Expression Pattern of WT1 Isoforms in Patients with Acute Myeloid Leukemia (AML), Myelodysplastic Syndrome (MDS) and Severe Aplastic Anemia (SAA)

The Wilms' tumor gene 1 (WT1) is highly expressed in a large proportion of human acute leukemias and other hematological malignancies. It has been demonstrated that WT1 protein is produced in more than 36 different isoforms. These variants have distinct, partially overlapping functions and their ratio is supposed to influence the final effect of WT1. However, only limited information on WT1 isoforms has been published so far and the relevance of their expression pattern remains unclear.

Our main aim was to determine the expression pattern of four WT1 isoforms characterized by the presence or absence of exon 5 and KTS insert (A[−/−], B[+/−], C[−/+], D[+/+]) in patients with AML, MDS and SAA.

We designed a unique qPCR system for detection and quantification of 4 major WT1 isoforms. Using this method we analyzed the ratio of WT1 isoforms in 8 leukemic cell lines (Kasumi-1, NB-4, K562, MV4;11, REH, NALM6, UOCB6, RS4;11) and diagnostic BM samples of 73 childhood AML, 30 adult AML, 20 childhood MDS, 9 childhood SAA as well as 23 healthy controls.

Median expression level of total WT1 in healthy donors was 29 WT1/ABL×10⁴ NCN. Childhood and adult AML expressed WT1 on significantly higher level compared to healthy controls (2058 and 3446 WT1/ABL×10⁴ NCN respectively; p<0.0001). The expression level of total WT1 in patients with MDS was 196 WT1/ABL×10⁴ NCN, whereas children with SAA had total WT1 comparable or lower than control samples (4 WT1/ABL×10⁴ NCN).

We found an excellent correlation between the total WT1 expression and the sum of WT1 isoforms (rho=0.916, p<0.0001). However, very low levels of total WT1 precluded detection of WT1 isoforms in majority of healthy donors and SAA samples, since we reached the limit of the sensitivity of qPCR method (as defined by limiting dilution experiments). For the same reason (very low total WT1 levels), 18 patients with AML (mostly AML M5) as well as 8 MDS samples were excluded from further analyses.

The analysis of WT1 isoforms ratio showed a diverse expression patterns of WT1 variants in the particular cell lines (p<0.0001). Interestingly, a similar pattern of WT1 isoforms was present in cell lines with MLL/AF4 rearrangement, independent of the lineage (myeloid - MV4;11 and lymphoid - RS4;11).

We could successfully determine the expression profile of WT1 isoforms in 1 healthy BM, which showed a substantial overexpression of isoform D with the ratio of 1.2: 1.8: 1.7: 5.3 for A, B, C and D variants, respectively. Rather surprisingly, we found a uniform expression pattern of WT1 isoforms (D>B>C∼A) in patients with childhood AML as well as adult AML (0.9: 2.5: 1.1: 5.5 and 1.3: 3.1: 1.3: 4.5 for isoforms A, B, C and D respectively). In contrast to AML samples, a different and more variable ratio of WT1 variants was found in children with MDS with predominance of variant D and similar levels of isoforms A, B and C (1.6: 1.5: 1.7: 4.6 for A, B, C and D respectively). Moreover, the unsupervised hierarchical cluster analysis according to the expression pattern of WT1 isoforms indicated that children with MDS belonged to a distinct cluster compared to childhood and adult AML.

This is the first report of the analysis of WT1 isoforms expression pattern in AML and MDS using a unique qPCR method for the detection of WT1 variants. Our data suggest that WT1 isoforms expression pattern is surprisingly uniform in pediatric and adult AML with predominant expression of Exon5[+] isoforms. To the contrary, the ratio of WT1 isoforms in childhood MDS significantly differ from AML indicating that the expression pattern of WT1 variants is related to the type of malignant cells.

Careful pre-analytical testing of the detection system parameters suggests the technical limitations in detection of WT1 isoforms in samples with very low total WT1 levels. Therefore, the previously reported differences in WT1 isoforms ratio between normal bone marrow and AML samples should be interpreted with caution.

Supported by MSM0021620813, GAUK 81709, IGA NS10488–3/2009

No relevant conflicts of interest to declare.