SEARCH

SEARCH BY CITATION

Keywords:

  • acute myeloid leukaemia;
  • Wilms tumour 1 gene;
  • relapse kinetics;
  • complete remission;
  • prognosis

Summary

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Authorship
  8. References
  9. Supporting Information

We hypothesized that Wilms tumour 1 gene (WT1) expression levels in acute myeloid leukaemia (AML) patients might have predictive value and reveal molecular relapse kinetics. WT1 level was determined at diagnosis, during therapy and post-therapy follow-up in 89 patients who reached first complete remission (CR1) (952 samples, median 8 samples/patient, range 2–38). CR1 bone marrow (BM) WT1 level above normal (based on 39 healthy donors) was an independent adverse prognostic factor regarding both disease-free survival [hazard ratio (HR) 4·46, = 0·001] and overall survival (HR 2·62, = 0·019). By grouping 34 BM and 99 peripheral blood (PB) complete remission samples in monthly intervals prior to clinical relapse, disease development was delineated and a simple mathematical model constructed, that allowed for the prediction of relapse detection rates (RDRs) as well as median times [tms] from WT1 positivity to clinical relapse. BM sampling was required to obtain RDRs above 93% and tms above 67 d. Acceptable RDRs and tms (81% and 44 d, respectively) could be acquired by bimonthly PB sampling. In conclusion, CR1 WT1 expression is an independent prognostic factor in AML. According to our model, BM is superior for relapse prediction, but PB samples are useful when shorter sampling intervals are possible.

Acute myeloid leukaemia (AML) is a heterogeneous disease, as has been abundantly shown by cytogenetic and molecular techniques (Estey & Dohner, 2006). From the growing understanding of the diverse events that initiate and maintain the malignant clone, cytogenetic and molecular markers are rapidly emerging as highly useful, and constitute a gold standard for prognosticating patients (Grimwade et al, 1998). Thus, from the pioneering studies by Rowley & Potter (1976) demonstrating that balanced translocations are a recurrent feature of this disease, a series of other aberrations, not detected by karyotyping, have shed light on the molecular heterogeneity of the disease (for a review, see Estey & Dohner, 2006).

Not only does molecular characterization enable the up-front prognostication of AML patients (Grimwade et al, 1998), but some of the aberrations can also be employed in determining minimal residual disease (MRD) at a given time-point during and after completion of cytoreduction. Indeed, longitudinal application of real-time quantitative polymerase chain reaction (RQ-PCR) techniques has demonstrated that relapses in some patients can be detected up to a year before they emerge clinically (Schnittger et al, 2003; Stentoft et al, 2006). Early intervention can be initiated before the patients have symptoms and the accumulated burden of therapy can be significantly reduced. Thus, in acute promyelocytic leukaemia, treatment based on rising PML-RARA fusion transcript levels resulted in an improved outcome (Lo Coco et al, 1999).

A caveat to considerations on molecular markers in AML is that a sizeable portion of patients lack fusion transcripts amenable to MRD detection. On the other hand, some genes – while not mutated – have been found to be overexpressed in AML (for a recent compilation, see Steinbach et al, 2006). Among these, we have been especially interested in the Wilms tumour 1 gene (WT1). While it is beyond the scope of this report to describe the extensive studies performed to determine the seemingly contradictory functions of this gene in health and disease (for a review, see Hohenstein & Hastie, 2006), it is clearly overexpressed in a number of solid tumours and haematological malignancies (Miyoshi et al, 2002; Ueda et al, 2003; Sotobori et al, 2006). Thus, in AML its role as a tumour marker has been investigated in a series of reports (Inoue et al, 1994, 1996; Cilloni et al, 2002; Trka et al, 2002; Garg et al, 2003; Ogawa et al, 2003; Cilloni & Saglio, 2004; Ogawa et al, 2004; Østergaard et al, 2004; Weisser et al, 2005; Lapillonne et al, 2006). Collectively, these authors found that WT1 was overexpressed in most patients at diagnosis and that the levels uniformly rose prior to clinical relapse. More specifically, Weisser et al (2005) and Lapillonne et al (2006) divided post-inducton therapy samples from 106 adult and 92 children, respectively, into 30–60 d intervals after diagnosis with 25–45 samples in each interval. WT1 overexpression in these intervals was significantly associated with both event-free survival (EFS) and overall survival (OS). As none of these studies took disease status into account at the time of sampling, no study has, to our knowledge, taken knowledge of the precise dates of regeneration after therapy into consideration or systematically evaluated the role of WT1 transcript levels in the subgroup of patients who do achieve complete remission (CR1).

We hypothesized that determination of WT1 at CR1 and at each visit during subsequent chemotherapy and post-treatment follow-up might yield information regarding the likelihood of sustaining the remission and reveal the relapse kinetics. Regarding the latter, we were particularly interested in elucidating to the extent to which WT1 determination could anticipate relapse to provide a window of opportunity, which would allow for alternative therapeutic options, e.g. a non-myeloablative stem cell transplantation.

