Secretion of Wnts is dispensable for hematopoiesis

In this issue of Blood, Kabiri and coworkers report the hematopoietic deletion of the endoplasmic reticulum–localized O-acyltransferase porcupine (PORCN), which is necessary for acylation of Wnts in the endoplasmic reticulum, enabling their secretion and binding to the frizzled receptors.¹  Unexpectedly, the absence of secreted Wnt factors does not have major effects on steady-state in vivo hematopoiesis or on long-term repopulating activity of Wnt-deficient hematopoietic stem cells.

Wnt signaling has been implicated to be a driving force in survival and proliferation in various types of cancer. As a result, many investigators are finding strategies to interfere with the activity of Wnt signaling intermediates. Targeting secretion of Wnt factors is one such promising strategy. Posttranslational acylation by PORCN is required for the secretion of virtually all Wnt factors.^2,3  The Wnt signaling pathway consists of a β- and γ-catenin–dependent canonical pathway and calcium ion– and small GTPase–dependent non-canonical pathways. The controversy surrounding the role of Wnts in hematopoiesis stems from observations in knockout mouse models, which showed that mediators of the Wnt pathway were essential for proper lymphocyte development. Surprisingly, however, normal hematopoiesis as well as lymphopoiesis occurs in the absence of both central mediators of canonical Wnt signaling: β- and γ-catenin.^4,5  In addition, transplantation of long-term hematopoietic stem cells (LT-HSCs) lacking both catenins show essentially normal hematopoietic regeneration, with normal development of all lymphocyte lineages, both in primary and secondary recipients. Nevertheless, expression of the activated S33Y form of β-catenin or upstream factors of the canonical Wnt signaling pathway shows that increased catenin activation interferes with normal progenitor production and maturation as well as decreases LT-HSC activity in a transplantation setting.^6,7  Some of the confusion regarding the lack of effect of catenin deletion and the hematopoiesis disruption at catenin activation was lifted in an elegant study in which Apc, a negative catenin regulator, was dosed. This study showed that LT-HSC is enhanced by low-level (twofold) catenin activation, whereas progressive activation impairs LT-HSC self-renewal in favor of enhanced myelopoiesis and lymphopoiesis, which, at further catenin activation, both show impaired differentiation.⁸ 

In this issue of Blood, Kabiri and coworkers took a different approach to investigate the impact of Wnt signaling on catenin activation and hematopoiesis. In several different conditional knockout models of the Wnt acyltransferase Porcn, they found that the number of different hematopoietic subsets, ranging from LT-HSCs to mature cells, did not change as a result of Porcn deletion (Porcn^Del).¹  Moreover, transplantation of Porcn^Del bone marrow cells showed normal hematopoietic reconstitution. Interestingly, expression of Wnt transcriptional targets, like Axin2, cyclin D1 (Ccnd1), and Myc, were unchanged in the Porcn^Del cells.¹  Considering reports clearly showing a dependence of hematopoietic reconstitution on Wnt factors like Wnt3a⁹  and Wnt5a,¹⁰  these results are highly surprising.

The absence of gross effects of Porcn deletion on hematopoiesis is good news for efforts aimed at inhibition of PORCN for clinical applications. The work shows that both systemic and targeted deletion of Porcn in hematopoietic cells does not impair hematopoiesis or regenerative capacity of LT-HSC.

However, the paper by Kabiri and coworkers also raises more fundamental questions about the nature of Wnt signaling, and its impact on cellular behavior. For instance, they find that despite deletion of the Porcn gene, transcription of the catenin-dependent Wnt targets Axin2, Ccnd1, and Myc, are still activated as if it is “business as usual.” Although (β- or γ-) catenin activation was not specifically studied in this paper, it seems likely that activation of catenin-dependent transcription does occur. It is possible that autocrine-acting Wnt factors do not need to be secreted to be active. For instance, intracellular Wnt5a expression localizing to polarized actin aggregations can be detected within LT-HSC¹⁰  and might activate signaling from within the cells. Another possibility is Wnt-independent catenin activation through other signaling pathways. For instance, catenin has been shown to be at the crossroads of several pathways, including the hedgehog, hippo, HIF-1a, FAK, Notch, and transforming growth factor β pathways. Thus, any of these pathways could crossactivate catenin and its targets in LT-HSCs and mature cells. Thus, the intriguing paper by Kabiri and coworkers not only resolves some of the questions regarding Porcn and secretion of Wnts in hematopoiesis but also fuels the catenin-Wnt controversy in hematopoiesis and suggests that catenin-dependent Wnt signaling occurs in the absence of secreted Wnt factors.