To the Editor

Abstract

It is well established that canonical Wnt signalling plays essential roles during embryogenesis and the self-renewal of adult tissues such as the skin and the gut. In addition, in a review in this journal, Staal and Sen have highlighted an important role of this pathway for haematopoiesis in general 1. Although we agree with this conclusion, we believe that it can currently not be deduced that canonical Wnt signalling within haematopoietic cells, as opposed to supporting non-haematopoietic cells, is essential for haematopoiesis in the adult. When considering the existing literature (as shown in Fig. 2 of 1), it is evident that the role of canonical Wnt signalling in haematopoiesis has been deduced to a large extent from gain-of-function experiments, i.e. the enforced expression of components of this pathway or ex vivo treatment of cells by Wnt ligands or inhibitors. Although some of these manipulations have yielded striking phenotypes, the relevance of these findings to normal haematopoiesis is uncertain. Indeed, as a precedent, while gain-of-function experiments suggested that Notch signals play a role in expanding hematopoietic progenitors, Notch signalling in HSC plays no essential role in the maintenance of HSC 2. There are few reports on loss-of-function experiments of genes implicated in the Wnt pathway that show clear-cut effects in the haematopoietic system. However, for most of these “clear-cut” phenotypes, it was not determined whether Wnt signalling was actually involved. This is important as many of these Wnt-signalling genes play roles outside the Wnt pathway. In addition, even when specific Wnt signalling components such as Fzd or Wnts were deleted, it was often not analysed whether the effects found involved canonical, β-catenin-mediated signalling or non-canonical, β-catenin independent signalling. Indeed, a recent study, which looked at this aspect in more detail, reported that Wnt4 affects haematopiesis via non-canonical signalling 3. Finally, in most cases where a phenotype was detected, it was not determined whether the defect was intrinsic to haematopoietic cells or whether it was indirect, e.g. due to a defect in stromal cells. The deletion of two nuclear effectors of canonical Wnt signalling, LEF-1 and/or TCF-1, resulted in impaired lymphocyte development, which was based on a defect(s) in the haematopoietic compartment 4. However, these transcription factors can interact with several regulatory proteins and repress transcriptional activity in the absence of Wnt signalling. It was thus possible that the loss of this repressive function caused the phenotype. Indicative of a role of canonical Wnt signal transduction in haematopoeisis, we have shown that thymopoiesis is dependent on the N-terminal domain of TCF-1. This domain binds β- and γ-catenin 5, 6, the two known transducers of canonical Wnt signals in vertebrates. However, we and others have recently reported that lymphopoiesis and haematopoiesis were normal in the combined absence of β- and γ-catenin in the hematopoietic compartment 5, 6. Importantly, we further showed that haematopoietic cells lacking β- and γ-catenin retained significant TCF/LEF reporter gene activity when stimulated ex vivo by Wnt ligands 5. We concluded that canonical Wnt signals can be transduced in the absence of β- and γ-catenin. However, based on these findings, it can currently not be concluded that canonical Wnt signalling in haematopoietic cells is essential for normal haematopoiesis. Although full-size β-catenin was absent, we reported the presence of a shorter β-catenin species upon cre-mediated deletion of exons III-VI of the Ctnnb1 gene specifically in haematopoietic cells 5. It was thus possible that TCF/LEF reporter gene activity in our mice was mediated by this truncated β-catenin protein. However, this β-catenin species does not bind to TCF-1 (5 and Fig. 1) nor does it activate a TCF/LEF reporter 5. Obviously, while it cannot be excluded that the truncated β-catenin protein has some other effects, our study 5 rules out the possibility that it transduces canonical Wnt signals. Truncated β-catenin does not bind TCF-1. HEK 293T cells were transiently transfected with myc-tagged TCF-1 alone (lane 1) or together with truncated β-catenin (lane 2) or with full-length β-catenin (lane 3). Anti-myc immunoprecipitates (i.p.) and whole cell lysates were analysed by Western blot (WB) specific for β-catenin. Based on these findings, we proposed the existence of an additional, currently unidentified transducer of Wnt signals in haematopoietic cells. Consistent with this possibility, an additional β-catenin-like Wnt transducer was indeed identified in Caenorhabditis elegans 7. However, even if an additional transducer of canonical Wnt signals is found in vertebrates, this still does not prove that canonical Wnt signalling plays a role in haematopoietic cells. Future studies, using a complete block of canonical Wnt signalling selectively in hamatopoietic cells, will be needed to determine the precise role of this signalling pathway for normal haematopoiesis in the adult