Materials and methods

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Authorship
  8. References
  9. Supporting Information

Patient cohort

Samples from 183 AML patients were received at the Laboratory of Immunohaematology from the Department of Haematology, Aarhus University Hospital (165 patients, diagnosed between October 1995 and December 2006) or the Department of Paediatrics, Aarhus University Hospital (18 patients, diagnosed between November 2000 and October 2006). No patients positive for RUNX1-RUNX1T1, CBFB-MYH11 or PML-RARA were included after August 2004 (Table I). Diagnosis of AML was based on morphology and immunophenotyping, the latter employing a standard panel of monoclonal antibodies. Treatment response was evaluated approximately 28 d after start of cytoreduction, or, in the case of poor responders, upon regeneration after salvage chemotherapy. Patients were considered in CR based on morphological and immunological findings. Adult patients were treated as previously described (Østergaard et al, 2004) or according to the AML15 protocol (after 1 January, 2004) (see http://www.aml15.bham.ac.uk/). Paediatric patients were treated according to the Nordic Society for Paediatric Haematology and Oncology (NOPHO) AML 93 protocol (Lie et al, 2003) or from January 2004 the NOPHO-AML 2004 protocol (ClinicalTrials.gov identifier NCT00476541, http://clinicaltrials.gov/ct2/show/NCT00476541). A summary of the treatment protocols used and a flow-chart depicting which patients participated in the different parts of this study can be found in the supplementary data [Appendix SI (incorporating Supplementary Figs S1–S4)]. Five children with Down syndrome treated at the Department of Paediatrics were not included in the study. All samples were obtained as part of diagnostic procedures and were used in accordance with protocols approved by the Ethical Committee for the County of Århus.

Table I.   Patient characteristics.
 All patientsWT1/ABL1 > 1/16WT1/ABL1 < 1/16P*
  1. *By chi-square test.

  2. Only patients treated with curative intent included in analysis.

  3. Only patients who attained CR1 included in analysis.

  4. §Only patients who attained CR2 included in analysis.

  5. By Wilcoxon′s rank-sum test.

  6. CR1/2, first/second complete remission; sAML, AML secondary to other haematological disease; tAML, therapy-related AML; ND, not done; NA, not applicable.

No. of patients18314142NA
 Treated with curative intent130102280·60
 Attained CR110683230·94
 After 1 course of cytoreduction8770170·43
 After > 1 course of cytoreduction191360·25
 Experienced relapse4732150·02
Allogeneic bone marrow transplant
 In CR113850·12
 In CR29810·13§
Diagnostic sample material
 Bone marrow156125310·02
 Average BM leukaemic blasts (%)6670630·03
 Peripheral blood5544110·53
 Male/female102/81 77/6425/170·57
Age distribution
 0–14 years181350·61
 15–59 years9679170·08
 60 + years6949200·13
 sAML/tAML24/3 17/3 7/00·44/0·34
Cytogenetics
 Favourable181440·94
  t(8;21)6240·01
  inv(16)8800·11
  t(15;17)4400·27
 Intermediate11490240·43
 Adverse181620·21
 ND3321120·01

RNA purification and cDNA synthesis

RNA purification was done on Lymphoprep (Axis Shields PoC AS, Oslo, Norway) separated mononuclear cells (MNC) as described earlier (Østergaard et al, 2004), on a MagNA Pure LC robot (Roche Diagnostics GmbH, Basel, Switzerland) or by using the TRIzol Reagent (Life Technologies, Frederick, MD, USA), as described by the manufacturers. cDNA synthesis was performed essentially as previously described (Pallisgaard et al, 1999) using 1 μg of TRIzol purified RNA or approximately 1 μg of MagNA Pure purified RNA, with the exception that 80 U of Moloney murine leukaemia virus reverse transcriptase (Invitrogen, Taastrup, Denmark) and one nanomole random nanomer primer was used.

RQ-PCR

Real-time quantitative polymerase chain reaction (RQ-PCR) was set up as previously described (Olesen et al, 2003) using the ABI Prism 7700 Sequence detector equipment (Applied Biosystems, Foster City, CA, USA) or the Mx3000P RQ-PCR System (Stratagene, La Jolla, CA, USA). WT1 primer and probe sequences have been described earlier (Østergaard et al, 2004). The control genes Abelson (ABL1) and ß-2-microglobulin (B2M), which have been validated in haematological malignancies (Beillard et al, 2003), were employed as follows: for comparison of diagnosis level between patients WT1 was normalized to ABL1 only, whereas MRD determinations were based on normalization to the average of ABL1 and B2M. B2M was omitted from the comparative analysis of the diagnostic samples as B2M, in contrast to ABL1, has been shown to be significantly different in normal and diagnostic bone marrow (BM) (Beillard et al, 2003). The B2M assay has been previously described (Pallisgaard et al, 1999). For ABL1 the sequences were: forward primer 5′ GGG TCC ACA CTG CAA TGT TT 3′, reverse primer 5′ CCA ACG AGC GGC TTC AC 3′, and probe 5′ FAM-TCA GATGCT ACT GGC CGC TGA AGG-TAMRA 3′. Assay set-up was done as described previously (Østergaard et al, 2004). All RNA samples were obtained from cryopreserved MNC lysed as described above. This resulted in less than 5% of the samples being rejected due to inferior RNA quality (identified as those with a threshold cycle number of B2M exceeding 25 and/or ABL1 exceeding 30).

