor loci in minimal residual disease studies according to their frequency and stability: IGK, TCRG, IGH, and TCRD (23, 30). Moreover, our data corroborate that preferentially two markers from different gene loci should be used. The remarkable instability of the individual IGH rearrangements can, to a large extent, be bypassed with the already widely applied approach to design primers that cover the DNJH region (26). Nevertheless, the lack of appropriate stable rearrangements together with the unpredictable clonal fluctuation prevents the use of the DNJH or any other rearrangement as an exclusive marker. Although V2-J and TCRB rearrangements have only been analyzed in a few cases, the preliminary results indicate that the ensuing configurations are also very heterogeneous as they are secondary rearrangements and ontogenetically younger (22, 40).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Presented in part at the 46th Annual Meeting of the American Society of Hematology, 2004, San Diego, California. This collaboration was done using the network of the Biology Group within the International Berlin-Frankfurt-Münster Group.
Received 6/ 8/05; revised 7/20/05; accepted 8/ 9/05.
|
References |
|---|
|
TopAbstractPatients, Materials, and MethodsResultsDiscussionReferences |
|---|
and T-cell receptor genes. Blood 1994;83:1892–902.2-J rearrangements are frequent in precursor-B-acute lymphoblastic leukemia but rare in normal lymphoid cells. Blood 2004;103:3798–804. gene rearrangements between presentation and relapse of childhood B-lineage acute lymphoblastic leukemia. Blood 1994;83:1355–62. deleting element rearrangements in precursor-B acute lymphoblastic leukemia are stable targets for detection of minimal residual disease by real-time quantitative PCR. Leukemia 2002;16:928–36.[CrossRef][Medline]
This article has been cited by other articles:
|
|
 |
|
  A. Bagg Malleable Immunoglobulin Genes and Hematopathology - The Good, the Bad, and the Ugly: A Paper from the 2007 William Beaumont Hospital Symposium on Molecular Pathology J. Mol. Diagn., September 1, 2008; 10(5): 396 - 410. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Fischer, G. Mann, M. Konrad, M. Metzler, G. Ebetsberger, N. Jones, B. Nadel, O. Bodamer, O. A. Haas, K. Schmitt, et al. Screening for leukemia- and clone-specific markers at birth in children with T-cell precursor ALL suggests a predominantly postnatal origin Blood, October 15, 2007; 110(8): 3036 - 3038. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Choi, M. J. Henderson, E. Kwan, A. H. Beesley, R. Sutton, A. Y. Bahar, J. Giles, N. C. Venn, L. D. Pozza, D. L. Baker, et al. Relapse in children with acute lymphoblastic leukemia involving selection of a preexisting drug-resistant subclone Blood, July 15, 2007; 110(2): 632 - 639. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
