Early prediction of treatment outcome in acute myeloid leukemia by measurement of WT1 transcript levels in peripheral blood samples collected after chemotherapy

Abstract

The Wilms’ tumor gene WT1 is a reliable marker for minimal residual disease assessment in acute leukemia patients. The study was designed to demonstrate the potential use of WT1 to establish quality of remission in acute leukemia patients for early identification of patients at high risk of relapse. A prospective study based on a quantitative Real–Time PCR (TaqMan) assay in 562 peripheral blood samples collected from 82 acute leukemia patients at diagnosis and during follow-up was established. The evaluation of WT1 in peripheral blood samples after induction chemotherapy can distinguish the continuous complete remission patients from those who obtain only an “apparent” complete remission and who could relapse within a few months. WT1 helps identify patients at high risk of relapse soon after induction chemotherapy allowing post-induction therapy in high risk patients to be intensified.

Introduction

Evaluation of minimal residual disease (MRD) in acute myeloid leukemia (AML) after initial chemotherapy is important to predict prognosis and may improve selection of the type and intensity of post-remission treatment.1 The mainstays of MRD studies in this setting include molecular tests, such as RTPCR amplification of chromosome translocations2,3 and multi-dimensional flow cytometry detection of aberrant phenotypes. 4–6 The major obstacle in MRD detection by RT-PCR or quantitative PCR (RQ-PCR) is represented by the limited percentage of AML patients presenting with detectable genetic aberrations.7 This prompted several investigators to validate the use of alternative markers for MRD detection suitable in the vast majority of AML patients, and in particular to test WT1 expression as a universal marker of leukemic cells.8–12 The Wilms’s tumor gene (WT1) codes for a transcription factor that has been shown to be highly expressed in several hematopoietic tumors including AML.8–12 Given the existence of background levels determined by WT1 expression in normal bone marrow, studies using qualitative RT-PCR have provided conflicting results on the clinical value of this marker,13,14 whereas most recent investigations by quantitative real-time RT-PCR (RQ-PCR) clearly distinguished WT1 transcript amounts related to AML cells, normal hemopoietic cells and post-chemotherapy regenerating normal bone marrow cells.15,16 Therefore, longitudinal RQ-PCR analysis of WT1 transcript amount may prove clinically relevant for AML monitoring. Furthermore, since WT1 expression in normal peripheral blood (PB) is about 1 log lower than in normal BM with the majority of normal PB samples scoring negative, we hypothesized that sequential RQPCR study of WT1 expression in PB might further improve the sensitivity of MRD evaluation in AML and might also favor compliance and sample availability. Finally, one of the main goals of MRD assessment is represented by the possibility of identifying, as soon as possible after induction chemotherapy the subset of patients who, although in CR, present a high risk of relapse. This means these patients can be treated with intensified chemotherapy protocols. To this purpose, in the present study, we analyzed WT1 expression levels at diagnosis and during follow-up in 82 AML patients treated with standard chemotherapy protocols. Our data indicate that this approach provides important prognostic information in AML by identifying patients at higher risk of relapse early after induction chemotherapy.

Design and Methods

After obtaining informed consent, 562 peripheral blood (PB) samples were collected from 82 AML patients. PB sampling was performed at diagnosis, after each cycle of chemotherapy and at sequential time intervals during follow- up and at relapse. All cases were classified according to FAB criteria, characterized at the cytogenetic level, and screened by RT-PCR for the presence of the most frequent fusion transcripts as previously described.17 The main clinical and biological characteristics of the cohort of analyzed patients are reported in Table1 (Online Appendix 1). Patients under 60 years of age were treated following standard protocols established by the GIMEMA Cooperative Group. (treatment details in Online Appendix 2) After each cycle of chemotherapy, a BM aspirate and biopsy were performed in order to assess the response to therapy. Complete remission was defined according to standard criteria.18 PB samples were collected at diagnosis, after induction and consolidation chemotherapy. PB samples were collected 1–5 days after the achievement of a neutrophil count 0.5×10/L. A mean of 6.8 samples per patient were available for RQPCR analysis (range 3–16). The median follow-up was 17 months (mean 20.7, range 5–66). Finally, as previously described,10 70 PB samples and 22 BM samples from healthy volunteers were used as control in order to define the normal range of WT1 expression in healthy subjects.