Definition of normality and MRD determination

The normal WT1 threshold was set to the mean normal WT1 level + 2 standard deviations. Normal WT1 levels were measured in 39 BM samples and 27 peripheral blood (PB) samples from either healthy volunteers or non-haematological patients undergoing orthopaedic surgery. This level was 6·8 and 1·4 WT1 copies/1000 ABL1 copies in BM and PB, respectively. Samples with WT1 expression levels above the normal levels were defined as ‘WT1 high’. The distribution of WT1 in normal BM samples can be seen in Fig 1. MRD and sensitivity levels were calculated as described earlier (Østergaard et al, 2004), using the ΔΔC(t) method (Beillard et al, 2003). Briefly, BM diagnosis WT1 level was defined as one and subsequent samples compared to this level using control gene normalized WT1 values.

image

Figure 1.  Wilms tumour 1 gene (WT1) expression levels in bone marrow (BM) in 156 AML patients and 39 normal donors. The difference in WT1 expression between two columns correspond to a two-fold reduction. WT1 was employed as a minimum residual disease (MRD) marker in patients with a WT1/ABL1 ratio of above 1/16 (Arrow). < 0·0001 (by Wilcoxon’s rank-sum test).

Download figure to PowerPoint

Time point definitions and statistic analysis

The prognostic value of high WT1 expression was evaluated: (i) after regeneration following the first course of chemotherapy (EV1) and (ii) at the time when morphological CR was first diagnosed (CR1).

Kaplan–Meier survival curves were generated using the stata statistical software package v. 9·2 (StataCorp, College Station, TX, USA). Patients were censored if an allogeneic BM transplantation was instituted. Groups were compared using a proportional hazards model. With just 89 patients available for survival analysis, no validation in an independent patient set was sought for. Likewise, covariate analysis was restricted by performing a univariate Cox regression analysis with the following covariates: age (divided in three groups, children, adults below 62, and adults above 62), sex, white blood cell counts at diagnosis, FLT3-ITD status, de novo/secondary leukaemia, cytogenetic risk group (according to Grimwade et al, 1998), or amount of chemotherapy necessary to achieve CR. Only the cytogenetic risk group and age above 62 were found to be significant covariates with a significance level of 0·20 and were used in the final multivariate analysis, where the normal significance level of 0·05 was used. The proportional hazards assumption was checked individually for both of these covariates.

Results

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Authorship
  8. References
  9. Supporting Information

Number of WT1 determinations performed

The WT1 expression levels at diagnosis were determined in all 183 patients and normalized to the ABL1 control as described above. In Fig 1, the highly significant difference in levels of WT1 expression in patients and normal donors is shown.

Assuming an adequate sample quality, the sensitivity of the WT1 assay depends on the degree of WT1 overexpression in the leukaemic blasts, and therefore the general practice in our laboratory has been to select only patients with a WT1/ABL1-ratio higher than 1/16 for WT1 follow-up. As seen in Fig 1, a clear separation was observed between the normal BM WT1 expression and the 80% of AML samples with a WT1/ABL1 ratio of more than 1/16.

A summary of the clinical characteristics and treatment appears in Table I. Notable differences are a higher incidence of RUNX1-RUNX1T1 positive leukaemias in the group with WT1/ABL1 <1/16 and a lower relapse rate in the group with WT1/ABL1 >1/16.

The WT1 level as a predictor of persisting CR

Previous studies on the value of WT1 as prognostic marker have relied upon WT1 levels determined at a fixed time-point(s) (Weisser et al, 2005; Lapillonne et al, 2006), where CR status was not known. When we used this approach, a clear correlation between course of disease and WT1 levels after the first course of cytoreduction (EV1) was observed [EFS, BM; hazard rate (HR) 3·24, 95% confidence intervals (CI) 1·38–7·62, = 0·007; EFS, PB; HR 2·21, CI 1·02–4·77, = 0·044; OS, BM; HR 3·45, CI 1·58–7·54, = 0·002; OS, PB; HR 3·80, CI 1·72–8·36, = 0·001] (Fig S5A–D).

The usefulness of molecular markers for stratifying AML patients after initial therapy should be greatest in those who are in morphological and immunological CR, but who may nevertheless have an underlying molecular MRD. In this setting, including patients at fixed time points after commencement of induction therapy [e.g. day 35–50 (Lapillonne et al, 2006) or day 16–60 (Weisser et al, 2005)] will result in the inclusion of patients not in CR after the first course, and this will hamper the value of the concept, as such patients will inevitably be included in the WT1 high group. We therefore decided to base the WT1 determination on samples obtained at CR1, irrespective of the number of prior chemotherapy courses. As can be seen from Fig 2, a significant effect on EFS and OS of high WT1 expression in BM in CR1 was observed (EFS, HR=4·45, CI 1·83–10·84, = 0·001; OS, HR = 2·90, CI 1·30–6·50, = 0·009). Similar results were obtained in PB (EFS, HR 2·52, CI 1·07–5·94, = 0·035; OS, HR 2·63, CI 1·17–5·91, = 0·019) (see Fig S5E–F).

image

Figure 2.  Event free survival (EFS) (A) and overall survival (OS) (B) according to Kaplan–Meier plots stratified for high WT1 bone marrow (BM) levels at complete remission (CR1). Data was adjusted for age above 62 years and cytogenetic risk group. EFS, hazard rate (HR) = 4·45, 95% CI 1·83–10·84, = 0·001; OS, HR = 2·90, 95% CI 1·30–6·50, = 0·009.

