The gene encoding Wilms Tumor 1 (WT1) is recurrently upregulated in Acute Myeloid Leukemia (AML). WT1 transcript burden is associated with primary refractory disease and with relapse when there is detectable WT1 minimal residual disease (Lambert et al., Oncotarget, 2014). We have recently described the gene regulatory networks in purified blasts from AML patients with various driver mutations (Assi et al., Nature Genetics, 2019) and found that WT1 was a key transcription factor node in all networks examined. Furthermore, in knockdown experiments as part of a shRNA depletion screen against transcription factors in vitro and in murine xenotransplantation experiments, we showed that WT1 was essential for leukemic maintenance (Martinez-Soria et al., Cancer Cell, 2018).
WT1 produces 8 primary isoforms, all of which are expressed in hematopoietic cells by employing one of two transcription start sites and alternative splicing in exon 5 and exon 9. To further understand how WT1 maintains leukemia, we cloned each of the 8 isoforms into a doxycycline-inducible vector and transduced them into a t(8;21) AML cell line, Kasumi-1, which well recapitulates primary t(8;21) AML (Ptasinska et al., Leukemia, 2014).
We carefully overexpressed each of the 8 WT1 isoforms to levels seen in AML patient samples and functionally and molecularly characterised them. Overexpression of isoforms lacking 3 amino acids in the zinc finger DNA binding domain (WT1 -KTS) significantly reduced cell growth and colony formation, increased apoptosis, caused a G1 cell cycle arrest and caused myeloid differentiation of leukemic blasts. However, overexpression of isoforms with the 3 amino acids (WT1 +KTS) in the DNA-binding domain had the opposite effects, thus having a major role in the maintenance of a leukemic phenotype. Also, we found that the expression of WT1 +KTS isoforms in purified blasts from patients was higher than WT1 -KTS isoforms and may therefore permit leukemic maintenance. We did not find any functional differences between isoforms employing alternate transcriptional start sites or between exon 5 splice variants.
We then performed the first ever WT1 chromatin immunoprecipitation (ChIP-seq) in AML and show that WT1 -KTS shows increased binding as compared to WT1 +KTS with 9451 unique WT1 binding sites in WT1 -KTS expressing cells and only 1355 sites in WT1 +KTS expressing cells with 2391 sites overlapping between the isoforms. This finding suggests that WT1 isoforms compete for binding at multiple sites and may explain why the ratio of expression between WT1 isoforms is critical (Calabrese et al., Blood, 2012). We also correlated WT1 ChIP-seq peaks to transcription as assayed by RNA-seq. We found that differentially bound cis-regulatory elements are associated with altered expression of genes involved in proliferation, cell cycle, apoptosis and differentiation pathways, explaining many of the functional alterations in cellular behaviour that we see.
Several vaccine studies and more recently a T cell receptor study targeted against WT1 have been undertaken but they have shown mixed efficacy in AML (Chapuis et al, Nature Medicine, 2019). Since we find that different isoforms of WT1 have antagonistic effects, we hypothesise that a more effective therapeutic strategy would be to selectively target only WT1 +KTS isoforms.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