Cytogenetic and molecular analysis

Cytogenetic and molecular analysis was performed in all patients included in the study following standard procedures.17,3 WT1 RQ-PCR reactions and fluorescence values were measured as previously described.10,19

Figure 1.WT1 expression at diagnosis and during follow-up of 27 patients in CR characterized by WT1 values within the normal range after induction chemotherapy. All these patients persisted in CR during follow-up. WT1 never increased above the normal range during follow-up. The broken line represents the upper normal

Results

As previously reported in other papers10,11 BM and PB samples obtained from healthy volunteers express very low levels of WT1. In our study, the majority of normal PB are negative for WT1 expression and the mean value in positive samples is 4±3 WT1 copies/10 ABL copies (median 0, range 0–20). BM samples of normal donors express a mean value of WT1 copies of 32±19 (median 28, range 0–90). A median value of 3,725 WT1 copies/10 ABL copies (mean value 7,222±12,496, range 68–95,549) was detected in the 82 PB samples collected at diagnosis and evaluated in the present study, and a median value of 30,212 (mean 57,830±22,980, range 416–122,714) in BM samples. Seventy-one out of 82 patients achieved complete remission (CR) and 11 patients were resistant to induction chemotherapy. Twenty-seven out of 71 patients attaining CR after chemotherapy persisted in remission with a median of 28 months of follow-up (mean 30.3 months, range 12–60)(Figure 1) and 44 relapsed after a median of ten months (range 5–66) during the follow-up period. (Figures 2 and 3). No significant differences were observed in WT1 transcript amount at diagnosis between the patients who persisted in CR and those who relapsed, either in PB (p=0.13) or in BM (p=0.27). No difference in the WT1 amount at diagnosis was detected in patients resistant to chemotherapy when compared with responders (p=0.05). Regression analysis demonstrates the absence of correlation between WT1 expression and WBC count at diagnosis (r=0.0008). There was no significant difference in WT1 transcript within the FAB and cytogenetic risk subgroups and in patients carrying ITD or point mutation of FLT3 when compared with the wild type FLT3 group (p=0.29). WT1 quantitative assessment was performed in PB samples obtained soon after recovery from induction chemotherapy induced aplasia. As shown in Figure 2, 23 patients out of 71, although in CR, displayed WT1 values above the normal range with a median value of 112 WT1 copies (mean 292±638, range 23–2840). By contrast, in 48 out of 71 patients who entered CR, the amount of WT1 transcript measured after induction treatment fell within the range detected in healthy controls with a median value of WT1 of 6 copies/10 ABL (mean=7.1±5, range 0.5–19). Interestingly, all patients showing WT1 values above the normal upper limit after induction chemotherapy relapsed after a median of seven months (range 4–16). As shown in Figure 3, 21 of the 48 patients with normal WT1 values after chemotherapy relapsed after a median of 12 months from diagnosis (range 6–44 months) and 27 patients persisted in CR after a median of 28 months of follow-up (range 12–60 months) (Figure 1). No significant differences were observed between WT1 transcript levels detected at CR after induction chemotherapy in the cohort of 27 out of 48 patients who subsequently persisted in remission when compared with the 21 who relapsed during follow-up (p=0.33). In all patients who reached a normal WT1 value after induction chemotherapy and later relapsed during follow-up, at least one abnormal value was detected before relapse. The detection of an abnormal WT1 value preceded by 1–6 months (mean 2.4 months) the hematologic relapse. In none of the 27 patients who persisted in CR were abnormal WT1 values detected during follow-up. Compared with the 21 patients who relapsed after achieving normal WT1 levels after induction therapy, the patients who showed abnormal WT1 levels after chemotherapy underwent disease relapse after a significantly shorter time interval from diagnosis to relapse (median of 7 vs. 12 months; mean 8.3±3.3 vs. 14.8±9.4; p=0.0033). In the 11 patients resistant to chemotherapy, the WT1 transcript amount persisted at very high levels after chemotherapy with a median value of 5,647 (mean 13,886±27,548) WT1 copies/10 ABL copies at diagnosis and a median of 3,180 (mean 10,232±19,103) after induction chemotherapy. This study shows that longitudinal quantitative analysis of WT1 expression in the PB of AML patients may provide relevant prognostic information by early identification of patients at highest risk of relapse. Doubts about the use of WT1 expression as a marker for MRD monitoring in leukemia have mainly been based on the existing background expression derived from normal hematopoietic cells.20 The recent advent of RQ-PCR technique for more precise and standardized quantification has allowed this problem to be partially overcome by identifying threshold values distinguishing the WT1 transcript amounts expressed in normal subjects from those of leukemic cells.10–16 We found that such discrimination, which is extremely relevant for the purpose of clinical studies of MRD, is best accomplished by using PB instead of BM, due to considerably lower WT1 expression levels and smaller individual variability in normal PB compared with BM. Whether WT1 amount at diagnosis in leukemia patients has any prognostic significance is still a subject of debate.21–23 In our study, the WT1 transcript level at diagnosis seems not to be correlated with patient outcome. The most important information derived from our study comes from WT1 expression analysis after induction chemotherapy, when the standard criteria define the patient as being in complete remission, and morphological criteria and flow cytometry on PB did not reveal the presence of circulationg blast cells. Since WT1 is overexpressed in the large majority of AML cases, it could represent the ideal marker for MRD evaluation. Recently, Lapillonne et al.24 demonstrated that WT1 quantitative assessment after a first course of induction treatment in BM samples represents the ideal tool to identify pediatric acute leukemia patients at high risk of relapse. Our data demonstrate that the sensitivity of WT1 analysis in PB is equal if not better than BM. Furthermore, it identifies a precise time point, in particular soon after the first cycle of chemotherapy for the evaluation of WT1 transcript with the intent of identifying patients at high risk of relapse. At this time point, WT1 copy number allows a better assessment of the quality of remission. In particular, the observation that abnormal WT1 values in PB after the first course of therapy in CR patients strictly correlate with relapse represents an important achievement as it would allow clinicians to intensify post-induction therapy and use more aggressive consolidation cycle to prevent relapse. From our study, however, we have seen that approximately half of the patients who reach normal WT1 values after induction chemotherapy relapse, although the reappearance of disease occurs later compared with those who do not reach the normalization of WT1. Therefore, only the abnormal WT1 values after induction treatment are unfailingly predictive of relapse. For the remaining patients, a stringent molecular follow-up post-remission is recommended since this may allow relapse to be predicted some months before its occurrence, when the conventional methods used are still unable to identify the reappearance of leukemic cells. The ongoing efforts to standardize real time methods of WT1 assessment and the introduction of rigorous, internationally accepted controls will enable RQ-PCR to become a robust and routine basis for diagnostic and prognostic procedures.