Download figure to PowerPoint

Leukaemia relapse detection by longitudinal WT1 determinations in CR

The WT1 follow-up data set included a total of 952 samples from 89 patients (median eight samples/patient, range 2–38) obtained up to 1794 d after diagnosis. In 32 patients 37 clinical relapses occurred, and in 29 of these at least one BM or PB sample was available during the 12-month period preceding the relapse (Fig 3). As we have earlier shown that relapse kinetics for MYH11-CBFB-positive leukaemias differ markedly from other AMLs by showing a slower MRD increment (Stentoft et al, 2006), we initially excluded those from this analysis. The remaining samples were grouped into monthly intervals preceding the relapse. Patients were considered negative or positive in all intervals between two negative or positive samples, respectively. When the number of positive patients in a given interval was depicted as a fraction of the total number of patients in that particular interval, we were able to identify differing relapse kinetics patterns in BM and PB (Fig 4). Thus, Fig 4 incorporates both information from patients where molecular relapse was detected (PB: Fig 3A, relapses 1–18; BM: Fig 3B, relapses 1–10) and patients where no molecular relapse was seen (PB: Fig 3A, relapses 19–31; BM: Fig 3B, relapses 11–23). In PB, we observed samples with normal WT1 expression quite close to the clinical relapse in some patients (relapse 12, 15 and 17 in Fig 3A), whereas in BM, no negative samples were observed in the period 3–4 months before relapse (Fig 3B). WT1 conversion graphs including CBFB-MYH11 can be seen in Supplementary Figs 6.

image

Figure 3.  Overview of patient samples and disease courses. x-axis: time before relapse, which is designated at time t = 0, y-axis: specific relapse number. Relapse numbers in bone marrow (BM) and peripheral blood (PB) does not correspond . (A) PB relapses 1, 2, 4, 7 and 13 are CBFB-MYH11 positive. six relapses with no remission samples omitted (B) BM Relapses 1, 11 and 20 are CBFB-MYH11 positive. 15 relapses with no remission samples omitted. Open circles: negative samples, Filled diamonds; positive samples. Dotted lines connect negative samples and mark the time span where patients are considered Wilms tumour 1 gene (WT1)-negative. Solid lines connect positive samples and mark the time spans where patients are considered WT1-positive.

Download figure to PowerPoint

image

Figure 4.  Wilms tumour 1 gene (WT1) positivity conversion patterns in bone marrow (BM) and peripheral blood (PB) prior to relapse. Number of relapses; total 24 (five CBFB-MYH11 censored as described in the text), BM 15 (three CBFB-MYH11 censored), and PB 22 (five CBFB-MYH11 censored). Median number of relapses in each interval; BM: seven (range 6–9), PB: nine (range 6–13).

Download figure to PowerPoint

Corroborating the notion that BM is superior to PB, we observed that in all patients in whom samples were available within 3 months prior to relapse, a median detection time of 4 months or 1 month before haematological relapse was seen in BM and PB, respectively (BM range 2–8, PB range 0–8) (Wilcoxon’s rank-sum test, P = 0·0086).

A mathematical model for optimal PB and BM sampling

Given these data, we next evaluated the optimal sampling interval based on a lack of data points, as must be accepted in a fraction of the patients (Fig 3), in particular after completion of cytoreduction. We showed that the areas below the graphs in Fig 4 could be interpreted as the total number of patients where relapse is detected. Given that the total number of patients is known, the likelihood of detecting a relapse by WT1 measurements as a function of the interval between PB and BM sampling can then be determined. Additionally, we employed an integral equation to determine the time point before relapse where half the molecular relapses had occurred. This can be interpreted as the median time to relapse for a patient with a positive WT1 sample [please refer to Appendix SII (incorporating Supplementary Figs 8 and 9) for a detailed description of the mathematical methods employed]. As expected, BM sampling provided higher detection rates and longer windows of opportunity before relapse, but employing bimonthly PB sampling, more than 80% of relapses were detected with a median time to relapse of nearly 50 d (Table II).

Table II.   Projected* relapse detection rates and median time to relapse in PB and BM depending on sampling interval.
Sampling interval (months)BMPB
RDRtm(days)RDRtm(days)
  1. *According to the mathematical model outlined in ‘Appendix SII’.

  2. PB, peripheral blood; BM, bone marrow; RDR, relapse detection rate; tm, median time to relapse.

 11·001130·9353
 21·00980·8144
 30·99840·6738
 40·93740·5435
 60·72700·3635
 90·50670·2435
120·37670·1835

Discussion

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Authorship
  8. References
  9. Supporting Information

In this report we have provided data suggesting that close monitoring of AML patients in CR overexpressing WT1 will add important biological information in a large subgroup of patients, which is not otherwise amenable to MRD determination, e.g. by fusion transcript determinations. Our data add knowledge to the value of WT1 as an MRD marker with the focus on its value in patients with established remission, in contrast to previous reports, which have dealt mainly with evaluations of levels at diagnosis followed by more limited determinations in CR. Several questions, both methodological and conceptual, need – however – to be considered in the light of our data set.

While the WT1 RQ-PCR assay has not yet been formally validated in a multicentre setting, as has those for a large number of fusion transcripts (Gabert et al, 2003), the methods employed here were all part of the Europe Against Cancer (EAC) BIOMED initiative (Gabert et al, 2003), as were the control genes chosen (Beillard et al, 2003). A drawback of our assay has been our findings in a single paediatric patient, in whom a clonal expansion of cells carrying a 8 bp deletion in exon 7 of WT1, where our reverse primer is placed, lead to a false negative reaction (Nyvold et al, 2006). In fact, the present series of patients singles out this patient as a very rare event (1/37 relapses). Thus, the risk of not detecting the relapse because of a mutation in the primer or probe sites is of minor concern compared to the simple risk of not detecting the relapse because of the time between sample acquisitions (as shown in Table II). Similarly, the presence of a single nucleotide polymorphism (NCBI reference SNP ID: rs16754) in the probe sequence is less important, as the frequency of the variant allele homozygosity is quite low (3·4%) (International Hapmap Project, http://www.hapmap.org) in persons of European ancestry. Even though probes and primers should optimally be designed so that they do not span known single nucleotide polymorphisms or in regions of recurrent mutations, we believe that these issues may have only very limited impact on our study.

Another methodological issue to be considered pertains to the threshold of gene expression applied for calculations. While this is indeed a feature of all RQ-PCR assays, it is especially relevant for a gene like WT1, which is also expressed in non-malignant cells. Earlier work has suggested that WT1 expression could be mainly ascribed to CD34+ progenitor cells (Ellisen et al, 2001), which raises the question as to the threshold level that should be applied in cytoreduced AML patients. We set our BM threshold at 6·8 WT1 copies/1000 ABL1 copies, which is lower than the BM threshold described by Cilloni et al (2002), who employed a threshold of 18 WT1 copies/1000 ABL1 copies. Our chosen threshold identified two patients who, despite the fact that they did not relapse, continually tested positive (WT1 follow-up curves for these patients can be found in Fig S7). These patients may comprise an interesting subgroup in which close follow-up seems warranted, for in contrast to patients with expanding leukaemic clones, where WT1 mRNA levels are rapidly rising, these patients displayed only very slowly increasing WT1 expression levels.

In the 81 remaining patient courses samples with a WT1 level above the threshold were observed in three patients. The remaining samples showed WT1 levels within the normal range, and none of the patients experienced clinical relapse. It cannot be ruled out that these sporadic elevated WT1 levels might reflect a transient, biologically relevant situation. On the other hand, if considered as false positives (i.e. WT1 level above normal), the detection rate compared to only 0·5% (as there were more than 600 WT1 determinations in CR) and was in effect, lower than could be expected from our definition of the BM threshold (mean + 2 standard deviations). In all three cases, retesting of the patients indicated that the sporadic positivity related to stochastic variation (data not shown). Based on our findings, we recommend that at least two consecutive samples should have a positive result, together with an evaluation of the relapse kinetics, as a prerequisite for taking action in terms of relapse treatment initiatives.

Despite these caveats, conceptual as well as methodological, it is now clear that the determination of WT1 after cytoreductive treatment allows for the identification of patients with a significant higher risk of relapse. Additionally, we were able to show that this is the case irrespective of whether WT1 expression is measured at a fixed time point after commencement of chemotherapy or whether the WT1 expression is determined when the patient is deemed to be in morphological and immunological CR. Thus, WT1 is a valid prognostic tool not only for the patients who achieve CR after initial cytoreduction, but also those in whom several courses of chemotherapy were necessary to attain CR. This is the first time that patients with WT1 expression below a defined normal threshold levels have been shown to have significantly longer EFS and OS than the group with high WT1 expression, even though all patients were in morphological and immunological CR. Adding to the applicability is the fact that BM and PB samples were comparable for these comparisons.

In the setting of longitudinal WT1 sampling during CR, our data adds considerably to previous studies employing WT1, which have usually employed less frequent testing or tested a smaller number of patients. Weisser et al (2005) tested 569 samples with a median of five samples/patient and were able to detect rising WT1 level in 16/42 relapses before clinical manifestation. By testing patients more often (952 samples, 83 patients, median of eight samples/patient) we were able to detect relapse in a somewhat higher proportion of relapses (21/37). More importantly, we incorporated the information from patients with samples from within the last year before relapse (available in 29/37 relapses). Thus, both samples with WT1 determinations above and within the normal range were used to delineate relapse kinetics.