Figure 2.WT1 expression at diagnosis and during follow-up of 23 patients in CR characterized by WT1 values above the normal upper limit in PB after induction chemotherapy. All these patients relapsed after a median of seven months from diagnosis. R: relapse. The broken line represents the upper normal limit.

Figure 3.WT1 expression at diagnosis and during follow-up of 21 patients in CR characterized by WT1 values within the normal range after induction chemotherapy. All these patients relapsed after a median of 13 months from diagnosis. The broken line represents the upper normal limit.

Footnotes

• Funding: this work has been supported by grants from AIRC (Associazione Italiana per la Ricerca sul Cancro), CNR (Progetto Finalizzato Oncologia), MURSTCOFIN 2003, AIL (Associazione Italiana contro le Leucemie), CRT, and by Regione Piemonte. EM is a fellow of the Gigi Ghirotti Foundation.
• The online version of this article contains a supplemental appendix.
• Authorship and Disclosures DC designed the study and wrote the manuscript; FM, FA and IDF collected the biological samples; EG, MF and DD performed RQ-PCR; SC, VR and RC performed RNA collection and sample storage; EM and PN collected clinical data; MS and FL-C revised the manuscript and GS provided the final approval. The authors reported no potential conflicts of interest.

• Received August 28, 2007.
• Revision received January 2, 2008.
• Accepted January 29, 2008.