Despite the inferior sensitivity of the WT1 assay compared within fusion transcript-based assays, we diagnosed relapses in WT1 patients several months prior to clinical, histological, and even immunological relapse. We consider this proof-of-concept for the use of gene overexpression as an independent detector of relapse in a single centre setting. It might be envisaged that the higher normal expression of WT1 in BM could hamper an early relapse detection in this tissue. However, in our data set, we were able to compare BM and PB relapse detection and striking differences were observed regarding the extent of positivity. Relapses were detected significantly earlier in BM, and the lag phase between BM and PB was as long as 2 months in some cases with paired samples (data not shown).

We have for the first time made an effort to define an optimal sampling interval for a molecular relapse detection in PB and BM. By employing a simple mathematical model we were able to establish that while BM is clearly superior to PB, bimonthly PB sampling of the patients could obviate the need for frequent BM aspirations. Such blood sampling is clearly an option, possibly at secondary centres or at the patients’ general practitioner, as samples for RQ-PCR analysis can be shipped overnight without detrimental effects on sample quality (van der Velden et al, 2004). Naturally, this model can be applied equally well to other MRD markers, e.g. fusion transcript determinations.

In conclusion, we have validated the WT1 assay as a powerful tool for determining the disease state in a large cohort of AML patients, who were otherwise not amenable to MRD diagnostics. While novel mutations such as NPM1 can be used as MRD markers, the use of WT1 offers the possibility to follow a larger proportion of patients with a single assay. Moreover, our data show that the WT1 assay provides a sufficient window of opportunity to instigate therapeutic actions in patients with impending relapse.

Acknowledgements

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Authorship
  8. References
  9. Supporting Information

This work was supported by grants to PH from the Danish Medical Research Council, the Danish Cancer Society, the John and Birthe Meyer Foundation, and the Karen Elise Jensen Foundation. We would like to thank Eigil Kjeldsen for help with the cytogenetic data, Caroline Juhl-Christensen, Anita Rethmeier, Lykke Grubach, Trine Silkjær, and Steffen Hokland for helpful discussions and Lone Siig Mikkelsen for expert technical assistance.

Authorship

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Authorship
  8. References
  9. Supporting Information

HBO did experimental work, analysed data, performed mathematical analyses and wrote the manuscript. KB and BLA handled the patient samples and did experimental work, CGN supervised the experimental work, IBO performed mathematical analyses, HH was responsible for the clinical characterization of and collection of samples from the pediatric patients, MØ and PH designed the study and wrote the manuscript. All authors reviewed, edited and approved the manuscript.

References

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Authorship
  8. References
  9. Supporting Information
  • Beillard, E., Pallisgaard, N., Van Der Velden, V.H., Bi, W., Dee, R., Van Der Schoot, E., Delabesse, E., Macintyre, E., Gottardi, E., Saglio, G., Watzinger, F., Lion, T., Van Dongen, J.J., Hokland, P. & Gabert, J. (2003) Evaluation of candidate control genes for diagnosis and residual disease detection in leukemic patients using ‘real-time’ quantitative reverse-transcriptase polymerase chain reaction (RQ-PCR) – a Europe against cancer program. Leukemia, 17, 24742486.
  • Cilloni, D. & Saglio, G. (2004) WT1 as a universal marker for minimal residual disease detection and quantification in myeloid leukemias and in myelodysplastic syndrome. Acta Haematologica, 112, 7984.
  • Cilloni, D., Gottardi, E., De Micheli, D., Serra, A., Volpe, G., Messa, F., Rege-Cambrin, G., Guerrasio, A., Divona, M., Lo Coco, F. & Saglio, G. (2002) Quantitative assessment of WT1 expression by real time quantitative PCR may be a useful tool for monitoring minimal residual disease in acute leukemia patients. Leukemia, 16, 21152121.
  • Ellisen, L.W., Carlesso, N., Cheng, T., Scadden, D.T. & Haber, D.A. (2001) The Wilms tumor suppressor WT1 directs stage-specific quiescence and differentiation of human hematopoietic progenitor cells. EMBO Journal, 20, 18971909.
  • Estey, E. & Dohner, H. (2006) Acute myeloid leukaemia. Lancet, 368, 18941907.
  • Gabert, J., Beillard, E., Van Der Velden, V.H., Bi, W., Grimwade, D., Pallisgaard, N., Barbany, G., Cazzaniga, G., Cayuela, J.M., Cave, H., Pane, F., Aerts, J.L., De Micheli, D., Thirion, X., Pradel, V., Gonzalez, M., Viehmann, S., Malec, M., Saglio, G. & Van Dongen, J.J. (2003) Standardization and quality control studies of ‘real-time’ quantitative reverse transcriptase polymerase chain reaction of fusion gene transcripts for residual disease detection in leukemia – a Europe Against Cancer program. Leukemia, 17, 23182357.
  • Garg, M., Moore, H., Tobal, K. & Liu Yin, J.A. (2003) Prognostic significance of quantitative analysis of WT1 gene transcripts by competitive reverse transcription polymerase chain reaction in acute leukaemia. British Journal of Haematology, 123, 4959.
  • Grimwade, D., Walker, H., Oliver, F., Wheatley, K., Harrison, C., Harrison, G., Rees, J., Hann, I., Stevens, R., Burnett, A. & Goldstone, A. (1998) The importance of diagnostic cytogenetics on outcome in AML: analysis of 1,612 patients entered into the MRC AML 10 trial. The Medical Research Council Adult and Children’s Leukaemia Working Parties. Blood, 92, 23222333.
  • Hoel, P.G., Port, S.C. & Stone, C.J. (1971) Continuous random variables. In: Introduction to Probability Theory, pp. 132133. Houghton Mifflin Company, Boston.
  • Hohenstein, P. & Hastie, N.D. (2006) The many facets of the Wilms’ tumour gene, WT1. Human Molecular Genetics, 15, R196R201.
  • Inoue, K., Sugiyama, H., Ogawa, H., Nakagawa, M., Yamagami, T., Miwa, H., Kita, K., Hiraoka, A., Masaoka, T., Nasu, K., Kyo, T., Dohy, H., Nakauchi, H., Ishidate, T., Akiyama, T. & Kishimoto, T. (1994) WT1 as a new prognostic factor and a new marker for the detection of minimal residual disease in acute leukemia. Blood, 84, 30713079.
  • Inoue, K., Ogawa, H., Yamagami, T., Soma, T., Tani, Y., Tatekawa, T., Oji, Y., Tamaki, H., Kyo, T., Dohy, H., Hiraoka, A., Masaoka, T., Kishimoto, T. & Sugiyama, H. (1996) Long-term follow-up of minimal residual disease in leukemia patients by monitoring WT1 (Wilms tumor gene) expression levels. Blood, 88, 22672278.
  • Lapillonne, H., Renneville, A., Auvrignon, A., Flamant, C., Blaise, A., Perot, C., Lai, J.L., Ballerini, P., Mazingue, F., Fasola, S., Dehee, A., Bellman, F., Adam, M., Labopin, M., Douay, L., Leverger, G., Preudhomme, C. & Landman-Parker, J. (2006) High WT1 expression after induction therapy predicts high risk of relapse and death in pediatric acute myeloid leukemia. Journal of Clinical Oncology, 24, 15071515.
  • Lie, S.O., Abrahamsson, J., Clausen, N., Forestier, E., Hasle, H., Hovi, L., Jonmundsson, G., Mellander, L. & Gustafsson, G. (2003) Treatment stratification based on initial in vivo response in acute myeloid leukaemia in children without Down’s syndrome: results of NOPHO-AML trials. British Journal of Haematology, 122, 217225.
  • Lo Coco, F., Diverio, D., Avvisati, G., Petti, M.C., Meloni, G., Pogliani, E.M., Biondi, A., Rossi, G., Carlo-Stella, C., Selleri, C., Martino, B., Specchia, G. & Mandelli, F. (1999) Therapy of molecular relapse in acute promyelocytic leukemia. Blood, 94, 22252229.
  • Miyoshi, Y., Ando, A., Egawa, C., Taguchi, T., Tamaki, Y., Tamaki, H., Sugiyama, H. & Noguchi, S. (2002) High expression of Wilms’ tumor suppressor gene predicts poor prognosis in breast cancer patients. Clinical Cancer Research, 8, 11671171.
  • Nyvold, C.G., Stentoft, J., Braendstrup, K., Melsvik, D., Moestrup, S.K., Juhl-Christensen, C., Hasle, H. & Hokland, P. (2006) Wilms’ tumor 1 mutation accumulated during therapy in acute myeloid leukemia: biological and clinical implications. Leukemia, 20, 20512054.
  • Ogawa, H., Tamaki, H., Ikegame, K., Soma, T., Kawakami, M., Tsuboi, A., Kim, E.H., Hosen, N., Murakami, M., Fujioka, T., Masuda, T., Taniguchi, Y., Nishida, S., Oji, Y., Oka, Y. & Sugiyama, H. (2003) The usefulness of monitoring WT1 gene transcripts for the prediction and management of relapse following allogeneic stem cell transplantation in acute type leukemia. Blood, 101, 16981704.
  • Ogawa, H., Ikegame, K., Kawakami, M. & Tamaki, H. (2004) WT1 gene transcript assay for relapse in acute leukemia after transplantation. Leukaemia & Lymphoma, 45, 17471753.
  • Olesen, L.H., Norgaard, J.M., Pallisgaard, N., Bukh, A. & Hokland, P. (2003) Validation and clinical implication of a quantitative real-time PCR determination of MDR1 gene expression: comparison with semi-quantitative PCR in 101 patients with acute myeloid leukemia. European Journal of Haematology, 70, 296303.
  • Østergaard, M., Olesen, L.H., Hasle, H., Kjeldsen, E. & Hokland, P. (2004) WT1 gene expression: an excellent tool for monitoring minimal residual disease in 70% of acute myeloid leukaemia patients – results from a single-centre study. British Journal of Haematology, 125, 590600.
  • Pallisgaard, N., Clausen, N., Schroder, H. & Hokland, P. (1999) Rapid and sensitive minimal residual disease detection in acute leukemia by quantitative real-time RT-PCR exemplified by t(12;21) TEL-AML1 fusion transcript. Genes, Chromosomes and Cancer, 26, 355365.
  • Rowley, J.D. & Potter, D. (1976) Chromosomal banding patterns in acute nonlymphocytic leukemia. Blood, 47, 705721.
  • Schnittger, S., Weisser, M., Schoch, C., Hiddemann, W., Haferlach, T. & Kern, W. (2003) New score predicting for prognosis in PML-RARA+, AML1-ETO+, or CBFBMYH11+ acute myeloid leukemia based on quantification of fusion transcripts. Blood, 102, 27462755.
  • Sotobori, T., Ueda, T., Oji, Y., Naka, N., Araki, N., Myoui, A., Sugiyama, H. & Yoshikawa, H. (2006) Prognostic significance of Wilms tumor gene (WT1) mRNA expression in soft tissue sarcoma. Cancer, 106, 22332240.
  • Steinbach, D., Schramm, A., Eggert, A., Onda, M., Dawczynski, K., Rump, A., Pastan, I., Wittig, S., Pfaffendorf, N., Voigt, A., Zintl, F. & Gruhn, B. (2006) Identification of a set of seven genes for the monitoring of minimal residual disease in pediatric acute myeloid leukemia. Clinical Cancer Research, 12, 24342441.
  • Stentoft, J., Hokland, P., Ostergaard, M., Hasle, H. & Nyvold, C.G. (2006) Minimal residual core binding factor AMLs by real time quantitative PCR-initial response to chemotherapy predicts event free survival and close monitoring of peripheral blood unravels the kinetics of relapse. Leukemia Research, 30, 389395.
  • Trka, J., Kalinova, M., Hrusak, O., Zuna, J., Krejci, O., Madzo, J., Sedlacek, P., Vavra, V., Michalova, K., Jarosova, M. & Stary, J. (2002) Real-time quantitative PCR detection of WT1 gene expression in children with AML: prognostic significance, correlation with disease status and residual disease detection by flow cytometry. Leukemia, 16, 13811389.
  • Ueda, T., Oji, Y., Naka, N., Nakano, Y., Takahashi, E., Koga, S., Asada, M., Ikeba, A., Nakatsuka, S., Abeno, S., Hosen, N., Tomita, Y., Aozasa, K., Tamai, N., Myoui, A., Yoshikawa, H. & Sugiyama, H. (2003) Overexpression of the Wilms’ tumor gene WT1 in human bone and soft-tissue sarcomas. Cancer Science, 94, 271276.
  • Van Der Velden, V.H., Boeckx, N., Gonzalez, M., Malec, M., Barbany, G., Lion, T., Gottardi, E., Pallisgaard, N., Beillard, E., Hop, W.C., Hoogeveen, P.G., Gabert, J. & Van Dongen, J.J. (2004) Differential stability of control gene and fusion gene transcripts over time may hamper accurate quantification of minimal residual disease – a study within the Europe Against Cancer Program. Leukemia, 18, 884886.
  • Weisser, M., Kern, W., Rauhut, S., Schoch, C., Hiddemann, W., Haferlach, T. & Schnittger, S. (2005) Prognostic impact of RT-PCR-based quantification of WT1 gene expression during MRD monitoring of acute myeloid leukemia. Leukemia, 19, 14161423.

Supporting Information

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Authorship
  8. References
  9. Supporting Information

Appendix SI. AML treatment protocols.

Appendix SII. Mathematical considerations on detection rates and time to relapse.

Fig S1. Patient sample flowchart.

Fig S2. AML15 protocol.

Fig S3. NOPHP93 protocol.

Fig S4. NOPHPO2004 protocol.

Fig S5. Supplementary Kaplan–Meier survival curves for the WT1 follow-up cohort.

Fig S6. WT1 positivity conversion patterns in BM and PB prior to relapse, including CBFB-MYH11 leukaemias.

Fig S7. WT1 follow-up graphs for two patients (#1 and #2) with continually high but stable WT1 expression and a relapsing patient.

Fig S8. Mathematical considerations on relapse detection rates.

Fig S9. Mathematical considerations on the first time positive relapse detection function.

Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
BJH_7132_sm_appendix si.doc35KSupporting info item
BJH_7132_sm_appendix sii.pdf279KSupporting info item
BJH_7132_sm_figs1.pdf51KSupporting info item
BJH_7132_sm_figs2.pdf92KSupporting info item
BJH_7132_sm_figs3.pdf53KSupporting info item
BJH_7132_sm_figs4.pdf78KSupporting info item
BJH_7132_sm_figs5.zip211KSupporting info item
BJH_7132_sm_figs5legend.doc25KSupporting info item
BJH_7132_sm_figs6.pdf97KSupporting info item
BJH_7132_sm_figs7.pdf50KSupporting info item
BJH_7132_sm_figs8.pdf41KSupporting info item
BJH_7132_sm_figs9.pdf41KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.